Annual Report 2015-2017 Annual Report 2015-2017 Annual Report 2015-20171

ForWindCenter for Wind Energy Research

ForWindCenter for Wind Energy Research

ANNUAL REPORT 2015-2017

– 1 – 2 Annual Report 2015-2017

GREETING

A unique R&D infrastructure is the fundamental basis for a relevant R&D programme for the industry and other sectors and sustains over a long span of time. I am happy to see that ForWind managed to extend its already unique wind energy research infrastructure during the period covered by this report. The research and test campus in Hannover-Marienwerder, which includes the test centre for support structures (TTH), the large wave flume, and the 3D wave basin has been expanded by the GeCoLab, a universal motor and generator test bench, which enables in depth investigation of electrical machines and converters. It facilitates a detailed investigation on both conventional and innovative converter and generator concepts including control and filter design methods. This includes investigations into dynamics and system stability, stationary and transient thermal loading, various methods of grid feed-in and control and the response to grid faults. This new generator converter laboratory was inaugurated in 2015 and is situated inside the TTH building. It is not only adding new test capabilities to For- Wind, but is also another example of how ForWind facilitates the necessary interdisciplinary research, stimulating electrical and civil engineers to research the interface problems of their respective disciplines.

Only two years later, another key research infrastructure was inaugurated, the WindLab at the Oldenburg location. This laboratory for turbulence and wind energy systems research adds a unique and badly needed wind tunnel facility to ForWind’s wind tunnel portfolio. A 30 m long measurement section, high wind speeds and especially the impressive self-developed active grid not only allow to reproduce realistically scaled atmospheric turbulence conditions in the constrained dimensions of a wind tunnel, but even to repeat them as often as necessary. This will change the way wind turbine models and their interactions in wind farm configurations can be investigated under controlled and defined external conditions.

I am also impressed by the welcoming and bright architecture of the WindLab building, which accommodates now all of ForWind’s wind energy researchers, technicians and management staff in Oldenburg. Furthermore it provides meeting and lecture rooms to host courses, conferences and workshops. This everything-under-one-roof concept creates the creative and innovative atmosphere, which leads to new ideas and solutions that so often originate from coffee-break discussions and spontaneous hall-way encounters.

At the Bremen location of ForWind, the research and test infrastructure extension continues. Already at the horizon is a laboratory that aims at investigating high-performance power electronic systems for wind turbines under realistic environmental and load conditions. It will be used for investigating their failure causes and to develop and to experimentally verify concepts for optimizing their robustness. I am pleased to see that the Federal Ministry for Economic Affairs and Energy is supporting this concept via the HiPE-WiND project, which was granted 12 million Euro in October 2019. This research infrastructure will also be used by ForWind’s prime research partner Fraunhofer IWES pursuing the successful and proven collaboration. Commissioning is scheduled for the end of 2019 and I am sure that this facility will be in great demand by both research establishments and industrial partners.

The realisation of these unique research facilities, their acquisition in competitive tenders and their successful implementation and use in projects reflect the success of ForWind as an interdisciplinary and internationally recognized center of expertise. Thus ForWind is excellently positioned to address the challenges of the global energy system transformation and to support the wind energy industry in doing so.

Dr. h.c. Ir. Jos Beurskens Chairman of the ForWind Advisory Board

– 2 – Annual Report 2015-2017 Annual Report 2015-20173

PREFACE

This report covers the period of 2015 to 2017. A period with extensions of ForWind’s research infrastructures, new member institutes and exciting new research projects and topics.

Of the many research projects carried out in the reporting period, we would like to highlight a few that are outstanding. First, there is "ventus efficiens – collaborative research to increase the efficiency of wind turbines in the energy system". This project involves a very large share of all ForWind members and forms the breeding ground on which many new project ideas are generated.

The joint research within the German Research Alliance Wind Energy (FVWE) – a unique strategic cooperation between the German Aerospace Centre (DLR), the Fraunhofer Institute for Wind Energy Systems (IWES), and ForWind – started with the project "Smart Blades – development and construction of smart rotor blades" in 2012. This project, which finished in 2016, was directly succeeded by the project "SmartBlades2 – construction, testing and development of smart rotor blades". The interest of industry, other stakeholders and media in this topic has been exceptional and clearly shows the relevance of such precompetitive research activities.

Despite the vast research infrastructures and laboratories, which are operated and used by the ForWind members, the comparison of research results with reality remains a challenge. Complete verification and validation under real external conditions is only possible to perfection in or on real plants. For several research questions this is impossible to do in commercially operated wind power plants or turbines. The development of the German Research Facility for Wind Energy (DFWind) is therefore a very exciting possibility. The research wind turbines as well as the many accompanying met masts will offer world class opportunities for the wind energy research community and the entire sector. ForWind, DLR, and IWES are working inexhaustibly to make this platform a reality, which hopefully will be the basis for many thrilling joint research projects yet to come.

We are thrilled to have the project "marTech – testing and development of maritime technologies for reliable energy supply" within the ForWind network. This project, which will tremendously increase the capabilities of the large wave flume in Hannover-Marienwerder, is the largest funded project yet. 35 million EUR have been granted by the Federal Government alone, co-funded by state money and investments by the University of Hanover. We are very much looking forward to using the improved large wave flume in research projects in the future.

All ForWind members are especially grateful to the Federal Ministry for Economic Affairs and Energy (BMWi), including the supporting Project Management Jülich (PtJ), and the Ministry for Science and Culture of Lower Saxony (MWK), who are the most prominent sponsors of ForWind’s research projects. In addition, the Federal Ministry for Education and Research (BMBF), the German Research Foundation (DFG), the European Commission (EC), and other regional and national sponsors supported us in tackling the challenges of and developing innovations for the wind energy sector.

ForWind is based on interdisciplinary cooperation and as such, we appreciate the stimulating teamwork with our national and international partners from research and industry. Together we will continue to support the entire sector in increasing the value of wind energy to the maximum.

Last but not least, we would like to thank all members of the ForWind Advisory Board for their always very valuable comments, counseling and suggestions as well as all ForWind staff for their unprecedented efforts and enthusiasm.

Prof. Dr.-Ing.Klaus-Dieter Thoben Prof. Dr.-Ing. Peter Schaumann Academic Speaker Deputy Speaker

– 3 – 4 Annual Report 2015-2017

CONTENT

Greeting Preface

The Organization

• Executive Board ______8 • Advisory Board ______9 • Participation Institutes ______9

Projects

WIND AS A RESOURCE

• Forwind – Small-Scale Wind Field Modelling / Large-Eddy Simulation ______10 • Parallel Computing Cluster for CFD and Wind Turbine Modeling ______12 • Analysis of the Shadowing Effects and Wake Turbulence Characteristics of Large Offshore Wind Farms by Comparison of alpha ventus and Riffgat (GW Wakes) ______15 • ClusterDesign – A Toolbox for Offshore Wind Farm Cluster Design ______19 • RealMe – Traceable Acquisition of Meteorological Data ______21 • WindScanner.eu – The European WindScanner Facility, Preparatory Phase ______22 • European High-performance Infrastructures in Turbulence (EuHIT) ______24 • PhD Programme on “System Integration of Renewable Energies” ______26 • Turbulence Resolving Simulations for Meteorological Assessment of Wind Energy Sites in Complex Terrain ______29 • German Research Facility for Wind Energy (DFWind) ______31 • NEWA – New European Wind Atlas Joint Programme ______35 • High Performance Computing (HPC) Cluster (WIMS-Cluster) ______38 • Further Development of a Two-dimensional Atmospheric Laser Cantilever Anemometer for the High-resolution Investigation of Turbulent Wind Flows for the Optimization of Offshore Wind Turbine Blades ______40 • Extended THETA for Site Assessment (ETESIAN) ______42 • Highly Accelerated Pitch Bearing Test ______44 • Lidar for Wind Energy Deployment (IEA Annex 32) – Phase II______46

SUPPORT STRUCTURES

• GROWup OPC– Grouted Joints for Offshore Wind Energy Converters under Reversed Axial Loadings and Upscaled Thicknesses filled with Ordinary Portland Cement (OPC) ______47 • INNWIND.EU – Innovative Wind Conversion Systems (10-20MW) for Offshore Applications ______50 • ProBeton – Development and Experimental Testing of Fatigue Resistant Concrete Foundation Structures for Offshore Wind Turbines ______52_ • HyConCast – Hybrid Substructure of High Strength Concrete and Ductile Iron Castings for Offshore Wind Turbines Cyclic Degradation of Axially Loaded Piles (2013) ______55 • Integration of Calculation Methods in the Software Tool IGtHPile ______57 • Ventus Efficiens – Joint Research for Raising the Efficiency of Wind Energy Converters within the Energy Supply System Subproject II – Efficiency Enhancement of Supporting Structures ______59

– 4 – Annual Report 2015-2017 Annual Report 2015-20175

• Bearing Behavior of Suction Bucket Foundations under Tensile Loading in Non-cohesive Soils ______63 • Development and Demonstration of a Rapid and Cost-Effective Installation Concept for Offshore Wind Turbines ______65 • WinConFat – Fatigue of Materials of Onshore and Offshore Wind Energy Structures Made of Reinforced and Prestressed Concrete under High Cycle Loading ______67 • ThermoFlight – Concept for the Development of an Opti-mized Maintenance and Inspection Method for Off-Shore Wind Turbines Using Thermography and SHM as Non-Destructive Testing Technologies in Combination with Unmanned Aerial Vehicles ______69

WIND POWER SYSTEMS

• Development of Nacelle-Based Lidar Technology for Measurement of Power Characteristic and Control of Wind Turbines (LIDAR II) ______70 • GeCoLab – Test Bench for Generators and Converters Systems ______72 • EVeQT– Increased Availability and Quality Optimization of Power Train Components and Gears for Wind Energy Systems 74__ • BladeMaker______75 • CompactWind: Increase of Wind Farm Power Output through Advanced Wind Turbine and Wind Farm Control Algorithms ______77 • Probabilistic Load Description, Monitoring and Reduction of Loads of Future Offshore Wind Energy Converters (OWEA Loads)______80 • Silicon Carbide Power Technology for Energy Effi cient Devices (SPEED) ______83_ • Innovation Cluster Power Electronics for Regenerative Energy Supply ______86 • Manufacturer-independent Retrofitting and Evaluation of the Remaining Life Time of Power Electronic of Wind Turbines with DFIG (WEA-Retrofit) ______88 • BisWind – Component Integrated Sensor System for Wind Energy Systems ______91 • Intermittent Wind Loads (InterWiLa) ______92 • HyRoS – Multifunctional Hybrid Solution for Rotor Blades Protection ______94 • Adaptive Operation and Control of Offshore Wind Farms Based on Specific Operational Strategies for Optimized Yield, Loads, and Grid Integration (RAVE – OWP Control) ______96 • Multi-Dimensional Stresses in High Power Electronics of Wind Energy Plants (HiPE-WiND) ______100 • PiB – Predictive Intelligent Operation System to Reduce the Risk of Icing of Wind Turbines ______102

WIND ENERGY INTEGRATION

• Advanced Wind Energy Systems Operation and Maintenance Expertise (AWESOME) ______104 • Resilience of Socio-Technical Systems exemplified at the Electricity Transport and Actor System ______106 • Influence of Short-time Dynamics of Renewable Energies on the Stability of Power Grids ______108 • GEOWISOL – Effects of the Geographical Distribution and Temporal Correlation of Wind and Solar Input on the Power Supply System ______109

STRUCTURAL HEALTH MONITORING

• IR-Vortex – Infrared Thermography as a Measurement Technique for Visualization of Vortex Structures on Rotating Rotor Blades ______111 • SHM_Rotorblatt: System for Early Damage and Ice Detection for Rotor Blades of Offshore Wind Turbines ______113 • Monitoring Suction Bucket Jacket – Development of feasible Numerical Simulation Methods for the Investigation of the Compression and Tension Load bearing Behaviour and the Prediction of Displacement Accumulation ______115 • Multivariate Structural Health Monitoring for Rotor Blades ______118

– 5 – 6 Annual Report 2015-2017

• Development and Validation of a Virtual Process Chain for Composite Structural Components Considering Imperfections with Application to a Rotor Blade Component (PROSIM R) ______121 146 148 Acoustics

• Inflow Noise ______123 • WEA-Akzeptanz – From the Source to the Perception ______124 150 152 Rotor Blades 154 156 158 • Generation of Realistic Turbulent in-Flow Conditions by an Active Grid, Stochastic Analysis of the Resulting Force 160 Dynamics on 2D Segments of Airfoils with Active Control Units ______128 161 • Advanced Aerodynamic Tools for Large Rotors – AVATAR ______131 • Hybrid Laminates and Nanoparticle Reinforced Materials for Improved Rotor Blade Structures (LENAH) ______134 • Wind Turbine Load Control under Realistic Turbulent In-Flow Conditions ______136

Cross-disciplinary Topics 162 164 166 • Smart Blades – Development and Design of Intelligent Rotor Blades ______139 168 • GigaWind life – Life time Research on Support Structures in the Offshore Test Site alpha ventus ______143 169 • Innovative Wind Conversion Systems (10-20 MW) for Offshore Applications (INNWIND.EU) ______162 • Integrated Research Programme on Wind Energy (IRPWIND) ______164 170 • Extreme Ocean Gravity Waves: Understanding and Predicting Breathers with Wave Breaking 171 and in Coastal Waters ______166 171 • SAMS Save Automation of Maritime Systems ______168 • Ventus Efficiens – AP 3 ______170 • Ventus Efficiens – Collaborate Research to increase the Efficiency of Wind Turbines in the Energy System ______173 • German Research Facility for Wind Energy (DFWind)______178 • Smart Blades 2 – Construction, Testing and Further Development of Intelligent Rotor Blades ______184 172 • Interdisciplinary Research Center on Critical Systems Engineering for Socio-Technical Systems II ______188 174 176 Qualification and Coordination 177

178 • Continuing Studies Programme Wind Energy Technology and Management (“Windstudium”) ______190 182 • European Wind Energy Master (EWEM) ______191 184 185 ForWind – Events 186

Events 2015

• Smart Blades Symposium in Oldenburg ______192 • Job and Education Fair “zukunftsenergien nordwest” ______193 • Hannover Messe 2015 ______194 • Visit of the Vice-President for the Energy Union an Energy Minister ______196 • Cooperation with Taiwanese University ______198

– 6 – Annual Report 2015-2017 Annual Report 2015-20177

Events 2016 188 190 • WindLab – New Research Building for Wind Energy Research ______199 198 • JGenerator Converter Lab (GeCoLab) ______201 199 Events 2017 202 202 202 • Minister President of Lower Saxony Stephan Weil visits ForWind ______203 • ForWind Symposia ______204

Documentation 208

216 PUBLICATION LIST

Peer Reviewed Articles ______206 Articles, Conference Proceedings, Conference Contributions ______210 Project Reports ______218 Public Lectures, Articles in Magazines and Other Publications ______220 Lectures ______221 PhD Theses ______222 Diploma, Master and Bachelor Theses ______223

LIST OF FORWIND STAFF MEMBERS ______228

IMPRINT ______238 144

– 7 – 8 Annual Report 2015-2017

FORWIND – CENTER FOR WIND ENERGY RESEARCH

The Organization

ForWind concentrates wind energy re- ForWind was founded in 2003 with the Together with the German Aerospace search in the northwest and connects support of the Ministry of Science and Cul- Center (DLR) and the Fraunhofer Institute 28 institutes and working groups of the ture of Lower Saxony (MWK). Wind energy for Wind Energy Systems IWES, ForWind universities of Oldenburg, Hannover and research at the universities of Oldenburg forms the Research Alliance Wind Energy. Bremen. Thus, ForWind forms a nation- and Hannover has since been combined wide unique research network and covers in ForWind. In 2009, the University of a broad spectrum of scientific topics. Re- Bremen joined as a new partner, enabling search focuses on engineering sciences, ForWind to significantly widen its research physics and meteorology, computational spectrum once again. ForWind is still sup- science and economics. ported by the MWK.

Prof. Dr.-Ing. Bernd Orlik Prof. Dr.-Ing. habil. Raimund Rolfes Prof. Dr.-Ing. Klaus-Dieter Thoben

Prof. Prof. Dr. Joachim Peinke Prof. Dr.-Ing. Peter Schaumann Dr. Martin Kühn

Executive Board

ForWind is lead by an Executive Board The current board members are Prof. Dr. Martin Kühn, consisting of two members each of the Prof. Dr.-Ing. Klaus-Dieter Thoben Prof. Dr.-Ing. Bernd Orlik, Carl von Ossietzky University of Olden- (1. spokesperson), Prof. Dr. Joachim Peinke, burg, the Leibniz University of Hannover Prof. Dr.-Ing. Peter Schaumann and Prof. Dr.-Ing. habil. Raimund Rolfes. and the University of Bremen. (2. spokesperson),

– 8 – Annual Report 2015-2017 Annual Report 2015-20179

Advisory Board Prof. Dr. Dr. Hans Michael Piper Institute for Geotechnical Engineering President of the Carl von Ossietzky Prof. Dr.-Ing. Martin Achmus Universität Oldenburg Chairman: Institute for Electrical Engineering and Prof. Dr.-Ing. Bernd Scholz-Reiter Measurement Technology Dr. h.c. Jos Beurskens President of the Universität Bremen Prof. Dr.-Ing. Heyno Garbe Director Institute for Machine Design and SET Analysis, Schagen (NL) Participating Institutes Tribology Prof. Dr.-Ing. Gerhard Poll

Members: The following institutes and working Institute for Building Materials Science Dr. Rita Dunker groups at the universities of Oldenburg, Prof. Dr.-Ing. Steffen Marx Advisor Hannover and Bremen are organized in Institute of Meteorology and Climatology Ministry for Science, Health and ForWind: apl. Prof. Dr. Siegfried Raasch Consumer Protection, Bremen Department for Universities and Institute for Steel Construction Universität Bremen Research Prof. Dr.-Ing. Peter Schaumann Bremen Institute for Metrology, Dr.-Ing. Jörg Hermsmeier Institute of Structural Analysis Automation and Quality Science Head of Research and Development Prof. Dr.-Ing. habil. Raimund Rolfes Prof. Dr.-Ing. habil. Andreas Fischer Flender GmbH Institute of Turbomachinery and Fluid Bremen Institute for Mechanical Oliver Hilgert Dynamics Engineering Head of Department Forming and Prof. Dr.-Ing. Jörg Seume Prof. Dr.-Ing. Bernd Kuhfuß, Application Technology Prof. Dr.-Ing. Kirsten Tracht Institute for Wind Energy Systems Salzgitter Mannesmann Forschung Prof. Dr.-Ing. Andreas Reuter GmbH, Duisburg Institute of Automation Prof. Dr.-Ing Kai Michels Institute for Business Informatics Moritz Horn Prof. Dr. Michael H. Breitner Managing Director Institute for Electrical Drives, Power Van Oord OWP Germany GmbH Electronics and Devices Ludwig-Franzius-Institute for Hydraulic, Prof. Dr.-Ing. Nando Kaminski, Estuarine and Coastal Engineering Dr. Sebastian Huster Prof. Dr.-Ing. Bernd Orlik Prof. Dr.-Ing. habil. Thorsten Schlurmann Head of Department Engineering Sciences and Knowledge Transfer Institute for Integrated Product Ministry of Science and Culture of Development Universität Oldenburg Lower Saxony, Hanover Prof. Dr.-Ing. habil. Klaus-Dieter Thoben Computing Science, Business Engineering Dr. Klaus Meier Institute for Microsensors, - actuators and Prof. Dr.-Ing. Axel Hahn Managing Director -systems Institute of Physics, Energy Meteorology wpd windmanager GmbH & Co. KG, Prof. Dr.-Ing. Walter Lang Dr. Detlev Heinemann Bremen MARUM – Center for Marine Institute of Physics, Turbulence, Matthias Schubert Environmental Sciences Wind Energy and Stochastics Managing Director Prof. Dr. Tobias Mörz Prof. Dr. Joachim Peinke Wyncon GmbH, Hamburg Institute of Physics, Wind Energy Systems Lars Weigel Universität Hannover Prof. Dr. Martin Kühn Managing Director Institute for Drive Systems and Power turboMech GmbH & Co. KG, Electronics Uplengen Prof. Dr.-Ing. Axel Mertens, Prof. Dr.-Ing. Bernd Ponick

Guests: Institute for Concrete Construction Prof. Dr.-Ing. Ludger Lohaus Prof. Dr. Volker Epping President of the Leibniz Universität Institute for Electric Power Systems Hannover Prof. Dr.-Ing. habil. Lutz Hofmann

– 9 – 10WIND AS A RESSOURCE Annual Report 2015-2017

Forwind – Small-Scale Wind Field Modelling / Large-Eddy Simulation

Carl von Ossietzky Universität Oldenburg Wind turbines interact with the flow in the ic flows for different stratifications, surfaces Institute of Physics, ABL. This means that not only the wind and topographies with wind turbines. Dif- Research group: Energy Meteorology turbine behavior depends on the flow in the ferent wind turbine parameterizations have ABL, but also the flow in the ABL is modi- been implemented into the LES code PALM Martin Dörenkämper, Sonja Drüke, fied when it passes a wind turbine. Wind and these parameterizations are continuous- Sonja Krüger, Isabelle Prestel, turbine wakes are characterized by reduced ly further developed. The final objective of Michael Schmidt, Lars Segelken, Gerald Steinfeld, Lukas Vollmer, wind speeds compared to the undisturbed the team is to transfer the knowledge gained David Wagner, Björn Witha, ambient flow and by additional turbulence. from simulations with the pure research tool Hauke Wurps, Detlev Heinemann This additional turbulence is caused e.g. by PALM into improved parameterizations for the strong shear of wind speed in the transi- models used in the wind energy business, Funding: Ministry for Science and Culture tion region between wake and ambient flow. such as wind farm parameterizations for of the Federal State of Lower Saxony mesoscale models or improvements for en- Measurements in the real atmosphere as gineering wake models. Ref.Nr. N.A. well as in the wind tunnel aim at obtaining a better understanding of turbulent flows. Duration: 12/2009 – 11/2015 Besides those two approaches turbulence Project Description resolving numerical models are widely used in order to gain a better knowledge Accounting for wind turbines in the on flow conditions in the ABL and within large eddy simulation model PALM wind farms. One limiting boundary condi- Since 2009 different wind turbine models tion to numerical modelling of ABL flows have been implemented into the LES model is the wide range of scales of turbulent mo- PALM by members of the small scale wind tions in the ABL. The wide range of scales field modelling group. The available mod- Introduction prevents an investigation with the means of els include different types of actuator disc direct numerical simulation for most ABL approaches, an actuator line approach, an The atmospheric boundary layer (ABL) is flows. Large-eddy simulation (LES), which enhanced actuator disc model with rota- the part of the atmosphere in which wind follows the approach of explicitly resolving tion and finally also an enhanced actuator turbines are operating most of their lifetime. the bulk of the turbulent motions in the ABL line approach, in which the aerodynamic It is the layer of the atmosphere in which while parameterizing the impact of smaller forces acting on the rotor blade segments the impact of the ground surface can be scale turbulence on the resolved scales, is are obtained by coupling the large-eddy felt in the flow conditions. The flow condi- the state-of-the-art approach for in-depth simulation model PALM with an aeroelas- tions in the ABL are characterized by tur- analyses of turbulent ABL flows. Due to the tic code such as FLEX or FAST. The latest bulence. The turbulent motions in the ABL limitation of computational resources today implementation of the small scale wind field cover scales from some millimeters to some in numerical simulations the flow at the modeling group into the LES model PALM thousands of meters. The convective ABL rotor blades cannot be resolved explicitly is an actuator sector method. This method is characterized by especially strong turbu- if also the whole ABL and its interactions provides still a coupling with the aeroelastic lence, while in situations with stable strati- with several wind turbines is simulated. code FAST, but is considerably faster than fication the turbulence in the ABL is char- Therefore, studies that aim at investigating the actuator line approach, as it allows for acterized by a more intermittent character. the interactions between the ABL and wind larger time-steps. Besides the stability the roughness of the farms require parameterizations in order to underlying surface and also the topography account for the effect of wind turbines on A major achievement of the members of of the terrain play a key role for the genera- the atmospheric flow. the small scale wind field modelling group tion of turbulence in the ABL. As turbulence within the reporting period was the integra- causes loads on wind turbines a better un- The small scale wind field modelling team tion of the actuator disc model with rotation derstanding of atmospheric turbulence can is a sub-group within the Energy Meteorol- and the simple actuator disc model into the help to optimize the design of wind turbines ogy group of ForWind. The members of the standard code of PALM. This means that with regards to the expected turbulence con- group use the LES model PALM [1] in order now also other groups can use the develop- ditions at a certain site. to study the interaction between atmospher- ments that have been done at ForWind by

– 10 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201711

just downloading a recent version of the for the intentional misalignment of the wind plant canopy model that is available in the PALM code. This has led e.g. to the appli- turbine are required to have the wake at the LES model PALM. The successful partici- cation of PALM for wind energy studies at same position for different atmospheric sta- pation in the benchmark provided further the TU Delft in the Netherlands and in other bilities. Furthermore, it could be shown that trust in the model PALM to apply it also to groups of ForWind. the concept of wake deflection by intention- other forested sites. al yaw misalignment works well for cases of The enhanced actuator disc model was e.g. stable stratification while the concept could used to study the impact of atmospheric sta- not successfully be applied under unstable Summary bility on the recovery of wakes within wind stratification [2]. farms. The wind turbine models imple- The LES model PALM has been used at mented in PALM are continuously verified Improvement of the engineering wake ForWind to study the flow conditions in the and improved. One example is e.g. the im- model FLaP ABL, the interaction between the atmos- plementation of different tip-loss correction ForWind’s in-house engineering wake mod- pheric flow and wind energy converters, the models into the actuator line and enhanced el, the Farm Layout Program (FLaP) [3] has flow conditions within as well as the flow actuator disc model. been further developed by coupling it to an conditions downstream of wind farms. The analytic wind farm wake model in the re- objective is to gain a deeper insight into The small scale wind field modelling group porting period. This means that FLaP can these flows and to transfer the knowledge of the Energy Meteorology group at For- now also be applied for the investigation of gained to improvements for fast, engineer- Wind is an active participant and coordi- the impact of the layout of new wind farms ing wake models, the development of wind nator of German activities in the IEA task inside existing wind farm clusters. farm parameterizations for larger-scale 31 “Wakebench”. The IEA task provides a models or simply better descriptions of tur- number of benchmarks for flow models for Flows in complex terrains bulent processes in mesoscale models. The wind energy applications. The participation In the reporting period members of For- wind turbine models implemented in PALM in these benchmarks is one of the main ap- Wind took part in the micro-scale model since 2009 have been further developed in proaches that are followed in order to ver- comparison at the moderately complex the reporting period. In the reporting period ify the wind turbine parameterizations that forested site Ryningsnäs. This benchmark intensive work on the verification of the de- have been implemented in PALM. offered an ideal opportunity to validate the veloped approaches was in progess.

Recently, the small scale wind field modelling group has also contributed to a field experiment that aimed at providing validation data for flow models for wind energy applications. The field experiment took part at a wind farm site consisting of two wind turbines in moderately complex terrain in North-Eastern Germany.

Another focus of the small scale wind field Actuator Line Actuator Sector Actuator Disc Model modelling team within the Energy Meteor- ology group was on the implementation of a Model (ALM) Model (ASM) with Rotation (ADMR) dynamic sub-grid scale model into the LES Figure 1: Actuator models code PALM.

Development of wind farm control strategies References In the reporting period the focus concerning [1] Raasch, S.; Schröter, M.: PALM – stabilities: an LES study, Wind Energ. the development of wind farm control strat- A large-eddy simulation model per- Sci., 1, 129-141, https://doi.org/10.5194/ egies was on the method of deflecting wind forming on massively parallel wes-1-129-2016, 2016. turbine wakes by intentional misalignment computers, Meteorol. Z. 10 (2001) [3] Lange, B.; Waldl, H.-P.; Gil Guerrero, of wind turbines with the wind direction. By [2] Vollmer, L.; Steinfeld, G.; Heine- A.; Heinemann, D.; Barthelmie, R.J.: numerical experiments with PALM it could mann, D.; Kühn, M.: Estimating the Modelling of Offshore Wind Turbine be shown that the atmospheric stability is a wake deflection downstream of a Wakes with the Wind Farm Program key parameter for the shape and position of wind turbine in different atmospheric FLaP, Wind Energy 6 (2003) the wind turbine wake. Different yaw angles

– 11 – 12WIND AS A RESSOURCE Annual Report 2015-2017

Parallel Computing Cluster for CFD and Wind Turbine Modeling

Carl von Ossietzky Universität Oldenburg The project ended on 31.05.2015, the final late the interaction between the atmospheric Research Group: Wind Energy report has been published [1]. boundary layer and wind turbine wakes. Meteorology Three turbine parameterizations have been Björn Witha, Stefan Albensoeder, Project Description implemented in PALM [3] (see also Fig. Wided Medjroubi, Detlev Heinemann, 1): A simple and fast actuator disk model Joachim Peinke HPC operation (ADM), a much more accurate but com- Funding: Federal Ministry for Economic The operational support of the HPC FLOW putationally expensive actuator line model Affairs and Energy (BMWi) system and the user support are included in (ALM) and an enhanced actuator disk mod- the project. This covers the installation and el with rotation (ADM-R) which considers a Ref.Nr. 0325220 optimization of software as well as inves- distribution of forces over the rotor disk and tigations in case of problems and introduc- the rotation of the blades as the ALM but Duration: 07/2010 – 05/2015 tion into HPC for users. Further tasks are which is much faster. the coordination between the IT-department and ForWind, and the extension of the HPC The turbine parameterizations have been system by new or additional hardware. enhanced by turbine control features like speed and yaw control in the framework of Introduction The FLOW cluster has been running very the project CompactWind. stably and operated at full capacity for most In wind energy research computational fluid of the time. Downtimes were in most cases In 2016 (after the end of the Parallel Com- dynamics (CFD) is an essential tool to im- planned, e.g. due to the move of the cluster puting Cluster project) the ADM-R includ- prove blades and complete wind energy tur- into the new facilities of the university IT ing the turbine controller have been imple- bines, to better understand wake effects and services. In early 2015 the operational sys- mented in the default version of PALM and the interaction of wakes in wind farms. To- tem was replaced and the job management is thus available for public use [4]. day’s simulations resolve more details and systems was re-configured leading to an im- enable the coupling of large and small scale proved performance of the cluster. In the final stage of the project the work on effects. However, to perform these simula- simulating large and infinite wind farms was tions, powerful computational resources are Wake simulations with the LES model continued. A joint publication for the Wake required. PALM Conference 2015 in Visby was written and A main task of the project was the further published together with Technical Universi- For this purpose, the ForWind/Fraunhofer development of wind turbine parameteriza- ty of Denmark and Uppsala University [5]. alliance has been supplied with an own tions in the LES model PALM [2] to simu- High-Performance Computing (HPC)-facility called FLOW (Facility for Large-Scale COmputations in Wind Energy Research). The HPC-cluster is designed for highly parallelized computations and contains more than 2200 cores with a high-performance interconnect and a large storage system.

In the scientific part of the project Large- Eddy Simulations (LES) with the software PALM are performed to investigate the interaction of the atmospheric boundary layer with the wakes of wind turbines. On smaller scales the computations of loads of wind turbines will be improved by simulations Figure 1: Schematic representation of the three turbine models implemented in PALM: with the software OpenFOAM. ADM (left), ADM-R (center), ALM (right)

– 12 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201713

Furthermore, the flow over complex terrain vided with the default PALM version. It has connection between the wake meandering has been investigated, starting with a typi- been complemented by scripts that convert and the vortex shedding behind the rotor cal idealized 2D sinusoidal mountain ridge data from WRF output files to the input for- disk was found. A connection between the test case. After verification with previous mat required in PALM and that allow aver- meandering and large eddies in the back- studies and sensitivity studies with differ- aging over several WRF grid points. ground flow seems to exist but could not be ent resolutions and aspect ratios of the hill, quantified. wind turbines have been placed on the hill A further objective of the scientific work in top (see Fig. 2). the project was the investigation of coher- Simulation of the flow around ent structures in offshore wake flows which wind turbine blades To account for realistic weather situations, have a strong effect on power production Different CFD methods have been com- a one-way coupling between PALM and the and loads. Wakes simulated by LES have pared for the laminar flow around two wind mesoscale models WRF and COSMO was been analyzed by means of power spectra turbines, the NREL Phase VI turbine and a tested. The coupling method itself is pro- and autocorrelations. For a single wake no multi-megawatt turbine: The conventional RANS (Reynolds Averaged Navier Stokes) method, the time-resolving URANS (Un- steady RANS) and DDES (Delayed De- tached Eddy Simulations) which combine the advantages of RANS and LES in one simulation. It could be shown that it is possible to simulate a full turbine with DDES with results similar to the URANS model. Although URANS could yield slightly more accurate global results for power and thrust, DDES showed benefits for the local pressure coefficients. To further increase the accuracy of the simulation, the CFD solver OpenFOAM [6] was coupled to a structure solver developed at ForWind allowing to consider flexible blade geometries typical for multi megawatt turbines (see Fig. 3).

Further work was dedicated to the testing of different turbulent inflow conditions. All tested methods have their pros and cons and no method is generally superior or infe- rior. Depending on the application different methods can be recommended. Results

figure 2: Flow across a sinusoidal mountain ridge with a wind turbine placed at hill top Figure 3: Wake flow behind a deformed for three different aspect ratios of the hill and a flat reference case. Shown is the horizon- flexible blade geometry for an inflow tal wind speed (u-component in m/s). speed of 10.9 m/s

– 13 – 14WIND AS A RESSOURCE Annual Report 2015-2017

of full turbine simulations have shown that of the investigated wind turbine model. around wind turbine blades which was sim- turbulence has a strong effect on the load The turbulent inflow results show not only ulated with OpenFOAM and investigated dynamics of a turbine. However, the effect a stronger fluctuation of the loads but also in terms of aerodynamic characteristics. of turbulence depends on many parameters, a reduction of the mean driving tangential The results stress the lack of generality of e.g. the turbine design, the turbulence in- force. The axial loads which are directly re- the knowledge concerning rotational effects tensity and the dynamics of the turbulence. lated to the thrust are affected in a similar and the need for extensive research in this Thus, future work is required to investigate way (see Fig. 4). Furthermore, a stronger field using full rotor simulations. the interaction between turbulence and wind mixing of the wake could be observed in the . turbines. turbulent case leading to an early recovery of the flow. Another objective of the scientific work in this project is to gain an improved understanding of the physical mechanisms of Summary flow separation at rotor blades and toim- prove the accuracy of numerical simulations Several wind energy projects benefit from References of separated flow. Of particular relevance is the HPC cluster FLOW that was installed as [1] Peinke, J. (Hrsg.): Parallel- the flow separation in connection with radial a part of this project. The cluster is operating rechner-Cluster für CFD und WEA flows and the Himmelskamp effect which at full capacity and performs very well. Modellierung, Abschlussbericht can lead to delayed flow separation. When des Forschungsprojektes, Laufzeit: investigating the effect of blade rotation on In the scientific part of the project, -differ 01.07.2010-31.05.2015, ForWind - Carl the aerodynamic coefficients, it has been ent wind turbine parameterizations in the von Ossietzky Universität Oldenburg, found that the effects of rotation on the drag LES model PALM have been enhanced and doi: 10.2314/GBV:867679581 (2015) force are airfoil-type dependent. Thus, the compared. They are used in this and other [2] Maronga, B.; Gryschka, M.; development of a general correction mod- projects to investigate the interaction be- Heinze, R.; Hoffmann, F.; Kanani-Süh- el is challenging. Significantly improved tween the atmospheric boundary layer and ring, F.; Keck, M.; Ketelsen, K.; Letzel, M. O.; Sühring, M.; Raasch, S.: The simulation results could be obtained after wind turbine wake. The scales reach from Parallelized Large-Eddy Simulation a modification of the Spalart turbulence single wakes up to very large and infinite Model (PALM) version 4.0 for at- model [7]. wind farms. Wakes in different types of ter- mospheric and oceanic flows: model rain are simulated, from offshore and flat to formulation, recent developments, and A comparison study between laminar and forested and hilly terrain. future perspectives, Geosci. Model turbulent inflow conditions disclosed sev- Dev., 8, 2515–2551, doi:10.5194/ eral qualitative and quantitative differences The second focus of the scientific work gmd-8-2515-2015 (2015) with respect to performance and loading in this project is on the details of the flow [3] Steinfeld, G.; Witha, B.; Dörenkämper, M.; Gryschka, M.: Hochauflösende Large-Eddy Simulationen zur Untersuchung der Strömungsverhältnisse in Offshore- Windparks. promet-meteorologische Fortbildung, 39, pp. 163-180 (2015) [4] Witha. B.: Description of the PALM wind turbine model, https://palm.muk. uni-hannover.de/trac/wiki/doc/tec/ wtm, last accessed on Dec. 03 (2018) [5] Andersen, S. J.; Witha, B.; Breton, S.-P.; Sørensen, J. N.; Mikkelsen, R. F.; Ivanell, S.: Quantifying variability of Large Eddy Simulations of very large wind farms, J. Phys.: Conf. Ser., p. 625, 012027, doi: 10.1088/1742- 6596/625/1/012027 (2015) [6] http://www.openfoam.com/ [7] Spalart, P. R.; Allmaras, S. R.: A one-equation turbulence model for aerodynamics flows, La Recherche Figure 4: Turbine wake (axial velocity component) for laminar and turbulent flow along Aerospatiale, 1(1), pp. 5-21 (1994). the transverse at r/R = 1 (blade tip).

– 14 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201715

Analysis of Shadowing Effects and Wake Turbulence Characteristics of Large Offshore-Wind Farms by Comparison of alpha ventus and Riffgat (GW Wakes)

Carl von Ossietzky Universität Oldenburg, Project Description very similar. TS-X has so far only been Institute of Physics compared using local point measurements Research Group: Wind Energy Systems Measurement campaigns: or modelled wind fields. Both measuring Research Group: Turbulence, Wind Energy Two offshore measurement campaigns systems can complement each other in and Stochastics were carried out by ForWind – University the future, since TS-X measurements can Research Group: Energy Meteorology of Oldenburg as part of the GW Wakes cover very large regions, but continuous joint research project: The first in the measurements are not possible (approxi- Teklehaimanot Zere Aman, David Bastine, German offshore test field alpha ventus mately every two to three days). Lidar of- Hauke Beck, Martin Dörenkämper, and the second in the offshore wind farm fers a comparatively smaller measurement Marijn Floris van Dooren, range, but can continuously scan the region Wilm Friedrichs, Paul Gerke Funcke, Riffgat, both in the German North Sea. Detlev Heinemann, Julian Hieronimus, In addition to meteorological parameters, around the instrument with a higher accu- Martin Kühn, Michael Schmidt, the wake flow behind wind turbines in the racy. Jörge Schneemann, Gerald Steinfeld, wind farm was measured with the remote Davide Trabucchi, Juan José Trujillo, sensing method Lidar (Light detection and Shading losses: Lukas Vollmer, Stephan Voß, ranging) and the surface temperature of the Due to the special arrangement of the wind Matthias Wächter, Björn Witha sea was measured with a measuring buoy turbines in the Riffgat wind farm (3 rows to derive atmospheric stability. In alpha with 10 turbines each), conditions arise for Partner: Fraunhofer IWES ventus, the existing measurement infra- certain wind directions which are to be ex- structure of the RAVE initiative (Research pected regularly in even larger wind farms. Funding: Federal Ministry for Economic at alpha ventus) could also be used and the For example, there are wind turbines for Affairs and Energy (BMWi) lidar systems, for example, compared with northeast and southwest wind that are Ref.Nr. 0325397A the existing Fino1 met mast. shaded by up to 9 turbines. In the course of the project, however, it became appar- Duration: 08/2011 – 09/2016 In Riffgat, measurements could then be ent that the greatest power losses do not taken in a larger wind farm in which up to occur at the rearmost wind turbines, but at tenfold superimposed wakes occur. This the turbines in the middle of the wind farm. is a typical scenario for the offshore wind The reason for this is that a sufficiently farms planned and built today. large wind farm acts on the flow like an Introduction almost solid obstacle and is surrounded by The data measured within GW Wakes will an external flow. The rear turbines already The goal of the GW Wakes project was to be used in particular for the development notice the acceleration of the flow on the better understand the wake effects in large of new models of wind turbine tracking leeward side of the obstacle. offshore wind farms in order to optimize and for the validation of numerical simula- the design and operation of offshore wind tions of the maritime boundary layer and In view of the frequent situation in the farms in the future with regard to shadow the wind farm flow contained therein. The North Sea in the future that another wind losses due to wake and wind farm turbu- data will continue to provide important in- farm is located downstream, it was shown lence. sights for future research. that the Riffgat wind farm influences the wind speed far more than 10 km down- For this purpose, laser-optical measure- Furthermore, the lidar data from GW stream. ments and high-resolution numerical simu- Wakes could also be used for a comparison lations of the wind farm flow were carried with wind measurements from the radar On the basis of these findings, the potential out. Furthermore, a new model of the wake satellite TerraSAR-X (TS-X). It was shown of wind energy utilisation in the North Sea flow and a software for optimizing the lay- for the first time that the spatial structures for the planned future expansion of off- out of new wind farms were developed. in the two simultaneously measured wind shore wind farms was calculated. fields of TS-X and the lidar systems are

– 15 – 16WIND AS A RESSOURCE Annual Report 2015-2017

Figure 1: Lidar system with large measuring range on the transition piece of a wind turbine in the Riffgat offshore wind farm

Figure 2: Wind field in the Riffgat offshore wind farm. Absolute wind speed with assumed constant wind direction (arrow). The underlying lidar scan used a constant elevation of 2.5°, so that the measuring height from the center to the outside constantly increases. In the southern white sector the view was blocked for the lidar system, the northern sector was hidden because the error is too high for the calculation of the absolute wind speed. The wake of the wind turbines in the wind farm can be clearly seen, they mix behind the wind farm to form the park wake.

– 16 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201717

Comparison of flow simulation with from kilometre-long atmospheric struc- tional changes of the wake (the so-called lidar measurements tures to small-scale turbulence that extends meandering) are again caused by kilometre- The newly developed method for model into the mm range. long structures of the atmospheric wind. validation enables the directly comparison of flow measurements with high-resolution For such small scales the investigations In order to describe the overall wake, it physical models for the first time. The good showed a surprising similarity to the ide- seems appropriate to divide the models ac- agreement (see fig. 5) strengthens the fur- alized turbulence in the laboratory, so that cording to the above-mentioned orders of ther application of these models as a data strongly simplified models can be consid- magnitude. In order not to develop models basis for models for wind farm planning. ered [2]. that are too complicated, the description must always keep the application in mind. Wake turbulence: The complicated shape changes in the cross- For example, a different description will be The wake flows of wind turbines (WTGs) section of the wake flow, on the other hand, suitable for the loads on a wind turbine than are very complex. In particular, there is could well be described as superposition of for the energy yield of an entire wind farm a very large range of relevant size scales different so-called modes. The slow direc- [2].

Figure 3: Simultaneous measurement with different sensor systems: Left: Wind field measured with lidar on offshore platform, right: Wind field determined from the data of the radar satellite TerraSAR-X, the data were corrected by an offset. Regions with very similar turbulence structures are marked in black ellipses [5].

Figure 4: Flow measurement in the Riffgat wind farm with two laser-optical remote sensing devices (Windlidar)

– 17 – 18WIND AS A RESSOURCE Annual Report 2015-2017

Wake model: ried out in the GW Wakes project, require complex simulations of individual wind For a more accurate prediction of wind immense computing capacity and time. turbines are incorporated into these calcu- farm performance and for more accurate For engineering consultants and wind farm lations. They were prepared in advance and results from software used to optimize the planners, neither the corresponding com- stored as a result table. layout of offshore wind farms, within GW puter hardware nor the required time is Wakes the "3D Shear Layer Wake Model" normally available, so in practice, highly Through this novel coupling of complex was developed. This is a new wake model simplistic calculation methods and models numerical calculations and fast modelling, which, based on physical principles, is bet- are used. the results of current research in the field ter suited for the complex wind flows with- of wind turbine simulation can be used for in a wind farm than the models normally The transfer from research to applied wind wind farm planning and optimisation tasks. used in engineering software for wind farm farm planning has been realized by Fraun- For validation purposes, the results were planning. hofer IWES within GW Wakes with the compared with standard tracking models software "flapFOAM". This is a wind farm on the one hand and with RANS and LES Technology transfer: and wake modelling software that can cal- simulation results of the Riffgat wind farm Turbulence-resolving Large Eddy (LES) culate a power prediction for a wind farm on the other [3]. wind farm simulations, such as those car- with little time expenditure. The results of

Figure 5: Comparison between model and measurement for a wake flow in alpha ventus on 20 February 2014. The model explicitly takes into account the meteorological conditions measured at the neighbouring measuring mast FINO1. [Figure: Lukas Vollmer, ForWind, University of Oldenburg] Further information: [Vollmer 2015]

References

[1] Bastine, D.; Wächter, M.; [4] Schneemann, J.; Bastine, D.; [6] Vollmer, L.; van Dooren, M.; Peinke, J.; Trabucchi, D.; Kühn, M.: v. Dooren, M.; Schmidt, J.; Steinfeld, G.; Trabucchi, D.; Schneemann, J.; Stein- Characterizing Wake Turbulence with Trabucchi, D.; Trujillo, J. J.; Vollmer, L.; feld, G.; Witha, B.; Trujillo, J.; Kühn, M.: Staring Lidar Measurements Journal Kühn, M.: “GW WAKES”: Measurements First comparison of LES of an offshore of Physics: Conference Series, IOP of wake effects in alpha ventus with syn- wind turbine wake with dual-Doppler Publishing, 2015, 625, 012006 chronised long-range lidar windscanners, lidar measurements in a German [2] Bastine, D.; Witha, B.; Wächter, Proceedings of the German Wind Energy offshore wind farm Journal of Physics: M.; Peinke, J.: Towards a Simplified Confernece DEWEK, 2015 Conference Series, 2015, 625, 012001 DynamicWake Model Using POD [5] Schneemann, J.; Hieronimus, J.; Jacob- [7] Abschlussbericht: Analysis Energies, 2015, 8, 895 sen, S.; Lehner, S.; Kühn, M.: Offshore wind https://doi.org/10.2314/GBV:886719402 [3] Schmidt, J.; Stoevesandt, B. The farm flow measured by complementary impact of wake models on wind remote sensing techniques: radar satellite farm layout optimization Journal of TerraSAR-X and lidar windscanners Journal Physics: Conference Series, 2015, 625, of Physics: Conference Series, 2015, 625, 012040 012015

– 18 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201719

ClusterDesign – A Toolbox for Offshore Wind Farm Cluster Design

Carl von Ossietzky Universität Oldenburg power plant operation models, load and developed to a wind farm cluster model. Institute of Physics performance models and related wind Benchmark runs carried out within the pro- Research group Energy Meteorology, farm control concepts. ject showed that the Ainslie model in FLaP Wind Energy Systems 2. Integration: The aforementioned models can deliver results that are in good agree- were integrated into a single toolbox al- ment with much more complex, but consid- Sonja Drüke, Robert Günther, lowing for an integrated wind farm clus- erably slower wake models. Due to its low Detlev Heinemann, Michael Schmidt, Gerald Steinfeld ter design. computational demand, FLaP is well-suited 3. Validation: The single models as well as for an evaluation of a huge number of dif- Funding: EU the toolbox as a whole were validated ferent layouts of a wind farm in order to se- against data obtained from measurement lect a number of promising ones that might Ref.Nr. FP7-Energy-2011 283145 / Cluster- campaigns within the project that took be worth of further to be investigated with Design place in operating wind farms. a slower, but more accurate wake model. 4. Valorisation: In the end, the results of the Such a combination of slower and faster Duration: 12/2011 – 05/2016 project were transformed into an industri- wake models might contribute to the reduc- al applicable solution ensuring a signifi- tion of computational time required for the cant impact of the developed solutions. optimization of a wind farm layout.

Introduction Within the project a unique high-quality Project Description LiDAR data set comprising several months A consortium consisting of the six partners of measurements of the flow conditions in 3E, ECN, Imperial College, RWE Innogy, Within the ClusterDesign project the main an offshore wind farm has been generated Senvion and ForWind with its working tasks of ForWind were the further develop- by applying a nacelle-based long-range Li- groups Energy Meteorology and Wind En- ment of wake models, the generation of a DAR system of ForWind. The value of ergy Systems, collaborated from 2011 to wind atlas for the Southern North Sea as this data set has already been shown in the 2016 in the EU project ClusterDesign [1]. well as the contribution to a field experi- framework of the ClusterDesign project, as The main objective of the project was the ment in an operating offshore wind farm the wake models of the partners ECN and development of a toolbox for the integrated with a nacelle-based LiDAR. ForWind could be improved based on the design of offshore wind farm clusters. For data from this measurement campaign. that purpose, different wind farm design Aiming at the development of a high-fidel- tools were combined in one toolbox. This ity tool for the simulation of flow condi- Within the project ForWind used the con- toolbox facilitates finding solutions that tions and loads in a wind farm a coupling cept of a dynamical downscaling of reanaly- meet multiple criteria such as a maximiza- between the large-eddy simulation model sis data with the mesoscale model WRF in tion of energy extraction, a reduced uncer- PALM and the aeroelastic code Flex5 was order to generate a wind and stability atlas tainty of energy yield estimates, a maximi- established in collaboration with load ex- (WASA) for the Southern North Sea. Before zation of the ratio between income and costs perts from Senvion. information on the atmospheric conditions as well as a maximization of the power sys- in that region with a horizontal resolution tem support capability. Due to porting of the code a considerable of 2 km for the time period between 1993 speed-up of ForWind’s in-house wake mod- and 2012 could be generated, some efforts The project was based on four main ele- el FLaP could be reached. Another modi- were spent on finding the most appropriate ments: fication that has been done to FLaP in the setup of the WRF simulations. Comparisons 1. Modelling: Existing models, each aiming project was the implementation of an ana- showed that the setup finally used for the at optimizing major wind farm function- lytical wind farm model that allows for tak- WASA generation was in better agreement alities related to design and control have ing the shadowing effects of neighbouring with the measurements than previous results been further developed. The functionali- farms on a target wind farm into account. obtained with WRF and published in litera- ties included power production, loads and Therefore, FLaP can now be used to calcu- ture. A major characteristic of the WASA is power system support. The set of models late the flow conditions in and energy out- that it provides time series of wind energy consisted of advanced wake models, put of wind farms that are in the wake of relevant parameters instead of merely statis- electrical connection models, virtual other wind farms, i.e. FLaP has been further tical information. This makes the data very

– 19 – 20WIND AS A RESSOURCE Annual Report 2015-2017

Figure 1: Share of situations with strongly unstable stratification in the period between 1993 and 2012 for the region of the Wind and Stability Atlas.

easy to be handled by wake models and pre- WRF. It contains advanced and fast wind with small distances between single wind serves the observed correlation between the farm wake models, i.a. ForWind’s further turbines. With regards to the wind and sta- different atmospheric parameters. developed wake model FLaP. Moreover, a bility atlas one of the major results was that turbine load response database and electri- according to the results obtained from WRF cal collection and grid connection models the atmosphere over the North Sea shows Summary as well as wind farm control strategies are more often an unstable stratification than part of the toolbox whose IT infrastructure the atmosphere over the land. Sites with a Six partners from academia and research in- allows for an easy implementation of other high percentage of situations with unstable cluding ForWind worked together in the EU models. The toolbox offers a user-friendly stratification might be favorable for the op- project ClusterDesign. The partners aimed graphical interface and infrastructure that eration of wind farms as due to the faster at developing a toolbox for the design of allows for linking all simulation outputs and recovery of wakes a better efficiency of the offshore wind farm clusters that gives ac- presenting the simulated absolute and com- wind farm can be expected. curate projections of energy production and parative costs. fatigue-life consumption, that allows for accounting for wind shadows caused by neigh- An extensive offshore measurement cam- References bouring wind farms and that allows for con- paign comprising load, met mast and na- trolling each farm with the aim to maximize celle-based LiDAR measurements aimed at [1] Woyte, A.: Final Report Summary - the production. The final toolbox comprises providing data for model validation. Moreo- CLUSTERDESIGN (A Toolbox for a meso-scale wind and stability atlas with a ver, ECN’s Active wake control concept has Offshore Wind Farm Cluster Design), resolution of 2 km for the Southern North been tested for the first time in a real operat- https://cordis.europa.eu/result/ Sea produced by ForWind that is based ing wind farm. It turned out that this control rcn/189357_en.html (2016) on 20 years of simulations with the model concept is especially useful in wind farms

– 20 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201721

RealMe – Traceable Acquisition of Meteorological Data

Universität Bremen Project Description ability of the measured data for evaluation Bremen Institute for Metrology, by developers, operators of wind turbines Automation and Quality Science (BIMAQ) Anemometers and tail vanes are wind sen- and service providers. For compliance with sors, which are sensitive measuring instru- enhanced quality-parameters, the sets of Michael Sorg ments installed to acquire meteorological measurement-data in comprehension of all data to calculate the energy yield of wind components of the measuring chain must be Funding: Federal Ministry for Economic turbines and wind farms. Various factors upgraded with digital signatures. This also Affairs and Energy (BMWi) affect the acquisition and transmission of prevents the manipulation of measured data measurement data. Among them are the which is required to ensure a correct assess- Ref.Nr. ID: 0325468A usual influences due to weather conditions ment of the site’s wind energy potential. Duration: 09/2012 – 04/2016 but also incidents distort the measurement Furthermore, for reasons of confidential- results. Thus, research partners from Am- ity the measured data must be encrypted. monit Measurement, Adolf Thies, Deutsche The partners evaluate appropriate meth- Windguard and efm together with scientists ods to digitize, digitally sign, and encrypt of the University of Bremen initiated the re- measurement values directly in the sensor. search project in order to improve both the Microelectronic circuits will be developed reliability and the accuracy of measured me- for the integration in meteorological sen- teorological values. sors which are connected by digital bus- systems for secure data transfer (see fig. 1). Goal of the cooperative research project The development of data plausibility checks is to develop and enhance measuring sys- requires additional sensors which will be tems for meteorological data. The results added to the meteorological systems for de- will increase reliability, security and trace- tecting influences on data quality.

Figure 1: Acquisition of meteorological data with an extension of the data record with encryption and signature, and including all components of the measurement chain

– 21 – 22WIND AS A RESSOURCE Annual Report 2015-2017

WindScanner.eu – The European WindScanner Facility, Preparatory Phase

Carl von Ossietzky Universität Oldenburg Introduction Project Description Institute of Physics, Research Group: Wind Energy Systems The "WindScanner Facility" should pro- The aim of the PP was to support the con- Partners: vide a distributed research infrastructure struction of the new Windscanner facility. • DTU, Denmark (coordinator) (RI) consisting of advanced remote sens- Besides it integrates and upgrades already • CENER-CIEMAT, Spain ing devices for mainly wind energy ap- existing wind energy and wind turbine test • ECN, Netherlands plications. It would be established all over centers as well as test sites in Europe. Due • CRES, Greece Europe as a new and innovative research to the establishment of the new mobile and • Fraunhofer IWES, Germany facility. Included in the European Strategy all over Europe distributed research infra- • SINTEF, Norway Forum on Research Infrastructures (ES- structure, it will be possible to coordinate • LNEG/INETI, Portugal FRI) for renewable energies in 2011, the and upgrade major European facilities with • University of Porto, Portugal "WindScanner Facility" should be able modern Lidar-WindScanners at the opera- • IPU, Denmark to provide new knowledge about detailed tional level enhancing in this way the Euro- three-dimensional atmospheric wind flows pean competitive advantage in wind energy Duration: 10/2012 - 09/2015 leading in this way to an increased devel- systems. opment of more efficient, stronger, smarter and lighter wind turbines. The concept of the preparatory phase was the coordination of common European ac- Even though the main objective is wind tivities and European energy research or- energy research the WindScanner Facility ganizations. The network was distributed is intended to be open for researchers and between WindScanner research demonstra- users from other disciplines as well. tion nodes and partners of the European En- ergy Research Alliance (EERA). The main objective of the new facility is to become a center of excellence formed by One of the major technical and economical leading European research organizations. tasks dealt with the equipment of highly trained personnel with mobile WindScan- The preparatory phase (PP) of the project ners. dealt with all technical, legal, financial and administrative fundaments required for the Furthermore innovation generating activi- construction of the research infrastructure. ties, personal trainings and exchange pro- . grams should support the establishment of the facility.

Structure The structure of the research infrastructure shall include the following elements: Ex- pected are 6-8 associate partners with new WindScanners, designed and created during the preparatory phase, at already existing test facilities in Europe.

A large European-level coordinated experimental research programme would be based on wind energy measurement campaigns with key contributions from Wind- Scanners.

– 22 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201723

Also a comprehensive database of wind Summary data from potential wind energy sites would be made available to researchers, The research facility WindScanner.eu modelers, wind turbine manufacturers and would consist of remote sensing measure- wind energy developers. ment equipment and concepts based on portable and easy deployable wind lidars The "WindScanner.eu hub" shall be set up and wind scanners. The new measurement in order to handle data flow and host serv- technology would be distributed and oper- ers and websites etc. ated at European nodes and be connected by fast, scientific computer networks. The Objectives of the Preparatory Phase obtained results would lead to improved Project computer models on multiple scales of wind The main objectives of the preparatory energy applications and allow an optimiza- phase were: tion of wind turbine designs from the blade to the whole wind farm. In conclusion, it • Establishment of the project manage- would lead to better located as well as bet- ment and logistics ter wind turbines thus reducing the cost of • Development of a financial scheme renewable energy. The preparatory phase of • Development and agreement on the the windscanner.eu facility project started in legal framework required for the estab- November 2012 and ended in 2015. lishment and operation of the facility • Draft planning of the scientific and technological work • Realization of the implementation steps for the disseminated regional Wind- Scanners • A joint planning for the regional expansion of WindScanner facilities

References

http://www.windscanner.eu.

– 23 – 24WIND AS A RESSOURCE Annual Report 2015-2017

European High-performance Infrastructures in Turbulence (EuHIT)

Carl von Ossietzky Universität Oldenburg ries, in order to significantly advance the Project Description Institute of Physics competitive edge of European turbulence research with special focus on providing the The work package 19 includes new sen- Research Group: Turbulence, Wind Energy knowledge for technological innovation and sor developments and was processed by 3 and Stochastics for addressing grand societal challenges. partners within the EuHIT consortium (Uni- versity of Oldenburg, SmartInst and Vision Jaroslaw Puczylowski, Joachim Peinke, Current members of EuHIT include 25 Michael Hölling Research Europe B.V.). The task of the Uni- research institutes and 2 industrial part- versity of Oldenburg was to further improve Funding: European Community ners from 10 European countries. Total the performance and reliability of the 2D- Framework Programme 7 14 cutting-edge turbulence research infra- LCA. For this purpose some requirements structures, most of which are developed on were defined. In particular, the 2D-LCA Ref.Nr. 312778 national funds and are run by EuHIT mem- should be able to operate on spatial scales ber institutes, consist the material basis of of about 100 μm and at a frequency of up to Duration: 04/2013 – 04/2017 EuHIT. A Networking Program and Joint 100 kHz and also become operable by inex- Research Activities within EuHIT intercon- perienced researchers. nect these infrastructures, together with the knowledge developed upon them. Besides modifications to the structural components, new sensing elements (canti- The main components of EuHIT may be levers) with increased directional sensitiv- summarized briefly as: ity have been developed. In contrast to the previous cantilevers, the new designs have a Introduction Integration: Not only the research activi- "stepped" vane that is aerodynamically opti- ties at the infrastructures, but also the train- mized and thus results in a better resolution EuHIT was created as advances in key ing programs, the sharing of data, and the of small angles of attack. Fig. 1 show the economical and societal issues facing Eu- interaction with numerical modeling and final design. rope, like ground-, air- and sea-transport, theoretical analyses are all integrated un- energy generation and delivery, processing der EuHIT. This facilitates access to the The conversion of the previous 2D-LCA to a in chemical industries, marine biosphere infrastructures and easy exchange of instru- more functional and user-friendly design has management, climate change impact, at- ments, techniques, data, and new ideas. been concluded successfully. Fig. 2 shows mospheric and marine pollution prediction, the new design of the 2D-LCA. The main and carbon capture and storage processes Innovation: New techniques, algorithms, amendments to the previous version are the are obstructed by the lack of understanding and next generation instruments are de- increased numbers of degrees of freedom of turbulence. The reason lies in the fact that veloped through Joint Research Activities (slide #10 for adjustment of the position turbulent flows underlie all macroscopic (JRAs) to maintain the EuHIT infrastruc- sensitive detector (PSD) and mount #10 for natural and technological flows as soon as tures at the leading edge. position adjustment of the cantilever) and mass transport is large. In the past 10 years the laser beam guidance (detailed illustra- Europe has surpassed all other nations in re- Dissemination: New tools and procedures tion in fig. 3). The adjustment possibilities search output and development of national are implemented to foster easy and open ac- allow the user to precisely align the laser infrastructures in turbulence research and its cess to data and techniques developed from onto the active area of the PSD. The modi- applications. However, due to a lack of Eu- EuHIT, including the transfer of knowledge fied laser path (see fig. 3) reduces ghosting ropean wide integration it has not reached from academia to industries. by means of a thin beam splitter plate and its full potential for innovation in the scienc- therefore leads to an improved signal qual- es and technologies. EuHIT was thus born The University of Oldenburg successfully ity. Other modifications involve improved to overcome these limitations. contributed in the EuHIT project by further electronics, cable connections and the sig- developing the so-called 2D-LCA (2D-La- nal-processing unit. EuHIT is a consortium that aims at inte- ser Cantilever Anemometer), which is a new grating cutting-edge European facilities for highly resolving anemometer particularly In order to improve the mobility of the turbulence research across national bounda- designed for turbulence research. whole system a portable calibration unit has

– 24 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201725

Figure 1: First realization of a cantilever with a “stepped” vane (a) and the final version (b)

Figure 2: Design of the 2D-LCA

been designed but not yet fully tested. In addition, based on the design of new cantilevers further cantilevers for other velocity ranges were manufactured and tested.

Summary

Within the Euhit project the further development of the 2D-LCA could be advanced considerably. The biggest achievement was the development of the cantilever, which significantly increases the performance of the sensor. Further design improvements, which led to easier handling and better signal quality, were also successfully implemented. Figure 3: Design of the laser beam guidance.

– 25 – 26WIND AS A RESSOURCE Annual Report 2015-2017

PhD Programme on “System Integration of Renewable Energies”

Carl von Ossietzky Universität Oldenburg Project Description mance and realization of the designed Institute of Physics, controller, under detailed time-varying Research Group: Wind Energy Systems A joint research cooperation is established interactions with boundary layer [6]. with Delft Center of Systems and Control Martin Kühn (Supervisor) (DCSC), TU Delft for investigating the Adjoint-based model predictive control Mehdi Vali (Candidate) closed-loop wind farm control in order to (AMPC) PhD project title: Wind farm control minimize the undesirable effects of wake The idea to maximize the power produc- framework for performance optimization interactions. Fig. 1 depicts the control archi- tion of wind farms in the presence of wakes with respect to dynamic flow conditions tecture of our investigated optimal control is to coordinate the control settings of in- framework for waked wind farms, consist- dividual turbines, by taking their wake in- Funding: Ministry for Sciences and ing of the following four main elements: teractions into account. Furthermore, in or- Culture of the Federal State of Lower der to balance power supply with demand, Saxony 1) a simplified model of wind farm to cap- a wind farm plant should respond to grid ture the dominant waked inflow dynam- requirements through control of its power Duration: 12/2013 – 12/2016 ics [2], production, the so-called active power 2) an adjoint-based model predictive con- control (APC). The non-unique solution trol (AMPC) to optimally adjust the of APC for wind farms is a resulting by- wind turbines degrees of freedom, e.g., product which can be exploited for optimal the axial induction factor and yaw angle power/load distributions among the wind [3], [4], turbines. 3) a state observer for flow field estimation Introduction against uncertainties and mismatches A model predictive control (MPC) has [5], been developed in order to optimally con- Control of turbines in a wind farm is chal- 4) a large-eddy simulation (LES) model of trol of wakes interactions with time-var- lenging because of the aerodynamic in- the wind farm in order to test the perfor- ying atmospheric and operational condi- teractions via wakes. The characteristics of a wake are reduced wind speed and increased turbulence. The former reduces the total power production of the farm and TSO

the latter leads to a higher dynamic load- dem Pk∣k+ N Adjoint MPC ing on the downstream turbines. In today p commercial wind farms, wind turbines are U * k∣k+ N u regulated greedy. By changing operating Optimizer points of upwind turbines, it is possible to * * ... * u1, k u2,k uN ,k influence the wake interactions, and thus t the performance of downwind turbines, WF plant possibly increasing overall power cap- Constraint Perf. Index ture or decreasing structural fatigue loads. Therefore, wind farm control has recently received significant attention to lower the y y ... y 1,k 2,k N t , k U k∣k+ N levelized cost of energy [1]. The inherent X^ p k∣k+ N p modeling uncertainties and time-varying Simplified Y X^ k inflow conditions, e.g., turbulent nature of Model k State observer wind, wind direction changes, and wake meandering demand for a closed-loop control approach to react fast enough against Figure 1: Schematic illustration of the closed-loop optimal control of wind farms. The sources of the wake interactions within a grey block contains the main components of the adjoint-based model predictive control wind farm. (AMPC).

– 26 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201727

tions. It relies on a 2D dynamic wind farm performance of interest, subjected to dis- wind farms. Fig. 2 plots the instantaneous model, the so-called WFSim [2], which turbances and uncertainties, with the same fields of the u-components of the wind at predicts the dominant inflows and wake computational cost of the model. Indeed, hub height of the wind turbines simulated interactions in a reasonably computation- adjoint variable associates the variations with both models. The total power produc- ally inexpensive manner. A constrained of the model and the cost function [3],[4]. tion of the wind farm and the wake impacts optimization problem is formulated with are well estimated using WFSim, which is respect to the wind turbines control inputs. Investigation of concepts a medium-fidelity wind farm model devel- The systematic structure of the optimal For practical perspectives of the approach, oped for closed-loop control of wind farms control problem allows us to consider it for it is demonstrated that WFSim is able to [2]. different wind farm control objectives. An capture the dominant dynamics of the re- adjoint approach is developed as a cost-ef- solved turbulent wakes of the PArallelized Fig. 3 demonstrates the performance of the fective tool to compute the gradient of the Large-eddy simulation Model (PALM) of adjoint MPC for power maximization and

2.7 U PALM 1 3 5 WFSim 2.4

2 4 6

2.1

1.8 Wind farm [-] U P C 1 3 5

1.5

2 4 6

1.2 100 300 500 700 900 1100 1300 time [sec] Figure 2: (left) The layout of the 2x3 wind farm example simulated with PALM (upper) and WFSim (lower). (right) The total power coefficient of the wind farm simulated with WFSim, compared with PALM.

1.05 Optimal ref. 1.3 Locally greedy AMPC (N T =200s) P s 1

1.2 0.95

0.9 1.1

0.85 Normalized total power [-] 1 Normalized total power [-] Ref 0.8 AMPC: RMS(e)=0.005 ° ° BaseL1: RMS(e)=0.091 8 yaw mis-alignment 0 yaw mis-alignment BaseL2: RMS(e)=0.016 0.9 0.75 600 800 1000 1200 1400 1600 1800 600 800 1000 1200 1400 1600 1800 time [sec] time [sec] Figure 3: AMPC performance for optimal control of the waked wind farm. (left) The maximized power production when wind direction changes. (right) Active power control for tracking a power reference demanded by the TSO.

– 27 – 28WIND AS A RESSOURCE Annual Report 2015-2017

providing active power services of waked Summary wind farms for optimally adjusting the wind turbine control inputs in order to real- An adjoint-based model predictive control ize wind farm control objectives, subjected (AMPC) has been developed for optimal to time varying atmospheric and opera- energy extraction of wind farms, consid- tional conditions. The left plot compares ering the aerodynamic wake interactions. the performance of the AMPC with the lo- The closed-loop wind farm control frame- cally greedy control with dynamic changes work relies on a medium fidelity wind farm in wind direction. The AMPC adjusts the model, the so-called WFSim, to predict the wind turbine control inputs to the corre- dominant wake dynamics in an inexpen- sponding optimal settings using feedbacks sive computational manner. AMPC ben- and thus the wind farm operates at optimal efits from an adjoint approach as a compu- points. The right plot illustrates the capa- tationally-efficient tool for computing the bility of the AMPC for optimally power gradient. Both features are computation- distributing among the individual wind tur- ally desirable for real-time control imple- bines while their total power productions mentations. The accuracy and reliability of follow a time-varying power reference, de- the model for computing the optimal tra- manded by the TSO (transmission system jectories are validated using an LES model References operator). Contrary to the baseline cases, of wind farm. The systematic formulation including a central open-loop control sys- of the optimal control problem make the [1] Knudsen, T.; Bak, T.; tem with two different power set-point AMPC compatible for different control ob- Svenstrup, M.: Survey of wind farm distributions, the AMPC improves signifi- jectives of waked wind farms, e.g., power control-power and fatigue cantly the RMS (root mean square) of the maximization, active power control, and optimization. Wind Energ., 18: power tracking error using feedbacks. optimal load distribution. pp. 1333–1351. (2015) [2] Boersma, S.; Doekemeijer, B.; Vali, M.; Meyers, J.; van Wingerden, J.-W.: A control- oriented dynamic wind farm model: WFSim, Wind Energy Science 3 (1) pp. 75-95 (2018). [3] Vali, M.; Petrovic, V.; Boersma, S.; van Wingerden, J.-W.; Kühn, M.: Adjoint-based model predictive control of wind farms: Beyond the quasi steady-state power maximization, in IFAC-PapersOnLine, vol. 50, no. 1, pp. 4510-4515, July 2017. [4] Vali, M.; Petrovic, V.; Boersma, S.; van Wingerden, J.-W.; L.Y. Pao; Kühn, M.: Model predictive active power control of waked wind farms, in American Control Conference, pp. 707-714, June 2018. [5] Doekemeijer, B.; Boersma, S.; L.Y. Pao; van Wingerden, J.-W.: Ensembel Kalman filtering for wind field estimation in wind farms, in American Control Conference, pp. 19-24, May 2017. [6] Vali, M.; Vollmer, L.; Petrovic, V.; Kühn, M.: A closed-loop wind farm control framework for maximization of wind farm power production, in 1st Wind Energy Science Conference (WESC2017), June 2017.

– 28 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201729

Turbulence Resolving Simulations for Meteorological Assessment of Wind Energy Sites in Complex Terrain

Leibniz Universität Hannover Introduction tested in an industrial environment for real Institute of Meteorology and Climatology, complex terrains, including vegetation and PALM research group The project is part of a larger joint research steep topologies, and is validated with field project coordinated by the German wind measurements from existing wind turbine Lennart Böske, Christoph Knigge, Sieg- sites. One major goal is to demonstrate the fried Raasch turbine manufacturer ENERCON. The main goal is to further develop new high applicability of the simulation tool in an industrial environment. Funding: Federal Ministry for Economic resolution simulation techniques from ba- Affairs and Energy (BMWi) sic and academic research in order to apply them for load and power forecasts of wind Ref.Nr. 0325719C turbines in complex terrain. The turbu- Project Description lent wind field is simulated with the large- Duration: 11/2014 – 10/2017 eddy simulation (LES) technique, where Wind energy sites in complex terrain are of- the large, main energy containing eddies ten exposed to very turbulent atmospheric are explicitly resolved by the numerical flow fields which generate heavy mechani- grid, while only the small scale turbulence cal loads on the rotor blades. The rapidly is parameterized. The numerical solver is changing loads may significantly reduce

Figure 1: Wind energy site in complex terrain (Source: Courtesy of ENERCON, Germany)

– 29 – 30WIND AS A RESSOURCE Annual Report 2015-2017

lifetime of the systems. Current simulation grid spacing, that minimize the computa- techniques used for site assessment studies tional demands of the simulations but guar- (Reynolds-Averaged Navier-Stokes models, antee sufficient accuracy of the results. RANS) only predict the mean flow features and can not account for turbulent fluctuations. Therefore, the project aims to apply an LES model for simulating the turbulent Summary atmospheric boundary layer flow above complex terrain with very high spatial The project aimed to demonstrate that tur- resolution down to a few meters. The high bulence resolving simulations of atmos- resolution data are used by project partners pheric flows are ready to be used for op- from the University of Stuttgart (Institut für erational site assessment studies in complex Aerodynamik und Gasdynamik, IAG) and terrain. High resolution simulations have the DLR Braunschweig (Institut für Aero- been carried out with the LES model PALM dynamik und Strömungstechnik) as input for existing sites and compared with field to engineering CFD models which calculate observations. the near field flow around the blades in -or der to determine mechanical loads.

Because of its high computational demands, the LES technique has been used mainly for basic research in the past, but the rapidly increasing power of computer systems nowadays allow to use it also for applied industrial applications. The PArallelized LES Model PALM [1] is the central tool for simulating the atmospheric turbulence in this project. It has been developed at IMUK, has a very high computational performance and is highly optimized to be used on massively parallel computer architectures. Such computer clusters are meanwhile operated even by medium-sized companies, demonstrat- ing that LES has or will become an option e.g. for site assessment studies.

The PALM code is also used by other For- Wind partners, mainly at the University of Oldenburg (Institute of Physics, Research Group Energy Meteorology).

The project has been finished in October References 2017. Model results for simple idealized topographies (e.g. Gaussian hills) have been [1] Maronga, B., Gryschka, M., Heinze, compared and validated with wind tunnel R., Hoffmann, F., Kanani-Sühring, F., data. Final simulations have been carried Keck, M., Ketelsen, K., Letzel, M. O., out for existing sites with very complex Sühring, M., and S. Raasch (2015): The Parallelized Large-Eddy Simulation Mo- topography like shown in Figure 1 and del (PALM) version 4.0 for Atmospheric compared with field measurement data. It and Oceanic Flows: Model Formulation, has been demonstrated that LES can be ap- Recent Developments, and Future Per- plied operationally for site assessment stud- spectives, Geosci. Model Dev. Discuss., ies. Therefore, sensitivity studies have been 8, 1539-1637 DOI: 10.5194/gmdd-8- done in order to determine optimum values, 1539-2015 e.g. for the computational domain size and

– 30 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201731

German Research Facility for Wind Energy (DFWind)

Carl von Ossietzky Universität Oldenburg to do research in the two multi-MW wind 3. First steps toward the development of a Institute of Physics, turbines in the windfarm. Moreover, phase soft- and hardware platform for testing of Research Group: Wind Energy systems 1 of the project aims at preparing the instal- experimental control concepts for wind (WESys), lation and the instrumentation of the two turbines and wind farms that might be Ressearch Group: Energy Meteorology wind turbines. Another objective of phase relevant to security. (EnMet), 1 is the execution of meteorological meas- 4. The further development of monitoring Research Group: Turbulence, Wind Energy urements in order to get an information on methods for wind turbines during meas- and Stochastics (TWiSt) the flow conditions at the site of the wind urement campaigns with changed opera- Martin Kühn, Detlev Heinemann, farm before the installation of the two wind tional parameters. Joachim Peinke, Martin Kraft, turbines. 5. The specification of a data acquisition and Gerald Steinfeld, Michael Hölling, a data management system for the utiliza- Vlaho Petrovic, Dominik Traphan, tion in the research wind farm. Abdulkarim Abdulrazek, Agnieszka Höllling, Tim Homeyer, Project Description Phase 1 of DFWind is structured in five Hauke Wurps, Agnieszka Hölling, work packages. Work package AP0 com- Leo Höning, Dominik Traphan, The contributions of ForWind-OL to prises the coordination activities for the Luis Vera-Tudela, Janna Seifert, DFWind can be attributed to the field of whole project as well as for the four sci- Marijn van Dooren, Róbert Ungurán wind physics. The researchers from Olden- entific work packages. Work package AP1 burg are aiming at providing the necessary deals with the complete wind turbine, while Funding: Federal Ministry for Economic Affairs and Energy (BMWi) measurement facilities that are required to work package AP2 aims at implement- get an improved understanding of the turing the instrumentation that is required for Ref.Nr. 0325936C bulent flow in the atmospheric boundary meteorological and acoustic measurements layer as well as the aerodynamic loads felt at the site of the wind farm. Work package Duration: 01/2016 – 12/2020 by the wind turbines. The objective is to AP3 deals with the preparation of measure- obtain data with an unprecedented degree ment technique for the rotor blade, while of spatial and temporal resolution. Another work package AP4 aims at preparing the goal of the researchers at ForWind-OL is the research wind farm for research concerning development of new control methods of the the supporting structure and geotechnics. wind farm with the aim to reduce loads, in- ForWind-OL is involved in the coordination Introduction crease power output and reliability of wind work package as well as the first three scien- turbines. These methods shall in later pro- tific work packages of the project. DFWind is a joint research project that jects be tested on the wind turbines in the is coordinated by the German Aerospace research wind farm. The following paragraphs summarizes the Center (DLR). Besides Carl von Ossietzky major contributions by ForWind-OL to the Universität Oldenburg (ForWind-OL) the The main objectives of the subproject of DFWind project in the years 2016 and 2017. further partners in the project are Leibniz ForWind-OL in phase 1 of DFWind can be Universität Hannover (ForWind-H), Uni- summarized as follows: WESys: versität Bremen (ForWind-HB) as well as In AP1 WESys supports DLR in the de- the Fraunhofer Institute for Wind Energy 1. The development of a concept and par- velopment of an "Advanced Protection and Energy System Technology (IWES). tial implementation of infrastructure for System" which shall assure the safe op- measurements of the meteorological am- eration of experimental plant controllers. The objective of phase 1 of the project bient conditions in the wind farm with WESys develops a LiDAR-based system is to deliver the basic infrastructure at the high spatial and temporal resolution. to measure the inflow wind speed for real- research wind farm that is funded by the 2. First steps towards the development of an time prognosis of the transient loads. For German federal state of Lower Saxony in aerodynamic measurement system for the this purpose, the use of a continuous-wave PROwind project. DFWind will include local inflow, the circulation and the wake LiDAR system, the so-called Spinner-Li- additional equipment such as sensors or a of rotor blade segments delivering data DAR is planned. It is placed in the wind data management system that is required with high resolution. turbine in the rotating spinner. The system

– 31 – 32WIND AS A RESSOURCE Annual Report 2015-2017

Figure 1: Free field test of short-range LiDAR System on top of Nacelle wind turbine AV04

measures the incoming wind on a spheri- Acquisition (SCADA) data from two wind is enough to allow predicting the fatigue cal surface at a focal distance of 10-150 m turbines from the Baltic I wind farm. Neu- loads of a wind turbine. at an opening angle of ± 30° in a temporal ral networks were designed to estimate the resolution of 400 points per second. First fatigue loads on the blade root. The most In AP2 WESys is responsible for the meas- models for forecasting the incoming wind suitable input variables for the estimation, urement of the near rotor inflow and down- field were created. The "Spinner-LiDAR" i.e. the most important statistical values of flow around the research turbines with rel- was successfully tested on the nacelle roof different sensors were identified. Depend- evant temporal and spatial resolution with on AV04 in the Alpha Ventus offshore wind ing on the location, wake situation or free LiDAR systems. Two coordinated ground farm as well as in the rotating spinner of an flow, the selection of the most suitable sig- based continuous-wave short-range Li- eno114 plant in the Brusow wind farm in nals was different. A generic model was DARs (WindScanner) will measure the open field trials. In comparison, the clearly created, considering both wind turbines. wind vectors with high resolution due to superior information content was shown Subsequently, the selection was updated the 5” prism of the systems. A first WindS- when placing the device in the rotating in case of failure of individual sensors or canner has already been delivered by Dan- spinner with a clear view to the front (see the lack of individual statistics. The re- ish Technical University. It has been tested fig. 1). sults showed that no additional statistics in the large wind tunnel of FW-OL in No- were chosen to replace the missing sensor vember / December 2017 with a focus dis- For the research wind turbines in the case, while single missing statistics were tance of approx. 30 m and showed already DFWind project, knowledge of the fatigue replaced by their closest complements. In a good resolution. [1-4]. loads sustained is relevant. The knowledge both cases a comparably good estimation is required regardless of currently avail- accuracy was achieved. This leads to the EnMet: able or possibly disturbed sensors. To en- conclusion that, for an accurate load esti- In AP1 the Energy Meteorology group was sure a reliable determination of the fatigue mate, no additional signals are needed be- involved in the process of defining the re- loads, investigations were carried out with sides the ones chosen. For normal operat- quirements for the data management sys- 10-minute Supervisory Control and Data ing conditions, the use of standard signals tem for the research wind farm. The data

– 32 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201733

management system must fulfil two main Moreover, several sensors have been ing wind field for the other wind turbine. tasks: Collecting all the measured data bought and put into operation in the labo- The arrangement of three met masts is in from the wind farm and making the data ratory during the first two years of the pro- such a way that the booms span a plane in available to users that have the permission ject. Two eddy-covariance systems have which the anemometers are located. The to access the data. The main contribution extensively been used during a measure- individual positions and the distance be- of the Energy Meteorology group was the ment campaign at the site of a wind farm in tween the cup- and ultrasonic-anemome- preparation and evaluation of a survey in Northern Germany. The objective of these ters should cover a wide range of scales in which the future users of the research wind measurements was to gain information on order to be able to investigate e.g. correla- farm were asked for their preferences and atmospheric stability at the site in order to tions within the wind field. Data analysis of requirements concerning the data manage- analyse the dependency of the wind tur- free field measurements of an offshore met ment system. Moreover, staff from the En- bine behaviour from atmospheric stability. mast shows that flow structures up to a few ergy Meteorology group accompanied the A major outcome of the work with the data hundred meters are self-similar. This im- process of the acquisition of a data man- from the eddy covariance stations at the plies that anemometers do not have to cov- agement system as a representative for the wind farm in Northern Germany was the er the area with the highest possible spatial future users of that system by taking part development of a postprocessing proce- resolution but can be positioned in a non- in regular meetings with the provider of dure for eddy covariance data that will fur- equidistant manner without losing infor- that data management system during the ther be used in the course of the DFWind mation. Fig. 3 shows a sketch of the three implementation phase. Furthermore, staff project. met masts with calculated positions for the from the Energy Meteorology group col- anemometers. In this concept the non-equi- lected data from sensors that were already TWiSt: distant distribution is applied in the verti- available at the project partners in order to Within DFWind, the research group TWiSt cal and horizontal direction for a few rows, provide it for early tests of the data man- is in AP2 responsible for the conceptional and columns respectively. A regular grid- agement system. design of the so-called met-mast array po- distribution has also been added in order sitioned between the two wind turbines at to provide data that can be easily used for In AP2 a major contribution of the Energy the research facility. The array will allow computer simulations. Besides the concep- Meteorology group was to develop a plan for measuring the wind field with high spa- tional design of the met-mast array TWiSt for the measurements and the layout of the tial and temporal resolution in either the investigated possible boundary conditions IECplus met mast. wake of the one turbine or the free incom- that might influence the measurements at

Figure 2: CAD technical drawing of a rotor blade segment with the aerodynamic glove (gray). The pressure taps are located in the green stripe where the microphones are distributed left of them. The 5-hole probe is located outside of the glove and will be mounted at a hard point on the blade directly.

– 33 – 34WIND AS A RESSOURCE Annual Report 2015-2017

the array e.g. blockage of the wind turbine operation with the DLR in Braunschweig is Summary located in the wake of the array. This topic involved in the design and development of was addressed in computer simulation as an aerodynamic glove allowing measuring Within the DFWind project ForWind Old- well as in wind tunnel experiments. the local flow conditions. Fig. 2 shows a enburg cooperates with partners to develop technical drawing with the basic design of an infrastructure that help researches to Knowledge of the local aerodynamic at the glove and the position of the different better understand the physics of wind. In the rotating rotor blade is very important sensors. Here, the 5-hole probe will meas- the first years of the project main compo- in order to better understand the complete ure the local flow velocity as well as the nents have been defined. This includes the performance of the wind turbine. In AP3 local angle of attack for the blade segment. unique characteristic met mast to measure Computer simulations are based on models Pressure taps in the green belt can measure the met mast array with several masts that that try to capture aerodynamic effect that the absolute pressure distributed along the will measure the turbulent wind field with can occur on rotor blades. For validating span of the profile where additional micro- high accuracy as well as the equipment to these models, the local inflow conditions phones will give access to the fluctuations characterize the atmospheric stability. The as well as the flow around the blade need to identify e.g. flow separation. methods to use the measuring devices have to be measured in the rotating system in the been improved. Some components could free field. Therefore, in AP3 TWiSt in co- already be delivered, as the first 5" cw- WindScanner, and tested in free field as the cw-Spinner-LiDAR.

References

[1] van Dooren, M. et al., Demonstra- tion of synchronised scanning Lidar measurements of 2D velocity fields in a boundary-layer wind tunnel, in Journal of Physics: Conference Series, 2016, vol. 753, no. 7, p. 072032: IOP Publishing. [2] Seifert, J.; Vera-Tudela, L.; Kühn, M.: Training requirements of a neural network used for fatigue load estimation of offshore wind turbines, Energy Procedia, vol. 137, pp. 315-322, 2017/10/01/ 2017. [3] Vera-Tudela, L.; Kraft, M.; Kühn, M.; How do missing data affect your data-driven model? Monitoring fatigue loads with SCADA, presented at the Offshore Windenergy 2017, London, UK, 2017. [4] Vera-Tudela L.; Kühn, M.: Analysing wind turbine fatigue load prediction: The impact of wind farm Figure 3: Sketch of met-mast array with three masts. The large circle represents the flow conditions, Renewable Energy, swept area of the wind turbines at the research facility. The colored symbols represent 2017. possible anemometer positions based on the calculations performed.

– 34 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201735

NEWA – New European Wind Atlas Joint Programme

Carl von Ossietzky Universität Oldenburg suitability based on a mesoscale to mi- its development. Thus, an updated version Research group: Wind Energy croscale model chain. with state-of-the-art mesoscale models, Meteorology • To create and publish a European Wind parameterization schemes and uncertainty Atlas database publicly accessible quantifications build a solid foundation for Björn Witha, Gerald Steinfeld, Stephan through a web interface. It will include the successful subsequent application of Voss, Jörge Schneemann a high-resolution mesoscale wind atlas the microscale downscaling. All this makes (3 km), as base of a microscale wind the New European Wind Atlas an invalu- Partners: 30 institutions from 8 countries able product for the wind energy industry, (Belgium, Denmark, Germany, Latvia, atlas (50 m) and joint associated uncer- Portugal, Spain, Sweden, Turkey), tainties. more so in the context of a shift in the en- coordinator: Technical University of • To define a verification and validation ergy production paradigm aimed to comply Denmark (DTU) framework for the model-chain based on with the Paris agreements. The information the experimental campaigns and means provided by this new atlas is necessary for Funding: Federal Ministry for Economic to quantify the uncertainties of the wind a realistic planning phase of wind energy Affairs and Energy (BMWi) atlas. development, which can last several years from strategic spatial planning, to site pros- Ref.Nr. 0325832B The last effort of a European wide wind at- pecting, to wind farm design and financing. las was conducted almost 30 years ago [3]. Detailed and robust information about the The region covered by this atlas is smaller, wind resource across an area is crucial for Duration: 03/2015 – 02/2019 the spatial resolution is lower and addition- the commercial evaluation of a wind farm ally, no mesoscale models were used for [4].

Introduction

The New European Wind Atlas (NEWA[1, 2]) Project is funded under the European Commission's 7th Framework Programme ERA-Net Plus that comprises nine funding agencies from eight EU Member States. The project aims to create a new and detailed European wind resource atlas, as well as collecting data from field experiments to generate a high-fidelity database of wind characteristics. The NEWA project is developing a new reference methodology for wind resource assessment and wind turbine site suitability based on a mesoscale-to-microscale model-chain. This new methodology will produce a more reliable wind characterization than current models, leading to a significant reduction of uncertainties on wind energy production and wind conditions that affect the design of wind turbines.

The overall objectives of NEWA are: Figure 1: Spatial coverage of the New European Wind Atlas, covering the EU, Turkey, • To develop a high-value data base of Norway and extending 100 km offshore. The location of high fidelity experiments are high-fidelity experiments. indicated by red dots and a line for a ferry-based profiling transect in the Baltic Sea. • To develop methodologies for wind re- Participating countries in the ERA-Net Plus project are in green: Belgium, Denmark, source assessment and wind turbine site Germany, Latvia, Portugal, Spain, Sweden and Turkey.

– 35 – 36WIND AS A RESSOURCE Annual Report 2015-2017

Project Description uring heights between 65 m and 275 m) in reference flow model that can link mes- the Southern Baltic was measured. While oscale and microscale wind climate condi- The project is divided into three main work mainly conducted by our project partner tions consistently. packages covering the field experiments, Fraunhofer IWES, ForWind was involved the development of a model chain and the in the planning of the experiment and in the ForWind is significantly involved in the wind atlas database. discussion of the results. preparation and implementation of the mesoscale wind atlas. Extensive studies Field experiments Further experiments conducted during with the mesoscale model WRF have been Within the NEWA project a number of full- NEWA but without contribution from For- performed in which different parameter scale field experiments have been prepared Wind: and model settings have been tested and and executed [5]. These include measure- compared. The primary goal of these tests ment campaigns over complex terrain as • RUNE: measuring the near shore flow is to find an optimal setup for the wind at- well as over flat terrain and offshore. For- with scanning lidars at Danish west las production runs. Wind has been involved in three of these coast experiments. Data from the experiments • Østerild: Measuring the flow in hub A further main task in NEWA is to develop will be released in a publicly accessible height over flat terrain in Northern Den- a probabilistic wind atlas which shall add database. mark with two horizontally scanning uncertainty information to NEWA. A key lidars mounted on tall masts component of the probabilistic wind atlas The Kassel forested hill experiment took • Hornamossen: Measuring the flow over is a multi-physics ensemble of WRF model place between October 2016 and January undulating, forested terrain in Sweden runs with different parameter settings, e.g. 2017 with accompanying long-term meas- with a 180 m met mast, ceilometers, li- different parameterizations of the atmos- urements from two tall met masts (140 m dar profilers and sodars pheric boundary layer and different forcing and 200 m) between November 2016 and • Alaiz: Flow over very complex terrain data. October 2017. During the extensive cam- in Northern Spain paign 9 long-range Doppler scanning li- To be able to assess the quality of certain dars and 6 vertical wind profilers (2 lidars, Model chain model runs, it is inevitable to compare the 4 sodars) were installed on a forested hill The NEWA model chain consists of two simulated results to measurements. For this site (Rödeser Berg) near Kassel. ForWind parts: the mesoscale model WRF and the purpose data from tall met masts (50 m – contributed with two scanning lidars. The microscale CFD model based on Open- 300 m height) was collected. An example experiment was optimized to measure the FOAM. The NEWA model chain will be of such a comparison to measurement data flow along the main wind direction (SW open to the public domain and serve as a can be seen in Fig. 2. towards NE) over the hill.

The Perdigao experiment is by far the largest of the NEWA experiments [6]. Al- together 55 met masts (between 2 m and 100 m height), 22 scanning lidars, 7 profiling lidars and several further instruments including radiosondes, radars and sodars have been installed on the site, a nearly perfectly aligned double ridge, in central Portugal. ForWind contributed with two scanning lidars. During 7 months, between December 2016 and July 2017, they have been measuring the flow over the double ridge.

In the FerryLidar experiment a Doppler lidar was placed on a ferry boat travelling on a regular route through the southern Baltic Sea between Kiel (Germany) and Klaipeda (Lithuania). Between February and June 2017 the vertical profile of the Figure 2: Comparison of simulated WRF data to measurement data for several stations. marine atmospheric boundary layer (meas- Courtesy of A. Hahmann (DTU)

– 36 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201737

A further task was to implement the analog stability, wind power) in several heights Summary ensemble method, previously used for between 50 and 500 m as well as statisti- wind forecasting, for wind resource assess- cal properties (e.g. wind speed distribution, A New European Wind Atlas (NEWA) is ment following the method of Vanvyve et wind rose) for each location in Europe on a being generated, open to the public and al. (2015)[7]. It is planned to use the analog 3 km x 3 km grid. based on mesoscale simulations with 3 km ensemble for the quantification of uncer- resolution which are statistically down- tainties in the mesoscale modelling. The microscale wind atlas will display the scaled to 50 m resolution. The wind atlas statistical properties on a 50 m x 50 m grid. will furthermore include information on ex- In the microscale modelling part of this The wind atlas will be open to the public. treme winds and uncertainties. Several large work package, ForWind has participated in End-users can view some of the data via a field experiments have been conducted in the Ryningsnäs benchmark with the LES web-interface and select data to be down- the framework of the NEWA project to be model PALM. A forested site in Sweden loaded from the NEWA database. used for model validation. was simulated and compared with measurements. Various domain configurations for the mesoscale wind atlas have been tested, and fi- Wind atlas database nally 10 domains over Europe have been A survey on stakeholders’ expectations to defined (see Fig. 3). NEWA has been conducted among potential end-users in the wind energy commu- For the final production runs and the multi- References nity. Based on the results of this survey, the physics ensemble an application request- final output parameters for the wind atlas ing computing resources to the amount of [1] www.neweuropeanwindatlas.eu, have been defined. 56 Mill. CPU hours has been submitted to [2] Petersen, E. L., Troen, I., Jørgensen, the PRACE consortium (Partnership for H. E., Mann, J.: Are local wind power resources well estimated? Environ. The mesoscale wind atlas will include time Advanced Computing in Europe) to be Res. Lett., 8, 011005, doi: 10.1088/1748- series of several quantities (e.g. wind speed spent on the MareNostrum supercomputer 9326/8/1/011005 (2013) and direction, temperature, atmospheric in Barcelona. [3]:Troen, I., Petersen, E. L.: European Wind Atlas, Roskilde: Risoe National Laboratory (1989) [4] Petersen, E. L., Troen, I.: Wind conditions and resource assessment, WIREs Energy Environ., 1, 206–17, doi: 10.1002/ wene.4 (2012) [5] Mann, J.; Angelou, N.; Arnqvist, J.; Callies, D.; Cantero, E.; Chávez Arroyo, R.; Courtney, M.; Cuxart, J.; Dellwik, E.; Gottschall, J.; Ivanell, S.; Kühn, P.; Lea, G.; Matos, J. C.; Palma, J. M. L. M., Pauscher, L.; Peña, A.; Sanz Rodrigo, J.; Söderberg, S.; Vasiljevic, N.; Veiga Rodrigues, C.: Complex terrain experiments in the New European Wind Atlas, Phil. Trans. R. Soc. A, 375, 20160101, doi: 10.1098/ rsta.2016.0101 (2017) [6] DTU Wind Energy: The Perdigao 2017 Field Experiment, Video on You- Tube, https://youtu.be/rd3MUG715Bs (2017) [7] Vanvyve, E.; Delle Monache, L.; Monaghan, A. J.; Pinto, J. O.: Wind resource estimates with an analog ensemble approach, Renew. Energ., 74, 761–773, doi:10.1016/j. Figure 3: Domain configuration for the mesoscale wind atlas production runs. The thick renene.2014.08.060 (2015) grey line delimits the area of the final wind atlas

– 37 – 38WIND AS A RESSOURCE Annual Report 2015-2017

High Performance Computing (HPC) Cluster (WIMS-Cluster)

Carl von Ossietzky Universität Oldenburg Using the powerful HPC cluster, this project Since the start of its regular operation, the Institute of Physics addresses three principal scientific ques- HPC cluster has been subject to multiple tions: It aims at an improved understanding planned updates and optimizations of the Joachim Peinke, Vlaho Petrovic, of the interaction of the atmospheric bound- installed software in order to guarantee opti- Gerald Steinfeld, Wilke Trei, Lena Vorspel, ary layer flow with wind turbines as well as mal scientific results. User support for about Sonja Krüger, Mehdi Vali, Marc Bromm, the impact on the wind turbine performance, 55-60 users has been provided continuously Hauke Wurps, Stefan Harfst, such as power generation and loads. Moreo- during the project, including the support for Stefan Albensöder, Shumian Zhao ver, the project tackles the improvement of migrating projects on larger scale HPC sys- Funding: Federal Ministry for Economic computational methods to allow for a holis- tems and investigations in the rare case of Affairs and Energy (BMWi) tic simulation of wind turbines problems. Therefor the close coordination between the IT-department and the ForWind Ref.Nr. FKZ 0324005 cluster operator was essential.

Duration: 11/2015 – 10/2019 Project Description Interaction of Wind Turbines with the Atmospheric Boundary Layer HPC Cluster operation The first scientific work package in this pro- The first phase of the project was dedi- ject addresses the interaction between the cated to the acquisition and installation of atmospheric boundary layer (ABL) flow the High Performance Computing cluster and the wind turbines. Particular emphasis EDDY, which is equipped with 5.856 CPU is given to the stable stratified atmospheric Introduction cores that are apportioned to 244 comput- boundary layer, which occurs in the south- ing nodes. All nodes share 21 TB of main ern North Sea region about 20% of the The main topics of wind energy research are memory and are connected via a non- time and in the northern German onshore the improvement of turbine designs and in- blocking FDR infiniband interconnect for region about 40% of the time [3]. The stable vestigations about the interaction of the tur- massive parallel computations. The cluster stratification causes significant wind speed bines with the atmospheric boundary layer. shares much of its infrastructure with the variations across the wind turbine rotor and Computational fluid dynamics (CFD) is a general-purpose scientific computing HPC results in asymmetric and long-lasting wake crucial tool for used for the investigation of cluster CARL of the Carl von Ossietzky flows, which are currently not represented this interaction in a time and resource effi- Oldenburg. in engineering models. cient manner. Starting in August of 2016, the housing of the two HPC clusters and the technical in- An optimized coupling of the large eddy To allow for CFD simulations providing the stallation was completed by the end of Oc- simulation (LES) model PALM [5] with necessary detailed resolution to address cur- tober 2016. the aeroelastic model FAST [4] has been rent research questions in the fields of wind realized to create a suitable simulation tool turbine design and wind resource estimation The installation phase ended with an ensu- chain for the investigation of interactions of on meso- and micro-scales, considerable ing official LINPACK [1] benchmark test, a wind turbine with a stable stratified atmos- computational resources are required. in which the two systems with a combined pheric boundary layer. The coupled method 571 computing nodes reached 457.2 Tflop/s has been assessed and optimized regarding Funded by the German Federal Ministry (tera floating point operations per second). computational costs and stability. The re- for Economic Affairs and Energy (BMWi), They were thus ranked 363th on the Top sulting improved LES model will be used to the High Performance Computing (HPC) 500 worldwide scientific supercomputers investigate wind farm control schemes and cluster EDDY (Heterogener Hochleistung- list in November 2016 [2]. to improve the mesoscale model WRF [6]. srechner für windenergierelevante Mete- orologie- und Strömungsberechnungen From November 2016 until February 2017 Wind Farm Control by a Coupled (WIMS-Cluster)) has been acquired and the installation of the scientific software and Aeroelastic Model with Large Eddy operated by ForWind (University of Olden- an in-depth testing phase took place. Subse- Simulations burg). quently, the HPC cluster was made available Different atmospheric conditions have a for scientific computations in March 2017. large impact on the properties of wind tur-

– 38 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201739

bine wakes such as the wake recovery and High Performance Code for Holistic improvements as well as improved grid pro- the wake expansion. However, the wake Simulations of Wind Turbines jection methods. Additional tasks include flow is also impacted by the wind farm CFD simulations have proven to be effective the adaption of such algorithms for imple- control itself, for example in yawed inflow in the development process of aerodynami- mentation on graphic cards, which is pro- cases or by means of so-called wake steer- cally critical components. A desired future jected to improve the overall performance ing. development in wind turbine design is a ho- per hardware cost ratio significantly. listic automatic optimization process. This The developed coupled LES model is used cannot simply be achieved by using mod- to simulate the wake of a wind turbine ex- ern high performance computing facilities, Summary posed to various different inflow situations but also requires an improved efficiency of – including different atmospheric stabilities the used CFD algorithms. New numerical The successful installation and operation of as well as different operational conditions strategies are thus investigated and their ef- the High Performance Computing cluster and control strategies of the wind turbine. ficiency and scalability is assessed. EDDY provides an important tool for mul- Wind fields of the wind turbine wake are ex- tiple research projects. Within the first half tracted from the LES to estimate the power A coupled algorithm, which solves the of the project WIMS-Cluster, capabilities performance and loads of downstream tur- equations for the pressure and the veloc- of several simulation methods have been bines using the aeroelastic tool FAST. Dur- ity components simultaneously, has been advanced, including the improvement of al- ing the project, the results will be validated implemented in the CFD code OpenFoam gorithms as well as the coupling of different against measurements from a field cam- [7]. The solver makes use of the so-called methods. This allows for the investigation paign in order to assess the feasibility of the LU-SGS algorithm in order to achieve of important research questions, such as the load estimation method. significantly reduced computational times impact of the atmospheric boundary layer compared to conventional solvers. Current and wind turbine control schemes on wind work focuses on stability and performance farm wakes.

References

[1] http://www.netlib.org/linpack [2] https://www.top500.org/site/50678 [3] Dörenkämper, M., Witha, B., Steinfeld, G., Heinemann, D., Kühn, M., 2015: The impact of stable atmospheric boundary layers on wind- turbine wakes within offshore wind farms. J. Wind. Eng. Ind. Aerod., doi: 10.1016/j.jweia.2014.12.011. [4] https://www.nas.nasa.gov/ Software/FAST/ [5] https://palm.muk.uni-hannover.de [6] https://www.mmm.ucar.edu/ weather-research-and-forecasting- model [7] https://openfoam.org/ Figure 1: Offshore Wind Farm Wakes in the German bight

– 39 – 40WIND AS A RESSOURCE Annual Report 2015-2017

Further Development of a Two-dimensional Atmospheric Laser Cantilever Anemometer for the High-resolution Investigation of Turbulent Wind Flows for the Optimization of Offshore Wind Turbine Blades

Even though the demand for high-resolution Carl von Ossietzky Universität Oldenburg quency is outside the measuring range. In Institute of Physics wind speed data from industry and the re- addition, a tracking system is to be devel- Research Group: Turbulence, Wind Energy search community is very high, there are oped so that the offshore measurements can and Stochastic currently no such measurements in the off- be carried out more independently of the shore area and only very few measurements wind direction and a sudden change in the Jaroslaw Puczylowski, Joachim Peinke, in the onshore area [3,4,5,6] This is due to wind direction does not result in the meas- Michael Hölling the lack of suitable measuring instruments. urement being aborted or even dropped out. For measurements in wind tunnels, hot In addition, the reliability of the measured Funding: Deutsche Bundesstiftung Um- wires are a widespread standard for high- values is to be further increased by tempera- welt (DBU) resolution measurements, but due to their ture control, which makes the sensor more nature (1-5μm thick) they are not unsuitable independent of temperature fluctuations and Ref.Nr. 32399/01 for long-term measurements in offshore ap- thus lessens the impact of environmental Duration: 12/2015 – 12/2016 plications. influences. In order to extend the service life of the measuring system as a whole, an Within the framework of a joint project be- additional cover is to be constructed from tween the Kiel University of Applied Sci- which the sensor is only moved out for ences and the University of Oldenburg, a the measurements. These measures are all new sensor has been developed to address aimed at making the 2D-ALCA more robust this challenge. The so-called 2D-ALCA and resistant to offshore conditions in the enables spatial and temporal high-resolu- North Sea. In order to assess the effective- tion measurements in the offshore area. The ness of these improvements in a meaningful Introduction measuring principle utilizes a laser-based way, free-field measurements is to be - car measurement of the deflection of a very ried out on the FINO3 in the North Sea and Since intermittent wind fields cause high small bending beam, which is deflected by the system is to be extensively tested. loads on wind turbines, which can lead to the flow. In comparison to hot wires, this damage, a very precise knowledge of the sensor element is very robust against exter- After successful tests of the renewals and structure of the wind field is necessary to nal influences and is therefore basically suit- improvements on the 2D-LCA, the measur- combat this danger, i.e. to know which vor- able for use in offshore applications. ing system was tested on the Fino3 platform tex sizes can occur. The investigation of the under real conditions. Fig. 1 shows the in- smallest vortex structures in the range of a stalled measurement system on the metmast few millimeters is also necessary. Even if Project Description at a height of 90m. these do not make a significant contribution to the loads, they are decisive for the The aim of this project is the further devel- Measurements were carried out and ana- aerodynamics of the rotors and thus for opment of the existing 2D-ALCA measure- lyzed on different days. The graph in fig. 2 the yield of the wind turbine. Furthermore, ment system in order to avoid the problems shows an example of a time series for the high-resolution data serve as a basis for the encountered in an earlier project [7], so that absolute velocity and flow direction for the development of turbulence models, which at the end of this project a functional proto- 2D-ALCA and an ultrasonic anemometer, in turn are essential for CFD calculations type is available, which is suitable for a pos- which was installed in the immediate vi- [1,2]. These simulations are often used in sible product development within the scope cinity. research and industry as a cost-effective al- of a cooperation with Sea & Sun GmbH. ternative to complex measurements and are becoming increasingly popular due to the In order to achieve a better temporal resolu- constantly increasing computing power of tion (>10 kHz), a weight-reduced cantilever computers. is to be developed so that its resonance fre-

– 40 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201741

Summary stable signal quality could be achieved enabled the measurement method to be through tracking and temperature control. validated directly by an anemometer. In ad- The aim of this project was the further de- Thus a considerable improvement of the dition, the service life of the system could velopment of 2D-ALCA to validate a robust system could be achieved and tested under be significantly extended, even if a detailed and for the offshore area suitable measure- offshore conditions. statement about the service life could not be ment method, which allows temporally and given due to external problems in the data spatially high-resolution measurements of Despite minor problems, the measurements connection. Besides the application in off- the atmospheric boundary layer. With the have shown that the atmospheric bound- shore turbulence measurement, further ap- development of a cover, the service life of ary layer can be measured with 2D-ALCA plication possibilities in other research areas the sensor could be significantly extended, under offshore conditions with high tem- are conceivable. and improved angular coverage and more poral resolution. Triggered measurements

Figure 1: The 2D-ALCA mounted at the Figure 2: Time series of the 2D-ALCA measurement in comparison with the ultrasonic metmast of the FINO3-platform. anemometer (section of measurement). a) Wind speed, b) Wind direction.

References

[1] Barthelmie, R. J.; Hansen, K.; [3] Schreck, S.; Lundquist, J.; [6] Siebert, Holger; Lehmann, Katrin; Frandsen, S. T. ; Rathmann, O.; Schepers, Shaw, W.: U.S. Department of Energy Shaw, Raymond A.: On the use of a hot J. G.; Schlez, W.; Phillips, J.; Rados, K.; Workshop Report: Research Needs for wire anemometer for turbulence measu- Zervos, A.; Politis, E. S.; Wind Resource Characterization, National rements in two-phase flows – A wind Chaviaropoulos, P. K.: Modelling and Renewable Energy Laboratory / NREL/TP- tunnel study. In: Journal of Atmos. Ocean. measuring flow and wind turbine wakes 500-43521 (2008) – Technol. (2007) in large wind farms offshore. In: Wind Forschungsbericht [7] Jeromin, Andreas; Schaffarczyk, Alois Energy 12 (2009), Nr. 5, [4] Neumann, T. ; Emeis, S.; Grigutsch, K.; P.: Messung und Analyse hochfrequenter pp. 431–444. – ISSN 1099–1824 Riedel, V.; Turk, M.: OWID Offshore Wind (>1 kHz) Turbulenzanteile im Offshore- [2] Mucke, Tanja; Harkness, Christi- Design Parameter. In: Wind zur Optimierung von aerodynami- ne; Argyriadis, Kimon: Offshore Wind DEWI Magazin Nr. 35 (2009) schen Blattprofilen / Abschlussbericht zu Turbulence - Model versus Measurement. [5] Sreenivasan, K. R.; Dhruva, B.: Is there Projekt 122-09-023, gefordert vom Land In: EWEA Offshore 2012, Copenhagen, Scaling in High-Reynolds-Number Turbu- Schleswig-Holstein (2012) – Forschungs- Denmark, 2012 lence? In: Progress of Theoretical Physics bericht Supplement No 130 (1998)

– 41 – 42WIND AS A RESSOURCE Annual Report 2015-2017

Extended THETA for Site Assessment (ETESIAN)

Carl von Ossietzky Universität Oldenburg Project Description ness, the effect of physical obstacles on the Wind Energy Meteorology surface like grass, buildings or trees have to A main objective of ForWind’s work in be represented. In general a surface rough- Renko Buhr, Siamak Akbarzadeh, ETESIAN is to modify the CFD solver ness is used similarly to engineering models. Johannes Munk, Björn Witha, THETA in such a way that it can simulate Forests are of special interest in wind ener- Gerald Steinfeld the atmospheric boundary layer and can be gy, as more and more wind parks are built in Partners: Fraunhofer IWES, DLR-AS-CAS, used in Enercon’s industrial site assessment. forested regions. Therefore the simple rep- WRD (lead) The enhancement of the code includes the resentation with a surface roughness is not implementation of a thermal solver that can sufficient. The Darcy-Forchheimer-Model Funding: Federal Ministry for Economic handle non-neutral situations, the represen- [1] was implemented into the THETA code Affairs and Energy (BMWi) tation of complex terrain in terms of rough- to account for sinks and sources of the for- ness, forest and orography as well as wake est in both turbulence and momentum. Two Ref.Nr. 0324000D modelling. A further focus of the project is different types of forests have been com- to automatize the calibration to important pared and the results were presented at the Duration: 01/2016 – 12/2018 site assessment parameters like roughness Wind Energy Science Conference 2017 [2]. and stability which is currently done in a Fig. 1 shows the effect of the forest on the manual process. flow. Within the forest, the flow is slowed down, above (and at the sides of the forest), Terrain roughness and forest a speedup effect is visible. Behind the forest Introduction A CFD solver to be used in site assessment the flow is forced back to the inflow profiles. needs to consider terrain roughness and spe- Several kilometers are necessary to reach Besides re-powering of wind turbines at cifically forests. Regarding terrain rough- free flow conditions. reliable high-yield locations, the placement of wind turbines in less promising onshore locations (complex and forested terrain) is of significant importance in the conversion of the German energy system to renewable energies. This makes the site assessment much more difficult, so new and more efficient numerical methods for site assessment are required.

The main focus of the joint project ETE- SIAN is to transfer high resolution numerical approaches from academic research to the industrial site assessment process. Our project partner DLR provides their well- established CFD solver THETA which has previously been used for industrial CFD applications. Together with Fraunhofer IWES ForWind will significantly enhance this code so that it can be applied to atmospheric flows. Project partner and consortium leader WRD will then integrate the enhanced solver into Enercon’s industrial site assessment process chain.

Figure 1: Flow over a forest (streamwise wind component)

– 42 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201743

Thermal Solver While in wind energy applications the atmosphere is often assumed to be neutrally stratified, this stratification is barely present in the planetary boundary layer. Therefore, additional source terms have been added to the model, following the Monin-Obukhov Similarity Theory. In detail the model proposed by project partner Fraunhofer IWES [3] was implemented in close collaboration with the project partners Fraunhofer IWES and DLR. A stable stratification dampens vertical movement and therefore turbulent exchange, leading to a stronger vertical wind shear, while a convective stratification enhances turbulent exchange with smaller vertical wind shear. The effect on vertical exchange can be seen particularly in Figure 2, in the wake effect behind a standardized cosine hill. Figure 2: Flow over a cosine hill: Turbulent kinetic energy (TKE) for stable (top), neutral Coriolis force (middle) and convective (bottom) atmospheric stratification The planetary boundary layer can be divided in two important parts, the Prandtl- and the Ekman-Layer. While the first is only influenced by the surface friction and no horizontal wind shears can be observed, the Ekman-Layer is additionally influenced by the free atmosphere and therefore by the Coriolis force. This leads to a rotation of A central part of the project work is to ex- the wind with height. While often neglect- plore the systematic and automatized cali- ed in engineering models an advanced nu- bration of two important parameters of the merical model for site assessment should thermal solver which is implemented in References include the Coriolis force. The implemen- the THETA code: the aerodynamic rough- [1] Bejan, A.: Convection in tation of the Coriolis force in THETA has ness length z0 and the atmospheric stability been fairly straightforward. Additionally, through the Obukhov length L. porous media in A. Bejan (Ed.), the turbulent length scale of the flow had Convection Heat Transfer, 4th edition to be limited in order to get reasonable re- (pp 540-542), Jon Wiley & sons, Inc., Hoboken, New Jersey, doi: sults. The results were in good agreement Summary 10.1002/9781118671627, 2013 with the well-known Leipzig dataset. [2] Buhr, R.; Aposporidis, A.; In ETESIAN the CFD solver THETA will Munk, J.; Steinfeld, G.; Heinemann, Automatized model calibration be significantly enhanced and modified so D.: RANS simulations of flows above Typically measurements over a certain that it can simulate the atmospheric bound- forest canopies of different period are available for prospective wind ary layer and can be used in industrial site topologies, Wind Energy Science farm sites. Depending on the size of the assessment. The enhancement of the code Conference 2017, Technical University wind farm these measurements have been includes the implementation of a thermal of Denmark, Lyngby Campus, performed at one or several locations with- solver that can handle non-neutral situa- June 26 - 29 June 2017. in the wind farm. CFD simulations, how- tions, the representation of complex terrain [3] Chang, C-Y; Schmidt, J; Dörenkämper, M; Stoevesandt, B: ever, depend on inflow conditions at the in terms of roughness, forest and orography A consistent steady state CFD boundaries of the simulated domain which as well as wake modelling. A further focus simulation method for stratified are not know a-priori. The accuracy of the of the project is to automatize the calibra- atmospheric boundary layer flows, J. simulations is massively depending on the tion to important site assessment parameters Wind Eng. Ind. Aerodyn., 172, 55-67, boundary conditions which have to be cali- like roughness and stability which is cur- 2018 brated with the met mast data. rently done in a manual process.

– 43 – 44WIND AS A RESSOURCE Annual Report 2015-2017

Highly Accelerated Pitch Bearing Test

Leibniz Universität Hannover Introduction Project Description Institute for Machine Design and Tribology (IMKT) Pitch bearings of wind turbines are large, The project splits up into two major parts: Fraunhofer-Institute for Wind Energy grease-lubricated rolling bearings that con- Development of calculation and simulation Systems (IWES) nect the rotor blades with the rotor hub. methods, carried out by IMKT, and devel- They are used to turn (pitch) the blades opment of test rig and test method, carried Matthias Stammler, Arne Bartschat, around their primary axis. Turning the out by IWES. Florian Schleich, Karsten Behnke, blades changes the amount of lift they gen- Oliver Menck, Fabian Schwack erate. Thus, pitching can be used for both IMKT uses a combined approach of test power and load control of the wind turbine. with type 7208 roller bearings and finite Funding: Federal Ministry for Economic The largest rotatory movement of a blade element simulation to understand the funda- Affairs and Energy (BMWi) is a 90° turn, which is used to change from mentals of wear damage modes under oscil- power production mode to idling or vice latory conditions [2, 3]. Ref.Nr. 0325918A versa. Under most circumstances, all move- Duration: 01/2016 – 09/2020 ments of pitch bearings are oscillatory, as IWES undertakes efforts to understand the the power output of the blades is at maxi- external loads and bearing movements [1, mum at the 0° position. As the wind speed 4] and transform the conditions of the wind and turbulence are of complex, stochastic turbine application to a test environment. The nature, this is also the case for the loads and name of test environment is BEAT6.1 (Bear- movements of blade bearings [1]. ing Endurance and Acceptance Test rig). IWES commissions the test rig in January The complex load situation in combina- 2019. Within the project, IWES will test six tion with the oscillating movements and blade bearings with a diameter of 5 m. the ever-increasing size of the bearings have made bearing tests more and more important. HAPT pursues two main goals: Summary Further develop the design methods of pitch bearings and provide a test infrastructure HAPT deals with pitch bearings of wind tur- for bearings of turbines of up to 10 MW, bines and aims at improving their reliabil- together with a comprehensible method ity. The collaboration between Fraunhofer for acceptance tests. IWES joined with the IWES, LUH IMKT and the bearing manu- IMKT and the bearing manufacturer IMO to facturer IMO has led to significant progress achieve these goals. in the field of simulation and test. The project will culminate in the endurance test of six pitch bearings with the size of 5 m at the newly built BEAT6.1 test rig.

– 44 – Annual Report 2015-2017 WIND AS A RESSOURCE Annual Report 2015-201745

Figure 1: 5 m bearings next to the BEAT6.1 test rig

References

[1] Stammler, M.; Reuter, A.; Poll, G.: [3] Schwack, F.; Prigge, F.; Poll, G.: Cycle counting of roller bearing oscillati- Frictional Work in Oscillating Bearings ons – case study of wind turbine – Simulation of an Angular Contact Ball individual pitching system Renewable Bearing under Dry Conditions and Small Energy Focus 25, pp. 40–47 (2018) Amplitudes, Beijing, China (2017) [2] Schwack, F.; Byckov, A.; Bader, N.; [4] Stammler, M.; Baust, S.; Reuter, A.; Poll, G.: Time-dependent analyses of wear Poll, G.: Load distribution in a roller- in oscillating bearing applications, Atlanta type rotor blade bearing. J. Phys.: (2017) Conf. Ser., p. 1037 (2018)

– 45 – 46WIND AS A RESSOURCE Annual Report 2015-2017

Lidar for Wind Energy Deployment (IEA Annex 32) – Phase II

Carl von Ossietzky Universität Oldenburg, The advisory board of the task is composed Summary Institute of Physics mainly of academics and industry repre- Research Group: Wind Energy Systems sentatives. The advisory board organizes IEA Task 32 Phase II is a task included in monthly conference calls and annual meet- the IEA Wind TCP. It aims at identifiying Davide Trabucci ings to bring the entire Task 32 community and mitigaring barriers to the use of lidars together. Over 250 participants contributed for different wind energy applications Funding: International Energy Agency to the meetings and workshops organized in including site assessment, power perfor- IEA – Task 32 – Phase II Phase II. mance testing, turbine controls and load Ref.Nr. IEA - Task 32: Wind Lidar Systems verifications and complex flow. Workshops for Wind Energy Deployment – Phase II Along with the annual meetings, different are organized regularly to address the chal- workshops are being conducted to solve the lenges set by the scientific committee. Duration: 01/2016 - 12/2018 challenges set by the community. During Results of the workshop are published in 2016-2017 the following workshops were different journals, workshop reports and organised: recommended practices.

• Workshop #1: Floating Lidar System: Introduction Current Technology Status and Require- ments for Improved Maturity. ORE Cat- IEA Task 32: is a task included in the Inter- apult, Blyth, UK. February, 23-24, 2016 national Energy Agency Wind Technology • Workshop #2: Optimizing Lidar Design Collaboration Programme (IEA Wind TCP), for Wind Turbine Control Applications. a cooperation that shares information and Boston Marriott Copley Place, Boston, research activities to advance wind energy MA, USA. July, 5, 2016. research, development and deployment with • Workshop #3: Lidar measurements for its member countries, under the support of wake assessment and comparison with the International Energy Agency (IEA). wake models. Technical University of Munich, Munich, Germany. October, 4, 2016 Project Description • Workshop #4: Power Performance: Round Robin for FDIS IEC 61400 12-1 References The Phase II of the IEA Task 32: Wind Li- Ed. 2. Calculation of Uncertainty for dar Systems for Wind Energy Deployment Lidar Application. University of Strath- [1] Bischoff, O; Würth, I.; Gottschall, started in 2016 and is focused on identifying clyde, Glasgow, Scotland Scotland. J.; Gribben, B.; Hughes, J.; Stein, D.; and mitigation barriers to the deployment of Round Robin August-December 2016. Verhoef, H.: IEA Wind RP 18. Floating wind lidar in four application areas: Final workshop: December, 14, 2016. Lidar Systems • Workshop #5: Elaboration of use cases [2] Sathe, A.; Banta, R.; Pauscher, L.; 1. Site assessment. in wake and complex flow measure- Vogstad, K.; Schlipf, D.; Wylie, S.: 2. Power performance testing. ments. University of Glasgow, Glasgow, Estimating Turbulence Statistics and Parameters from Ground- and Nacelle- 3. Turbine controls and load verification. Scotland. June, 19-20, 2017 Based Lidar Measurements. 4. Complex flow. • Workshop #6: Power Performance Meas- [3] Clifton, A.; Boquet, M.; Burin, E.; urement Using Nacelle Lidars. DONG 5. Out of the box: Introduced in 2017, it Hofsäß, M.; Klaas, T.; Vogstad, K.; addresses new developments such as the Energy, Gentofte, Denmark, September, Clive, P.; Harris, M.; Wylie, S.; Osler, E.; use of lidar forecasting of wind speed 27, 2017 Banta, R.; Choukulkar, A.; Lundquist, and power, the "Open Lidar" modular • Workshop #7: Lidar Campaigns in Com- J.; Aitken, M.: Remote Sensing of lidar concept. plex Terrain. University of Stuttgart, Complex Flows by Doppler Wind Lidar: Stuttgart, Germany. November, 8, 2017 Issues and Preliminary Recommendations.

– 46 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201747

GROWup OPC– Grouted Joints for Offshore Wind Energy Converters under Reversed Axial Loadings and Upscaled Thicknesses filled with Ordinary Portland Cement (OPC)

Leibniz Universität Hannover Project Description As a result of the experimental investiga- Institute for Steel Construction tions it has been shown that OPC grouts (project leader) The development of offshore wind turbines are generally suitable for grouted joints of Institute of Building Materials Science in Germany and worldwide is becom- offshore wind turbines. Using a slurry mixer (project co-leader) ing more efficient and cost effective. The sufficient properties regarding the pump- substitution of a high-performance mortar ability and the flowability can be achieved. Peter Schaumann, Anne Bechtel, grout typically used with a standardized In order to achieve a sufficiently high early Alexander Raba, Joshua Henneberg (IfS) OPC-grout could have significant economic and final strength of the grout, especially Ludger Lohaus, Michael Werner, Dario Cotardo (IfB) advantages. The aim of the investigations Portland cements with a strength class of presented was to compare typical cement 52.5 are suitable, cf. Fig. 2 [1]. Using OPC Funding: Federal Ministry for Economic grouts with an established high-perfor- grouts, an increased hydration heat develop- Affairs and Energy (BMWi) mance mortar grout and, furthermore, to ment must be expected. In order to avoid proof the usability of different OPC-grouts temperature-induced expansion phenom- Ref.Nr. 0325290 as a grout material in grouted joints for off- ena, the use of cements with increased sul- shore wind turbines. A large-scale test facil- phate resistance is recommended. Duration: 06/2011 – 07/2017 ity for the simulation of filling processes for the in situ assembly of grouted joints was In addition material tests, analyses of fill- developed. In addition, a submerged test ing processes and the fatigue behaviour of facility was developed, cf. Fig. 1, and was different grout materials were determined used for investigations to quantify the hy- using small- and large-scale test facilities Introduction dration heat development under nearly real- for cyclic axial loading. With special focus istic conditions. on submerged conditions and the occurring Connections between the foundation piles failure mechanism. and the support structures of offshore wind turbines are executed by filling the annulus between the two steel tubes with a cementi- tious grout material. In general, specialized high-performance mortar grouts are used in Germany with respect to the certification body and BSH. Usually, high-performance mortar grouts are cost-intensive and are more sensitive to fluctuations in the water content as well as to varying temperatures, both aspects influence the flowability. To this end, the application of an ordinary Portland cement grout (OPC-grout), which represents a standardized building material, could provide an economic and executional benefit. Indeed, OPC-grouts are already used worldwide and established not only in the oil and gas industry. However, the application of OPC-grouts for offshore wind turbines in Germany is not usual and is only accepted in individual cases with respect to the approval procedure. Figure 1: Test facility for hydration heat development under submerged ambient conditions

– 47 – 48SUPPORT STRUCTURES Annual Report 2015-2017

Figure 2: Relative compression strength distribution of filling test setup

Due to shear keys on the inner surfaces of the grout strength, which is one reason why the grouted connection the interlocking be- OPC-grouts lead to reduced fatigue resist- tween steel and grout increases. This leads ances [5]. As OPC-grout is generally suita- to different failure mechanisms under dry ble for grouted connections, higher connec- and submerged ambient conditions. In both tion strengths can be reached by increasing cases grout material nearby the shear keys grout lengths. crushes as consequence of locally concentrated load transfer. Crushed material leads to notch smoothening effects around the shear key and stops further damage under dry ambient conditions. Further connection resistances can be activated. A grouted connection in dry ambient conditions finally fails after additional load increase due to shear cracking (so called compression strut failure) [2]. Contrarily, submerged ambient conditions lead to water ingress into the connection [3]. Due to pumping effects, caused by cyclic loading and elastic deformations, the crushed grout material is washed out, cf. Fig. 3. Over time, the grouted connection fails by pushing the pile continuously though the grout material [4]. Under more realistic submerged ambient conditions the fatigue resistance of axially loaded grouted connections decreases significantly.

Small-scale tests under varying load frequencies show lower fatigue resistances with higher load frequencies cf. Fig. 4. Figure 3: Fatigue failure mechanism of Due to local grout failure the resistance of grouted connection under submerged grouted connections directly depends on ambient conditions

– 48 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201749

Figure 4: SN-curve showing influence of load frequency on fatigue resistance

Summary

All in all, numerous experimental and numerical investigations are carried out within the BMWi-research project GROWup OPC, which led to two main results. Firstly, OPC- References grout is generally suitable for offshore use in grouted connections. Due to lower strength [1] Cotardo, D.; Lohaus, L.; Werner, M.: [4] Raba, A.: Fatigue behaviour of of the material itself more grout length is Practical Performance of OPC-grout for submerged predominantly axially Offshore Wind Turbines in Large-scale loaded grouted connections. Hannover, necessary to reach comparable resistances Execution Tests, Proceedings of the 27th Germany, Leibniz Universität Hannover, to grouted connections with high strength International Offshore and Polar Enginee- Institute for Steel Construction. grout. Secondly, submerged ambient condi- ring Conference (ISOPE), San Francisco, Dissertation (2018) tions lead to a different failure mechanism USA, June 2017 [5] Schaumann, P.; Henneberg, J.; under cyclic loading and thereby to a sig- [2] Bechtel, A.: Fatigue Behaviour of Axi- Raba, A.: Axially loaded grouted nificant decrease of fatigue resistance of the ally Loaded Grouted Connections in Jacket connections in offshore conditions using grouted connection. Structures. Hannover, Germany, Leibniz ordinary portland cement, Universität Hannover, Institute for Steel 12th International Conference on The financial support by the Federal Minis- Construction. Dissertation (2016) Advances in Steel-Concrete Composite [3] Schaumann, P.; Raba, A.; Bechtel, A.: Structures, Valencia (2018) try for Economic Affairs and Energy as well Zum Ermüdungsverhalten von Grout- as the support by Fraunhofer IWES, Det Verbindungen bei unterschiedlichen Norske Veritas Germanischer Lloyd Group, Umgebungsbedingungen, in Stahlbau 85, Senvion SE, RWE Innogy and Strabag Off- 12/2016, pp. 679-688. shore is kindly acknowledged.

– 49 – 50SUPPORT STRUCTURES Annual Report 2015-2017

INNWIND.EU – Innovative Wind Conversion Systems (10-20MW) for Offshore Applications

Leibniz Universität Hannover Project Description and substructure that simplifies and unifies Institute of Structural Analysis, turbine structural dynamic characteristics at Institute for Building Materials, The overall objectives of the INNWIND. different water depths. Institute for Steel Construction EU project are the high performance inno- The LUH is involved in work package 4 Jan Häfele, Andreas Ehrmann, vative design of a beyond-state-of-the-art (offshore foundations and support struc- Niklas Scholle, Luka Radulovic 10-20MW offshore wind turbine and hardware demonstrators of some of the critical tures). Here, the objective is the preliminary Funding: European Union’s Seventh components. These ambitious primary ob- design of a hybrid jacket as shown in Fig. 1 Programme for research, technological jectives lead to a set of secondary objec- that comprises classical steel and so called development and demonstration tives, which are the specific innovations, sandwich tubes. Sandwich tubes usually – EC FP 7 new concepts, new technologies and proof have a non-metallic core which might be of concepts at the sub system and turbine an elastomer, grout or concrete, enwrapped Ref.Nr. Grant Agreement No. 308974, level. The progress beyond the state of the with steel faces at the inner and outer diam- INNWIND.EU art is envisaged as an integrated wind tur- eter. Due to the considerably better buckling behaviour of sandwich tubes compared to Duration: 11/2012 – 10/2017 bine concept with a light weight rotor having a combination of adaptive characteris- steel tubes it seems possible to reduce the tics from passive built-in geometrical and entire costs for the substructure and hence structural couplings and active distributed offshore wind energy in this way. But there smart sensing and control, an innovative, are many issues that have to be solved below-weight, direct drive generator and a fore an application of sandwich tubes for offshore structures is possible. Introduction standard mass-produced integrated tower

The research project with a total of 27 Euro- pean partners is an ambitious successor for the UpWind project, where the vision of a 20MW wind turbine was put forth with specific technology advances that are required to make it happen. The overall objectives of the INNWIND.EU project are the high performance innovative design of a beyond- state-of-the-art 10-20MW offshore wind turbine and hardware demonstrators of some of the critical components.

Figure 1: INNWIND.EU reference jacket design with sandwich braces

– 50 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201751

Figure 2: Bottom-fixed support structure concepts

Different jacket concepts have been devel- A central core of the research is the analy- For jacket substructures a large cost reduc- oped to support the INNWIND.EU 10 MW sis of cost reduction potential considering tion potential through design optimisation, reference wind turbine in a water depth of either innovative support structure concepts use of cost-effective materials and improved 50 m. Designs are based on conceptual de- or innovations on component level as well -automated-manufacturing and installation sign level using a reduced number of (gov- as optimized state of the art solutions, such practices could be shown. erning) integrated design load calculations as the application of new materials and load in order to assess the fatigue limit state and mitigation concepts. At the beginning of Overall the LCOE of the 20 MW offshore ultimate limit state. The developed solutions the project, a target cost reduction value of wind turbine was reduced by more than 30% are compared with an initially designed 20% was defined. The final cost calculations as compared to the 2012 EWII basis of the 5 reference jacket considering the structural show a reduction potential between 12%- MW wind turbine. This evaluation of LCOE mass and associated manufacturing costs. 43%. It should be noted that the technology- was primarily based on direct CAPEX sav- The reference jacket is a classical 4-leg steel readiness level and the level of development ings and increase in AEP from the innova- jacket concept with pre-piled foundation differ significantly between the different tions combined with the larger capacities. and is developed in the beginning of the IN- proposed concepts. Consequently, a direct Further savings in the LCOE due to OPEX NWIND.EU project. In parallel, automated comparison is rather difficult and the result- reduction from lower fatigue loading and design optimization procedures have been ing cost estimates of novel concepts still ease of maintenance is also expected. More developed and successfully applied, con- have high uncertainty. details can be found in the final report [1]. necting the design changes with the design and cost assessment in a loop. The following jacket solutions are addressed: Summary

1. Modular 4-leg and 3-leg jackets with INNWIND.EU has investigated offshore classical prepiled foundation and tubular wind turbines between 10 MW – 20 MW tower; capacity with the development of several References 2. Full-lattice tower concept; component innovations and demonstrations 3. Hybrid jacket using sandwich materials; of some of the key technologies. Over the [1] http://www.innwind.eu/news. 4. 4-leg jacket variant with suction buckets period of its 5-year progress, it has satisfac- October 2017. LCOE Reduction for to substitute the piled foundation. torily addressed all objectives that were tar- the Next Generation Offshore Wind geted. Several of the results can be directly Turbines – Outcomes from the INN- A sketch with reference to the points 3. and used by the industry. WIND.EU Project 4. is given in Fig. 2.

– 51 – 52SUPPORT STRUCTURES Annual Report 2015-2017

ProBeton – Development and Experimental Testing of Fatigue Resistant Concrete Foundation Structures for Offshore Wind Turbines

Leibniz Universität Hannover behavior for this high number of load cy- ble long beyond the previously defined Institute of Concrete Construction cles, especially for concrete under water. life time for numerous OWT in the North (project management) The uncertainty in determining the fatigue and Baltic Sea. Steffen Marx, Christoph von der Haar, resistance significantly affects the eco- Jörg Diederley nomic application of concrete foundation A drastic reduction of electricity production structures for OWT. costs would be possible for planned and ex- Institute of Building Materials Science isting offshore wind farms. Ludger Lohaus, Julian Hümme Project Description Fatigue behavior of concrete under off- Partner: shore-specific conditions Ed. Züblin AG – Central Engineering Division The collaborative research project ProBeton A special test rig was built to investigate the is funded by the German Federal Ministry influence of different environmental condi- Funding: Federal Ministry for Economic for Economic Affairs and Energy. This pro- tions on the fatigue behavior of concrete, cf. Affairs and Energy (BMWi) ject focuses on the fatigue behavior of large fig. 1. In this test rig small-scale test speci- concrete foundation structures for OWT. mens are tested dry, wet and alternating wet/ Ref.Nr. 0325670 Within ProBeton, typical fatigue loaded dry with respect to their fatigue behavior. construction details like shell connections, Due to these experimental tests, SN curves Duration: 11/2013 – 10/2016 changing cross sections and transition struc- could be derived for high strength concrete tures are investigated. under offshore-specific conditions. Further- more, the influence of fatigue-related dam- For the first time large-scale experimental age on the chloride migration is analyzed. tests and also small-scale tests under offshore-specific conditions are used to inves- Experimental component tests Introduction tigate the high cycle fatigue bearing capac- A swinging system is constructed to inves- ity. The experimental tests are accompanied tigate large stress ranges with high load- Current foundations of offshore wind tur- by intensive numerical simulations. ing frequencies in large-scale experimental bines (OWT) are typically built as steel tests. Two unbalance exciters generate dy- structures. Compared to a steel design, A realistic, experimentally verified descrip- namic loadings with a frequency nearby the concrete constructions offer advantages in tion of the fatigue behavior of concrete first eigenfrequency of the specimen within terms of manufacturing and maintenance structures for OWT provides two major ad- the swinging system. In this way, it is possi- costs as well as durability. This can be vantages: ble to investigate even large structural parts verified by the history of civil engineering or sections effectively in respect to their fa- structures. For example bridges and tanks • Concrete foundation structures can be tigue behavior. In this process test frequen- are initially built as steel constructions and produced much cheaper and more ef- cies up to 14 Hz were conducted. The speci- now mainly built as reinforced and partial- ficient. In deeper water they also occur mens were prepared with strain gauges and ly prestressed concrete constructions. Even in competition with conventional steel ultrasonic transducers. These sensors were in established offshore areas (e. g. oil and structures. used to measure the strain development and gas rigs) this trend can been observed, in • The residual lifespan of existing con- deterioration processes of the concrete due particular with increasing system sizes. crete support structures of wind turbines to fatigue loading, cf. fig. 2 and 3. Different can be fundamentally re-evaluated. Nu- stress levels and stress ranges were consid- For OWT the fatigue resistance is of great merical studies on a gravity foundation, ered in these experimental investigations. importance. Combined stresses by wind, which take into account the redistribu- waves and dynamic structural behavior tion of stresses in the structure, indicate Numeric structural investigations causes load cycles up to 109. There are no a multiple of the current design life, cf. The obtained knowledge of the experimen- reliable findings on material and structural [1]. A power production could be possi- tal tests was used within the numerical in

– 52 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201753

vesti

Figure 1: Under water test rig (© Institute of Building Materials Science)

– 53 – 54SUPPORT STRUCTURES Annual Report 2015-2017

gations. They are performed on a design of and the influence of environmental condi- a heavyweight foundation. A damage model tions (water, chlorides). The research results derived from the experimental tests was im- were used to optimize concrete foundation plemented in numerical simulations. structures. The objective target of this project was to describe the fatigue behavior of concrete foundations more realistic and to Summary ensure the economical use of concrete in the field of offshore wind energy. The results Based on previous numerical investigations of this research project are published in [4] on wind turbines [2] and small scale experi- among others. mental tests on high strength concrete specimen [3], this project investigates the fatigue behavior of real structural parts or sections

References

[1] Urban, S.; Strauss, A.; Macho, W.; Bergmeister, K.; Dehlinger, C.; Reiterer, M.: Zyklisch belastete Beton- strukturen, in Bautechnik 89, issue 11, pp. 737-753, Ernst & Sohn (2012) [2] Göhlmann, J.: Zur Schädigungsberechnung an Beton- konstruktionen für Windenergieanla- gen unter mehrstufiger und mehraxi- aler Ermüdungsbeanspruchung, Dissertation, Institut für Massivbau, Leibniz Universität Hannover (2009) [3] Grünberg, J.; Oneschkow, N.: Gründung von Offshore- Windener- gieanlagen aus filigranen Betonkon- Figure 2: Vibrating test specimen (© Institute of Concrete Construction) struktionen unter besonderer Beob- achtung des Ermüdungsverhaltens von hochfestem Beton, Abschlussbericht, Hannover, 2010 [4] Marx, S.; Hansen, M.; von der Haar, C.; Diederley, J.; Lohaus, L.; Hümme, J.: Entwicklung und versuchstechnische Erprobung von ermüdungsfesten Gründungskonstruk- tionen aus Beton für Offshore-Wind- energieanlagen, Abschlussbericht zum BMWi-Forschungsvorhaben ProBeton (FKZ 0325670), 2017, DOI: 10.2314/ GBV:889255482 Figure 3: Set up of large-scale fatigue tests (© Institute of Concrete Construction)

– 54 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201755

HyConCast – Hybrid Substructure of High Strength Concrete and Ductile Iron Castings for Offshore Wind Turbines

Leibniz Universität Hannover The overall objective of this project is to thin walled concrete precast pipes. In order Institute of Concrete Construction assess the feasibility and applicability as to their required load bearing capacity the (project management) well as to investigate the necessary basics piles are made of high strength or ultra-high Steffen Marx, Christoph von der Haar, for planning, design and construction of a strength concrete in less stressed areas of Marina Stümpel hybrid substructure. Transport and instal- the structure. Filled up with normal-strength lation concepts will be developed, the risk mass concrete the substructures are maneu- Institute for Steel Construction of scour on the seabed will be analyzed vered to its final position, aligned locally Peter Schaumann, Luka Radulovic and the structural behavior of the installed and finally lowered. Franzius-Institute for Hydraulic, Estuarine components and connections will be inves- and Coastal Engineering tigated. These knots of ductile iron can be produced Torsten Schlurmann, Arne Stahlmann, in Germany as well as the high strength Kim Mario Welzel precast concrete pipes in excellent quality Project Description and with own resources in large quantities. Partners: Thus, they have an excellent potential for grbv Ingenieure im Bauwesen GmbH & The research project is funded by the Feder- OWT. Co. KG al Ministry of Economic Affairs and Energy. SSF Ingenieure AG Several institutes of the Leibniz University Project course Siempelkamp Giesserei GmbH Hannover, several industrial companies and To achieve the objectives the research pro- Max Bögl Bauservice GmbH & Co. KG engineering consultants are involved. ject is divided in three work packages. In Funding: Federal Ministry for Economic the first work package the project partners Affairs and Energy (BMWi) The innovative concept of the novel, hybrid use their expertise to develop a substruc- Project Management Jülich (PTJ) substructure bases on combination of large- ture. The aim is to define the conditions that sized, thin-walled ductile iron casting knots affect the design of the hybrid substructure Ref.Nr. 0325651 with high-strength, lightweight precast con- in terms of serial production, economy, crete pipes. The relevant design objective flexible use and stability. The results of Duration: 01/2014 – 12/2017 is a construction optimally exploited to the the individual works were collected and properties of the used materials. To ensure presented on the screening workshop and high cost efficiency the entire process chain compiled in an optimized version, called is account from the production of individual base construction (cf. fig. 1) for the follow- components through the inland transport, ing work packages. pre-assembly, the offshore transport, instal- Introduction lation, completion up to the operation of the In the second work package numerical load finished support structure. simulations and structural studies were carried out on the base structure. Experimental Current substructures of offshore wind Used materials investigations of some small scale test spec- turbines (OWT) are typically built as steel The mechanical properties of ductile iron, a imen lead to better specifications within the structures. Compared to steel structures, nodular graphite cast iron material, are simi- structural design. concrete constructions offer advantages in lar to steel. However, a significant improve- terms of manufacturing and maintenance ment of this material is the manufactur- In the third work package the examinations costs as well as durability. Therefore, the ability and increased corrosion resistance. concentrated on special areas of the base research project "HyConCast - Hybrid This material provides large-scale knot construction and in particular on their con- substructure of high strength concrete and structures with wall thicknesses accord- nection elements. For these parts numeri- ductile iron castings for offshore wind ing to the flow of forces and deliver higher cal and experimental investigations were turbines" deals with the development of fatigue resistance than large scale welded conducted. In addition to the numeric in- a novel, hybrid substructure for offshore steel components. vestigations large-scale experiments were wind turbines for water depth up to 50 m performed. and outputs of ≥ 6 MW. The mainly claimed to multi-axial bending knots are connected with low-cost,

– 55 – 56SUPPORT STRUCTURES Annual Report 2015-2017

Summary

The overall objective of the project is to assess the feasibility and applicability of the design concept and to investigate the necessary basics for planning, detailed design, and construction of this hybrid substructure. Transport and installation concepts are being developed for the substructure, the risk of scouring at the seafloor of the installed substructure is being analyzed, and the structural behavior of the installed components and connections is be examined on the basis of numerical and physical models in different levels of detail.

References

[1] http://www.hyconcast.uni-hannover.de [2] Göthel, O. (2014): Forschungs- vorhaben HyConCast – Hybride Substruktur aus hochfestem Beton und Sphäroguss für Offshore- Windenergienalgen, 4. Ingenieurtechnisches Kolloquium des Ingenieurbüros grbv, September 18, 2014, Hannover [3] Stümpel, M.; Marx, S.; Schaumann, P. Seidl, G. Göhlmann, J.: An Innovative Hybrid Substructure for Offshore Wind Turbines, Proceedings of the 27th International Offshore and Polar Engineering Conference (ISOPE) June 25-30, 2017, San Francisco, USA [4] Stümpel, M.; Marx, S.: An Innovative Hybrid Substructure Made of High-Strength Concrete and Ductile Cast Iron for Offshore Wind Turbines, DEWEK 2017 - 13th German Wind Energy Conference, October 17./18., 2017, Bremen Figure 1: Base construction [3, 4]

– 56 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201757

Integration of Calculation Methods in the Software Tool IGtHPile

Leibniz Universität Hannover Project Description proposed by Georgiadis (1983) for making Institute for Geotechnical Engineering corrections in layered soils has been suc- The enhancement of the state of knowledge cessfully introduced in the modular source Mauricio Terceros, Martin Achmus on the load-bearing behavior of pile-sup- code of the software tool IGtHPile. Based ported structures is accomplished through on the definition of the equivalent depth, Funding: Ventus efficiens; Lower Saxony the development of three individual issues the overlay method attempts to balance Ministry for Science and Culture (MWK) that from the author’s point of view are abrupt changes in the course of the maxi- Ref.Nr. ZN3024 considered of crucial importance, among mal bedding resistance of the layered soil others. These issues and the role played by over depth in all existing layers below the Duration: 01/2014 – 12/2019 the software tool IGtHPile are briefly ex- top layer as can be appreciated in Fig. 1. plained as follows. By doing so, the prediction effects of layered soil can be contemplated using homo- For the design of laterally loaded piles, geneous soil p-y approaches. the p-y methods are commonly applied Introduction according to guidelines such as API and For the purpose of obtaining an integrated DNVGL. However, these approaches were analysis of the load-bearing behavior of The offshore wind energy converters have calibrated on small-diameter piles in a ho- the horizontally loaded piles, the cohesive been mainly founded on pile-supported mogeneous soil. Several experimental and soils must also be thoroughly examined. In structures. The load-bearing behavior of numerical investigations refute the validity fact, the monopile foundations cannot be the axially and horizontally loaded piles of such p-y approaches for the applica- completely founded on soft soil due to its has not yet been fully clarified. An inten- tions on large-diameter piles as monopiles low resistance. Nevertheless, layered soils sive research is being carried out to enhance in a nonhomogeneous soil. On this basis, are frequently composed at least partially the geotechnical knowledge of the capabil- an evaluation of p-y approaches is carried by cohesive soils. For such reason, studies ity of pile foundations. Precise calculation out through comparative studies by using on the validity of the currently predomi- methods which have been developed by three-dimensional finite element simula- nantly used p-y approach for soft clay are the Institute for Geotechnical Engineering tion to verify the reliability of their appli- carried out, making a clear overview of (IGtH) are highly required to standardize cation on monopile foundation in layered the state of the art p-y methods possible. design procedures for an optimal treatment sand. In this sense, the overlay method An extensive assessment of six static p-y of the relevant effects that exist in offshore foundation structures. For such purpose, the software tool IGtHPile is developed simultaneously to facilitate the study of different subjects by implementing analytical methods for proposed or current calculation procedures.

Figure 1: Calculation procedure for consideration of soil layering according to Georgiadis (1983)

– 57 – 58SUPPORT STRUCTURES Annual Report 2015-2017

Figure 2: Calculation method for determination of the load-dependent stiffness matrixes according to Terceros (2015)

approaches has been conducted to find the tangent stiffness terms of the matrices are Engineering (IGtH) are possible through deviations compared to three-dimensional generated using the pile design software the availability of an independent software numerical simulation with finite element tool IGtHPile which can be calculated by tool such as IGtHPile in which the source methods. The investigated p-y methods the calculating procedure as shown in Fig. code can be freely modified to obtain ef- are replicated with the help of the software 2. For integrated dynamic simulations of ficient implementations. tool IGtHPile as such the diameter-depend- the whole structure, the results produced ent formulation proposed by Steven & Au- by IGtHPile can be considered to obtain dibert (1979) for consideration of large a significant simplification of the calcula- diameter pile foundation. Anyhow, the pre- tion process in relation to the foundation sented results are used to formulate a new, stiffness. generally valid p-y approach for laterally References loaded piles in soft clay. Summary [1] American Petroleum Institute (API): Recommended Practice 2GEO, In addition to the usual geotechnical de- Geotechnical and Foundation Design sign proofs of the load-bearing capacity In this report, three research issues con- Considerations, 1th Edition, April 2011 and permissible deformation at the pile sidered in the scope of the Ventus effi- and Addendum October 2014 head calculated using the p-y methods, the ciens project are presented with respect [2] Det Norske Veritas and foundation-soil system should have a suf- to horizontally loaded piles in which the Germanischer Lloyd (DNV GL): ficient stiffness to ensure that the eigen- load-bearing behavior is calculated using Offshore Standard DNVGL-ST-0126, frequencies of the overall structure remain three-dimensional numerical simulations. Support structures for wind turbines, within the acceptable excitation limits to Thereby, systematic comparative analyses Edition April (2016) avoid resonance effect resulting from the are carried out to evaluate the reliability [3] Georgiadis, M.: Development of dynamic and cyclic loads. The prediction of the analytical solutions. Based on the p-y curves for layered Soil, Proc of foundation stiffness is generally derived object-oriented programming technique, Geotech Pract Offshore Eng, Austin, TX, USA, ASCE, pp. 536-545 (1983) from the result of the p-y methods. A sim- the implementation of proposed or current [4] Steven, J.B.; Audibert, J.: plification has to be taken into account by calculation procedures has been success- Reexamination of p-y curve means of a single "nonlinear support node" fully undertaken within the software tool formulation, 11th Offshore Technology due to the high calculation effort gener- IGtHPile developed by the Institute for Conference, Paper No 3402 (1979) ated by the overall dynamic simulations. Geotechnical Engineering (IGtH). [5] Terceros, M.; Schmoor, K.; It means that the foundation stiffness is Thieken, K.; Achmus, M.: Software represented in the form of load-dependent The conclusion can be drawn that the IGtHPile Version 3 & Calculation stiffness matrices, which describe the re- evaluation and analysis of analytical meth- Examples Version 3, http://www.igth. spective influence of the load components ods proposed either by the state of the art uni-hannover.de/downloads (2015) on the deformation. The load-dependent design or by the Institute for Geotechnical

– 58 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201759

Ventus Efficiens – Joint Research for Raising the Efficiency of Wind Energy Converters within the Energy Supply System Subproject II – Efficiency Enhancement of Supporting Structures

Leibniz Universität Hannover The investigations carried out in SP II – Ef- ing from the fatigue behavior of welded Institute for Steel Construction ficiency enhancement of supporting struc- joints (WP 1) and grouted joints (WP 2), Peter Schaumann, Jan Kulikowski tures – concern the support structures of via bearing capacity of the joints of con- on- and / or offshore wind turbines. SP II is crete tower segments (WP 3) up to the soil- Institute of Building Materials Science divided into five work packages (WP) (see foundation interaction (WP 4). Beside the Ludgar Lohaus, Corine Otto, Kerstin Fig. 1). studies of the different single components, Elsmeier a numerical investigation of the whole sys- Institute of Concrete Construction The support structures together with the tem based on multibody system simulations Steffen Marx, Michael Hansen, wind turbine itself build a system subjected is carried out (WP 5). Large-scale experi- Steffen Hartwig to continuous cyclic and random loads. SP mental tests regarding the topics of the work II addresses different crutial points of such a packages within SP II are performed in the Institute for Geotechnical Engineering highly dynamically loaded structures start- Test Center for Support Structure. Martin Achmus, Mauricio Terceros

Institute for Structural Analysis Raimund Rolfes, Cristian G. Gebhardt, Christian Hente

Funding: Ventus efficiens; Lower Saxony Ministry for Science and Culture (MWK)

Ref.Nr. ZN3024

Duration: 12/2014 – 12/2019

Introduction

The ForWind research project ventus efficiens – Joint research for raising the efficiency of wind energy converters within the energy supply system is funded by the Ministry of Science and Culture of Lower Saxony over a five–year period. The project tackles the main challenges facing the improvement of efficiency for wind energy converters. It is subdivided into four subprojects (SP) covering energy conversion (SP I), supporting structures (SP II), drive train and grid connection (SP III) and the integration of the whole system (SP IV). Figure 1: Project Summary: Subproject II

– 59 – 60SUPPORT STRUCTURES Annual Report 2015-2017

Project Description Work Package 2 of high-strength grouts, a first approach to High-strength grouts are often used for the time-efficient testing method was devel- Work Package 1 production of connections in wind turbines oped. The fatigue limit state represents a driving e. g. grouted joints. For such constructions, factor in design of steel support structures. which are exposed to high-cyclic loadings Therefor the test frequency is increased in The fatigue design in current standards and and must be capable of bearing several hun- a fixed time interval. First results with the guidelines is based on S-N curves, which dred million load cycles during their service testing method show that the increase in are mostly based on fatigue tests up to 107 life, the verification at the fatigue lime state frequency and the time of each frequency load cycles. The range beyond 107 cycles is the decisive factor. depend on the used material. and the question of the existence of an endurance limit is of particular interest of on- The aim of the investigations is to deter- Work Package 3 going research [1]. mine fatigue behavior of high-strength In work package 3, the joint carrying capac- grouts without interferences by testing tech- ity of circular prestressed ring segments of Experimental investigation in the very high- nology. For this purpose, the testing inter- wind turbine towers is investigated by nu- cycle range (above 107) requires testing ferences have to be determined. Based on merical analysis and experimental tests. The facilities able to perform this high number these results, a time-efficient testing method current design rules [4], consisting of the of cycles within a reasonable time frame, for fatigue tests in an adequate testing time classical torsion theories according to Saint- allowing simultaneously tests on relatively will be developed. The following influences Venant for thin-walled closed and open large scale specimens to limit possible scale could be identified as essential influences cross-sections in combination with Cou- effects. For this reason a new testing device on the fatigue behaviour of high-strength lomb's static friction law, are not correct. operating at a frequency of up to 200 Hz grouts. Depending on loading situation this leads to has been developed. This device is able to both, unsafe results as well as unused potential. The aim of the investigations is to perform tests on axial loaded plate speci- Maximum stress level Smax: For stress levels develop new theories, as the current theories mens with a thickness of up to 20 mm and a lower than Smax < 0.75, specimens of high- width of up to 40 mm. It allows testing with strength grout reached significantly lower do not provide satisfactory results. a maximum stress amplitude of 70 kN and numbers of cycles to failure than expected. a stress ratio of 0.1 ≤ R ≤ 0.5. The operat- During these fatigue tests, a considerable in- The first experimental investigations dealt ing principle of the device is based on the crease in temperature up to 70° C has been with the determination of the coefficients resonance at the level of the first natural fre- observed [3]. of friction of smooth segment joints, which quency of the system composed of the test- confirmed the assumption of the currently valid value from the standard. In addition, ing device and the specimen [2]. The device Load frequency fP: Higher load frequencies, coefficients of friction were determined in will be used for fatigue tests of butt welded e. g. fP = 10 Hz result in: Higher numbers steel specimens and ductile cast iron speci- of cycles to failure for maximum stress lev- dynamic tests. For these tests, the normal stress was run dynamically during the test mens with and without notch effects in the els Smax ≥ 0.75 and lower numbers of cycles 7 9 and a coefficient of friction was determined range of 10 to 10 load cycles. At present to failure for maximum stress levels Smax < the validation of the developed testing tech- 0.75 [3] several times. The results showed an in- nology is performed. crease of the coefficients of friction after a Environmental conditions: The storage and dynamic load. With regard to the load-bear- The planned fatigue tests within WP 1 will testing under water has a significant influ- ing capacity in the whole structure, numeri- be used to investigate the endurance limit of ence on the fatigue behaviour. The numbers cal investigations as well as experimental fatigue strength. The influence of the devel- of cycles to failure of specimens stored and tests were carried out. The numerical inves- oped testing technology on the test results tested under water are lower than those of tigations illustrate the need for action with is examined. For this purpose, the obtained the specimens which were tested in dry en- regard to a realistic description of the tor- results are compared with the test results vironment. The numbers of cycles to failure sional bearing capacity. The investigations carried out with the conventional testing ma- of sealed specimens tend to be lower than prove that the currently valid theories only chines. Furthermore numerical and analytical those of unsealed specimen. inadequately describe the load-bearing ca- models for prediction of fatigue behaviour of pacity. An analytical formulation of the ex- welded steel components will be developed. Specimen dimensions: Further investiga- isting joint bearing capacity also illustrates The influence of the thickness effect and the tions are necessary because the results from the inadequacy of the current approaches welding process as well as the use of high- literature and from own experiments on this [5]. To confirm the identified deficiencies in strength steels will be studied. These models influence are not yet clear. the design, small-scale tests were carried out can eventually be used in the design process on segmented plastic pipes and a large-scale and will allow further optimization of sup- Based on the results of the identification of test was carried out on a segmented rein- port structures for wind energy converters. essential influences on the fatigue behaviour forced concrete tower at the Test Centre for

– 60 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201761

Support Structures (TTH). In the test with cessfully introduced in the software tool electronics, supply and others. From a the segmented plastic pipes with a diameter IGtHPile. structural point of view, it should be empha- of 20 cm and a total length of 1.2 m the nu- sized that some components like the blades merically determined load carrying capacity For the purpose of obtaining an integrated experience very complex movements with values could be approximately confirmed, analysis of the load-bearing behavior of the large displacements and rotations, and in ad- however the extremely pronounced creep horizontally loaded piles, the cohesion soils dition multiple interactions take place. The behavior made the test difficult. The results must also be thoroughly examined. In fact, standard numerical tools are very restricted clearly show that the current theories are not the monopile foundations cannot be com- due to the inability to predict critical and applicable to this construction. The large- pletely founded on soft soil due to its low supercritical behaviour. The main objective scale reinforced concrete test was carried resistance. Nevertheless, layered soils are of work package 5 is the development of a out on an approximately three-meter-high frequently composed at least partially by parametric simulation model for offshore tower with a diameter of 60 cm. This cor- cohesive soils. For such reason, studies on wind turbines, which is based on the in- responds to a scale of 1/10 to a real tower of the validity of the currently predominantly house-code DeSiO. The idea is to consider a wind turbine. The measured load capaci- used p-y approach for soft clay are carried the whole system comprised by the struc- ties also confirm that the current theories are out, making a clear overview of the state of ture, soil, wind, water, controls and addi- not to be used to describe the torsional load the art p-y methods possible. An extensive tional involved sub-systems as elements of a capacity. However, the results also show a assessment of six static p-y approaches has single coupled dynamical system. With this difference between the carrying capacities been conducted to find the deviations com- model, it is expected to analyze controlled from the numerical calculations and the pared to three-dimensional numerical simu- and uncontrolled unsteady cases as well as experimental tests. The differences are cur- lations. The investigated p-y methods are non-linear phenomena. This is intended to rently attributed to possible crack formation implemented in the software tool IGtHPile. carry out numerical simulations in the time in the segments. domain and relies upon: i) the development In addition to the usual geotechnical design of a structural model based on multibody The previous investigations on segment proofs of the load-bearing capacity and per- systems and the finite element method; ii) joints of wind turbines in work package 3 missible deformation at the pile head, the the implementation of a soil model; iii) the have shown that the current theories are not foundation-soil system has to produce a adjustment and adaptation of a model based correct and that new analytical solutions sufficient stiffness to ensure that the eigen- on the boundary element method for the have to be found. frequencies of the overall structure remain aerodynamics and hydrodynamics; and, iv) within the acceptable excitation limits to the development of a multibody representa- Work Package 4 avoid resonance effect resulting from the tion for the drivetrain. From 2015 to 2017, Intensive research is being carried out to en- dynamic and cyclic loads. The prediction the focus was set on the most complex hance the geotechnical knowledge of the ca- of foundation stiffness is generally derived components, i.e. the structural module and pability of pile foundations. In this respect, from the result of the p-y methods. A sim- the soil. Regarding the structural module, the software tool IGtHPile is developed si- plification by a single "nonlinear support a very important aspect was the inclusion multaneously to facilitate the study of dif- node" has to be considered due to the high of geometrically exact structures within a ferent subjects. calculation effort generated by the over- multibody framework, including beam and all dynamic simulations. It means that the shell models for composite-multilayer-hy- For the design of laterally loaded piles, the foundation stiffness is represented in the perelastic materials. The resulting semi-dis- p-y methods are commonly applied ac- form of load-dependent stiffness matrixes, cretized equations of motion and kinemati- cording to guidelines such as [6] and [7]. which describe the respective influence of cal restrictions are solved in time domain However, these approaches were calibrated the load components on the deformation. by means of an energy-conserving/decay- on small-diameter piles in a homogeneous The load-dependent tangent stiffness terms ing, momentum-conserving integration. soil. Several experimental and numerical of the matrixes are generated using the pile The flexible members of a wind turbine are investigations refute the validity of such p-y design software tool IGtHPile. then connected with the multibody formal- approaches for the applications on large- ism. Different analysis types to perform diameter piles as monopiles in a nonhomo- Work Package 5 dynamic and static calculations have been geneous soil. On this basis, an evaluation of There is a steadily growing interest in considered, as well as the determination of p-y approaches is carried out through com- characterizing the dynamical behaviour of eigenfrequencies and eigenforms multibody parative studies by using three-dimensional offshore wind turbines. The complexity of system with an adapted modal analysis. finite element simulations to verify the re- such mechanical system is high and very liability of their application on monopile different fields are involved, e.g. structural The calculation software was validated on foundation in layered sand. In this sense, the dynamics (blades, hub, nacelle, tower, pile the basis of examples from the relevant lit- overlay method proposed by [8] for making and drivetrain), soil dynamics, aerodynam- eratures and published in high-ranking jour- corrections in layered soils has been suc- ics, hydrodynamics, controls, generation, nals [9, 10]. The next steps are to consider

– 61 – 62SUPPORT STRUCTURES Annual Report 2015-2017

the interaction of the surrounding fluid with technical Engineering are possible through the structure using a strong coupling scheme the availability of an independent software between the multibody system, finite ele- tool such as IGtHPile. ment method and the boundary element • A parametric model for the whole system method. wind turbines including interactions with the environment is further developed Summary

The optimal efficiency enhancement of support structures for wind energy converters can be reached only by considering the whole system (see Fig. 2). The three-year work on the project have shown the importance of close cooperation between the different research fields. In this period the following key results have been reached:

• The new testing device operating at a frequency of up to 200 Hz for fatigue tests of butt welded steel specimens and ductile cast iron specimens has been developed • The essential influences on the fatigue behaviour of high-strength grouts for a time- efficient testing method could be identified • The previous investigations on segment joints of wind turbines have shown that the current theories are not correct and that new analytical solutions have to be developed. • The evaluation and analysis of analytical methods proposed either by the state of the art design or by the Institute for Geo- Figure 2: Wind turbines with interactions

References

[1] Schaumann S.; Steppeler S.: Fatigue [5] Hartwig, S., Marx, S.: Zum Torsions- [8] Georgiadis, M: Development of p-y tests of axially loaded butt welds up to tragverhalten extern vorgespannter Kreis- curves for layered Soil, Proc Geotech very high cycles, Procedia Engineering 66 segmente mit trockenen Fugen. Beton- Pract Offshore Eng, Austin, TX, USA, (2013), pp. 88-97. und Stahlbetonbau, Jahrgang 112, Heft 11, ASCE, pp. 536-545 (1983) [2] Alt A.: Patent DE 10204258 B4: Ernst & Sohn, 2017, S. 740–746. [9] Gebhardt C.G.; Rolfes R.: On the Prüfvorrichtung zur Dauerschwing- [6] American Petroleum Institute (API): nonlinear dynamics of shell structures: prüfung von Prüflingen 2007. Recommended Practice 2GEO, Geotechnical Combining a mixed finite element [3] Elsmeier, K., Hümme, J., Oneschkow and Foundation Design Considerations, 1th formulation and a robust integration N., Lohaus, L. (2016): Prüftechnische Edition, April 2011 and Addendum October scheme, Thin-Walled Structures (2017), Einflüsse auf das Ermüdungsverhalten 2014 vol. 118, pp. 56-72. hochfester feinkörniger Vergussbetone, [7] Det Norske Veritas and Germanischer [10] Gebhardt C.G.; Hofmeister B.; Beton- und Stahlbetonbau, Jahrgang 111, Lloyd (DNV GL): Offshore Standard Hente C.; Rolfes R.: Nonlinear dynamics Heft 4, Ernst & Sohn, 2016, S. 233-240. DNVGL-ST-0126, Support structures for of slender structures: a new object- [4] Grünberg, J., Göhlmann, J. (2013): wind turbines, Edition April (2016) oriented framework, Computational Concrete structures for wind turbines. Mechanics (2018), pp. 1–34. Ernst & Sohn a Wiley brand, Berlin, Germany

– 62 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201763

Bearing Behavior of Suction Bucket Foundations under Tensile Loading in Non-cohesive Soils

Leibniz Universität Hannover loading of suction buckets. This is why Institute for Geotechnical Engineering a new testing facility was erected at the IGtH, where the tensile bearing behavior Martin Achmus, Patrick Gütz can be investigated with regard to differ- Funding: German Research Foundation ent loading conditions (Gütz et al., 2017).

Ref.Nr. 266046182 In Fig. 1, results of selected model tests under various heave rates indicate a strong Duration: 09/2015 – 08/2019 dependency of the tensile resistance (solid lines) and negative pressure (dashed lines) on the applied heave rate. These results Figure 1: Vertical resistance vs. displace- confirm the findings of previous model ment for various heave rates (L=250 mm, Introduction tests (Bye et al. 1995, Kelly et al. 2006). D=260 mm) However, the maximum tensile resistance The Institute for Geotechnical Engineering at higher loading rates requires larger dis- (IGtH) conducts research in the field of ten- placement of the suction bucket, which sile bearing behavior of suction bucket foun- possibly exceeds the allowable heave in dations. This study aims at the improvement terms of serviceability requirements. of the applicability of multipod foundation for offshore wind turbines (OWT). A main Besides monotonic loading, the behavior focus in this research project is the assess- under tensile cyclic loads is investigated ment of the partially drained loading con- (cf. Fig. 2). The applied swell load has a dition, where negative pressure beneath the mean load being equal to the drained ca- suction bucket’s lid contributes to the total pacity of the suction bucket and a load tensile resistance. Until now, there is no de- amplitude of 50% of the drained capacity

sign concept available allowing the estima- (tensile load in the range of 0.5 Fdrain and tion of the negative pressure, which depends 1.5 F ). The loading causes a continuous drain Figure 2: Displacement vs. number of on the loading (magnitude and duration), heave of the model. Furthermore, it can be cycles in physical model (Exp.) for cyclic permeability of the soil and the geometry of seen that the increasing number of load cy- loading exceeding the drained capacity the suction bucket. If the negative pressure cles leads to an accumulation of negative (L=500 mm, D=510 mm) is not taken into account, the economic and pressure beneath the suction bucket’s lid, ecologic advantages of the suction bucket which results in corresponding force Δu∙A foundations are diminished. However, the (cf. Fig. 3). mobilization of the negative pressure requires certain displacement of the founda- Numerical back calculation and extra- tion so that the design of suction bucket polation foundations in partially drained condition is A finite element (FE) model was estab- more driven by serviceability specifications lished for investigating the suction buck- rather than the maximum resistance. et’s tensile bearing behavior adopting a coupled pore water-stress-analysis (cf. Achmus & Gütz, 2016). First, the model Project Description was used for the back calculation of the physical model tests. Fig. 4 and Fig. 5 Physical model tests show two back calculations of partially Although several publications in terms of drained monotonic model tests. The results Figure 3: Vertical force vs. number of physical modeling of suction bucket foun- are in acceptable agreement. Further back cycles in physical model (Exp.) for cyclic dations are available, these studies are calculations are conducted for the verifica- loading exceeding the drained capacity not comprehensively covering the tensile tion of the FE model. The investigation of (L=500 mm, D=510 mm)

– 63 – 64SUPPORT STRUCTURES Annual Report 2015-2017

multiple model scales under comparable Summary loading conditions provides insight in the scale effect. Based on these findings, the The research project on the tensile bearing numerical model can be used for extrapo- behavior of suction bucket foundations in lation. non-cohesive soil provides a comprehensive overview on a complex geotechni- Parametric studies cal issue and adopts several methods for Parametric studies with arbitrary loading investigating the versatile interactions of conditions are executed with the verified this type of foundation. Based on physi- FE model. For instance, perfect cyclic swell cal model tests in conjunction with finite loads considering various load amplitudes element simulation, the bearing behavior

(from 0 kN to Fmax) and frequencies are Figure 4: Vertical resistance vs. displace- can be understood properly and these new investigated. Fig. 6 summarizes some of ment in physical model (Exp.) and nume- insights will be used to develop a straight- these simulations for prototype dimensions rical back calculation (FE) for monotonic forward rheological model, which simpli- (L=D=9 m). Obviously, load amplitudes heave rate (L=500 mm, D=510 mm) fies the design process of suction bucket less or equal to the drained capacity cause a foundations. vertical displacement during one cycle, but induce no continuous heave of the suction bucket. On the one hand, the suction bucket experiences an ongoing heave, as the load exceeds its drained capacity. On the other hand, the amount of displacement is significantly affected by the surplus load. In References case of high load amplitudes, the interface at the suction bucket’s skirt behaves plastic [1] Bye, A.; Erbrich, C.; Rognlien, B.; and is therefore predominantly defining the Tjelta, T. I.: Geotechnical Design of linear trend of accumulating heave. Bucket Foundations, Proceedings of Figure 5: Vertical resistance vs. displace- the Offshore Technology Conference (OTC), pp. 869–883 (1995) A qualitative comparison of Fig. 2 and ment in physical model (Exp.) and numerical back calculation (FE) for monotonic [2] Gütz, P; Achmus, M.; Thieken, K.; Fig. 6 confirms that a suction bucket foun- heave rate (L=250 mm, D=260 mm) Schröder, C.: Model testing of suction dation experiences ongoing heave for ten- bucket foundations under tensile loa- sile loads in excess of the drained capacity. ding in non-cohesive soil, Porceedings It is necessary to state that in case of the of the 8th International Conference on physical model test shown in Fig. 2 the Offshore Site Investigation and Geo- load on the suction bucket is not totally re- technics (OSIG), pp. 578–585 (2017) leased causing a permanent tensile force, [3] Kelly, R. B.; Houlsby, G. T.; which is in contrast to the exemplary re- Byrne, B. W.: Transient vertical loading sults of the numerical calculations shown of model suction caissons in a in Fig. 6. pressure chamber, Géotechnique 56(10), pp. 665–675 (2006) [4] Senders, M.: Suction caissons in Rheological model sand as tripod foundations for offs- Finally, the findings of the physical model hore wind turbines, Computers tests and numerical simulation are trans- Figure 6: Displacement vs. number of Ph.D. thesis. The University of ferred to a rheological model, where 4 cycles in numerical simulation for various Western Australia, Australia (2008) springs and 1 damper represent the re- cyclic loading amplitudes (L=D=9 m) [5] Achmus, M. & Gütz, P. (2016): sistances of a suction bucket foundation Numerical modeling of the behavior (Senders 2008). The complex interaction of bucket foundations in sand under of these components requires the defini- cyclic tensile loading, Proceedings of tion of individual non-linear equations maximum resistance under certain loading, the 6th international conference on for every component of the model. After the rheological model describes the entire structural engineering, mechanics and verification of the rheological model, it can load displacement behavior. This aspect is computation, SEMC 2016, Cape be used as a tool for the design of suction essential with regard to the stiffness of the Town, South Africa, pp. 2085-2091, https://doi.org/10.1201/9781315641645. bucket foundations. In comparison to other foundation or the assessment in terms of methods available, which only predict the serviceability criteria.

– 64 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201765

Development and Demonstration of a Rapid and Cost-Effective Installation Concept for Offshore Wind Turbines

Universität Bremen Project Description cludes weather data, vessel data as well as Institute for Integrated Product all logistical process times and costs, is used Development (BIK) Offshore wind energy is a key component to evaluate the economic efficiency of the of the German Energiewende and stead- logistics concept. Klaus-Dieter Thoben, ily more capacity will be installed in the Andreas F. Haselsteiner, future. Average wind velocities and con- The goal of the project “SKILLS” is the Marco Lewandowski, sequently the generated power are higher proof of a feeder ship concept, which could Stephan Oelker, Jan-Hendrik Ohlendorf offshore than onshore. Nevertheless, at replace the current concept with the shut- tling installation vessel. If the project suc- Funding: Federal Ministry for Economic the moment offshore electricity generation Affairs and Energy - Homepage (BMWi) costs are higher. Major cost drivers are the ceeds, the installing offshore wind turbines logistics processes for installing the wind will become economically more efficient Ref.Nr. 0325934B turbines. Currently, the big and heavy tur- and consequently the electricity generation bine components can be lifted off a floating costs will fall. Duration: 09/2015 – 12/2019 transport vessel only at very good weather conditions. Consequently, companies do not Current Project Status and Next Steps charter inexpensive transport vessels to feed We performed an economic analysis us- the components to the jacked-up installation ing discrete event simulation to determine vessel (“feeder ship concept”). Instead, the the costs of the feeder ship concept. The more expensive installation vessel shuttles analysis showed that under the assumed between the wind farm and base port (“all- model assumptions the feeder ship concept in-one concept”). can save costs compared to an all-in-one concept [1]. Further, we performed a mo- The feeder ship concept would become eco- tion analysis of the feeder ship to determine nomically effective if the lifting operations the expected motions of the loaded compo- could be performed at harsher environmen- nents. These values served as requirements tal conditions. Then, there would be more in the development of novel lifting devices. and longer weather windows. Therefore, Lastly, we systematically analyzed available in the joint research project “SKILLS” re- technical solutions to describe and improve searchers from the University of Bremen the state-of-the-art lifting process [2]. Next, and practitioners from the companies Sen- we will perform offshore test. We will ana- vion GmbH, Amasus Offshore B.V. and Jan lyze the motions of the components blade, de Nul NV develop technical solutions to nacelle and hub during the lifts. lift wind turbine components off a floating transport vessel.

To model the lifting processes, the dynamics of the transport vessel, the turbine components and the crane are simulated numerically. These dynamics are governed by loads induced by waves and wind. We use the software ANSYS AQWA to simulate the hydrodynamic excitation of the vessel. Based on these results we analyze the lifting process using multi-body dynamics simulations. A probabilistic simulation, which in-

– 65 – 66SUPPORT STRUCTURES Annual Report 2015-2017

Figure 1: Copyright: Senvion GmbH 2014

References

[1] Oelker, S., Ait-Alla, A., Lütjen, M., Lewandowski, M., Freitag, M., Thoben, K.-D.: A simulation study of feeder-based installation concepts for offshore wind farms. In Proceedings of the 28th International Ocean and Polar Engineering Conference (ISOPE 2018), International Society of Offshore and Polar Engineers (ISOPE), pp. 578–583 (2018.) [2] Haselsteiner, A. F., Ohlendorf, J.-H., Oelker, S., Ströer, L. Thoben, K.-D., Wiedemann, K., De Ridder, E., Lehmann, S: Lifting wind turbine components from a floating vessel: A review on current solutions and open problems. Accepted for publication in Journal of Offshore Mechanics and Arctic Engineering.

– 66 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201767

WinConFat – Fatigue of Materials of Onshore and Offshore Wind Energy Structures Made of Reinforced and Prestressed Concrete under High Cycle Loading

Subproject – Operational Effects on Concrete Fatigue

Leibniz Universität Hannover Project Description documented in the literature. Fatigue tests Institute of Concrete Construction, carried out on prismatic specimens show Institute of Building Materials Science The aim of the "WinConFat" subproject that with eccentric load application, stress "Operational effects on concrete fatigue" is rearrangements take place in the concrete Steffen Marx, Ludger Lohaus to close existing gaps in knowledge about compression zone. Furthermore, it can Funding: Federal Ministry of Economic the fatigue behavior of concrete for wind be seen that with increasing eccentricity Affairs and Energy (BMWi) turbines and to gain a better understand- the numbers of cycles to failure increase. ing of the fatigue processes in concrete These results are all based on studies on Ref.Nr. 0324016A under cyclic loading. Only in this way the normal-strength concrete and are not read- existing constraints in the design of con- ily transferable to high-strength concrete. Duration: 11/2016 – 10/2020 crete support structures for wind turbines In addition, there are only few results in the can be reduced and a safe and economic Very High Cycle Fatigue range (N > 2∙106), design can be guaranteed. In the respective which is particularly important for wind tur- work steps of this subproject influences bines. In this subproject, experimental and from stress redistribution, from different numerical studies on the influence of stress Introduction loading frequencies and velocities as well redistribution on the fatigue resistance of as the influence of water saturation and real concrete components of wind turbines Structures for wind turbines are subjected to chloride transport on the strain develop- are derived. strong dynamic loads due to wind and wave ment and the numbers of cycles to failure loading, which make their foundation and of high strength concretes are investigated. The loading frequency and loading veloc- tower constructions susceptible to fatigue. Another fundamental element for the fa- ity have an influence on the fatigue resist- At present, towers for wind turbines, and tigue design is the determination of the ance of concrete. As the loading frequency

for offshore wind turbines even the founda- basic value of the fatigue strength fcd,fat. In increases, usually the fatigue resistance also tions, are still predominantly built as steel addition, recommendations for monitoring increases. Due to this frequency depend- constructions. The efficient application of methods based on ultrasonic measurements ence, test results with different loading reinforced concrete and prestressed con- are developed which can help to evaluate frequencies are not fully comparable with crete for those structures is currently being the remaining service life of structures for each other. Furthermore, wind turbines are prevented by a conservative design concept wind energy turbines. These studies are loaded in the low-frequency range, whereas against fatigue, especially for high-strength indispensable for a better transferability of fatigue tests in the laboratory are usually concrete. In order to improve this design laboratory-based small-scale specimens to carried out at higher loading frequencies. concept, which dates back to the 1990s, ex- the behavior of real structures. Thus, this Increased loading frequencies are required tensive fatigue tests on concrete, reinforcing sub-project provides the basis for low-cost, in order to be able to generate the load cy- steel and their bond behavior are required. safe and efficient structural support struc- cles to be expected within the service life This is made possible by the 48-month joint tures in concrete, which can play an impor- of the wind turbine. This can lead to unsafe research project "WinConFat" with the co- tant role for an environmentally friendly and results compared to real conditions. For this operation of seven university institutes, the resource-saving, reliable and affordable en- reason, this subproject investigates the in- Bundesanstalt für Materialforschung und ergy supply in Germany. fluence of loading frequency on the fatigue -prüfung (BAM), the German Committee behavior so that it can be taken into account for Structural Concrete (DAfStb) and the For the influence of stress redistribution in the interpretation of test results. German Society for Concrete and Construc- on the fatigue behavior of concrete there is tion Technology (DBV). only a very limited amount of test results

– 67 – 68SUPPORT STRUCTURES Annual Report 2015-2017

The determination of the basic value of the accumulated damage sum. Whether this Summary

the fatigue strength fcd,fat together with the additional reduction is justified is unclear. definition of the S-N-curves has a decisive There is a lack of knowledge on the fa- On the basis of the coordinated joint re- influence on the fatigue design. The basic tigue behavior of water-saturated concretes, search project "WinConFat" the various, as value of fatigue strength includes a reduc- which will be closed in this subproject by yet insufficiently researched, influences on tion term dependent on the concrete com- experimental and theoretical investigations. the fatigue behavior of high-strength con- pressive strength. This additional reduction cretes will be experimentally determined in fatigue stress was presumably taken from At present, there are no methods for analyz- and scientifically described. In addition, the static design, which was intended to en- ing the fatigue condition of wind turbines. the investigations are carried out up to the sure the relationship between the concrete In other academic disciplines continuous load cycle range of N = 107, which has great strength tested in the laboratory and that is monitoring is an established tool for life cy- importance for wind turbines. Furthermore, actually present in the structure. But this is cle analyses. The monitoring of strains and recommendations for monitoring methods now solved by other factors. In addition, deformations is also increasingly used. Due based on ultrasonic measurements are de- there was a lack of practical experience with to the low durability of the sensors as well veloped which can help to evaluate the re- high-strength concretes in determining the as the impossibility of describing degrada- maining service life of structures for wind basic value of fatigue strength. Concretes tion processes for life-time monitoring, the energy turbines. The expected research re- with compressive strengths greater than mentioned measured variables are only par- sults should serve as the basis for innova- C80 were therefore not included in the de- tially suitable. On the other hand, ultrasonic tive, cost-effective and sustainable develop- velopment. In modern high-performance measurement methods in various scientific ments of tower and foundation structures concretes the additional reduction has a disciplines (e.g. medicine, biology) already made of high-strength concrete and ensure dramatic effect on the basic fatigue strength offer an enormous performance potential. longer service lives of existing and new

fcd,fat. The performance of modern high-per- But also in other construction-related ar- concrete structures for wind turbines. formance concretes can therefore hardly be eas (e.g. steel construction, natural stone used and their use, especially in the area of testing) or in the field of geosciences, ul- high cyclic stresses, is thereby constrained. trasonic measurement technology is used Another aim of this subproject is the statisti- successfully. The application is clearly un- cal determination of the strength-dependent derdeveloped for concrete constructions,

reduction of fcd,fat. This knowledge is of but especially in laboratory experiments, great importance for the economical use of the method shows that it is ideally suited for concrete in wind turbines. condition assessment and damage monitoring. But in real component, the temperature Results about the fatigue behavior of water- and humidity as well as stress conditions saturated and chloride-stressed concrete of the component and the posthardening of are limited. The available results show that the concrete can influence the measurement water saturation without chlorides strongly results. There are no universal approaches influences the fatigue resistance of concrete. that describe the extent of these influences. There is also a dependence of the fatigue re- An assessment for the remaining service sistance on the loading frequency. It is fur- life is not yet possible. For this purpose, ther assumed that the compressive strength, theoretical and experimental investigations the porosity and the tightness of the con- are carried out in order to be able to assess crete influence its fatigue resistance. Also, the fatigue process on the basis of ultrasonic the basic mechanisms of the fatigue behav- measurements and to conclude by moni- ior of water-saturated concretes are not clar- toring measures on the remaining service ified. In this context, any scale effects and life of wind turbines. This not only brings the influence of seawater must be described advantages in the design of structures for to ensure the transferability of laboratory new wind energy turbines, but also creates results on small-sized specimens to real the basis for the continued use of existing components. The S-N-curves used for the support structures after the 20-year design fatigue design of offshore wind turbines are life and thus also has a positive effect on the based on experimental investigations under costs of wind energy. dry conditions. S-N-curves based on fatigue tests under typical offshore conditions do not exist. These conditions can be considered only simplified by a reduction factor in

– 68 – Annual Report 2015-2017 SUPPORT STRUCTURES Annual Report 2015-201769

ThermoFlight – Concept for the Development of an Optimized Maintenance and Inspection Method for Off-Shore Wind Turbines Using Thermography and SHM as Non-Destructive Testing Technologies in Combination with Unmanned Aerial Vehicles

Universität Bremen Bremen Institute for Metrology, Automation and Quality Science (BIMAQ)

Christoph Dollinger

Funding: BIS Bremerhavener society for investment promotion and urban development mbh

Ref.Nr. 59203/4-ZB

Duration: 01/2017 – 08/2018

Project Description

The planned expansion of offshore wind energy in Germany requires the maintenance Figure 1: Thermographic image of the inner structure of a rotor blade and operation to be efficiently organized both economically and ecologically for at least 25 years for a growing number of wind energy turbines. The maintenance and For the nondestructive testing of the inner testing teams are confronted with new chal- structure of offshore wind turbine rotor lenges offshore. This is due to short time blades the potential of thermographic im- windows as a result of difficult weather con- ages taken from unmanned aerial vehicles is ditions as well as high safety requirements investigated. The resulting requirements in and regulations. terms of weight and power supply limit the variety of suitable thermographic cameras Especially the rotor blade tests by industrial and due to that affect the available spatial climbers are difficult to plan under these and thermal resolution. harsh conditions. With the objective of min- References imizing the use of personnel for inspections In order to characterize the method, thermo- and the resulting downtimes of the offshore graphic measurements, both with high-end [1] Dollinger, C.; Balaresque, N.; wind turbines, the use of non-destructive and light-weight thermographic systems, Sorg, M.; Goch, G.: Thermographic testing methods and structural health moni- in standstill for deep structural (see fig. 1) measurement method for turbulence toring is investigated. Especially in combi- and on the running wind turbine for surface boundary layer analysis on wind nation with unmanned aerial vehicles, these near defects are performed [1]. The objec- turbine airfoils. In: AWEA Wind Power technologies can contribute to an efficient, tive is to compensate the observed technical 2014 Conference and Exhibition, safe, energy and material-optimized rotor limitations by the use of image processing Las Vegas, May 5-8, 2014. blade inspection process. in terms of a contrast enhancement.

– 69 – 70WIND POWER SYSTEMS Annual Report 2015-2017

Development of Nacelle-Based Lidar Technology for Measurement of Power Characteristic and Control of Wind Turbines (LIDAR II)

Carl von Ossietzky Universität Oldenburg of wind energy converter systems (WEC). cedures were derived for the reconstruction Institute of Physics Cooperating partners in the project were of wind fields from lidar measurements Research Group: Wind Energy Systems ForWind – University of Oldenburg, Stutt- and for the evolution of the wind field be- Research Group: Marine Physics Group, gart Wind Energy (SWE) – University of tween measurement and rotor plane from Research Group:Turbulence, Wind Energy Stuttgart, and Adwen GmbH. DEWI GmbH both theoretical turbulence description and Stochastics and the Federation of German Windpower and measurement campaigns. Moreover, a e.V. (FGW) participated in the project as comparative procedure was developed to Paul G. Hofmeister, Martin Kühn, Martin Kunze, Rainer Reuter, contractors. The project was headed by evaluate lidar measurements and optimize Matthias Wächter Prof. Dr. Martin Kühn, first based in the the scanning configuration of lidar devices. (Universität Oldenburg), University of Stuttgart, from April 2010 on This procedure was validated using real- in ForWind – University of Oldenburg. world measurements by estimating the ro- Po Wen Cheng, Steffen Raach, tor-effective wind speed with the help of an David Schlipf, Ines Würth observer design from measurement data of (Universität Stuttgart) Project Description a wind turbine.

Funding: Federal Ministry for Economic To exploit nacelle- or spinner-based lidar Based on these findings, different lidar- Affairs and Energy (BMWi) measurements for the operation of a WEC, based control strategies were developed and firstly different statical and dynamical pro- examined with respect to their potential and Ref.Nr. 0325216A

Duration: 11/2010 – 01/2015

Introduction

The joint research project Development of nacelle-based lidar technology for power performance measurement and control of wind turbines (LIDAR II) aimed at the primary goal of developing crucial technological building blocks for future large- scale multi-MW wind turbines in offshore wind farms of power plant scale, using new and comprehensive control and monitoring strategies. It was carried out from 1.11.2010 to 31.1.2015 in the context of the accompanying research performed at the German offshore test site alpha ventus. Primary research goals have been the development of a robust and industrial-suited nacelle-based wind lidar, the development of lidar-based control strategies for load reduction and energy yield improvement as well as providing methods for lidar-based measurement Figure 1: Measurement principle of a nacelle-based lidar. With the main direction towards the inflow, wind velocities are scanned along a trajectory. As an example a circular and monitoring of the power performance trajectory is shown.

– 70 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201771

Figure 2: The robust and cost-efficient lidar system “Whirlwind 1” in the laboratory (left) and mounted offshore on AV07 in the alpha ventus wind farm, together with the SWE scanning lidar (right)

applicability. Especially successful was the ing in an offshore-suited prototype. With its control concept of predictive pitch control robustness, low manufacturing costs, and its which showed evidence of substantial re- suitability for integration into the spinner of duction of structural fatigue loads in simula- a WEC, it offers a unique combination of tion studies based on an Adwen WEC. To features. Industrialization of the lidar has this end a simulation technique was devel- started directly after the project finished. oped which enable a realistic reconstruction of the measurement situation with different Furthermore, dedicated methods for the configurations, using simultaneously meas- lidar-based measurement of a wind tur- ured lidar and turbine data. Furthermore the bine's power performance were developed, potential of energy yield increase by reduc- which provide the power curve following tion of yaw misalignment and improved the principles of IEC 61400-12-1 on the control of rotational speed (“lambda-opt one hand and the dynamical, instationary tracking”) were investigated. For rotational power characteristic based on 1 Hz meas- speed control, real-world measurement data urements on the other hand. It was shown of a research wind turbine from a different that direct measurement of the wind inflow project could be used. The investigations by a nacelle-based lidar results in a shorter and experiments confirm applicability and measurement time and a significant reduc- advantages of the newly developed concepts tion of uncertainty in both cases. Based on and thus provide important technology ele- the dynamical power characteristic, further- ments for future cost-optimized multi-MW more a monitoring approach was developed References wind turbines. allowing of a continuous surveillance of the power performance as well as quick and [1] Chen, P. W.; Hofmeister, P. G.; sensitive detection of deviations. Kühn, M.; Kunze, M.; Raach, S.; Summary Reuter, R.; Schlipf, D.; Wächter, M.; Würth, I.: LIDAR II Final Report, https://www.tib.eu/de/suchen/ Substantial results of the project LIDAR II download/?tx_tibsearch_search%5Bd are on the one hand the development and ocid%5D=TIBKAT%3A860329518&cH testing of the robust, cost-efficient, and ash=755cb360db4ee2b2f5d44e74f067 industrial-suited wind lidar Whirlwind 1 4e6b#download-mark starting from conceptual design and result-

– 71 – 72WIND POWER SYSTEMS Annual Report 2015-2017

GeCoLab – Test Bench for Generators and Converters Systems

Leibniz Universität Hannover tion phase in 2013, technical infrastructure increasing complexity of electrical grids Institute for Drive Systems and Power and electrical components were supplied in and rapid emerging of frequency converter- Electronics 2014. Power electronics, transformers and based generators encourage a much deeper power distribution followed by the end of study of the converter-generator interaction. Bernd Ponick, Axel Mertens 2014. With the hall nearly filled to capac- GeCoLab has been created to deal with ity, the installation phase started in January these challenges. Funding: Federal Ministry for Economic Affairs and Energy (BMWi) 2015. Everything arrived in time and fitted into place. The last items – the electrical ma- The system diagram (see fig. 1) gives an Ref.Nr. 0325398 chines – arrived in early summer 2015, then impression on the universal possibilities of- we could start off with the operation step by fered by the test bench facility, and this not Duration: 01/2012 – 09/2016 step. Finally, first start-up tests under power only for the test machines and converters were successfully made in autumn 2015. specifically equipped for testing, but also for specimens, be it machines or converters, provided by customers. An impression of Introduction Project Description the test bench shows the fig. 2.

It is done! In the last two years, our vision The GeCoLab is a universal motor and Equipment parameters turned into reality: A new large-scale Uni- generator test bench which enables a deep - rated voltage up to 690 V versal test bench GeCoLab to investigate investigation of electrical machines and - permanent magnet synchronous machine steady-state and dynamic properties of elec- converters. The study of megawatt genera- with respective full converter -1 trical machines and converters including tors with or without the power electronic [PN = 1.2 MW, nN = 375 min converter/machine interactions is almost is an ongoing important topic due to the (0 - 750 min-1)] completed. GeCoLab stands for Generator- very different phenomena and uncertain- - doubly-fed induction machine with Converter-Laboratory. After the construc- ties of large electrical drive systems. The respective wind converter

Figure 1: Single line diagram

– 72 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201773

-1 generator interactions and their influence [PN = 2.08 MW, nN = 1780 min ] With the test bench it is possible - converter-based grid modelling - Test your motor or generator prototype on other system components, such as bearings and gearboxes. [SN = 4.4 MVA; unbalance, freely adjust- with different types of converters for able frequency and harmonics] fault diagnostics, performance valida- - maximum component weight 20 t tion, and analytical modelling and design - base area of span 10 m x 4.3 m method development Summary - Carry out investigations on our genera- Tests tion (PMSM and DFIG) or your com- GeCoLab offers an optimal solution for re- - steady-state and transient operating ponents using our sensor types, such as search on the electrical drive train of wind

conditions (Ψ, PV, M, U, I, ECD, ...) torque, position, speed, voltage, current, turbines and Hydrogenerators. A quick - temperature rise and loss distribution flux, vibration, temperature, etc. change of components is made possible by - diagnostic methods - Carry out investigations on both con- a tensioning field and on-site 20t crane. The ventional and innovative converter and machine foundation is decoupled from the on electrical machines: generator concepts including control building by means of pneumatic spring ele- • flux distribution, end-winding fields and filter design methods. This includes ments. In addition, a readily accessible ter- • current displacement, circulating investigations into dynamics and system minal box is installed to replace the inverter currents stability, stationary and transient thermal with a test object. Researchers on the men- • Additional losses through harmonics loading, various methods of grid control tioned topics result in an improved valida- • Winding defects and their diagnosis and the behaviour of grid faults, such as tion of analytical modelling, diagnostic pro- voltage dips, symmetrical and asymmet- cedures and advanced simulation model for rical short circuits. electrical and mechanical components for a - Carry out investigations on the converter- better design of generator and converter.

Figure 2: Installed test bench References

[1] https://www.tth.uni-hannover. de/159.html

– 73 – 74WIND POWER SYSTEMS Annual Report 2015-2017

EVeQT– Increased Availability and Quality Optimization of Power Train Components and Gears for Wind Energy Systems

Universität Bremen Bremen Institute for Metrology, Automation and Quality Science (BIMAQ)

Jan F. Westerkamp, Dirk Stöbener

Funding: Federal Ministry for Economic Affairs and Energy (BMWi)

Ref.Nr. 0325490A

Duration: 09/2012 – 01/2017

Project Description

Damages of the gear box in wind energy systems (WES) are one of the major reasons for WES downtimes. These damages are caused in part due to insufficiently manu- Figure 1: Coordinate measuring machine measuring the large gear standard factured gear box components (gears and bearings).

The quality inspection of these components lacks of fast geometry sensors and of ad- References equate sensor systems for quality criteria of the surface integrity. Additionally, even [1] A. Günther, F. Balzer, I. Lindner, D. if sufficient sensors are available, the trace- Stöbener, J. F. Westerkamp, G. Goch: ability of the geometry measurements is not • new strategies and evaluation methods Einsatz von Koordinaten-messgerä- ensured due to a worldwide missing large for areal gear measurements. ten (KMG) an Großverzahnungen. In: gear standard. VDI-Berichte, Bd. 2243, pp. 139-154, These goals were mostly achieved: A 2014 Therefore, the goals of the project focus on large gear standard was manufactured and [2] F. Balzer, M. Schäfer, I. Lindner, A. the development of: calibrated by the Physikalisch-Technische Günther, D. Stöbener, J. F. Wester- Bundesanstalt (PTB). Measurements of the kamp: Recent advances in optical • a gear-like large gear standard BIMAQ during a national measurement in- gear measurements. In: International Conference on Gears, Garching, • optical sensors for coordinate measur- tercomparison resulted in small deviations Oct. 5-7, 2015, pp. 655-666 ing machines (CMM) to acquire gear (<1 µm) from the calibrated values. Addi- [3] A. Günther, D. Stöbener, G. Goch: geometries tionally, interferometric probes for CMMs Self-Calibration method for a ball • CMM-mountable sensors for surface and a new strategy for the areal gear meas- plate artefact on a CMM. CIRP integrity measurements (roughness, sur- urement were developed and successfully Annals 65(1): pp.503-506, 2016 face damages due to heat treatment) tested by the project partners.

– 74 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201775

BladeMaker

Universität Bremen Introduction Project Description Institute for Integrated Product Development (BIK) Rotorblades for wind energy plants are pre- In the Direct Textile Placement (DTP) subproject, the Institute for integrated Prod- Klaus-Dieter Thoben, Jan Franke, dominantly still manufactured by hand. In uct Development (BIK) is concerned with Marvin Richrath, Martin Rolbiecki, order to achieve the goal of an industrial Jan-Hendrik Ohlendorf production, the project partners from indus- automating the process of manufacturing try and research in the joint project Blade- the rotor blade shell. For this purpose, dry Funding: Federal Ministry for Economic Maker develop solutions to automate the multiaxial fabrics are deposited directly into Affairs and Energy (BMWi) rotor blade production. A major goal of the the rotor blade forming tool in a continuous Project Management: project is the construction and the start up deposition process. The special challenge in Projektträger Jülich (PTJ) of the BladeMaker demo center to test the the development of handling technology is Ref.Nr. 0325435E industrial production processes and devices the flexible material characteristics of the for automated production developed by technical textiles used. Therefore, the ma- Duration: 10/2012 – 03/2018 the project consortium and thus to achieve terial behaviour during deposit and drap- a higher and reproducible manufacturing ing in multi-axially curved contours is first quality at lower manufacturing costs. Fur- analysed in realistic preliminary tests. At the thermore, the demo center is to be estab- same time, virtual and experimental product lished as a national and international contact development methods are used to develop point for research and development in rotor and evaluate various technical solutions and blade manufacturing [1]. concepts.

Figure 1: From the methodical approach to the initial operation

– 75 – 76WIND POWER SYSTEMS Annual Report 2015-2017

The further development of the solution Summary process steps of the automated continuous approaches under consideration of the later non-crimp fabric depositing and consist of integration of the devices into the Blade- As part of the "BladeMaker" project, the the base frame, the material storage unit, the Maker demo center and the adaptation of BIK developed a device, the DTP-Effector, gripping and draping unit as well as a large the concepts to the rotor blade design will for the continuous placement of non-crimp number of sensors and actuators and the as- be combined in process demonstrators and fabric layers directly into the complex sociated peripherals. The BladeMaker pro- initially tested and optimized in an indus- shaped rotor blade mould. The DTP-Effec- ject is funded by the German Federal Minis- trial environment in the laboratories of tor thus becomes one of the central tools for try for Economic Affairs and Energy within the BIK at the University of Bremen. The the gantry robot system of the "BladeMak- the “6th Energy Research Programme” un- handling technology developed at the BIK er" demo center for automated rotor blade der the grant number 0325435. We grate- will then be transferred to the BladeMaker production. The components of the DTP- fully acknowledge this support [3]. demo center and integrated into the gantry Effector are conditioned by the individual robot system [2].

References

[1] Rolbiecki, M. (Institut für integrierte Produktentwicklung – BIK, Uni Bremen); Braun, R. (Fraunhofer IWES): BladeMaker: Industrialisie- rung der Rotorblattproduktion, Ingenieurspiegel 4 2014, Public Verlag [2] Franke, J.; Richrath, M.; Thoben, K.-D. (Institut für integrierte Produktentwicklung – BIK, Uni Bre- men): Forscher automatisieren Rotorblattfertigung, Maschinen- Markt 35 2016, Vogel Buchverlag [3] Richrath, M.; Franke, J.; Ohlendorf, J.-H.; Thoben, K.-D. (Institut für integrierte Produktent- wicklung – BIK, Uni Bremen): Effektor für die automatisierte Direk- tablage von Textilien in der Rotor- blatt-fertigung, lightweight design 5 2017, Springer Vieweg Verlag

– 76 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201777

CompactWind: Increase of Wind Farm Power Output through Advanced Wind Turbine and Wind Farm Control Algorithms

Carl von Ossietzky Universität Oldenburg more economic use of the limited number The higher level research objective of Com- Institute of Physics of onshore sites that guarantee high energy pactWind was the more economic, efficient Research Group: Wind Energy Systems, yields. The project was carried out in coop- and nature-friendly use of limited onshore Research Group: Energy Meteorology eration with the component manufacturer wind turbine sites by developing and test- “Robert Bosch GmbH”, the turbine manu- ing new control strategies for wind turbines Marc Bromm, Detlev Heinemann, Martin facturer and wind farm developer “eno en- and farms to reduce wake induced loads and Kühn, Vlaho Petrović, Andreas Rott, Ger- ergy systems GmbH” and the “Wind Energy ald Steinfeld, Marijn van Dooren, Lukas energy losses in wind farms. Vollmer, Stephan Voß Institute” of the TU München. In front of this background, the sub-goals Partners: included the study of possibilities to place Robert Bosch GmbH (coordinator), Project Description wind turbines closer to each other and there- eno energy systems GmbH, by reducing land occupation and increasing Technische Universität München – Wind In CompactWind, the partners worked to- energy yield per unit area. Energy Institute gether on six main work packages: WP1 "Wake control", WP2 "Multifunc- Focus of the work was the development of Funding: Federal Ministry for Economic tional IPC-Control", WP3 "Wind turbine new interconnected control strategies at tur- Affairs and Energy (BMWi) design", WP4 “Integrated wind turbine and bine level and at wind farm level to increase Ref.Nr. 0329592A-D park control”, WP 5 "Validation", and WP6 energy output and to reduce wake induced "Preparation for turbine design, wind park loads on individual wind turbines. The im- Duration: 12/2012 – 11/2016 planning, operation and the energy supply pact of the obtained results has been studied system". ForWind scientists mainly focused with respect to the economic and reliable on the characterization of wake behavior design of wind turbines and wind farms as and on the demonstration of wake deflec- well as its impact on operation. tion in high fidelity numerical simulations (WP1) and in field experiments (WP5). The results of CompactWind are based on Introduction Both showed a great potential for wake de- high-fidelity simulations which reflect the flection, and thus for wind farm control, but interaction between wind turbines and the During the design and the operation of a also a significant dependency of the wake atmospheric boundary layer in detail, ex- wind farm, it is necessary to consider aero- behavior on the prevailing atmospheric con- periments using state of the art scaled wind dynamic interactions among turbines. Wake ditions [1,2]. In order to track and charac- turbine models in an atmospheric bound- flow conditions are leading to substantial terize wakes in the field, ForWind scientists ary layer wind tunnel, as well as field ex- losses in the energy yield as well as to in- employed lidars to measure wind velocity periments on a 3.5 MW wind turbine of type creases in structural loads due to larger in- in two field campaigns [4], as well as in a eno114. homogeneity in the inflow. The effects are wind tunnel campaign (WP5) [3]. For the particularly strong at short distance behind analysis of wind turbine loads during wake Important results of CompactWind include: the wake generating turbines and limit the deflection techniques, they developed a minimum distances between turbines posi- simulation environment by coupling a high- WP1: Wake control tioned in a wind farm. fidelity flow simulation code with an aeroe- • The possibility to deflect wakes by yaw lastic simulation code (WP2) [2]. In WP4, misalignment has been verified based on The joint research project CompactWind ForWind scientists developed a robust Large Eddy simulations (LES) and wind focused on the development of turbine and wind farm controller [5], which takes into tunnel experiments. LES show that the wind farm control concepts for influencing account uncertainties in the wind direction atmospheric stability has a crucial influ- the flow in wind farms to reduce the impact measurements, and which was successfully ence on the characteristics of the deflected of wake flow conditions. The aim was to tested in the high fidelity simulations. Final- wake and the amount of deflection. maximize the energy yield of a wind farm ly, possibilities and challenges of applying • Individual Pitch Control (IPC) was prov- with respect to its area without a dispro- the obtained results for the development of en to be unsuitable for wake deflection. portional additional loading of the turbines. offshore wind farm control was discussed in • A robust control system to increase energy This should allow a more efficient and WP6. output taking into account the inherently

– 77 – 78WIND POWER SYSTEMS Annual Report 2015-2017

dynamic inflow variation (esp. direction) fully supported the wind park control al- developed robust control strategy have has been developed. gorithm during wind tunnel tests. been tested successfully in LES for differ- • A short-range Lidar system has been used ent ambient conditions for wake measurements of scaled turbines WP3: Wind turbine design • Test of a real time, applicable, robust feed in a wind tunnel first time. • Assessment of qualification for IPC inte- forward control structure for operation gration on wind turbine eno114 with yaw offset in the field WP2: Multifunctional IPC-Control • Requirement analysis of pitch systems • Coupling of Large-Eddy code PALM with for IPC WP5: Validation the aeroelastic code FAST • Assessment of sensor concepts for IPC to • Field experiments could confirm wake • Model based and inherent stable con- control loads in waked operation deflection through yaw misalignment. In troller for IPC that is active below rated • Study of necessity of adaptation of yaw addition, the impact of wind veer on wake power drive for operation with IPC and yaw offset development has been verified. • Wind field and wake observer based on • IPC has been integrated and tested on rotor loads have been validated in BEM WP4 Integrated wind turbine and park wind turbine type eno114. simulations and in wind tunnel experi- control • Load measurements in standard, yaw off- ments. The wind field observer success- • Active wake deflection by yawing and the set and IPC operation

Figure 1: Wake flow measured by lidar behind an inclined wind turbine (10 minute mean value). In the wake, the wind speed is reduced to between 5 and 7 meters per second (violet to orange) compared to the undisturbed flow of about 9 meters per second (light yellow). By yawing the turbine (see dashed rotor axis), the center line of the wake (white dotted line) is deflected from the wind direction (black horizontal line). The horizontal and vertical dimensions are shown in multiples of the rotor diameter (D = 114 meters).

– 78 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201779

• Power measurement in standard and yaw offset operation • Wind tunnel experiments showed that yaw misalignment can reduce loads and increase wind farm power output. For a wind farm consisting out of three wind turbines an automated control algorithm could increase total power significantly.

WP6: Preparation for turbine design, wind park planning, operation and the energy supply system • The prevailing atmospheric stability at the wind farm location, inflow and wake characteristics of individual turbines in the wind park, and the state resp. operating point of the wind turbines have been identified as essential inputs for a wind farm control system based on wake deflection. • Due to the still high cost for procurement and operation of necessary measurement Figure 2: Installation of a long-range Lidar measurement system on nacelle of eno 114 systems and due to the absence of fully wind turbine functional and ready to use alternative solutions, further steps in development have to be taken to achieve an economic wind farm control system based on wake deflection especially for small and medium size wind parks. • It has been shown that the results can be transferred to offshore wind farms with a potential increase of energy output. The industrial application of wake deflection is still facing challenges similar to those onshore.

Summary References [1] Vollmer, L.; Steinfeld, G.; fields in a boundary-layer wind tunnel, Within the joint research project Com- Heinemann, D.; Kühn, M.: Estimating Wind Energy Science 2, pp. 329-341, pactWind, the flow in wind farms and how the wake deflection downstream of a 2017. it can be influenced by different turbine wind turbine in different atmospheric [4] Bromm, M.; Rott, A.; Beck, H.; control concepts to increase the overall stabilities: An LES study, Wind Energy Vollmer, L.; Steinfeld, G.; Kühn, M.: Field farm energy output and reduce the structural Science 1, pp. 129-141, 2016. investigation on the influence of yaw loads on the individual turbines is studied. [2] Bromm, M.; Vollmer, L.; Kühn, M.: misalignment on the propagation of New control strategies for the integration in Numerical investigation of wind wind turbine wakes, Wind Energy 21, pp. 1011-1028, 2018. commercial wind farm control systems are turbine wake development in directionally sheared inflow, [5] Rott, A.; Doekemeijer, B.; Seifert, derived and validated in large-eddy simula- Wind Energy 20, pp. 381-395, 2017. J-K.; van Wingerden, J-W.; Kühn, M.: tions, wind tunnel experiments and in free- [3] van Dooren, M.F.; Campagnolo, F.; Robust active wake control in conside- field flow and load measurements at multi- Sjöholm, M.; Angelou, N.; Mikkelsen, T.; ration of wind direction variability and megawatt turbines. Kühn, M.: Demonstration and uncer- uncertainty, Wind Energy Science 3, tainty analysis of synchronised scanning pp. 869-882, 2018. lidar measurements of 2-D velocity

– 79 – 80WIND POWER SYSTEMS Annual Report 2015-2017

Probabilistic Load Description, Monitoring and Reduction of Loads of Future Offshore Wind Energy Converters (OWEA Loads)

Carl von Ossietzky Universität Oldenburg turbine by control algorithms that do not termittency large wind speed changes which Research groups:Wind Energy Systems; lead to an extensive increase of loads in might result in large loads are more frequent Energy Meteorology; other components have been developed than predicted by a wind model that pro- Turbulence, Wind Energy and Stochastics and tested. The overall objective of all the duces a Gaussian distribution of fluctua- work done in the project was to reduce the tions. The second approach aiming at im- Martin Kühn, Detlev Heinemann, financial risks in the exploitation of offshore provements of the first part of the modelling Sebastian Ehrich, Pedro Lind, wind energy resources by an improved pre- chain was to use lidar measured wind fields Binita Shrestha, Gerald Steinfeld, Juan José Trujillo, Luis Vera Tudela, diction of the life expectancy of turbines in the framework of the numerical model Matthias Wächter, Hauke Wurps and extreme loads. in order to improve the agreement between measured and simulated wind fields. This Funding: Federal Ministry for Economic approach resulted in the development and Affairs and Energy (BMWi) Project Description implementation of a new turbulence recy- cling method in PALM. However, it turned Ref.Nr. 0325577B It was one of the objectives of OWEA out that it requires still considerable efforts Loads to gain an improved understanding of both on the measurement and the modelling Duration: 12/2012 – 08/2016 the loads on offshore wind turbines. One of side to make this approach beneficial. the approaches that was followed to reach this objective was to further develop the Another part of the project dealt with the interdisciplinary modelling chain. The de- question how nacelle-based Lidar measure- Introduction velopment of this chain started in the previ- ments can help to better understand and pre- ous project OWEA. It allows for a detailed dict wind turbine loads. A technique for the The four project partners University of investigation of the interaction between the tracking of wakes by Lidar measurements Stuttgart, Adwen GmbH, Senvion and atmospheric flow and a wind turbine in spe- was developed that carries out a fitting of University of Oldenburg worked together cific situations. The modelling chain con- a two-dimensional Gaussian function to a in the joint research project OWEA Loads sists of a part that aims at providing an ac- measured wind field. The position of the [1], which was part of the accompanying re- curate description of the atmospheric flow, symmetry axis of the fitted function is then search at the German offshore test site alpha taking into account the synoptic condition assumed to be the wake position. Results ventus. Subcontractors were the University and resolving the bulk of the turbulence in of the wake tracking analysis were then of Hannover, the University of Tübingen the atmospheric boundary layer. The flow brought together with measured loads in or- and DNV GL. field that is obtained from this first part of der to analyze how dynamic loads depend the modelling chain is then used to provide on the wake position. In that framework it The project aimed at describing the aerody- initial and boundary conditions to a second was also shown that the newly developed namic, hydrodynamic and operational loads flow model that is able to give very accurate wake tracking algorithm can be used for on offshore wind energy converters. It was information on the response of the wind tur- an improvement of the Dynamical Wake intended to use the comprehensive data ob- bine to the incoming flow. In OWEA Loads Meandering model (DWM). Here, it was tained from measurements at alpha ventus two different approaches were followed to shown that it is important to feed the DWM and the analysis of that data in previous achieve improvements in the first part of the with data obtained from Lidar measure- research projects, e.g. OWEA. Moreover, modelling chain. One approach was to use ments analyzed in a moving frame of refer- the project dealt also with new questions the stochastic Continuous Time Random ence instead of a fixed frame of reference that have to be answered in the process of Walk (CTRW) model in order to provide to get a good agreement between measured designing new generations of offshore wind a turbulent inflow to the CFD code Open- and simulated loads. energy converters. One of these aspects was FOAM. In contrast to other wind field mod- the description of the stochastic character- els (e.g. the Mann model) the CTRW model Another task of ForWind dealt with the istics of loads by the means of probabilis- can reproduce even higher order statistics of development of a stochastic approach that tic methods. Different concepts aiming at real wind fields and especially the intermit- allows for predicting time series of specific reducing the loads in one component of a tency of the wind fluctuations. Due to the in- loads based on time series of another pa-

– 80 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201781

rameter. This method should also be able to reproduce higher order statistics observed in measured load time series. Here, the loads were modelled by a stochastic differential equation, a so called Langevin equation. By this method measured time series of the rotor torque and tower base bending moment, respectively, could successfully be reconstructed based on a given time series of the wind speed measured by a nacelle anemometer. Moreover, it could be shown that the stochastic model derived from data measured at one wind turbine in a wind farm can be used to accurately estimate the load time series of another wind turbine of the same type in the farm when the wind speed measured at that wind turbine is available as input. This might be of special interest as it means that only one heavily instrumented wind turbine in a wind farm is required to supervise the loads felt by all wind turbines of a wind farm.

The results of the stochastic model were compared with those of an artificial neural network built up from the same data as that used for the derivation of the stochastic method. Mean values and small fluctuations are well-predicted by both methods while larger fluctuations could only be predicted by the stochastic model. Figure 1: Measured (black) and reconstructed (red) torque. (a) time series of torque, (b) Zoom-in in (a), (c) PDFs (probability density function) of torque increments for different Besides attempting to improve the possi- temporal scales. bilities to predict and understand wind turbine loads, OWEA Loads dealt also with This allows for applying a damping only still be reduced to 78% of the possible load the question how by the application of con- in those situations in which the oscillation reduction without limitation of the pitch trol algorithms loads in certain components frequency is close to the eigenfrequency. It activity. In case that the standard deviation can actually be reduced without increasing was shown that this approach led to 80% of of the generator torque is not allowed to in- the loads in other components of the tur- the reduction that would be possible if the crease by more than 60%, the loads at the bine too much. standard control were continuously active, tower base can be reduced to 80% of that while resulting in only 75% of the increase what is possible when there are no limits Two different approaches were followed of the pitch activity. to the application of the active generator with regards to this objective. The first ap- torque control. Thus, achieving a signifi- proach aimed at achieving an active tower The second approach aims at a trade-off- cant load reduction with simultaneously damping by coupling of a control algorithm analysis between the reduction of fatigue relative small collateral effects is possible. with an oscillation analysis. The second ap- loads in the tower and an increase of col- proach attempted to gain an active reduc- lateral effects in other components. Three Another part of OWEA Loads followed tion of lateral and longitudinal loads by a different controllers in different combina- the path that it is still necessary to increase selective choice between different control tions were studied for this analysis. With the knowledge on the atmospheric ambient concepts. In the first approach the standard a multi-objective optimization method the conditions to which offshore wind turbines control loop for the tower load reduction of most effective control concept for a certain are exposed in order to come to improved a turbine is coupled with a time-frequency- situation was found. If a limit is set that the load estimates and subsequently adapted analysis technique, which allows for esti- pitch activity must not increase by more wind turbine designs. mating the tower oscillation frequencies. than 60%, the loads at the tower base can

– 81 – 82WIND POWER SYSTEMS Annual Report 2015-2017

Therefore, a new device for probing the are suspected to be able to cause high loads environment and by developing and further marine atmospheric boundary-layer up on wind turbines, has been studied for the improving models that provide the response to heights of 500 m has been applied for North Sea region. For that purpose, data of the turbine to these ambient conditions. the first time in an offshore measurement from long-term simulations with the mes- Moreover, the project investigated several campaign. The Unmanned Aerial Vehicle oscale model WRF for the year 2009 have options how loads in one component of the (UAV), developed and controlled by staff been analyzed. It turned out that due to the wind turbine can be reduced by the means of the University of Tübingen, was applied decreased probability of stable stratifica- of a controller without increasing the loads in 5 measurement campaigns and carried tions LLJs seem to occur less often over in other components of the wind turbine too out 71 flights, each with a length of about the North Sea (10.5% of the time) than over much. All three working groups of ForWind 30 to 45 minutes. Due to the strict regula- the surrounding land. Interestingly, the dif- Oldenburg have been involved in the pro- tions for aerial vehicles it was not possible ferences in the number of LLJ events with ject. Highlights of the project results are: to carry out the originally planned fully-au- cores at wind energy relevant heights is tonomous flights. In order to nevertheless however small. Most of the LLJ cores are - that a stochastic, intermittency-inducing be able to measure atmospheric conditions found in the height range between 200 and inflow model based on continuous time over the open sea, the flights took place in 300 m above the ground. It was found that processes was implemented into the sighting distance from the island of Heligo- the core height is correlated with the LLJ CFD code OpenFOAM. land. It turned out that the usage of UAVs maximum wind speed. Moreover, the core - that the benefits of an algorithm devel- can be helpful to gain a detailed insight in height is correlated with the atmospheric oped in the project with the objective to the marine atmospheric boundary layer in stability. Onshore a stronger stability is re- switch the fore-aft and side-side damp- single specific situations. However, for a quired in order to get an LLJ core at the same ers on or off to minimize the cost func- continuous observation of offshore condi- height as at an offshore site. The analysis of tion while maximizing the reduction in tions met masts are still needed. the numerical results was also accompanied tower bottom bending moments could by an analysis of in-situ and remote sens- be shown. An aspect concerning the ambient condi- ing measurements over the North Sea. This - that actually observed load fluctuations tions of offshore wind turbines that was analysis supported the results obtained from could be well reproduced with a sto- studied in the framework of the OWEA the simulations with the WRF. chastic modelling approach and by an Loads project was the determination of artificial neural network approach. The the risk of icing in the North Sea. For that The results presented above have shown stochastic modelling approach is even purpose, a model has been developed that that mesoscale simulations can be a valuable to reconstruct small-scale fluctua- allows for the evaluation of the ice accre- able tool for the analysis of offshore con- tions well. tion. This model can be fed with data de- ditions. Due to the fact that only recently - that for the first-time measurements of rived from simulations with the mesoscale data from measurement campaigns over wind and turbulence profiles were -car model WRF. In order to get a conservative the open sea has become available, e.g. the ried out over the ocean with the help of estimate of the icing risk over the North planetary boundary layer parameterization an unmanned aerial vehicle. Sea, long-term simulations have been car- in the mesoscale model is solely based on - that a methodology has been created ried out with the mesoscale model WRF data from onshore measurements. Within with which it is possible to determine covering a winter with more ice days than OWEA Loads it was therefore intended to the time series of icing at a certain site in an average winter. According to the improve the results of WRF for offshore based on input from a mesoscale simula- WRF-ice-accretion-modelling-chain icing environments by deriving an improved tion model. should have occurred in 5-10% of the time planetary boundary-layer scheme based on - that both, numerical and experimental at a height of 100 m over the North Sea in measurements and large-eddy simulations investigations have shown that low-lev- the winter 2010/11. Over the Baltic Sea the for offshore sites. As a result, the adaptation el jet events are of relevance at heights icing risk was found to be higher, as the of some constants in the Mellor-Yamada- that affect offshore wind turbines. occurrence of icing was predicted during Nakanishi-Niino model was recommended. 10% of the time. The analysis made clear that icing might become an even more important problem for new generations of Summary References wind turbines, as the icing risk was found [1] Bartsch, J. et al.: Probabilistische to increase considerably with increasing The OWEA Loads project attempted to Lastbeschreibung, Monitoring und height above ground. increase the knowledge on offshore wind Reduktion der Lasten zukünftiger turbine loads by an analysis of data from Offshore-Windenergieanlagen (OWEA Finally, also the probability of the occur- the German offshore test site alpha ventus Loads), Abschlussbericht, Stuttgart, rence of large wind speed maxima at low by improving the knowledge about the at- 374 pages (2017) heights, so called low-level jets (LLJ) that mospheric ambient conditions in the marine

– 82 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201783

Silicon Carbide Power Technology for Energy Efficient Devices (SPEED)

Universität Bremen Project Description electronic device under combined thermal Institute for Electric Drives, Power and electrical stress [5]. An HTRB test is Electronics and Devices (IALB) The work presented in this report focuses conducted with novel 3.3kV SiC-JBS-Di- on the aspects of reliability of HV power ode chips mounted on a test substrate and Felix Hoffmann, Nando Kaminski devices which is the main part of the work potted with silicone gel. Funding: Seventh Framework Programme, package covered by IALB in this project. European Commission The results of this work are also published The structure of the diodes under test is in [2]. Reliability is a major concern for shown in Fig. 1. It consists of a p-stripe de- Ref.Nr. 41V7736 power systems design. Particularly, the sign with a relatively large area of 5x5 mm². Duration: 01/2014 – 12/2017 integrity of the junction termination is The chip design features a JTE-based edge- crucial for HV devices with voltages of termination with an n+ doped channel stop- 3.3kV and above. For this reason, a long- per similar to the design described in [3]. term high temperature reverse bias test had been conducted with a novel design of The surface of the edge termination is cov-

Introduction 3.3kV SiC Junction Barrier Schottky (JBS) ered with an SiO2/polyimide stack. A total diodes developed within the scope of the of 8 substrates with 4 chips each had been The development of a new generation of SPEED project. packaged for testing. Fig. 2 shows a test high power semiconductor devices, able to substrate populated with 4 diode chips cov- operate above 10 kV, is crucial for reduc- SiC-JBS-Diodes are well-established in ered with globtop. The chips are soldered ing the cost of power electronics (PE) for commercial applications up to 1.7kV and to a direct bonded copper (DBC) substrate power transmission and renewable energies. devices for higher voltages are under de- and mounted on a copper baseplate to in- The material properties of Silicon Carbide velopment [3,4]. However, higher voltage crease thermal capacitance and avoid ther- (SiC) clearly are superior to those of Silicon also increases electrical stress on the edge- mal runaway during testing. Two different (Si) for high voltage (HV) devices. Pooling termination. The High Temperature Re- types of package insulation were tested. An world-leading manufacturers and research- verse Bias (HTRB) is the standard test pro- overview of the test substrates is given in ers, SPEED aims at a breakthrough in SiC cedure to survey the integrity of a power Table I. technology along the whole supply chain:

• Growth of SiC substrates and epitaxial- layers. • Fabrication of power devices in the 1.7/>10 kV range. • Packaging and reliability testing. • SiC-based highly efficient power conversion cells. • Real-life applications and field-tests in close cooperation with two market- leading manufacturers of HV devices.

The main targets are cost-savings and superior power quality using more efficient Figure 1: Structure of devices under test power converters that exploit the reduced power losses of SiC. To this end, suitable SiC substrates, epitaxial-layers, and HV de- Substrate S1 S2 S4 S5 S6 S7 S8 vices shall be developed and implemented Globtop cover Yes Yes Yes No No No No in two demonstrators, a solid-state trans- Chip count 1 2 2 3 4 3 2 former and a windmill power converter. For details about the SPEED-project see [1]. Table 1

– 83 – 84WIND POWER SYSTEMS Annual Report 2015-2017

Whereas all substrates received a silicone covering, whereas the lower graph shows sary to remove infant mortality sensible gel potting, the chips on substrates S1 to the current of the substrates with silicone diodes. Those chips were not tested and S4 had been coated with an additional gel only. To verify thermal stability, the test also disregarded for further test analysis. globtop covering before the silicone gel was initialized with a slow increase in tem- The blocking curves indicate that almost potting was applied. The globtop covering perature from 85°C to 125°C for 500 hours. all substrates exhibit a change in their re- used for this test is an epoxy-based com- Thus, the leakage current also increases verse blocking characteristics i.e. a leak- pound. It is applied to avoid direct contact during that phase. This initialization phase age current increase in a certain voltage of the silicone gel with the chips’ surface is shaded in yellow in Fig. 3. One chip on range. The blocking curves are shown in and prevent interference of possible charg- substrate S2 (red line, upper graph), which Fig. 2. Both test splits trending towards the es in the silicone gel with the electric field was already suspicious during initial static same direction concerning these changes. applied during HTRB. As a reference, the blocking measurement, failed during this Hence it can be concluded that there is no substrates S5 to S8 were covered with sili- test initialization and was removed from the significant impact of the insulation mate- cone gel potting only. The test is conducted test batch. After the initial phase, the test rial, or respectively its embedded charge, with an ambient temperature of 125°C and was conducted for more than 5000 hours on HTRB performance. a reverse bias of 2640V, corresponding to under steady state reverse bias of 2640V

80% of the diode’s nominal voltage (VAK,max) and constant temperature of 125°C. Hence surface charges from the silicone [5]. The leakage current for each substrate gel are not emerging as an important factor logged over the course of the test is shown Some of the tested diodes showed intrin- influencing the lifetime of a device under in Fig. 3. The upper graph shows the current sic deficiencies in blocking capability. It HTRB stress. of the substrates with additional globtop shows that an initial screening is neces-

Figure 2: Blocking curves during initial measurement, after 2500h and 5500h of HTRB.

– 84 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201785

The current logs in Fig. 3 clearly indicate tion or failure of a chip. Also, the interme- application. For stable and robust devices, that both test splits show a significant diate measurements of reverse blocking a major challenge of silicon carbide mate- general decrease in leakage current over characteristics showed that no significant rial is the edge-termination due to its al- the course of the test at high temperature, degradation of blocking capability could most ten times higher critical field strength which is consistent with other publications. be observed. Hence, all substrates passed compared to silicon, and the results of this However, the root cause of this behavior the HTRB test without any failures. long term HTRB test show that an appro- could not be unveiled during this test. priate edge-termination design can meet the necessary stability requirements and The test was terminated after more than Summary can achieve reliable devices. 5000 hours of constant HTRB stress. Dur- ing our test none of the substrates had a sig- This result suggests that the quality of nificant increase in leakage current, which silicon carbide HV epi material approaches would have been an indicator for degrada- the maturity level required for commercial

101 init S1 (CH02) phase Intermediate Δ measurement S2 (CH03) 0 10 S4 (CH05) Θ Planned downtime ) A m (

n t 10-1 urr e C

10-2 Θ ΘΘΔ Δ ΔΘ ΘΘ Δ Θ Θ Θ ΘΘ Θ Θ

10-3 0 1100 2200 3300 4400 5500 time (h)

101 init Intermediate S5 (CH06) phase Δ measurement S6 (CH07)

0 Planned downtime 10 Θ S7 (CH08)

) S8 (CH09) A m (

n t 10-1 urr e C

10-2 Θ ΘΘΔ Δ ΔΘ ΘΘ Δ Θ Θ Θ ΘΘ Θ Θ

10-3 0 1100 2200 3300 4400 5500 time (h) Figure 3: Leakage current log during HTRB

References [3] A. Mihaila et al., ‘Experimental investigation of SiC 6.5kV JBS diodes [1] www.speed-fp7.org safe operating area’, in ISPSD 2017, pp. [2] F. Hoffmann et. al., ‘Long Term 53–56. High Temperature Reverse Bias (HTRB) [4] J. Millán et al., ‘High-voltage SiC Test on High Voltage SiC-JBS-Diodes’, devices: Diodes and MOSFETs’, in CAS in ISPSD 2018, pp. 435-438. 2015, pp. 11–18. [5] cf. JEDEC JESD22-A108

– 85 – 86WIND POWER SYSTEMS Annual Report 2015-2017

Innovation Cluster Power Electronics for Regenerative Energy Supply

Leibniz Universität Hannover This subproject focuses on ageing mecha- cal parameters of the IGBT which change Institute for Drive Systems and Power nisms in IGBT modules. Fig. 1 shows the depending on the junction temperature. In Electronics (IAL) vertical design of an IGBT module. double pulse tests, various TSEPs were analyzed with respect to their temperature Marcel Moriße, Simon Weber Due to conduction and switching losses sensitivity. On-state voltage and turn-off delay time were selected as appropriate Funding: Ministry of Science and Culture during operation, the junction tempera- parameters. For the detection of both pa- of Lower Saxony ture in the power semiconductor (SI chip) rises. Since the temperature changes cycli- rameters, a low-cost electronic circuit was Ref.Nr. VWZN2989 cally, there are ageing processes at the wire developed. The total component costs of bonds and the solder joints. the circuit are less than six € per piece for a Duration: 09/2014 - 08/2017 quantity of 1,000 pieces. To predict the lifetime of the modular design by means of lifetime models, it is nec- At an open IGBT module, the IGBT’s sur- essary to know the junction temperature. face temperature as well as the two TSEPs were measured in the inverter mode. Based Within this subproject, an approach has on the measured data depicted in Fig. 2, been developed to measure the junction conclusions can be drawn on the junction temperature while the drive is operating temperature. When the junction tempera- based on temperature-sensitive electrical ture is known, lifetime consumption is pre- parameters (TSEP). TSEPs are electri- dictable.

Project Description

WP2: Condition Monitoring, Remaining Lifetime Prediction and Preventive Maintenance Strategies Within the scope of the Innovation Cluster Figure 1: Vertical design of IGBT module [1] Power Electronics, one research topic was to investigate the reliability of electrical converters in wind turbines. According to the research results of the project RELI- AWIND (2011), the main failure sources are the pitch system and the frequency converter. Ageing mechanisms of the IGBT modules and insulation defects are the most important failure causes in frequency converters. The aim of the meanwhile completed Innovation Cluster Power Electronics, which was funded by the Land of Lower Saxony and by Fraunhofer-Ge- sellschaft, was to identify the causes and develop approaches for improvements and failure prediction. Within the project, IAL has worked in close cooperation with Fraunhofer IWES in Hannover. Figure 2: Measurement of on-state voltage and turn-off delay time

– 86 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201787

WP3: Drive Train Modelling to Deter- One research result proves that typical and main converter can be expected from ac- mine Harmful Influences on Converters partially vast simplifications are permis- tive torsional damping (ATD) of the drive in Wind Turbines sible when modelling the overall system, train through additional generator torque. In this subproject of the Innovation Clus- in order to determine the thermal stress However, since the main converter as a ter, the aim was to identify harmful influ- on the power semiconductors. Operating significant cost driver is normally rather ences during operation by the development point dependent load cycles due to differ- not excessively oversized, it is important of wind turbine models with typical drive ent wind speeds in the range of several sec- to consider the effective impact of ATD in train topologies and converter concepts onds have a considerably larger impact on the design phase. Another point of research which consider the specific properties of thermal module stress than high-frequency was the consideration of control methods the subsystems (rotor, drive train and elec- wind turbulences or drive train vibrations. to mitigate the thermally particularly criti- trical system including grid) in an overall In contrast, active control strategies can cal generator synchronism in systems with model. have a more relevant influence on lifetime. doubly-fed induction machine. Their consideration can play an important role when In the next step, a lifetime algorithm (Fig. Individual pitch control (IPC) to reduce designing power modules for the generator 3) was applied to evaluate the effects on magnitude-dependent asymmetrical bend- converter. Fig. 4 shows the avoidance of the power semiconductors. The algorithm ing moments at the rotor blade bearings the largest lifetime-significant temperature is based on an accurate thermal model of increases for example the load on the differences of a diode when the magnetiz- the converter modules and on current em- pitch converter. Due to typically oversized ing reactive power for the generator during pirical investigations concerning their ex- pitch drives, these surplus loads might not synchronism is partly supplied by the grid pected lifetime depending on specifically be crucial for a premature failure of the converter. para-meterizable power cycles. pitch converters. Similar effects on the

Figure 3: Schematic procedure when calculating the lifetime consumption of the single converter semiconductors

Figure 4: Thermal relief of a diode of the generator converter by reactive power supply of the grid converter in generator synchronism References

[1] M. Schulz, Vervielfachung von Lebensdauer und Leistungsdichte mit IGBT5 & .XT, ELEKTRONIKPRAXIS Leistungselektronik & Stromver- sorgung, September 2015)

– 87 – 88WIND POWER SYSTEMS Annual Report 2015-2017

Manufacturer-independent Retrofitting and Evaluation of the Remaining Life Time of Power Electronic of Wind Turbines with DFIG (WEA-Retrofit)

Universität Bremen tenance cost [1]. The experience of wind project, the core responsibility of IALB is Institute for Electric Drives, Power farm operators shows that downtimes due the development of appropriate measure- Electronics and Devices (IALB) to failing power electronics start dominat- ment equipment, condition monitoring in ing failure statistics. To avoid unplanned the field, reliability testing in the lab, and Holger Groke, Wilfried Holzke, Alexander downtimes a prediction of the remaining lifetime modelling. Brunko, Nando Kaminski, Bernd Orlik, lifetime would be extremely helpful. Christian Zorn The required field-data is logged over years in an active, commercial wind farm Funding: German Federal Ministry of in northern Germany between Bremen Economic Affairs and Energy Project Description and Hamburg. Seven measurement units Ref.Nr. 0325758A–D The aim of the project is to increase the and special electronic systems, developed service life of the power electronics in by the IALB at the University of Bremen, Duration: 12/2014 – 08/2018 wind turbines and to enable the prediction were used to measure and analyse the stress of failures. Based on an online evaluation of the power electronics of seven wind en- of electrical parameters measured on wind ergy plants with high resolution 24 hours a turbines in operation, a model-based calcu- day and 365 days a year. The huge amount Introduction lation of the expected remaining lifetime of data is recorded and processed by a of the power electronics shall be achieved. combination of DSP, FPGA and high-end While predicted maintenance for drive The concept developed for the preventive IPC devices. One of the built up measure- trains and mechanical parts is already a maintenance of power electronics is then ment systems is shown in Fig. 1. Currently, state-of-the-art procedure to reduce down- tested on a functional model as a labora- the flexible and modular system allows times, this procedure is currently not avail- tory setup by retrofitting older systems sampling of up to seven channels simulta- able for the converter system. Failures due with new system services and avoiding re- neously with 16 Bit at 50 kHz or up to 18 to degradation of power electronics, as part certification of the system, independent of channels if lower frequency is sufficient. of the converter, cause significant main- the manufacturer. Within the scope of this An extension board with a FPGA, con-

GSM / UMTS (optional)

HDD 2 x 1 TB IPC

EtherCAT

DSP 1 Trigger TMS320- In- and F28335 Output

fA <= 50 kHz 2 Trigger 1 16 Bit 1 Trigger

FPGA SPI Single-board LPF Spartan 6 computer fA >= 2 MHz Figure 1: Measurement system developed at IALB (left), 12 Bit structure of the measurement system (right) RAM

– 88 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201789

nected to a fast RAM, with seven channels with 12 Bit at 10 MHz, was added for real- time signal analysis and data compression. Furthermore, there are four channels for humidity and temperature measurements available. All raw data is transferred to the IPC and stored with a timestamp.

There are different approaches to extend the measurement system towards a monitoring system. Methods based on online signal measurements use e.g. the Collec- tor-Emitter-voltage [2] or modified gate drivers [3] to estimate the virtual junction temperature Tvj, as the most meaningful parameter. In this work, a model-based approach has been chosen. As described in Fig. 2, information about the used power Figure 2: Model to estimate the remaining lifetime electronic devices (Insulated Gate Bipolar Transistor (IGBT), diode) and the structure of the whole package including the cooling system is necessary. The required param- ing a Cauer-Model is straight forward, a Fig. 3 shows current and temperature eters were determined by offline laboratory real-time capable version of the rain-flow measurements of a wind energy plant for a measurements similar to the approach in count, to analyse the temperature cycles, is period of two weeks, recorded by the data [4]. Mains and DC-link voltages, currents not trivial. Finally, the models can reveal the acquisition system. The first plot shows and temperatures have to be measured in thermo-mechanical and the electro-chemi- one phase of the grid current, the second real-time. While look-up tables and solv- cal status of the semiconductor module. shows the heat sink temperature in front of

Figure 3: Current and temperature measurements for a period of two weeks

– 89 – 90WIND POWER SYSTEMS Annual Report 2015-2017

and behind an IGBT device as well as the Summary temperature in the cabinet. A flexible measurement system for long- As expected, the cooling system keeps term data recording and processing was the temperature of the devices at a given established. The recorded data is utilised to set point while the wind energy plant is in analyse the stress of power electronics of operation. If the wind energy plant is of- wind energy plants connected to the grid. fline, e.g. 1st to 4th of June (box in Fig. Based on this input and reliability models 3), the temperature of the heat sink follows of the power semiconductors the remain- the temperature of the environment. These ing lifetime of the electronics can be pre- temperature cycles must also be considered dicted. The accuracy of the results strongly for calculating the lifetime. While the high- depends on the precision of the model pa- est temperature at the junction is the result rameters. Therefore, the parameters have of the conduction losses, the minimum to be adjusted for each new IGBT based on temperature is applied from the outside extensive laboratory experiments. How- (ambient) of the module. This leads to the ever, the results obtained so far suggest requirement that the condition monitor- that neither the thermo-mechanical nor the ing system has to run independently of the electro-chemical degradation nor special wind energy plant, technically speaking for events in the grid limit the useful service the whole lifetime of the power electronic life of the investigated wind energy plant. device. This has to be taken into account This would direct the search for the root when selecting a digital signal processor causes of indeed occurring failures away and implementing the rainflow count, due from the classical failure mechanisms to- to limited memory for storing the cycles. wards “new” effects, but this is subject to further evaluation. Feeding the measured data like in Fig. 3 into the lifetime estimation models yields useful lifetimes of far beyond 100 years, i.e. the modelled degradation mechanisms would hardly lead to any failure. Admit- tedly, the wind park under investigation is rarely running at rated power and the temperature cycles are rather small. Further- more, no special events that would have References caused significant live time consumption [1] Fischer, K.; Wenske, J.: Towards were detected. Thus, the operational condi- reliable power converters for wind tions in the investigated wind park have to turbines: Field-data based be considered as rather smooth and are not identification of weak points and cost ideal to challenge the power electronics. drivers, EWEA 2015, Paris [2] Bęczkowski, S.; Ghimre, P.; de Vega, A. R.; Munk-Nielsen, S.; Rannestad, B.; Thøgersen, P.: Online Vce measurment method for wear-out monitoring of high power IGBT modules, EPE 2013, Lille [3] Denk, Bakran, M. M.: Health- Monitoring of IGBT power modules using repetitive half-sinusoidal power losses, PCIM Europe 2016, Nuremberg [4] Zorn, C.; Kaminski, N.: Temperature–humidity–bias testing on insulated-gate bipolartransistor modules, IET Power Electronics, vol. 8, iss. 12, pp. 2329-2335 (2015)

– 90 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201791

BisWind – Component Integrated Sensor System for Wind Energy Systems

Universität Bremen Bremen Institute for Metrology, Automation and Quality Science (BIMAQ)

Michael Sorg

Funding: Federal Ministry for Economic Affairs and Energy (BMWi)

Ref.Nr. 0325891D

Duration: 01/2015 – 07/2019

Project Description

Drive trains of wind energy systems experience a broad range of dynamic loads. Tran- sient torque reversals originate in power loss and emergency stops, start cycles and in sheer winds and turbulence. The subsequent failure of bearings and gearboxes result in over 50 % of wind energy. To improve the design of drive train components with precise load cycles, precise and long-term measurements are required.

Torque sensors are currently used only sporadically and not in volume production. Direct measurements of loads are not available for most parts of the drive train, especially from the inside of the gearbox. Data over the lifetime are scarce and correlations to failure events are thus limited to a few cases. Figure 1: Research wind energy system of the University of Bremen

The cooperative research project develops a component-integrated measuring system. The key design aspects are measurement and electrodes directly on shafts for the of torque, temperature, vibration and rota- durable sensor itself. To be self-sufficient References tional speed with a sensor that is resistant newly developed AIN and AIScN based [1] Tracht, K.; Goch, G.; Schuh, P.; to aging and aggressive media, and is self- piezoelectric structures have to provide the Sorg, M.; Westerkamp, J. F.: sufficient. energy for the sensor module which in turn Failure probability prediction based will be assembled on a cylindrical low tem- on condition monitoring data of wind The scientific and technical objectives perature co-fired ceramics. This subproject energy systems for spare parts cover a broad range beginning with the investigates both the suitability and the per- supply. CIRP Annals 62(1): process development for direct coating formance of the measuring system for ap- pp. 127-130, 2013 and structuring of resistance structures plication in wind turbines.

– 91 – 92WIND POWER SYSTEMS Annual Report 2015-2017

Intermittent Wind Loads (InterWiLa)

Carl von Ossietzky Universität Oldenburg Research Group: Turbulence, Wind Energy and Stochastics

Matthias Wächter, Joachim Peinke ForWind – University of Oldenburg,

Partner: ) Alexandros Antoniou, Philipp Thomas, τ Fraunhofer IWES (Coordinator) p( δ u

Funding: Federal Ministry for Economic 7 Affairs and Energy (BMWi) 10

Ref.Nr. 0325798B

Duration: 01/2015 – 12/2017 1e−11 1e−08 1e−05 1e−02 −10 −5 0 5 10 δuτ , στ

Figure 1: PDF of wind speed increments δuτ = u(t+τ)-u(t) for a time lag of τ = 3 s, measured at the FINO 1 platform in the German North Sea (blue line). Additionally a Gaussian PDF Introduction is shown with identical mean and standard deviation (black line). The probability of a 7 σ event is underestimated by a factor of 107 by the Gaussian assumption. It is well known that real inflow at wind turbines exhibits complex and strongly non-Gaussian statistics, see, e.g., [1]. This observation is commonly called intermittency and holds especially for two-point statistics such as wind speed increments

δuτ = u(t+τ)-u(t), i.e., wind speed differences over a fixed time lag, see Fig 1. Nev- ertheless, current industry standards such as IEC 61400 [2] recommend to consider only completely Gaussian wind statistics in the design procedure of wind energy systems.

While a striking difference between this Gaussian assumption and the real-world wind measurements is obvious (cf. Fig. 1), its impact especially on wind turbine loads is still unclear and not understood. The project InterWiLa was designed to tackle this question from a combined numerical and experimental approach. Project partners were Fraunhofer IWES in Bremer- haven (leader) and ForWind – University Figure 2: Test stand for beam samples at Fraunhofer IWES in Bremerhaven. A hydraulic of Oldenburg. actuator is applying force time series to the right end of the beam.

– 92 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201793

Project Description

In a numerical approach, damage equivalent loads at different locations were derived from wind turbine simulations using both Gaussian and (more realistic) intermittent wind fields. As an industry level tool, GH Bladed 4.7 was used with the IWES 7.5 MW reference turbine [3]. For further investigations, load time series were extracted from these simulations at different locations Figure 3: Damage equivalent loads (DEL) as results of the numerical load simulations. Pre- in the turbine. From these load time series, sented are DEL for wind inflow generated by Kaimal and Mann models (using an inherent damage equivalent loads (DEL) were de- Gaussian assumption as recommended by IEC [2]) and by the Continuous Time Random rived for a first evaluation of the damage Walk (CTRW) model (incorporating realistic temporal intermittency). For the example of the yaw moment (left), the DEL are higher with intermittency, while for the blade root related to both types of wind inflow. As a bending moment (right) no significant difference is visible. highly loaded location in a wind turbine, a typical rotor blade structure at about 25% of the blade length was selected.

In an experimental approach, those load wind fields with identical mean and stand- significantly reduced. We take this result as time series were used to perform experi- ard deviation are compared, cf. Table 1. The a clear hint for the existence of an impact ments on lifetime of typical composite blade time until total failure shows the reverse of wind intermittency on damage at wind material and glued joints. To this end, beams behavior, however, in real-world situations, turbines. More detailed investigations, also reflecting a typical rotor blade structure the first cracks are expected to initiate fur- on other components besides rotor blades, were manufactured and tested in a beam ther degradation of the structure, which is could be the topic of further projects. test stand at IWES in Bremerhaven. The not reproduced in the laboratory. initiation and development of cracks were observed until total failure of the beams. Summary Results of the numerical simulations indi- References cate slightly larger DELs for some of the The project “Intermittent Wind Loads” load channels for intermittent wind, com- (InterWiLa) has investigated the impact of [1] Wächter, M.; Heißelmann, H.; pared to Gaussian wind fields as recom- wind intermittency on the damage at typi- Hölling, M.; Morales, A.; Milan, P.; mended by IEC, see Fig. 2. However, this cal wind turbine composite materials and Mücke, T.; Peinke, J.; Reinke, N., is not the case for all of the load channels. structures. Numerical simulations did not Rinn, P.: The turbulent nature of the allow for clear conclusions on whether such atmospheric boundary layer and its impact on the wind energy conversion Results of the experiments, on the other impact exists or not. On contrary, in experi- process, Journal of Turbulence 13, hand, indicate a stronger damage (in terms mental investigations it was found that un- 2012 der intermittent inflow conditions the time of occurrence of the first crack) for loads [2] International Electrotechnical generated by intermittent wind, if inflow until the first damage of beam samples is Commission (IEC), International Standard 61400: Wind Turbines, International Electrotechnical Inflow with Commission, 2005 Inflow without [3] Popko W., Thomas P., Sevinc A., intermittency intermittency Rosemeier M., Bätge M., Braun R., Meng F., Horte D., Balzani C., Bleich Beam no. 3 5 7 9 4 6 8 10 O., Daniele E., Stoevesandt B., T [h] 9,0 41,0 1,3 1,5 N/A 96,0 5,3 1,5 Wentingmann M., Polman J. D., A Leimeister M., Schümann B., Reuter T [h] 45 828 98 188 71 739 122 76 A.: (2018), IWES Wind Turbine IWT- total 7.5-164 Rev 4, Fraunhofer Institute for Wind Energy Systems IWES, Bremer- haven, https://doi.org/10.24406/IWES- Table 1: Time until occurrence of the first crack (TA) and time until total failure of the beam samples in the test stand. The inflow wind fields had pairwise identical mean and N-518562 standard deviation for beams 3 and 4, 5 and 6, 7 and 8, and 9 and 10, respectively.

– 93 – 94WIND POWER SYSTEMS Annual Report 2015-2017

HyRoS – Multifunctional Hybrid Solution for Rotor Blades Protection

Universität Bremen Project Description methods for a systematic and methodical Institute for Integrated Product approach. The selection of potentially suit- Development (BIK) As part of the joint project "HyRoS" a mul- able methods will be based on the require- tifunctional protection of rotor blades of ments analysis. The methods of Design of Klaus-Dieter Thoben, Michael Brink, Experiments will be applied throughout the Jan-Hendrik Ohlendorf wind turbines will be developed on the basis of a hybrid material solution. The func- scientific and methodical development and Funding: Federal Ministry for Economic tionality includes the protection of the rotor implementation of the test series. For ex- Affairs and Energy (BMWi) blade surface against erosion as well as an ample, the determination of suitable mate- integrated de-icing system. This will be rial parameters and significant factors that Ref.Nr. 0325937F achieved through a new material combina- have an impact on production processes. tion of non crimp fabric and thermoplastic. These offer a systematic approach to iden- Duration: 12/2015 – 05/2019 Hence, an erosion of the rotor blades can be tify relationships of causes and effects and reduced. The new leading edge will also be make them quantifiable. The project goal, to equipped with a rotor blade heater, which relevantly reduce rotor blade erosion, offers will be operated with an intelligent energy- a number of advantages. For example, in- efficient control system. This is intended to creasing the efficiency of a wind turbine by Introduction prevent or reduce icing of the rotor blades in reducing the air turbulence at the surface of critical regions. Ice formation on the blades a rotor blade, and the risk of blade damage Rain, hail and frost are damaging the rotor is a particular threat in winter at inland lo- reduced, even at higher tip speeds envisaged blades of wind turbines. This is particularly cations and usually leads to the turbines for the future. In addition, the integrated evident at the leading edges, the front edges being shut down. From an economic point heating system provides new installation of the rotor blades directed against the wind. of view, these shutdown phases have a par- areas. A project successful in developing Ice can quickly build up there and erosion ticularly negative impact because in winter a multifunctional hybrid solution to protect damage is more frequent. All this considera- there are particularly good wind conditions the rotor blades, and proving its ability on a bly reduces the efficiency of a wind turbine. for electricity generation. Tests conducted demonstrator, will be a great contribution At present, the problem is mostly addressed by a turbine manufacturer in Sweden have to an environmentally friendly, reliable and with special paint systems and coatings. The shown that a turbine with rotor blade de- affordable energy supply. Institute for Integrated Product Develop- icing can generate at least 25 percent more ment (BIK) at the University of Bremen and electricity during the winter months than an partners in the joint project “Multifunctional identical turbine without de-icing. Summary hybrid Solution for Rotor Blades Protec- tion” (HyRoS) are focusing on new material The work package of the Institute of In- Wind turbines are exposed to extreme loads combinations in rotor blade production and tegrated Product Development (BIK) en- and environmental influences. Harmful ero- on integrated de-icing systems. compasses researching and developing the sion and weather-induced ice build-up oc- application technology, i.e. the production cur especially at the leading edges of the processes and devices to reliably apply the rotor blades. This can lead to considerable multifunctional rotor blade guard into the aerodynamic performance losses of the tur- forming tool. We are also responsible for the bine. As part of the "HyRoS" joint project, scientific and methodological implementa- multifunctional protection of rotor blades is tion and support of the test series. Based on being developed on the basis of a hybrid ma- present and future scenarios the joint project terial solution. On the one hand, this should will carry out a comprehensive requirements include protection of the nose edge against analysis for the technologies, materials, pro- erosion by a novel material combination cesses and devices to be developed. These of multiaxial fabrics and a thermoplastic, scenarios include the production of rotor and on the other hand it should prevent ice blades and hybrid rotor blade protection, as build-up by an integrated de-icing system. well as the operation of the rotor blades. The BIK will use various product development

– 94 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201795

Figure 1: Damage to wind turbines such as this one at the leading edge of a rotor blade requires extensive servicing and the use of special forces suitable for high altitudes. (Source: MB Bladeservice Einbeck)

– 95 – 96WIND POWER SYSTEMS Annual Report 2015-2017

Adaptive Operation and Control of Offshore Wind Farms Based on Specific Operational Strategies for Optimized Yield, Loads, and Grid Integration (RAVE – OWP Control)

Carl von Ossietzky Universität Oldenburg Project Description • Determination and monitoring the fa- Institute of Physics, tigue loads to set up a load envelope for Research Group: Wind Energy Systems The project pursues two main objectives – design and subsequent operation based Research Group: Turbulence, Wind Energy from a technical-scientific and strategic de- on a physical model of the dynamic wake and Stochastics velopment point of view – to help operating flow (WP C). New methods such as the Research Group: Energy Meteorology offshore wind farms more economically. so-called "Wake Meandering Model" are to be further developed using Lidar and Martin Kühn, Joachim Peinke, Detlev Heinemann, Andreas Rott, Technically and scientifically, the planned load measurements in the alpha ventus Vlaho Petrović, Isabelle Prestel, project aims at the methodical development wind farm and tested in industrial design Gerald Steinfeld, Matthias Wächter, of different operating strategies for yield, practice (WP D.3). This exact determina- Pyei Phyo Lin load and grid optimisation in offshore wind tion of the individual loads also forms the farms by regulating the whole wind farm. In basis for the various operating strategies Partners: Stuttgart University, addition to a novel, modular control concept previously to be developed in the second Stuttgart Wind Energy (SWE); for the active power of the individual wind sub-target, which are to use the load re- Global Tech I Offshore Wind GmbH; turbines, this is to be achieved by more pre- serves of the individual turbines in the German Lloyd Industrial Services GmbH cise determination of the fatigue loads of the wind farm economically. (DNV-GL) individual wind turbines in the wind farm • The concepts developed jointly by universities and industry are to be tested in Funding: Federal Ministry for Economic and their load reserve in relation to the de- Affairs and Energy sign loads. WP D using physical wind farm models and in industrial design practice for fa- Ref.Nr. BMWi (FKZ 0324131A) The second main objective is of a strategic tigue loads of future wind turbines on the developmental nature. It lies in the combi- one hand and tested regarding economic Duration: 03/2017 – 02/2020 nation of academic research on new control efficiency (WP D.3) on the other, in order and operation strategies as well as load mod- to be subsequently transferred to utilisa- els with current industrial developments for tion. the design and more economical operation Introduction of future large offshore wind farms. The second high-level project goal with its strategic development approach is to avoid The expansion of offshore wind energy use The first main - technical-scientific -re the major delays that have occurred in the is progressing rapidly. By 2015, more than search objective is pursued by the four tech- past, particularly in the practical, industrial half of the German government's target for nical work packages (WP) A to D and their application of new regulatory procedures in 2020, 6.5 GW installed nominal capacity sub-packages with the following sub-goals: wind energy. For this purpose for the offshore sector, has already been • Development of fast industrial wind farm • tasks are worked on cooperatively and achieved. Despite this increasing impor- simulations based on a flow-structure there is an intensive exchange between tance for the power supply, there is still coupling (WP A.1) to develop and test university and industrially managed sub- great potential for technical and economic control and operation concepts (WP B) tasks (WP A-D), improvement. This is the background to and calculations of individual fatigue • the new, risky control approaches for a the RAVE project, in which various new loads (WP C and D). This requires large- real wind farm with 80 wind turbines can simulation and control approaches are to eddy simulations as well as Lidar and be validated on the basis of real SCADA be developed jointly by universities and load measurements in the alpha ventus data (WP D.1), industrial partners and tested close to in- wind farm as a reference. • improved calculation methods for fatigue dustry. • Development of a modular, application- loads are being tested in industrial design oriented wind farm controller for differ- practice for future wind turbines (WP ent operating strategies for yield, load D.2) and grid optimization in offshore wind farms (WP B.1).

– 96 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201797

Use of the alpha ventus and Global is not representative of large offshore wind with validated models, but is not suitable for Tech I offshore wind farm data farms. In particular, multiple wakes (over- validation as such, as load measurements The many-sided subtasks rely on a data ba- lay of four or more wakes), as they occur are missing. sis that is obtained on the one hand from the in large offshore wind farms, do not occur alpha ventus test field and on the other hand in alpha ventus due to the layout. The data Brief description of the work packages from the Global Tech I offshore wind farm, from a wind farm with only 6 turbines ac- Fig. 2 illustrates the project structure. Above which went into operation in September cessible for the project is too small to vali- the block diagram, the two scientific work 2015. Previous RAVE projects have created date the short-term forecast model in WP objectives are mentioned, which mainly re- an extensive database for alpha ventus on B.1. For this reason, in addition to the meas- late to work packages B and C. An overview which to build and expand in this project. urements in alpha ventus, data from the of the four technical work packages is given The comprehensive sensor measurement Global Tech I wind farm with the Adwen below. In the short description of the result- system installed on the wind turbines in al- AD 5-116 turbine type are also collected. ing 10 subtasks, the leading project partner is pha ventus form the basis for some of the With 80 wind turbines, Global Tech I offers named in each case. concepts being investigated in this project. both good possibilities for the transferability In addition, the FINO1 measuring station of results from investigations on the largely A. Reference data wind farm flow offers the possibility of recording mete- identical turbines in alpha ventus and a Work package A includes data collection orological data required for the validation suitable reference for large offshore wind and the necessary measurement campaigns of measurement data and the calibration of farms. However, since only the standard from Global Tech I. Measurements at alpha measuring systems. However, the 3x4 tur- SCADA data is available in Global Tech I, ventus have been cancelled due to the re- bine layout of the alpha ventus wind farm this wind farm can be used for comparison cent incident at AV7. The data serves to

Figure 1: Offshore wind farms alpha ventus (top) and Global Tech I (bottom)

– 97 – 98WIND POWER SYSTEMS Annual Report 2015-2017

Figure 2: Project structure and higher-level project goals

validate the simulations and is used in all tal and wind farm conditions. At the same In A.3, high-frequency SCADA data are other work packages. time, the findings are also used to improve recorded by GT I, supported by ForWind, an industrial wind farm model of Global in the Global Tech I wind farm and meas- In A.1, SWE, supported by ForWind, pro- Tech I (GT I). urements of the inflow in the rotor centre vides a reference model environment that (spinner) are carried out on four wind tur- is used for aero elastic simulation of less Experimental validation of the above- bines. The SCADA data should be pro- complex wind farm test cases with spe- mentioned simulation environments in cessed and made available in a 1 Hz reso- cial external environmental conditions. A.1, SWE is done using measurements at lution in order to support and validate the In addition, this work package develops a Senvion turbines in A.2 in the alpha ventus models to be developed in B.1 and C.1. resource-saving, fast simulation environ- test field. These consist of a reduced Lidar ment that also provides reliable accuracy measurement campaign compared to the B. Development Wind farm controller and is intended to function as a wind farm original application, in which the inflow In the sub-package B.1, ForWind is de- test environment in D.1 for testing and conditions of two wind turbines are meas- veloping a modular wind farm controller verifying the new modular wind farm con- ured synchronously; this configuration rep- based on standard SCADA data to increase troller. ForWind uses the coupling of large resents a novelty in alpha ventus. In addi- yield, reduce load or support grid stability. eddy simulation and aeroelastic model. tion, the loads occurring in these systems, Using a shortest time forecast model, the This Coupling has been developed and val- which are equipped with additional load data are used to determine the condition of idated in other projects as new computa- sensors, are recorded and synchronized the wind farm in advance over the next few tional time efficient flow models in order to with the Lidar measurements. Results are minutes, taking into account various set correctly map the influence of atmospheric used to validate the simulation environ- point specifications. A cost function is used stability for specific selected environmen- ment in A.1. to evaluate the loads, yield and uniformity

– 98 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-201799

of the feed-in and on this basis the set point D. Validation and industrial testing specifications of the individual system per- Finally, the control and operating strategies formance can be optimised. The modular from WP B and the methods for load cal- wind farm controller is designed to extend culation from WP C are industrially tested the functionality of an existing wind farm and evaluated. controller. The qualitative characteristics of the mod- C. Development for calculation of fatigue ular WP controller are verified in D.1 by loads SWE supported by ForWind. SWE uses Work package C deals with both stochas- the wind farm test environment developed tic modelling of operational loads and the in WP A.1. Both test cases for the repre- analysis of individual fatigue loads in the sentative operating states and a test specifi- wind farm. cation for special cases requiring increased simulation effort are created. In C.1, ForWind develops stochastic models for operational loads and analyses them D.2 deals with the comparison of fatigue regarding uncertainties, transferability to loads between different simulation methods wind turbines of the same type and depend- (1. effective turbulence, 2. DWM, 3. wind encies on different environmental condi- farm test environment) and with measure- tions. In alpha ventus, stochastic models ment data based on C.3. The evaluation of for operational loads are to be derived from the fatigue calculation or the quantification available measurement data at the research of the error in the fatigue calculation us- data base. For verification of these models ing simple tracking models (effective tur- a generic numerical aero-elastic turbine bulence, DWM) enables an improvement model shall be used to ensure a reliable of the industrial design process and is to procedure. be included in guidelines and standards for certification. The task is managed by DNV- In Global Tech I this procedure shall be GL and is significantly supported by SWE. used to derive stochastic load models from The economic evaluation of the developed available operational data of that wind strategies and concepts is carried out in farm. For verification of these Global D.3 by the wind farm operator GT I with Tech I load models again the generic tur- the help of ForWind. In addition to the bine model shall be used. This way no financial estimation of the long-term use, intellectual properties of Adwen are trans- this also includes in particular the evalua- ferred from alpha ventus to Global Tech I. tion with regard to the provision of tertiary This scheme of work is illustrated in fig. 2. control power.

In C.2, SWE uses the Lidar measurements at Senvion turbines to perform time series validation. The aim is to optimize the existing simulation model regarding the repre- sentability of high-resolution environmental conditions. This is a prerequisite for the analysis of the individual fatigue loads at Senvion turbines by SWE with the support of DNV-GL in C.3. In this sub-task the fatigue loads are evaluated with different complex simulation environments for different environmental and wake situations. The partial shading is to be integrated into the fatigue load calculation and the load increase is to be analysed as a function of the wake situation.

– 99 – 100WIND POWER SYSTEMS Annual Report 2015-2017

Multi-Dimensional Stresses in High Power Electronics of Wind Energy Plants (HiPE-WiND)

Universität Bremen Project Description volved representing the industry in the Institute for Electric Drives, Power joint project. Electronics and Devices (IALB) The aim of the HiPE-WiND research project is to investigate high-performance The participating institutes are closely Holger Groke, Johannes Adler, power electronic systems for wind turbines linked with the wind turbine manufactur- Bernd Orlik, Nando Kaminski, ers as well as the manufacturers of the Christian Zorn under real environmental and load conditions, to investigate their failure causes frequency converters and components. Funding: 6th Energy Research Programme and to develop and experimentally verify By cooperating with the institutes as inde- of the Federal Government (BMWi) concepts for optimising their robustness. pendent bodies, all necessary information This requires tailored test-methods expos- and data can be collected, anonymised and Ref.Nr. 41V7736 ing entire converter systems to different evaluated. Furthermore, the databases of stresses simultaneously. The purpose of wpd and previous studies regarding field Duration: 10/2017 – 09/2020 these investigations is to obtain indications data acquisition serve as a basis for the de- of weak points in the system hardware velopment of application-relevant test se- through accelerated aging and to analyse quences and campaigns. In addition to the how to achieve an optimisation of the in- realistic multimodal load scenarios, a pro- verter service life by specifically influenc- cedure for accelerated testing is conceived ing the electrical loads. in order to achieve suitably shortened test Introduction periods compared to the 1:1 replication of The testing of specimen ranging from the field-typical load profiles. Further analyses The loss statistics of major German insur- power electronic component to the overall on component level in conjunction with ers as well as the "ReliaWind"-Study and inverter system requires a powerful test component simulations support the activi- field studies of the Fraunhofer Institute for system that allows for the application of ties towards increasing the service life of Wind Energy Systems (IWES) prove that typical electrical loads of a wind turbine, the whole inverter system under extreme the frequency converters of wind turbines simulated disturbances and system interac- conditions. have high failure rates and are among the tions in a reproducible and repeatable man- most frequently failing components of wind ner and as often as desired. Such a com- turbines [1,2]. plex test facility with the necessary load Summary functions and adaptable in performance In wind energy, the power electronics is and operating voltage is not yet available In recent years, there has been no signifi- exposed to particularly harsh conditions. It throughout Europe and is developed, set up cant reduction in the comparatively high has to sustain wind and net loads as well as and put into operation within the project. failure rate of power electronics in wind environmental stress. Wind loads are low It enables for manufacturer-independent turbines. On the contrary, in some cases the loads, overload or strong alternating loads research and an appropriate investigation failure rate has even increased. As the fail- due to fluctuations in the wind characteris- of the power electronics of modern wind ure causes have not yet been conclusively tics. Loads from the grid result from over- turbines. investigated, the problems are far from be- voltages and current surges due to switching solved. ing operations, short circuits or lightning In order to guarantee reliable and, in par- strikes. In addition, there are environmental ticular, application-relevant results, the The test facilities of the HiPE-WiND effects such as temperature gradients, high researchers at HiPE-WiND work closely project now enable detailed research of relative humidity, salty atmospheres etc. together with their industrial partners. the durability of power electronic systems The critical factor for the service life of the In addition to the Institute for Electrical used in state-of-the-art wind turbines. The power electronics, however, is the combina- Drives, Power Electronics and Devices examination of failure mechanisms under tion of environmental and electrical operat- (IALB) and the Fraunhofer Institute for realistic multimodal environmental condi- ing stresses to which the converter systems Wind Energy Systems (IWES), Enercon tions as well as load conditions promotes are exposed. (Wobben Research and Development the development of technical improve- GmbH), Breuer Motoren GmbH and wpd ments increasing the robustness of the con- windmanager GmbH & Co. KG are inverter system.

– 100 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-2017101

Figure 1 - Schematic representation of the planned test facility

References

[1] J. B. Gayo, „Final Publishable Sum- mary of Results of Project ReliaWind“, 2011 [2] A. Bartschat, C. Broer, D. Corona- do, K. Fischer, J. Kučka, A. Mertens, R. Meyer, M. Moriße, K. Pelka, B. Tegtmeier, S. Weber und J. Wenske, “Zuverlässige Leistungselektronik für Windenergieanlagen, Abschlussbericht zum Fraunhofer- Innovationscluster Leistungselektronik für regenerative Energieversorgung”, Stuttgart: Fraunhofer-Verlag, 2018

– 101 – 102WIND POWER SYSTEMS Annual Report 2015-2017

PiB – Predictive Intelligent Operation System to Reduce the Risk of Icing of Wind Turbines

Universität Bremen Introduction Fig. 2 shows a distribution of icing alarm Institute for Integrated Product durations created from SCADA-data of dif- Development (BIK) Wind turbines (WT) and in particular their ferent wind turbines acquired in the project. blades operating in cold climate areas are As we see, for some turbines, even single Klaus-Dieter Thoben, frequently facing icing problems during icing alarms can, in fact, last as long as al- Jan-Hendrik Ohlendorf, Stephan Oelker, winter operation, especially at low tempera- most a month. Markus Kreutz, Kamaloddin Varasteh tures and high humidity. Icing represents Funding: Federal Ministry for Economic a significant threat to the performance and To overcome the icing problem, two ap- Affairs and Energy (BMWi) durability of wind turbines [1]. Based on proaches exist. While de-icing approach Project Management: the researches in the project WECO [2] the tries to remove the ice after it has been al- Projektträger Jülich (PTJ) icing probability of the plants for Northern ready formed on the blades, the anti-icing Germany is between 7 to 14 days and in approach tries to anticipate the occurrence Ref.Nr.: 0324220A higher altitudes it can even reach values of of the ice and start preventive measures to up to 30 days per year (see fig. 1). Thus, ic- restrain the formation of the ice. In this re- Duration: 12/2017 – 11/2020 ing is a relevant factor for the profitability of gard, the main aim of the research project wind turbines. is to investigate and develop an intelligent management system that is able to forecast and simultaneously deal with ice accretion on rotor blades of wind turbines.

Figure 2: Early result – histogram of icing alarm durations from SCADA-data of several wind farms

Figure 1: Iced wind turbine, adopted from Project WECO [2]

– 102 – Annual Report 2015-2017 WIND POWER SYSTEMS Annual Report 2015-2017103

Project Description • Improvement of energy efficiency by the Summary possibility to heat the parts of a rotor blade The project aims to develop a system for differently and only when needed An intelligent, predictive operation system predictive and intelligent operation man- for reducing and avoiding icing downtimes agement that will significantly reduce the The project is in its initial phase, where the for wind turbines that shares its knowledge risk of icing of wind turbines. This system different types of data have been gathered not only across units of the same wind farm, consists of a data-mining component that from the different project partners and are but also across several wind farms, is being forecasts the icing events, and a heating sys- processed and filtered in a common data- developed in this project. tem in rotor blades that avoids the formation base with a uniform interface. This step also of icing. involves analyzing data that is not numeric The predictive model will be developed in nature and thus difficult to handle. using methods of machine learning. The In fact, the approach will be developed training data comes from a common data- based on data mining and data analytics Once the common database is completed, base and will include a great variety of data using all data resources including mete- the development of the model will begin received from project partners, such as his- orological measurements, SCADA data, using methods from machine learning. The torical SCADA-, weather- and plant-data life-cycle based, historical and live data. goal is to predict ice few hours before it ac- from several wind farm sites. As a part of the project, weather forecasts tually occurs, so that de-icing systems like and data corresponding to installed heating heating on the rotor blades can activate be- After developing the model and integrating systems in rotor blades will be taken into forehand in a manner that the icing can be it with de-icing systems on the wind tur- account. Moreover, the developed system is avoided altogether while keeping the de-ic- bines, a real world validation of the opera- not restricted to one wind turbine or wind ing activation time to a required minimum. tion system will be performed farm only, but also should make a linkage Additional measures to confirm the validity with further wind farms possible. Finally, of the icing prediction in situ are also being the various strategies will be developed in developed in parallel. order to confront icing before ice formation, which is considered as a particular feature At a later stage, after the completion of the of the project. model development, the integration of the model with de-icing systems on the wind The following specific scientific and techni- turbines to an intelligent management sys- cal objectives are the goals of the project: tem will be performed and the functionality • Increase the availability of a wind turbine of the system will be validated at a test site by means of an adapted operation man- in a cold region in Germany. agement and control of a heating system, even in weather conditions with a high risk of ice formation • Better planning basis for energy grid operators in Germany through a more accurate forecast of the availability of wind turbine capacity • Increase the ice prediction accuracy and possible ice formation and thus obtain an increase in energy yield References • Develop a system for predicting the risk of ice formation based on meteorological [1] Parent, O. ; Ilinca, A. : Anti-icing weather models and plant data and de-icing techniques for wind • Transfer of forecasts of ice formation turbines: Critical review. Cold Regions from individual wind turbines to other Science and Technology, Vol. 65, Issue wind turbines and entire wind farms (net- 1; 2011. pp. 88-96 [2] Tammelin, B. ; Cavaliere, M. ; working) Holttinen, H. ; Morgan, C. ; • Develop a comprehensive approach con- Seifert, H.; Säntti, K.: Wind Energy sidering heterogeneous data sources such Production in Cold Climate (WECO): as SCADA data, weather and environ- FMI, ENEL, VTT, GH, DEWI & FMI, mental data (regional and national) and 1998 plant data (design, location, history)

– 103 – 104WIND ENERGY INTEGRATION Annual Report 2015-2017

Advanced Wind Energy Systems Operation and Maintenance Expertise (AWESOME)

Carl von Ossietzky Universität Oldenburg, Project Description wind turbine (WT) power based on lidar Institute of Physics measurements. Research Group Wind Energy Systems One of the eleven young researchers is a 3. Develop a methodology to forecast WF Research Group Energy Meteorology PhD student at the research group Wind power based on radar measurements. 4. Develop a model to detect and forecast Laura Valldecabres Sanmartin, Energy Systems (WESys) from ForWind – WP ramp events based on radar Martin Kühn University of Oldenburg. Under the Work Package 3: "Wind Farm O&M Planning" measurements. Funding: EU the PhD student aims at developing very- short term wind power forecasts based on Structure Ref.Nr. EU – 642108 (ETN-ITN Projekt, remote sensing measurements to improve 1. Objective 1: Develop techniques to pre- Marie Skłodowska-Curie Actions) wind farm operation and the grid stability. dict very short-term wind speed with lidar measurements. Duration: 01/2015 - 12/2018 By increasing the accuracy of very short- In cooperation with DTU, Denmark, the term wind forecasts (< 30 minutes) the PhD student should develop techniques associated costs can be reduced. Reli- to forecast wind speeds based on lidar able forecasts would also benefit wind farm observations in a very short-term hori- (WF) operators since electricity markets zon of five minutes. are becoming more flexible with the use of 2. Objective 2: Develop a methodology to Introduction intraday gate closure times as short as five forecast wind turbine (WT) power based minutes. Very short-term wind power (WP) on lidar measurements. The increasing integration of wind energy forecasts are usually based on statistical pro- Using large eddy simulations, conducted into the grid, mainly driven by an increase cessing of historical data, so they are unable at the University of Oldenburg, the PhD in the installation of offshore farms, togeth- to detect unexpected WP changes. Long- student should develop techniques to er with the aging of existing onshore parks range lidars (light detection and ranging) forecast the wind power of an offshore has brought the focus on optimizing the op- and radars are capable of measuring wind wind turbine in different short-term hori- eration and maintenance (O&M) activities, speeds and partly direction up to 30 km and zons. The performance of the forecasting in order to ensure a cost-effective exploita- present a proper spatio-temporal resolution technique should be assessed under dif- tion of wind farms. O&M costs represent, for WF applications. ferent meteorological conditions. over the lifetime of a wind turbine, an aver- 3. Develop a methodology to forecast WF age of 20-25% of the total levelised cost of The WESys group has an extensive experi- power based on radar measurements. energy per kWh generated. Besides that, the ence on the use of lidars for different wind In cooperation with the energy com- European wind energy industry is having energy applications such as wake detection pany Ørsted and based on dual-Doppler difficulties in finding skilled professional and resource assessment. The global aim measurements the PhD student should to define new strategies for O&M activities. of the PhD project is to develop an accu- develop techniques to forecast determin- rate methodology for a very-short term WP istically and probabilistically the power With the aim of designing a training pro- forecast in an offshore WF based on remote of offshore wind turbines. gram that faces the main R&D challenges sensing measurements. 4. Develop a model to detect and forecast related to O&M and deals with the shortage WP ramp events based on radar meas- of highly-skilled professionals on this area, Objectives urements. the Advanced Wind Energy Systems Op- The development of an accurate methodol- The PhD student should focus of detect- eration and Maintenance Expertise (AWE- ogy for a very-short term WP forecast in ing and forecasting ramp events based SOME) project was conceived. The AWE- an offshore WF based on remote sensing on dual-Doppler radar measurements. SOME project (EC-GA contract no 642108) measurements is based on the achievement aims at educating eleven young researchers of the following four goals: in the field of wind power O&M while 1. Develop techniques to predict very building a training network that gathers the short-term wind speed with lidar whole innovation value chain. measurements. 2. Develop a methodology to forecast

– 104 – Annual Report 2015-2017 WIND ENERGY INTEGRATION Annual Report 2015-2017105

220

200

180

160

140

120

100

20 6266 0 6264 438 440 6262 442 6260 444 446 448 450 6258

Figure 1: Trajectories conducted with long-range lidars in the RUNE experiment [4]

References Kühn, M.: Optimizing LIDAR Summary measurements for very short-term [1] Gonzalez, E.; Nanos,E. M.; Seyr, H.; power forecasting using a LIDAR Long-range lidar and Doppler radar obser- Valldecabres, L.; Yürüşen, N.Y.; simulator, EAWE PhD Seminar, vations are investigated to develop very Smolka, U.; Muskulus, M.; Melero, J. J.: Cranfield, United Kingdom, short-term forecasts of wind power in off- Key Performance Indicators for Wind Poster presentation (2017) shore wind farms. These include the devel- Farm Operation and Maintenance, [4] Valldecabres, L.; Peña, A.; Energy Procedia, Volume 137, 2017, opment of forecasts of wind speed, power Courtney, M.; L von Bremen, L.; pp. 559-570, https://doi.org/10.1016/j. of single wind turbines and the aggregation Kühn, M.:. Very short-term wind speed egypro.2017.10.385. forecast of coastal flow by dual-Doppler of several wind turbines. Deterministic and [2] Valldecabres, L.; Friedrichs, W.; scanning lidar. WESC 2017, DTU, probabilistic approaches are investigated. von Bremen, L.; M Kühn, M.: Spatial- Copenhagen, Denmark. An important focus is put on analysing the temporal analysis of coherent offshore Oral presentation. forecast of ramp events. The outcome of wind field structures measured by scan- [5] Valldecabres, L.; Friedrichs, W.; the research activities is to be published in ning Doppler-lidar, Journal of Physics: von Bremen, L.; Kühn, M.: Analysis of several journal publications. The research Conference Series, 753(072028), doi: Wind Field Fluctuations on Wind Farm is aimed at reducing the uncertainty of the 10.1088/1742-6596/753/7/072028 (2016) Power Output, 12th EAWE PhD Seminar wind power predictions in minute scale ho- [3] Valldecabres, L.; Steinfeld, G.; on Wind Energy in Europe. rizons and to contribute to improve the wind Trujillo, J.-J.;L von Bremen, L.; Oral presentation. (2016) farm operation and the grid stability.

– 105 – 106WIND ENERGY INTEGRATION Annual Report 2015-2017

Resilience of Socio-Technical Systems exemplified at the Electricity Transport and Actor System

Carl von Ossietzky Universität Oldenburg wind power or photovoltaics are fluctuating These measurements were compared to Institute of Physics and difficult to predict. With an increasing freely available power feed-in data provid- Research Group: Turbulence, Wind Energy share of renewables, it is a challenge to sus- ed by ENTSO-E. As a result we see that in and Stochastics (TWiSt) tain a reliable energy supply system. times of large wind power feed-in especially many extreme grid frequency fluctuations Hauke Hähne, Matthias Wächter, are recorded, see fig. 1. Joachim Peinke Project Description Partners: Next, we pinpoint the impact of wind pow- Ulrike Feudel (Coordinator, Institute for The resilience of the electricity transport er injection on the grid frequency by the Chemistry and Biology of the Marine and actor system as an example of a socio- analysis of conditioned increment PDFs

Environment) technical system was investigated, based p(Δτ f |Pw). Thus, we learn how likely a

Klaus Eisenack, Thorsten Raabe on preliminary work in the fields of power frequency increment Δτ f is if an amount (Department of Economics) grids, non-linear dynamics, and turbulence Pw of wind energy is fed to the grid. We [1-5]. Here we showed evidence that wind show this PDF for a short (τ = 200ms) and Funding: Ministry for Science and Culture power feed-in impacts the short-time fluc- a long (τ = 10s) time scale for different of the Federal State of Lower Saxony tuations of the grid frequency in a measur- ranges of Pw in figs. 2(a) and (b). First, we able way. To this end a measurement setup observe that on the short scale, that the tails Ref.Nr. ZN3045 was developed which recorded high-fre- deviate from the normal distribution (gray Duration: 04/2015 – 08/2019 quency time series of the frequency in the reference curve), whereas the increment public electricity grid. Such measurements PDF is very close to normal on the long were taken from November 2016 to March scale. Second, for the long time scale, the 2017, and allowed for a derivation of the PDFs are almost identical, irrespectively

grid frequency in 5 Hz temporal resolution. of Pw . On the short time scale, however, Detailed information is available in ref. [6]. we observe a broadening of the distribution

Introduction

Socio-technical systems provide essential services for society. Important examples are supply systems (like urban water supply, ag- riculture and forestry, electricity systems) or critical infrastructure (like telecommunica- tions, harbors, and transport systems). Many socio-technical systems currently undergo major transitions or are under pressure from global environmental or demographic change. It is thus of utmost importance to understand how, under changing conditions, such systems can be socially managed in order to remain resilient to disruptions.

As an example of a socio-technical system we study electricity transport and actor sys- Figure 1: Large short-term increments accumulate on days with a high share of wind tems. This system is of crucial interest for power fed to the grid. Left axis and violet boxes: histogram of occurrences of large in- the German Energiewende, where an increments |Δτ f | > 2 mHz (τ = 200 ms) binned for two days for the first 70 days of our creasing share of renewable energy supply measurement. Right axis and orange curve: amount of onshore wind power fed to the is fed in. In contrast to conventional energy grid in Germany. Production data are taken from [11] and smoothed with moving average supply, important renewable energies like of two days.

– 106 – Annual Report 2015-2017 WIND ENERGY INTEGRATION Annual Report 2015-2017107

Figure 2: Wind energy feed-in affects the short-term term statistics of the power grid frequency. (a) PDF of increments Δτ f on the time

scale τ = 200 ms for different intervals of Pw (color-coded) normalized by the standard deviation σ1 of the smallest Pw interval. Increasing

feed-in Pw broadens the increment distribution resulting in a tenfold higher probability for a 5σ1 event (black arrow). (b) Increment PDF for a larger time scale, τ = 10 s. The increments follow the same, almost Gaussian, distribution; independent of the amount of wind

energy Pw fed to the grid. (c) Variances of increment distributions plotted against Pw for different time scales (color-coded). On the

shortest time scale, τ = 200 ms, the distribution becomes broader with increasing Pw . This effect diminishes with increasing time lags τ .

with increasing Pw . A statistical measure tational inertia will be much lower than to- sign and control strategies are not properly of the amplitude of frequency fluctuations day. This will lead to even faster frequency adapted, such frequency fluctuations may be- is the variance of these increment PDFs. In dynamics with larger amplitudes. If grid de- come highly critical for the grid stability [10]. fig. 2(c), we show that the variance of the

increment PDF increases with Pw for the smallest time lag τ = 200ms. For increasing time lags τ, this effect quickly diminishes. On time scales of τ = 800ms and above, References the variances show no clear trend with P . w [1] Giovanni, F.; Hejde Nielsen, A.;Falsig [6] Haehne, H.; Schottler, J.; Wächter, Pedersen, N.: Analysis of a power grid M.; Peinke, J.; Kamps, O.: The footprint using a Kuramoto-like model. The Euro- of atmospheric turbulence in power grid Summary pean Physical Journal B 61.4 pp. 485-491 frequency measurements. EPL (2008) (Europhysics Letters), 12, p. 30001 (2018) We have shown that wind power feed-in [2] Schmietendorf, K. et al. Self-organi- [7] ENTSO-E Operation Handbook impacts the power grid frequency on time zed synchronization and voltage stability Policy 1, https://www.entsoe.eu/filead- scales that lie below one second. The time in networks of synchronous machines. min/user upload/ library/publications/ range up to approximately one second is The European Physical Journal Special entsoe/Operation Handbook/ interesting because it lies in the range of Topics 223.12, pp. 2577-2592 (2014) Policy 1 .final.pdf activation of primary frequency control [3] Milan, P.; Wächter, M.; Peinke, J.: [8] Zhang X.-P.; Rehtanz C.; Pal B.: Flexible AC Transmission Systems: [7]. This suggests that fluctuations by wind Turbulent character of wind energy. Physical review letters 110.13, p. 138701 Modelling and Control, Springer Science power injection on longer time scales are (2013) & Business Media (2012) successfully compensated. Power quality is [4] Mehrnaz, A. et al.: Short term [9] Liang X.: IEEE Trans. Ind. Appl. 53 a key challenge for the grid integration of fluctuations of wind and solar power p. 855 (2017) renewable generators [9]. Although the ab- systems. New Journal of Physics 18.6 , [10] Schäfer, B.; Matthiae, M.; Zhang, X.; solute size of the fluctuations we consider is p. 063027 (2016) Rohden, M.; Timme, M.; Witthaut D.:

small (Δτ f < 20 mHz for τ = 200 ms), a pre- [5] Schmietendorf, K.; Peinke, J.; Phys. Rev. E 95: 060203 (2017) cise knowledge of the fluctuation statistics Kamps, O.: The impact of turbulent [11] ENTSO-E Transparency Platform, is essential to correctly estimate the proba- renewable energy production on power https://transparency.entsoe.eu/ bility of large, possibly critical, increments. grid stability and quality. The European In future power grids with a high share of Physical Journal B 90.11 (2017) p. 222. renewable energy sources, the amount of ro-

– 107 – 108WIND ENERGY INTEGRATION Annual Report 2015-2017

Influence of Short-time Dynamics of Renewable Energies on the Stability of Power Grids

Carl von Ossietzky Universität Oldenburg Project Description network of N synchronous generators and Research Group: Turbulence, Wind Energy motors transforming mechanical power and Stochastics (TWiSt) To capture the main characteristics of real into electrical power, or vice versa. The wind feed-in, we generate intermittent time coupled dynamics of the phase angles δ Katrin Schmietendorf, Matthias Wächter, i series on the basis of a Langevin-type model and magnitudes E of the complex nodal Joachim Peinke i and impose a realistic power spectrum. For voltages are given by Partner: comparison, we use correlated Gaussian Universität Osnabrück, Department of noise of the same spectrum and Gaussian Physics, Prof. Dr. Philipp Maaß white noise. The stochastic feed-in is implemented into a Kuramoto (KM)-like power Funding: Deutsche Forschungsgemein- grid model, capturing frequency and voltage schaft (DFG) dynamics in the order of seconds. The KM approach is based on electrical engineer- Network topology Ref.Nr. PE 478116-1 ing standards [2] and has also been used We first consider a two-machine (G-M) in nonlinear dynamics and control theory. system as the basic component of any Duration: 08/2015 – 07/2018 With this study, we investigate the impact power network, Fig. 1(a). Secondly, the of stochastic feed-in with realistic proper- complex network topology is based on an ties (temporal correlation, realistic power IEEE test system with 33 generators, 40 spectrum and intermittent increment statis- consumers and 108 links, Fig. 1(b). The

tics). We clarify which characteristics have machine parameters are set equal to γi =

to be taken into account for the likelihood of 0.2, αi = 2.0, Ci = 0.993, χi = 0.1. Pm,i and Bij Introduction noise-induced desynchronization given by are specified below. the average escape time. Feed-in fluctuations induced by renewables Implementing noise are one of the key challenges to the stabil- The Dynamical Model The increment probability density func- ity and quality of electrical power grids. In In power system dynamics, a synchronized tions (PDFs) of measured wind power data particular short-term fluctuations disturb state with constant frequencies, voltages significantly deviate from Gaussianity and the system on a time scale on which load and stationary power transfer is the desired its power spectrum S(f) displays 3/5 -decay balancing does not operate. Wind and solar mode of operation. Within the KM-like ap- with some discrepancy in the high frequen- power are known to be strongly non-Gauss- proach, the power grid is represented by a cy range (see Fig. 2). ian with intermittent increment statistics in these time scales. We investigate the impact of short-term wind fluctuations on the basis of a Kuramoto-like power grid model considering stability in terms of desynchronization. Details of these investigations are found in [1]. We present a procedure to generate realistic feed-in fluctuations with temporal correlations, Kolmogorov power spectrum and intermittent increments. By comparison to correlated Gaussian noise of the same spectrum and Gaussian white noise, we found out that correlations are essential to capture the likelihood of severe outages.

Figure 1: (a) Two-machine system: generator G with input Pm,1 transferring electrical po-

wer Pe,12 to a motor M with output Pm,2 . (b) IEEE topology: orange/red squares denote generators, blue circles motors.

– 108 – Annual Report 2015-2017 WIND ENERGY INTEGRATION Annual Report 2015-2017109

Based on this, we consider three types of synthetic feed-in noise. For the most realistic scenario, we generate intermittent power time series by use of the Langevin- type model

Secondly, we drop the intermittency feature and use Gaussian noise of the same power spectrum and standard deviation. By

enforcing the spectrum we induce temporal Figure 2: Escape scenario in the two-machine system. (a) Generator input P̃ 1(t) = Pm,1

correlations. This type of noise is referred + px(t) (D = 2.0 and p = 0.7), generator frequency ω1 and motor frequency ω2, the gray ̄ as Gaussian53 in the following. Thirdly, lines indicate the location of the stable limit cycle. Escape at Tout = 283. (b) Tout as a we put aside the applicational requirements function of D for p = 0.55, 0.6, and 0.7, calculated from M = 10 000 trajectories. Circles denote ̄ for Gaussian53 noise. and consider plain Gaussian white noise. Tout

System outages due to desynchronization in the two grid topologies

We define the first-escape time Tout as the time at which the first machine desyn- chronizes. Fig. 2 and 3 show examples of desynchronization scenarios for both grid topologies. For the spatial correlations of the noise we restrict ourselves to the two limiting cases, namely independent power feed-in P̃ i (t) and global noise P̃ i (t) = P ̃ (t). For the two-machine system, we observe in ̄ fig. 2(a) that atT out the system leaves its basin of attraction and enters the limit cycle Figure 3: Escape scenario in the IEEE grid: (a) Phase coherence r and frequencies ωi for region. Obviously, escaping is not simply a ̄ the first twelve machines, that fall out. Line colors indicate the machine type. (b) Tout as matter of the actual magnitude of the feed- a function of Nfl . Strongly intermittent (D = 2.0), weakly intermittent (D = 0.1) and Gaus- in noise, but essentially depends on the sian53 noise, each individual and global, calculated from M = 25 000 trajectories. For

system’s location in phase space. Fig. 2(b) global noise and Nfl < 10 stable configurations occur, which survive simulation intervals ̄ of T > 30 000. shows average escape time Tout as a function of intermittence strength D and for Gaussi- an53 noise for different penetrations p. The realistic power spectrum and intermittent on power systems with fluctuating feed-in higher the percentage of fluctuating input increments into a KM-like power grid. By should address spatial correlations, optimal and the intermittency strength D, the lower comparison to Gaussian correlated noise embedding, smart storage and control. the system stability with respect to noise- and Gaussian white noise we identified induced desynchronization. Results for the which feed-in characteristics have to be IEEE grid in fig. 3 follow similar trends. taken into account for investigations on References However, the differences between strongly power grid stability. In a complex transmis- intermittent and Gaussian53 noise, as ob- sion grid with a single node, e.g. represent- [1] Schmietendorf, K.; Peinke, J.; Kamps, O.: The impact of turbulent re- served in the two-machine system, are not ing a large offshore wind farm, severe out- newable energy production on power apparent here. ages may happen in form of noise-induced grid stability and quality, Eur. Phys. J. desynchronization. These are mainly due to B 90: 222 (2017), DOI: 10.1140/epjb/ correlation of the feed-in and the intermit- e2017-80352-8 Summary tency feature by itself plays a minor role. [2] Kundur, P.; Balu, N.J.; Lauby, M.G.: In summary, we showed that the character- Power system stability and control , We implemented intermittent realistic feed- istics of real wind power potentially desta- McGraw-Hill (1994) in fluctuations with temporal correlations, bilize the power grid. Subsequent studies

– 109 – 110WIND ENERGY INTEGRATION Annual Report 2015-2017

GEOWISOL – Effects of the Geographical Distribution and Temporal Correlation of Wind and Solar Input on the Power Supply System

Universität Bremen Bremen Institute for Metrology, Automation and Quality Science (BIMAQ)

Volker Renken, Michael Sorg

Funding: Federal Ministry for Economic Affairs and Energy (BMWi)

Ref.Nr. ID: 0325695B

Duration: 09/2015 – 06/2017

Project Description

The rising penetration of renewable energies became an important issue in the Ger- man electricity sector within the past years. For some regions the renewable generation is already surpassing the power demand. Because of the geographically and temporal varying distribution of the generation of re- Figure 1: Mean wind energy generation in MW 2014 of 95 zip code regions newable energy, a geographical distribution of the energy is inevitable. In order to plan the required infrastructure for the energy distribution, a detailed knowledge about the References geographical and temporal power generation is crucial. However, the data availabili- [1] V. Renken: Auswirkung der geo- ty for the distribution of the renewable pow- graphischen Verteilung und zeitlichen er generation in Germany is insufficient due Korrelation von Wind- und solarer to the complexity of the energy system there Einspeisung auf die Stromversorgung are only simulation based studies available. (GEOWISOL). Konferenz „Zukunftsfä- based data filling algorithms are introduced hige Stromnetze“, Berlin, Sep. 22-23, 2016, (oral presentation) For this reason, a real measuring data based and compared to conventional interpolation [2] V. Renken, M. Sorg, A. Fischer: comparison between the renewable power techniques [2]. As a result, the data filling Regionale Bewertung der Einspei- generation and the electricity demand is algorithms are validated and the power gen- sung erneuerbarer Energien auf Basis conducted within GEOWISOL [1]. The eration is shown to be very heterogeneous von Messdaten und einer modell- data is given as time series of 15 minutes over space and time. Due to the generated basierten Datenergänzung. DFMRS average values for each zip code region measurement based data set, infrastructure Windenergietagung, Bremen, March for wind, solar and demand quantities. For questions regarding the energy system can 16, 2017, (oral presentation) enhancing the still incomplete data, model- be answered with higher reliability.

– 110 – Annual Report 2015-2017 STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017111

IR-Vortex – Infrared Thermography as a Measurement Technique for Visualization of Vortex Structures on rotating Rotor Blades

Carl von Ossietzky Universität Oldenburg testing is done often visually or manually in ried out on a flat plate and on an airfoil Institute of Physics, a quite time-consuming way. Recently, non- model with artificial defects, here obstacles Research Group: Turbulence, Wind Energy destructive testing methods have gained of different sizes. These measurements are and Stochastics (TWiSt) increasing attention. Especially infrared performed in the wind tunnel to be able to thermography (IRT) is a very promising generate different wind speeds and Reyn- Gerd Gülker, Dominik Traphan, candidate. olds numbers. As a main result of the PIV Ivan Herràez, Sebastian Ehrich, measurements different flow regimes could Joachim Peinke be identified describing the effects depend- Partners: Fraunhofer Wilhelm-Klauditz- Project Description ing on the defect-height related Re-number Institute for Wood Research (WKI) Reh [1]. It shows, that below a certain critical

Braunschweig; The main goal of this joint project, which is value of Reh the turbulent wedges are main- O. Lutz, Authorized Expert on Fibre carried out in cooperation with the partners ly localized without spreading and without Composite Materials, Bundorf; Fraunhofer Wilhelm-Klauditz-Institute for destabilizing the boundary layer. The situa- KENERSYS GmbH Münster; Wood Research (WKI), O. Lutz Authorized tion changes above the critical value of Reh, RENERCO Renewable Energy Concepts Expert on Fibre Composite Materials, KEN- leading to a spreading of the wake and to AG München. ERSYS GmbH and RENERCO Renewable a detached boundary layer, which may be Energy Concepts AG, is to investigate the unwanted and can lead to undesired loads Funding: Federal Ministry for the Environ- recorded signatures within thermograms and losses. ment, Nature Conservation, Building and Nuclear Safety (BMU) and to relate these patterns to the associated flow situation and to identify the potential Ref.Nr. 0325401A for damage recognition and unfavored operational conditions. Duration: 11/2011 – 06/2015 To demonstrate the capabilities of IRT a typical thermogram of a real rotor blade in operation is shown in Fig. 1. The thermal signature clearly shows two main details, Introduction which are related to above mentioned problems. In the lower part turbulent wedges Due to increasing number of wind energy generated probably by surface defects lo- converters (WEC) new efficient monitoring cated near to the leading edge of the airfoil and diagnostic methods are required, which can be recognized. These turbulent regions can identify damages and failures remote- appear colder as the rotor blade is warmer ly during continuous-running operation. than the surrounding air, which is often the This is particularly important in offshore case in the early morning sun. Further on, a wind parks. Modern rotor blades represent clear line is visible which separates a warm- highly loaded parts of WEC worsened by er from a colder part of the airfoil and iden- progressive lightweight design and increas- tifies the natural laminar-turbulent transition ing length. Thus, defects within the glass of the boundary layer. or carbon fibre reinforced plastic as well as uncontrolled boundary layer transition have To determine, which temperature signature Figure 1: Thermogram of a rotating to be avoided. Consequently, quality control represents which flow situation, combined rotor blade show turbulent wedges and regular on-site testing of rotor blades is measurements with IRT and high speed due to defects and the transition line in essential for the economic operation. So far, particle image velocimetry (PIV) are car- darker colors

– 111 – 112STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017

Contrary to IRT measurements with PIV at operating WEC are difficult. Thus, the measurements simultaneously performed with IRT and PIV in the wind tunnel are compared to find out if both methods yield the same information about the flow situation within the boundary layer. In Fig. 2 results of both methods obtained in a turbulent wedge behind an obstacle of height h=1.75mm are shown color-coded at Figure 2: Comparison of PIV (foreground) and IRT (background, false colours) measure-

Reh=1150. ments of a turbulent wedge on a flat plate show very good agreement

In the background the thermogram is shown overlaid by an excerpt of the corresponding velocity field measured by PIV. The temperature is depicted color coded in arbitrary units to match the color range of the velocity field for easier comparison. Both measurements show very good agreement. This im- plies that any particular flow phenomenon near the surface can be potentially mapped in a thermogram.

To get an indication of the impact of the artificial defects on the airfoil performance, lift Figure 3: Wall shear stress from RANS simulation of a rotor blade with 10 turbulators force measurements were also conducted on an DU 91-W2-250 profile without and with five obstacles of diameter d=2mm each and for different angles of attack (AoA). Sur- prisingly we found, that these tiny modifications already lead to a decrease in lift force ing our wind tunnel measurements exhibits turbulent transition can be identified from at maximum region, which corresponds to that the simulation predicts very similar the temperature signature recorded by an an AoA of 10°, of about 5%. In addition, temperature structures representing turbu- easy to use thermographic camera. CFD particularly in the post stall range the fluc- lent wedges behind obstacles within the simulations of a whole WEC model with tuations of lift forces increase pronounced boundary layer. Moreover, it indicates the artificial defects on the rotating rotor blades by more than 50%, which lead to higher bigger influence of the outer disturbances show nearly the same structures as the ther- dynamic loads and may influence the pre- compared to the inner ones. In conform- mograms and can thus be used to get a better dicted lifetime of the rotor blades. ity with the findings in the lab experiments understanding of the underlying principles. this can immediately be explained by an in- Finally, for a better insight in the underlying crease of the velocity in a rotating system basic consequences of rotor blade damages from the root to the tip, which leads to an

and for validation a RANS simulation of a increasing Reh. Hence, outer obstacles are References complete WEC with DU 91-W2-180 airfoils already above the critical Reh, while the in- was accomplished using a k-kl-ω transition ner ones are below that value. model. In fig. 3 the result of the simulation [1] Traphan, D. et al.: High-speed for the outer half of an airfoil equipped with measurements of different laminar- turbulent transition phenomena on 10 surface defects in total is shown. Summary rotor blades by means of infrared thermography and stereoscopic PIV, In fig. 3 the wall shear stress on the surface The results of the present study show that Proc. 10th Pacific Symp. on Flow of an airfoil is shown which according to the IRT is a well suited method for rotor blade Visual. and Image Proc., Naples, 2015 Reynolds Analogy [2] is known to be close- inspection. Combined measurements of IRT [2] Kakac, S. and Yener, Y.: Convective ly connected to the temperature signature of and PIV have proven, that details of the flow Heat Transfer, Second Edition, CRC a thermogram. A comparison with fig. 1 and structure within the boundary layer e.g. tur- Press, 2nd edn. (1994) with the experimental results obtained dur- bulent wakes of small defects or laminar-

– 112 – Annual Report 2015-2017 STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017113

SHM_Rotorblatt: System for Early Damage and Ice Detection for Rotor Blades of Offshore Wind Turbines

Leibniz Universität Hannover have to be developed, in order to detect sification for changing environmental and Institute of Structural Analysis, damage at an early stage and prevent the operational conditions (EOCs) and hypoth- Institute for Information Processing total loss of the blade structure. esis testing by using the acceleration signals of six measurement positions that are Stavroula Tsiapoki, Thomas Krause, Structural health monitoring contributes to distributed over the blade length. A residue Stephan Preihs early damage detection and to the improve- from the stochastic subspace identification ment of competitiveness by preventing high (SSI) method and a residue from a vector Funding: Federal Ministry for the Environ- maintenance and repair costs. Despite the autoregressive (VAR) model were used, in ment, Nature Conservation, Building and existence of various monitoring systems, order to obtain two CPs. These are used as Nuclear Safety (BMU) the visual inspections are still necessary, indicators for changes in the response of since the existing systems have not reached the structure. The airborne sound acoustic Ref.Nr. 0325388 a satisfying level of reliability. mission damage detection approach moni- tors the blade with three fiber optical mi- Duration: 12/2011 – 06/2016 crophones. A model of the cracking sound Project Description was developed, which describes characteristics of these sounds in the time-frequen- Within the preceding project „Adaption cy-power domain. A detection algorithm and application of an early damage detec- uses these characteristics to detect dam- Introduction tion and load monitoring system for com- ages, to estimate their significance and to posite rotor blades of wind turbines“(FKZ handle environmental noise. Both methods The rotor blades of a wind turbine form a 0327644B) a method based on the propor- were applied on data from a fatigue test of vital part of the overall structure and their tionality of the velocity in the position of a 34 m rotor blade, which was harmonical- continuous monitoring is therefore of great maximum displacement and maximum dy- ly excited for over one million load cycles significance. The production of rotor blades namic stress was developed [2]. in edgewise direction, leading to a signifi- is still mostly performed manually and leads cant damage at the trailing edge. to a varying quality. Defects due to fabrica- In this project, a vibration-based SHM- tion, fatigue and extreme loads are the main scheme (see Fig. 1) and an acoustic emis- Further, the potential of combining the reasons for the beginning and propagation sion (AE) approach based on airborne two complementary approaches could be of damage. Among the most often encoun- sound were developed and tested for dam- shown. tered damages are bondline failure between age detection at wind turbine rotor blades. the upper and lower shell of the blade and Mechanical and acoustic data were gained delamination of the composite material. both under test facility conditions and at Summary Both damage types can influence the load operating state. For this purpose a fatigue bearing behavior of the rotor blade and po- test on a 34 m rotor blade (see Fig. 2) as The results of a SHM concept that consists tentially disturb the operation of the wind well as measurements on a 50.8 m blade of of the three steps of machine learning, the turbine. a wind turbine in operation were conduct- calculation of CPs and hypothesis testing ed. The fatigue test included an eigenfre- were compared to the results of an airborne The annual failure rate of rotor blades is quency test, an ice-accretion detection test, sound damage detection algorithm for the relatively low compared to other compo- as well as a fatigue test in the edgewise di- fatigue test of a 34 m rotor blade. Four blade nents of the turbine, such as the electrical rection until the development and propaga- states were documented during the fatigue system, but an occurrence of damage on the tion of damage. During the measurements test: undamaged state, existence of fatigue, rotor blades causes a longer period of down in operation the mechanical and the acous- occurrence of a significant failure at the time that can reach up to 5 days [1]. The tic system will be installed in a 3.4 MW trailing edge and damage propagation. standstill of the wind turbine results in great wind turbine, in order to obtain data that costs, which are even higher in the case of correspond to the real operating conditions. For the SHM scheme, two CPs resulting offshore wind turbines. Therefore, the rotor The vibration-based approach includes the from the SSI method and VAR models were blades should be continuously monitored estimation of condition parameters (CPs), presented for different settings, such as two and structural health monitoring methods machine learning by means of data clas- types of classification, confidence intervals

– 113 – 114STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017

Figure 1: Modular SHM framework: Training and subsequently testing data instances are analyzed through a combination of a machine learning algorithm, condition parameters and probabilistic models

and assumptions for the CP distributions. Both provided detection of the damaged state of the blade for both cases of structural changes due to fatigue, damage occurrence at the trailing edge and propagation. The airborne sound damage detection algorithm detects the occurrence of the structural damage without a false detection. Events before the damage and events, where damage propagation took place, were detected without false alarms. The observation of the results of the two approaches showed that the comparison and Figure 2: Fatigue test on a 34 m rotor blade and instrumentation combination of different damage indicators derived from different methods and physical quantities can be beneficial for SHM. The two approaches function complemen- References [3] Tsiapoki, S.; Krause, T.; Rolfes, R.; tary, since the CPs of the SHM-scheme Ostermann, J.: Erweiterung und provide information concerning the state [1] ISET, Germany, http://www.iset. Erprobung eines Schadensfrüherken- of the blade (healthy or damaged), while uni-kassel.de/pls/w3isetdad/www_ nungs- und Eisdetektionssystems für the acoustic emission approach provides iset_new.main_page Rotorblätter von Windenergieanlagen [2] Gasch, R.: Eignung der – SHM.Rotorblatt: Abschlussbericht instantaneous information about dam- Schwingungsmessung zur Ermittlung zum Forschungsvorhaben der Gottfried age events such as damage occurrence or der dynamischen Beanspruchung in Wilhelm Leibniz Universität Hannover, propagation. Investigations for the appli- Bauteilen, PhD Dissertation, Institut für Statik und Dynamik (ISD), cation of both approaches under operating Technische Universität Berlin, Institut für Theoretische Nachrichten- conditions belong to the future work of the Faculty of Mechanical Engineering, technik und Informationsverarbeitung authors. A more detailed description of the Berichte aus der Bauforschung (tnt), Leibniz Universität Hannover. results be found in the final report of the Nr. 58 (1968) (2016) project [3].

– 114 – Annual Report 2015-2017 STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017115

Monitoring Suction Bucket Jacket – Development of feasible Numerical Simulation Methods for the Investigation of the Compression and Tension Load bearing Behaviour and the Prediction of Displacement Accumulation

Leibniz Universität Hannover kum Riffgrund 1 in the German North Sea. Project Description Institute for Geotechnical Engineering The turbine was manufactured by Siemens Martin Achmus, Christian Schröder, Klaus and has a power of 4 MW (Type SWT In this project both the installation and the Thieken, Jann-Eike Saathoff, Tim Gerlach 3.6/4.0 120 m). To date there is a lack of bearing behavior of the foundation and the practical experience with such foundation overall structure were observed intensively Institute of Structural Analysis structures under the typical load situations by a monitoring system. The measurements Raimund Rolfes, Tanja Grießmann, Niko- for offshore wind turbines (large horizontal were used to validate the calculation meth- lai Penner, Andreas Ehrmann and moment loading at relatively low static ods and structural models developed in the vertical loads with possibly large recurring project. The joint project Monitoring SBJ Funding: Federal Ministry for Economic is divided into the three sub projects of the Affairs and Energy (BMWi) tension loads). Consequently, there are neither standardized methods of calculation Leibniz University of Hannover (LUH), Ref.Nr. 0325766A nor proof concepts available. Also validat- the company DONG Energy Renewables ed structure models are missing that take Germany GmbH (DONG) and the Federal Duration: 08/2014 – 02/2017 the nonlinear and load history dependant Institute for Materials Research and Test- soil structure interaction into account and ing (BAM). The Federal Maritime and Hy- with whom the bearing behavior due to cy- drographic Agency (BSH) is an associated clic and dynamic loads can be calculated. partner.

Introduction

The central eco-political goal of the Fed- eral Government is to increase the share of renewable energies of the gross power consumption to about 80 % up to the year 2050. At present Germany is still in the beginning of extensive construction activity in the North Sea and Baltic Sea. More and more it has to be assumed that several construction projects have to be carried out simultaneously to accelerate the development of offshore wind energy and to reach the ambitious goals.

DONG Energy Wind Power has developed an innovative foundation structure in which an optimized industrially manufactured jacket is combined with suction buckets (hereafter called "SBJ"). The SBJ (see fig. 1) is designed for 5 - 8 MW turbines and waterdepths of up to 60 m and was installed in August 2014 in OWP Bor- Figure 1:Design of the Suction Bucket Jacket of DONG Energy Wind Power [1]

– 115 – 116STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017

LUH, DONG and BAM worked closely After a description of the state of the art, model with load-dependent stiffnesses. In together on the subprojects under the co- investigations were presented concerning a further step, the geotechnical modelling ordination of LUH. The LUH subproject the influence of a loose layer embedded in of the SBJ foundation was considered by was executed by the Institute for Geotechni- dense sand. It was shown, that especially for 3D numerical models. The modelling of cal Engineering (IGtH) and the Institute of large heave rates, a loose layer nearby the the overall model proved to be inappropri- Structural Analysis (ISD). tip has a significant influence on the evolu- ate, since the very high number of elements tion of suction pressure and therefore on the requires very large computing power. As a bearing behaviour. Additionally, the SBJ result, individual bucket models were de- Summary soil profiles were considered. veloped that take the influence of the rising jacket structure on equivalent restraining To reach the goals of this project several Within the investigations on heave accu- conditions into account. Two approaches work packages were defined. The results of mulation due to cyclic tensile loading, the with different accuracy were developed. In these work packages are summarized briefly used material model hypoplasticity with its the standard approach, the restraining ac- in the following. More information is avail- extension intergranular strain led to implau- tion of the jackets is realized by a torsion able in the final report [1]. sible results. Investigations with a simpli- spring, whereby the rotation of the jacket is fied material behaviour indicate that cyclic neglected. Improved result of the behaviour Installation process (IGtH) loading with a load magnitude greater than can be achieved by an advanced approach The soil exploration data delivered by the drained bearing capacity leads to an ac- with an influence on the torsion spring. The DONG were checked and evaluated to de- cumulation of heave. Thereby, the accumu- suitability of both models could be shown. rive a geotechnical soil model containing lation rate is dependent on the load magni- Subsequently, an extensive parameter study the relevant soil parameters for further cal- tude. was used to investigate the main parameters culations. Soil parameters were derived by influencing the bucket foundation behav- CPT correlations and yielded best estimate, Foundation stiffness (IGtH) iour. lower and upper bound soil profiles. Flow The nonlinear soil-structure interaction rep- net calculations considering the former soil resents an essential part both for the design After applying the load-dependent stiffness profiles were carried out to calculate the of the foundation as well as for the entire matrices, using both approaches to the ex- critical suction for certain boundary condi- structure. Only a sufficiently accurate es- ample of the SBJ prototype, the considerations like soil loosening or anisotropic water timation of the foundation stiffness in the tion of the foundation reaction was extended permeabilities. The critical suction is a lim- overall model allows the sustainable de- to un- and reloading conditions. It results in iting value for the admissible suction to pre- sign of the structure by utilizing the exist- a lower nonlinearity, which cannot be as- vent hydraulic failure during an installation. ing soil resistances. At present, there are no sumed as a completely linear behaviour. Parameter studies were carried out with approved spring approaches for describing This is different with a mono-foundation. three different analytical methods to simu- the load-dependent foundation stiffness of The foundation reaction appears to be ap- late the installation process considering the bucket foundations. In order to determine proximately linear when reloaded and the previously obtained results. Back calcula- the system-dependent resistances, it is load-dependent foundation stiffnesses for tions using the measured data during the therefore necessary to rely on very compu- monopiles consequently are dispensable. installation of the SBJ show that the recom- tationally expensive numerical continuum It should be emphasized that the developed mended bandwidths of calibration param- models. Furthermore, the rotational fixing interface can also transfer the post-cyclic eters for the respective analytical methods dependence of the rising jacket structure re- load-bearing behaviour of any geotechni- can be regarded as suited for the considered sults in a high complexity. cal model to the overall model and thus installation. Based on these needed features, the devel- has more significance far beyond the work opment of an innovative interface between scope. Bearing capacity and deformation the geotechnical model and the overall accumulation (IGtH) model is presented, which allows the inte- Measurement data analysis and system The objectives within this work package gration of the nonlinear foundation stiffness identification (ISD) were the investigation of the bearing behav- of any geotechnical model into the overall Within the scope of the work package ex- iour of suction buckets under tensile load- model without the use of an iterative proce- tensive measurement data were analysed. ing in homogenous and especially layered dure. The derived load-dependent stiffness The analysed data were used to identify subsoil. Besides pure axial tensile loading, matrices can then be integrated in the over- the dynamic behaviour of the structure and inclined loads should be accounted for. In all model by means of so-called superele- to gain an improved understanding of the addition it was unclear, if a cyclic tensile ments. It could be shown that the nonlinear global load-bearing behaviour. loading package may lead to an accumula- foundation behaviour can be completely tion of heave. defined with a few load-controlled calcula- Data preprocessing could provide an impor- tions and can be transferred to the overall tant contribution to quality assurance by

– 116 – Annual Report 2015-2017 STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017117

using efficient algorithms on the basis of with six degrees of freedom and considered statistical values (metadata) and the com- as linearized for a certain operating point. parison of redundant or correlated informa- A preliminary study examines a more pre- tion. cise numerical modeling of the tubular joint stiffness in terms of its influence on the To assess the influence of environmental structural eigenfrequencies. and operational conditions on the global dynamic behaviour of the structure, a classi- Based on this, a three-step global sensitivity fication was performed. The essential influ- analysis (one-at-a-time analysis, regression ences were taken into account by means of analysis, variance-based sensitivity analy- wind speed, generator speed and pitch angle sis) is conducted to detect those design pa- of the rotor blades. rameters, which affect the regarded eigenfrequencies the most. The identified eigenfrequencies and mode shapes were representing the system pa- Finally, a model updating (or model cali- rameters, which were identified by the op- bration) is carried out with this reduced set erational modal analysis method Frequency of parameters to adjust the lower eigenfre- Domain Decomposition. In addition to the quencies and mode shapes of the numerical usual identification based on acceleration model according to the results from meas- measurements, all strain measurements urement data of the real structure in idling were successfully used. position. Here, an evolutionary algorithm is applied for optimization. The histograms and probability density functions were used for the evaluation of Hydrosound measurements (ISD) the identified eigenfrequencies. All correc- The goal of this work package was the tion methods were considered separately to provision of the experimental proof that allow comparisons. In case of the second the installation process of the SBJ is a eigenfrequency of the suction bucket jacket, low-noise construction method. The hy- the influence of the environmental condi- drosound measurements and evaluations tions could be removed to a high extent by were conducted by itap GmbH (Institut für calculation. technische und angewandte Physik GmbH, Oldenburg). The noise levels were meas- With the experience gained from the measured at three positions before, during and uring system, the suitability of the existing after the installation process. A comparison measuring system can be assessed and im- with measurements of the background noise provements can be made. before the installation shows that the perma- nently existing background noise is clearly Efficient modeling and model increased by the noise of the accompanying validation (ISD) construction ships but the actual installation The simulation of the entire offshore wind process of the SBJ does not lead to a further turbine, consisting of rotor-nacelle-assem- increase of hydrosound in a distance of at bly, tower, and jacket substructure with least 500 m. The hydrosound measurements suction bucket foundation, is an important thus confirm the assumption of a low-noise enhancement in addition to the detailed in- construction method. vestigation on the bucket. Here, effects of the foundation on the behavior of the whole structure can be regarded within the under- References lying model assumptions. [1] Achmus, M. : Verbundprojekt: At first, the structural model is described Monitoring Suction Bucket Jacket – based on the available design documents. Monitoring SBJ. Final Report, Leibniz The soil-structure-interaction of the in- University of Hannover, Institute for stalled buckets, located in the North Sea, is Geotechnical Engineering, (2017) represented as a reduced model at one node

– 117 – 118STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017

Multivariate Structural Health Monitoring for Rotor Blades

Leibniz Universität Hannover of wind turbines. The investigations are car- a detailed numerical model. Therefore, the Institute of Structural Analysis ried out first on a destructive fatigue testing integration of the fully available rotor blade of a large-scale rotor blade in a laboratory data into the development of SHM process- Marlene Bruns, Aamir Dean, Christian setup (cf. Figure 1) and subsequently on an es represents such a unique characteristic of Gerendt, Tanja Grießmann, Eelco Jansen, operational offshore wind turbine. Thereby, MultiMonitorRB. In most research projects, Helge Jauken, Dorian Pache, a unique characteristic of this joint research this knowledge is not fully available, since Raimund Rolfes project includes the full knowledge of the it is normally strongly protected by the wind Funding: Federal Ministry for Economic geometry and the material properties as well turbine manufacturer. Affairs and Energy (BMWi) as information about the blade manufacturing of the exploratory rotor blade. Based on the design data of the rotor blade, a Ref.Nr. 0324157A detailed numerical model is created. On the one hand, this model enables the application Duration: 03/2017 – 02/2020 Project Description of model-based SHM methods. On the other hand, it is used for the simulation of mate- Monitoring systems are an important re- rial fatigue via damage evolution as well as quirement for increasing system availability for the assessment of residual strength. In and reducing the costs. Still, damage moni- addition, a central new idea in the project toring systems are not yet well established, is the combination of numerical simulation Introduction because there do not exist clear limits that models with data-driven SHM methods. detect damage and can thus give a clear As one of the main renewable energy re- warning. Furthermore, there exist many dif- In order to estimate the material fatigue oc- sources, wind energy has gained an impor- ferent rotor blade designs that have differing curring in the rotor blade during the fatigue tant role in the generation of sustainable properties. Even blades of the same produc- testing, a so-called Fatigue Damage Model energy. For this reason, the aim to achieve a tion show differing structural behavior due (FDM) is implemented. The model bases high degree of utilization as well as the aim to the manual manufacturing process. This on the idea of [2], which expands the linear to enhance the durability of wind turbines is why fundamental investigations of the damage accumulation theory (cf. Palmgren- became vital research topics. Therefore, the damage development mechanisms are nec- Miner approach [3]) with a non-linear fa- ability to identify structural damage and essary, which can only be carried out with tigue model. The measured data recorded consequently prevent component failure is a significant tool of interest in relation to flexible service intervals and condition- based maintenance of wind turbines. Since the rotor blades represent a centerpiece of a wind turbine, their failure means downtime for weeks or even months, resulting in high repair costs. This is why rotor blades are counted among the leaders in terms of the duration of the downtime and the repair costs caused by their failure, although they are not leading in terms of damage frequency [1].

Essential goals of the project “Multivari- ate Structural Health Monitoring for Rotor Blades” (MultiMonitorRB) are the development, the combination and the testing of global as well as local structural health Figure 1: Testing of a large-scale rotor blade at the Fraunhofer Institute for Wind monitoring (SHM) methods for rotor blades Energy Systems in Bremerhaven

– 118 – Annual Report 2015-2017 STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017119

during the testing then serve to validate the simulation results.

A preliminary simulation of the fatigue testing was performed under defined test conditions, including an edgewise excitation up to one million cycles. Results showed that a matrix-dominated fatigue damage of areas located at the leading and trailing edge is to be expected. Especially the areas at the trail- Figure 2: Matrix-dominated fatigue damage at the trailing edge ing edge between 5 and 10 m from the hub- end (cf. Fig. 2) as well as at the leading edge located 6 m from the hub (cf. Fig. 3) show significant fatigue damage.

The definition of different load scenarios as well as different test conditions is to simulate different damage scenarios, which gives additional information on the sensitivity of different features with respect to the occurred damage. This information is used to investigate and develop data-driven methods that identify condition parameters used Figure 3: Matrix-dominated fatigue damage at the leading edge to monitor the structural health. In addition, knowledge about the sensitivity of the different features is used to identify suitable features for model-based SHM.

First steps concerning model-based SHM were conducted on a simple beam model of the large-scale rotor blade. Since the testing was not yet performed, a reference model and a model with an artificially induced damage were created. The damage was assumed to be material fatigue, simulated by a stiffness reduction. An initial study showed that an increase of design variables results in an objective value space with many lo- Figure 4: Parameterization of the damage distribution function along the blade length cal minima, making numerical optimization for a two-dimensional model unfeasible [4]. Therefore, approaches are needed that minimize the amount of design variables and thus, the dimension of the optimization problem. Most common methods use the assignment of one design variable to a group of finite elements having similar mechanical properties [5, 6]. Within this project, an alternative approach is analyzed that uses a damage distribution function, which can be described only by few variables. Fig. 4 illustrates a damage distribution function designed for two-dimensional beam structures, resulting in three design variables. Of course, the goal is to apply Figure 5: Parameterization of the damage distribution function along the blade length this damage distribution function to a three- for a three-dimensional model

– 119 – 120STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017

dimensional shell model, resulting in an additional fourth design variable, representing the peripheral direction (cf. Fig. 5).

Due to the use of various SHM systems within this joint research project, a harmo- nization and temporal synchronization of the differing measuring systems have to be planned and carried out. Only if these conditions are met, a comparability of the acquired measurement data and thus, the detection results is ensured among the project partners. At ISD, an electrical and a fiber optical measurement system as well as electrical and fiber optical accelerometers were purchased and are currently being prepared.

Summary References The aim of the project MultiMonitorRB is the development, combination and testing [1] Hahn, B.: Schadensstatistik von of SHM methods, that are able to capture WEA und präventive Schadensvermei- different features and damage indicators. dung, Windenergie NRW, November In sense of a multivariate procedure, dif- 29-30, 2012, Bad Driburg, Deutschland [2] Krueger, H.: Ein physikalisch basier- ferent vibration-based and acoustic SHM tes Ermüdungsschädigungsmodell zur approaches are considered. These SHM Degradationsberechnung von Faser- methods are to guarantee an automated and Kunststoff-Verbunden, PhD thesis, reliable detection, localization and classifi- Institute of Structural Analysis, Leibniz cation of relevant damages during the early Universität Hannover, Deutschland, stage. 2012 [3] Miner, M. A.: Cumulative Damage The project sets an important milestone in Fatigue, Journal of Applied by uniting data-driven SHM methods with Mechanics 12 (1945), pp. 159-164 numerical simulation methods using fully [4] Bruns, M.; Hofmeister, B.; known rotor blade design data. This pro- Pache, D.; Rolfes, R.: Finite element model updating of a wind turbine blade cedure promises to gain knowledge, very – a comparative study, in: EngOpt 2018 likely resulting in better damage detection Proceedings of the 6th International and localization procedures. Conference on Engineering Optimization, Springer International Publishing, 2019, pp. 569-580 [5] Schroeder, K.; Gebhardt, C. G.; Rolfes, R.: A two-step approach to damage localization at supporting structures of offshore wind turbines, Structural Health Monitoring 17 (5), pp. 1313-1330, (2018) [6] Levin, R.; Lieven, N.: DYNAMIC FINITE ELEMENT MODEL UPDATING USING SIMULATED ANNEALING AND GENETIC ALGORITHMS, Mechanical Systems and Signal Processing 12 (1), pp. 91-120 (1998)

– 120 – Annual Report 2015-2017 STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017121

Development and Validation of a Virtual Process Chain for Composite Structural Components Considering Imperfections with Application to a Rotor Blade Component (PROSIM R)

Leibniz Universität Hannover Project Description are mentioned as examples here, the appro- Institute of Structural Analysis priate data is gathered at the handling and A major issue in the industrial application at the infusion stage respectively. As an out- Raimund Rolfes, Benedikt Daum, of fiber-reinforced composites is the estab- come of the manufacturing simulation, all Eelco Jansen, Gerrit Gottlieb lishment of a reliable, fast and cost-effective the necessary information about the relevant manufacturing process. Production-related imperfections are available for the final me- Universität Bremen manufacturing defects can significantly re- chanical strength analysis of the composite Institute for Integrated Product duce the laminate quality. In order to opti- component. In the mechanical analysis, the Development (BIK) mize the production process, it is therefore influence of manufacturing process-related Klaus-Dieter Thoben, essential to determine precisely the influ- defects on the linear and non-linear struc- Jan-Hendrik Ohlendorf, Likith Krishnappa ence of manufacturing defects on the mate- tural-mechanical behavior is then quantified rial properties. via systematic numerical investigations. Funding: Deutsche Forschungsgemeinschaft (DFG) In a consortium of three project partners, the Since imperfections occur at fundamental- research project, PROSIM R aims to inves- ly different length scales, both the process Project Nr. 329147126 tigate essential parts of the process chain of simulations as well as the mechanical anal- a rotor blade by numerical and experimen- ysis are based on a multi-scale approach. Duration: 08/2017 – 08/2020 tal techniques. The following sub-processes Following a sequential homogenization are considered: handling and draping of the multiscale scheme, mechanical simulations reinforcing flexible textile, infusion and are first performed for the fiber-matrix mi- impregnation with matrix resin and curing cromechanical model in order to predict and finally the mechanical analyses of the the effective material behavior of a rov- Introduction reinforced fiber composite. The results from ing (micro scale). The material parameters each process step are passed on from one determined by homogenization of the mi- Rotor blade development is driven by in- process simulation to the next (see Fig. 1). cro scale models are then used at the next creasing performance and cost require- Thus, the project PROSIM R extends the scale level defined by the roving architec- ments. In this context, the shortening current state-of-the-art models which only ture (meso scale). The material properties of cycle times during production, quick considers individual process-steps and not resulting from the second homogenization adaptation of the production process to process simulation as a whole. This strategy are then available as effective properties for changed materials and quality-control of allows the project, PROSIM R to overcome structural simulations at component scale rotor blades are becoming increasingly rel- a the major drawback which is the actual (macro scale). In the course of this multi evant. These issues are of particular con- representation of imperfections (e.g. voids scale process, imperfections are represented cern, since manual labor is still the main or geometric deviations of the fiber path) in on each of the three scale levels, as neces- method of production. It is thus essential to mechanical analysis due to the lack of ac- sary. acquire a deeper understanding of how the curate data. In the project PROSIM R, the manufacturing process affects the resulting imperfection data is predicted via the manu- Many imperfections can be characterized mechanical behavior. These issues define facturing processes, rather than making use only by statistical distributions of param- the scope of the research project, ‘PRO- of some assumptions. Parameters charac- eters, rather than single-valued determin- SIM R’. The main aim of this project is to terizing different kinds of imperfections, istic parameters. For example, the volume predict imperfections and process fluctua- such as roving undulation or the local resin occupied by a void is random for every void tions in composite components via a series saturation, are predicted at the appropriate instance, however, the void-volume still of linked process simulations. Results from stage in the simulation chain and are passed follows a certain frequency distribution. these process simulations then form the ba- on to the next stages of simulation. For rov- In order to ensure a correct representation sis for subsequent mechanical analyses. ing undulations and resin saturation, which of such statistical properties, probabilistic

– 121 – 122STRUCTURAL HEALTH MONOTORING Annual Report 2015-2017

analyses are performed from information Summary cally reduces the number of necessary tests. which can be obtained from the manufac- Furthermore, the developed methods can turing process simulations. The project is application oriented. Expected account for the influence of manufacturing results from the project will improve the ac- imperfections on the structural-mechanical Validation of the numerical models will curacy of static strength prediction of practi- behavior of a finished component and can be performed at each scale level, so that cal, and thus imperfect, rotor blades based help to determine the fault tolerances during the possible modelling errors can be iso- on information available from the manu- the manufacturing processes. Respective lated and subsequent improvements are fa- facturing process. Moreover, the project is criteria and design rules for fault tolerances cilitated. Also, in the case of the micro and expected to illuminate the relative and abso- could help in reducing scrap and facilitate meso scale models, validation of statistical lute impact of typical imperfections on the the certification processes. quantities will be attempted by testing a se- static strength of fiber reinforced composite ries of appropriately designed coupons. At components in general. This is expected to component scale, only a small number of lead to further advances in material develop- tests can be carried out due to financial con- ment, especially for the integration of new straints, thus, one or a few static tests will materials into existing processes. Currently, be carried out for a representative structural complex test series are necessary to adapt component, i.e. a segment of the rotor blade manufacturing processes when changing spar-cap. materials. A virtual process chain dramati-

Figure 1: Representation of the process simulation of rotor blade production planned in the PROSIM R project

– 122 – Annual Report 2015-2017 ACOUSTICS Annual Report 2015-2017123

INFLOW-Noise

Carl von Ossietzky Universität Oldenburg Project Description Summary Institute for Physics, Research Group: Turbulence, Wind Energy During the previous reporting period most The measurements carried out to validate and Stochastics of ForWind’s work within this project has the results of the CFD simulations were processed and a report on the measurements Joachim Peinke, Michael Hölling, been finished. This included mainly the de- Ingrid Neunaber sign and assembly of a setup to fulfill the fol- was written. Results were presented at the lowing requirements: First, all experiments DAGA 2015. Funding: Federal Ministry for Economic should to be executed in turbulent flows Affairs and Energy (BMWi) with atmospheric-like properties to ensure the applicability to wind energy. Second, the Ref.Nr. 0325459B background sound pressure level coming from the setup had to be low enough to be Duration: 10/2012 – 09/2016 able to decompose the leading and trailing edge noise from the background. Third, the flow characteristics of the final setup had to be characterized. Additionally, all the measurements have been performed.

During the current reporting period, a detailed report on the measurements and the data handling was written. It contains de- Introduction tailed information on the setup, the measurements from July 2014 and the methods The project INFLOW-Noise aims at the used to process and analyze the data. Ad- investigation of the sources of leading and ditionally, the results were presented at the trailing edge noise of rotor blades in turbu- DAGA 2015 in Nuremberg in a talk, “Ex- lent flows. Reducing the noise emission is perimentelle Untersuchung des Vorder- und of major concern to support the acceptance Hinterkantenlärms eines Profils in Turbu- of wind energy converters (WECs) and to lenter Strömung”. fulfill noise regulations at new sites. In cooperation with the Fraunhofer IWES, the The acoustic measurements that were car- Institute of Aerodynamics and Gas Dynam- ried out by the itap were further processed, ics (IAG) of the University of Stuttgart, the and a report on the measurements was pre- Institute for Technical and Applied Physics sented. The results indicate a localization of (itap) and GE Wind Energy, the noise sourc- noise sources ranging from about 630 Hz es are examined with computational fluid to about 2500 Hz at the leading edge of the dynamics (CFD) simulations that are vali- airfoil. This is in accordance with the re- dated with wind tunnel experiments on an sults from the measurements taken with the References airfoil at the University of Oldenburg. The sound level meter, as in this frequency range the sound pressure level of the setup with experiments include acoustic measurements [1] Wächter, M.; Heißelmann, H.; with a directional microphone array and a airfoil exceeds the sound pressure level of Hölling, M.; Morales, A.; Milan, P.; sound level meter and flow measurements the setup without airfoil. The angular reso- Mücke, T.; Peinke, J.; Reinke, N.; with hot wires as well as measurements of lution of the directional microphone array is Rinn, P.: The turbulent nature of the the pressure and the pressure fluctuations on not sufficient to distinguish between differ- atmospheric boundary layer and its the surface of the airfoil. ent angles of attack. impact on the wind energy conversion process, Journal of Turbulence, Vol. 13, No. 26, p. 1-21 (2012)

– 123 – 124ACOUSTICS Annual Report 2015-2017

WEA-Akzeptanz – From the Source to the Perception

Leibniz Universität Hannover Institute of Structural Analysis

Raimund Rolfes, Tobias Bohne, Jasmin Hörmeyer, Susanne Martens

Funding: Federal Ministry for Economic Affairs and Energy WEA-Akzep anz

Ref.Nr. 0324134A Logo of the project WEA-Akzeptanz Duration: 04/2017 – 03/2020

ses the perception at the immission site. In logical and wind turbine parameters under Introduction cooperation with the partners an acoustic different environmental conditions. The in- overall model will be developed. The model fluence of operational parameters and noise Public acceptance of wind turbines is es- will include sound generation and sound reduction technologies are evaluated. Using sential for the successful transition of en- propagation, as well as the psychoacoustical the validated model, the noise effect on the ergy resources towards renewable sources. assessments. The latter is especially impor- residents will be objectified and made pre- Due to the expansion of wind power near tant for the public acceptance. In order to dictable. The overall acoustic model could to populated areas, the public acceptance validate the different models and the over- be also used in the planning stage of wind of wind turbines has decreased in the last all model, extensive field measurements are farms. few years. The social development towards performed to monitor acoustical, meteoro- wind turbines needs to be positive in order to achieve the objectives in the area of energy system transformation. Since local residents complain in particular that they feel disturbed by noise immission, it is important to make citizens aware of actual noise effects of wind turbines.

The project “WEA-Akzeptanz” addresses this requirement and aims at the prediction and objectivity of noise effects on the residents. For this, an interdisciplinary approach is applied linking the sound generation, radiation and propagation of wind turbines with the psychoacoustic evaluation at the immission site (see fig, 1). For this purpose, the industrial manufacturer Senvion examine the sound generation and radiation. The Institute of Structural Analysis and the Institute of Meteorology and Climatology of the Leibniz Universität Hannover investigate the sound propagation under realistic atmospheric conditions, whereas the Insti- tute of Communications Technology analy- Figure 1: Concept of the overall acoustic model

– 124 – Annual Report 2015-2017 ACOUSTICS Annual Report 2015-2017125

Project Description psychoacoustic evaluation of the sound turbine noise, including simulations of immissions of a wind turbine is of greater various weather and operating conditions. First, the challenges of the entire transmis- importance and thus ultimately decisive. Subsequently, the individual submodels sion chain from sound generation to the The emotional perception is also closely are coupled to form a complete acoustic psychoacoustical part are described. Then, linked to the visual perception of the wind model. The overall acoustic model can be the overall acoustic model and measure- turbines. validated with the help of the measurement ments are presented. campaigns carried out. Model Transmission chain of the sound The overall acoustic model serves to pre- Measurements The generation of noise emissions caused dict and objectify the noise effects of wind Due to many circumstances, measurements by aeroacoustic phenomena on the rotor turbines on local residents. In the model, in the field and near to wind turbines are blade is already very complex. In addi- the entire transmission chain from the challenging. The biggest challenge is the tion, several noise sources emit individual sound source to the sound receiver (point capability of measurement data to vali- tones due to rotating components (gears, of immission) will be linked. Because of date the complex model. The data needs fans, servomotors, pumps). Aspects such different requirements and the complex- to provide the quantification of different as aging or production-related tolerance ity of the transmission chain, the overall physical and range-dependent effects. Dif- deviations ultimately ensure that each acoustic model will be separated into sub- ferent ground and weather conditions, the wind turbine has its own individual "sound models. These submodels includes a model topology, and the wind turbines with their footprint". Even small deviations - such as of the sound generation in the near field, a individual source characteristics affect the a ventilation hatch in the tower or produc- model of the sound propagation in the far sound. For this reason, the project includes tion deviations of a few centimeters in the field and a model of the psychoacoustic an- four measurement campaigns monitoring rotor blade or an alternative manufacturer noyance at the point of immission. acoustic, meteorology and wind turbine of critical components - can lead to signifi- parameters. The first experimental cam- cant shifts in the resulting sound spectrum. The development of the near-field model paign will be carried out in 2018, lasting Critical sound phenomena are often only for sound generation is divided into two seven weeks in the northern part of Germa- detected during field measurements. As a areas. This includes the model of aero- ny. This test side is characterized by simple result, actions against these phenomena dynamic sound generation, which shows environmental conditions i.e. flat landscape can only be carried out very late in the aging phenomena, special operating con- and less natural cover. Later sites will in- design process or even in the field and are ditions and wind farm effects. In addition, clude complex environmental conditions therefore associated with correspondingly the near-field model includes structural and different types of wind turbines. An high costs. vibrations and the resulting sound radia- overview of an overall measurement setup tion as well as the stimulation and trans- is shown in fig. 2. The measurement equip- Due to the large distances between the resi- mission of structure-induced sound by ment for psychoacoustic investigations and dential areas and the wind farms, the sound wind turbine specific components. For the examination of wind turbine sound propa- generated by a wind turbine is not equal to development of the far field model, impor- gation is seen in fig. 3 to 6. the sound perceived as noise from citizens. tant atmospheric parameters influencing Depending on topography and changing the sound propagation (temperature, wind, Due to the requirement to measure under atmospheric conditions (i.e. wind and tem- turbulence, etc.) are defined on the basis free field conditions, acoustic measure- perature distribution), the sound propaga- of a meteorological model. Subsequently, ments present further challenges. Besides tion is frequency and location-dependent. these atmospheric parameters as well as non-technical challenges such as finding the effects of topography and vegetation a suitable site with low background noise, The transmission chain ends with the di- are included in the far-field l model so the wind turbine itself is a complex source rectly affected residents who have their that sound propagation can be evaluated and hence hard to measure. At greater own subjective perception of the "sound under realistic atmospheric conditions. A distance to the source, the separation of footprint". This is connected to the asso- psychoacoustic model will be developed background noise and sound emission ciated psychoacoustic evaluation of sound at the point of immission to investigate the from wind turbines is identified as a further immissions. If residents feel subjectively loudness and annoyance of wind turbine challenge. It is hard to get a high signal- disturbed by the sound of wind turbines, operating noises and to evaluate and op- to-noise ratio due to the fact that the wind the critical values which are defined in timize the noise development with regard turbine sound and the background noise is guidelines are irrelevant for them. The to minimum annoyance in various work increasing with higher wind speeds. The guidelines determining sound immissions and life situations (sound design). In addi- background noise is characterized by noise insufficiently consider phenomena such tion, the immission site will be created as from leaves and vegetation in general as as sound modulations. For the residents, an audiovisual 3D simulation environment well as by wind-induced noise on micro- the emotional perception and thus the for the psychoacoustic evaluation of wind phones.

– 125 – 126ACOUSTICS Annual Report 2015-2017

To meet the presented challenges during acoustic measurements, a suitable measurement setup was developed. To minimize the influence of wind-induced noise a standard windscreen and a specific developed secondary wind screens will be used reducing the atmospheric turbulence incident on the microphone. Moreover, acoustic measurement station will be placed as far as possible but at least 10 m away from trees or buildings. An example of an acoustic measurement station including a secondary wind screen and a solar panel for external power is shown in fig. 3.

Summary

Noise emissions caused by wind turbines receive a lot of attention in politics and is of public interest. Public acceptance of wind turbines is essential for the successful Figure 2: Schematic measurement setup for psychoacoustic investigations and transition of energy resources towards reexamination of wind turbine sound propagation newable sources. The wind energy project “WEA-Akzeptanz” addresses this requirement, and aims the development and validation of an overall acoustic model, which includes the sound generation, radiation and propagation from wind turbines, and the psychoacoustical assessments at the immission site. Using the validated model, the noise effect on the residents will be objectified and made predictable. The overall acoustic model could be used in the planning stage of wind farms. More information about the project “WEA-Akzeptanz” and the ongoing research activities are found at the website www.wea-akzeptanz. uni-hannover.de.

Figure 3: Acoustic measurement station to monitor wind turbine noise

– 126 – Annual Report 2015-2017 ACOUSTICS Annual Report 2015-2017127

Figure 4: Mast for long-term measurement of wind speed and direction, temperature and humidity

Figure 5: Multicopter and sodar to measure vertical Figure 6: Measurement setup for the psychoacoustic evaluation of wind-profiles wind turbine noise

– 127 – 128ROTOR BLADES Annual Report 2015-2017

Generation of Realistic Turbulent in-Flow Conditions by an Active Grid, Stochastic Analysis of the Resulting Force Dynamics on 2D Segments of Airfoils with Active Control Units

conditions on wind turbine blades. In the Carl von Ossietzky Universität Oldenburg, Project Description Institute of Physics, following example, we generated two in- Research Group: Turbulence, Wind Energy Preliminary work flow situations varying the intermittency of and Stochastics (TWiSt) Locally at the rotor blades, the resulting inflow angle fluctuations, while keeping its inflow is influenced by fluctuating wind mean value and standard deviation almost Gerrit Kampers, Agnieszka Hölling, speeds and directions and the rotational constant. Fig. 2 shows probability density Michael Hölling, Joachim Peinke movement of the blades. In the first phase of functions (PDF) of measured lift coefficient this project, this local inflow was estimated increments of the ACP for low intermittency Funding: German Research Foundation based on an analysis of offshore wind data. (left) and high intermittency (right), each (DFG) The fluctuations of the local inflow angle of for the adaptive camber (dashed lines) and the fixed case (solid lines). Both cases are Ref.Nr. PE 478/15-1 attack γ were analysed by means of increment statistics, where the increment normalized by the standard deviation of the increments in the fixed case. The angle of Duration: 01/2013 – 12/2015 ∂ α= α(t+τ)-α(t) τ attack with respect to the mean inflow is 8°. describes changes of α(t) on the time lag τ. The different inflow statistics are captured It could be shown, that the intermittent na- in the lift. This indicates, that mechanical ture of the wind data is directly transferred loads of wind turbines likely also feature Introduction to the local inflow at the turbine’s blades. intermittent statistics. For both inflow cases, An experimental setup was built to test air- the Adaptive Camber mechanism clearly at- Wind turbines work within the atmospheric foils in different reproducible inflow condi- tenuates lift fluctuations on different time- boundary layer (ABL), which is dominated tions created by an active grid. The setup scales. by turbulence. In such turbulent flows, gusts was successfully tested for periodic inflow Stochastic analysis occur much more frequently on short time situations. Furthermore, the stochastic Lan- scales than estimated by Gaussian statistics, gevin approach was applied to lift data of an To analyse the dynamical response of the which are used in industry standards [1]. airfoil in turbulent inflow in a first test. ACP to turbulent inflow in a quantitative This effect is called intermittency, which way, we apply the stochastic Langevin ap- has a big impact on mechanical load fluc- Measurement of aerodynamic proach. The Langevin equation quantities tuations and is considered to increase wind turbine failure rates [2]. Active and passive In the second phase of the project, a Self- decomposes fluctuations of C into a deter- flow control elements represent a promising Adaptive Camber equipped Clark-y airfoil L approach for the reduction of these chang- segment (ACP) is tested in different un- ministic and a stochastic part, where Γ(t) ing forces and loads. In the present project, steady inflow conditions. The profile fea- is Gaussian δ-correlated white noise. The (1) (2) a new method for wind tunnel experiments tures coupled leading- and trailing edge drift and diffusion coefficients D and D is adopted, that allows an advanced inves- flaps, allowing it to passively adapt its cam- describe the deterministic relaxation of the (1) tigation of the dynamics of aerodynamic ber [3]. The mechanism can be blocked to system to its desired steady state D =0 and forces acting on airfoil segments with and stay in the Clark-y shape (fixed case). the stochastic part of the system and can be without control units with reproducible obtained directly from measurement data turbulent inflow conditions. Based on an The experimental setup was modified, us- [4]. analysis of wind data, dominant features are ing only the vertical axes of the active grid. reproduced in the wind tunnel by means of With this setup, quasi 2D angle of attack To filter mechanical vibrations in the setup an active grid. To study the resulting force fluctuations can be created without moving from measured lift data, we apply an Em- dynamics, we use a stochastic method, that the airfoil itself, which prohibits influences pirical Mode Decomposition and exclude allows for a quantitative analysis of the dy- of inertia on the ACP mechanism. A sketch modes containing periodic oscillations from namical performance of airfoils with and of the setup is shown in fig. 1. our signals prior to the stochastic analysis. without flow control elements. This method has been successful in a similar The active grid allows to reproduce differ- problem [5]. ent features of the estimated local inflow

– 128 – Annual Report 2015-2017 ROTOR BLADES Annual Report 2015-2017129

active grid ⤻ force / torque sensor

u ACP

force sensor stepper motor

Figure 1: The ACP in the closed test section of the wind tunnel downstream of the active grid. 8 8 10 10 6 6 10 10 4 4 10 10 PDF / a.u. PDF / a.u. 2 2 10 10 0 0 10 10 − 4 − 2 0 2 4 − 5 0 5

δCLτ / σ(δCLτfixed) δCLτ / σ(δCLτfixed) Figure 2: PDF of lift increments of the ACP (dashed lines) and fixed case (solid lines). The graphs are vertically shifted. From top to bottom Γ takes the values of 0.007, 0.014, 0.042, 1 and 3s. 20 10 10 5 1 − 1 − s / / s 0 ) ) 1 1 ( 0 ( D D 10 5 − − 20 10 − − − 0.2 0 0.2 − 0.2 0 0.2 0.4

CL CL Figure 3: Deterministic drift function of the measured lift force dynamics of the ACP (black) compared to the fixed case (blue).

– 129 – 130ROTOR BLADES Annual Report 2015-2017

Fig. 3 shows the drift coefficient of the lift Summary for the previously shown low intermittency (left) and high intermittency (right) inflow In the present study, we experimentally cases for the ACP (blue) compared to the tested a Self-Adaptive Camber airfoil, that fixed case (black). has the potential to alleviate turbulence induced lift fluctuations. The airfoil was tested The deterministic drift is clearly increased in turbulent inflow, generated by an active by the ACP, indicating a stronger tendency grid. to its fixpoint when perturbed by fluctuating inflow, resulting in a smaller probability for To analyse the effect of the adaptive camber fluctuations. The system has a stronger re- on the dynamics of the lift fluctuations, we sistance to an external excitation, for exam- applied a stochastic approach based on the ple induced by a gust and responds quicker Langevin equation, that provides additional to a deviation from its fixpoint. Consequent- information about the resistance of the sys- ly, the fluctuations of the lift are reduced by tem to turbulent excitation. The response means of standard deviation and extreme time of the system to excitation is captured. values, without significantly affecting the This information could become helpful for underlying dynamics. future studies on fatigue loads.

References

[1] Mücke, T.; Kleinhans, D;, Peinke, J.: Atmospheric turbulence and its influence on the alternating loads on wind turbines, Wind Energy 14, p. 301 (2011) [2] Taver, P.; Qiu, Y.; Korogiannos, A.; Feng, Y.: The correlation between wind turbine turbulence and pitch failure, EWEA 2011, p.149 [3] Lambie B.: Aeroelastic Investigation of a Wind Turbine Airfoil with Self- Adaptive Camber, doctoral thesis, Technical University of Darmstadt (2011) [4] Friedrich, R.; Siegert, S.; Peinke, J.; Lück, S.; Siefert, M.; Lindemann, M.; Raethjen, J.; Deutschl, G.; Pfister, G.: Extracting model equations from experimental data, Phys. Lett. A, Volume 271, Issue 3, pp. 217-222, 200 [5] Hadjihosseini, A.; Peinke, J.; Hoffmann, N.P.: Stochastic analysis of ocean wave states with and without rogue waves, New Journal of Physics, Vol.16 (2014)

– 130 – Annual Report 2015-2017 ROTOR BLADES Annual Report 2015-2017131

Advanced Aerodynamic Tools for Large Rotors – AVATAR

Carl von Ossietzky Universität Oldenburg Introduction Project Description Institute of Physics, Research group TWiSt Modern wind turbines are constantly in- AVATAR was set to validate different engi- creasing in size, i.e. rated power, rotor neering tools for aerodynamic calculations Hendrik Heißelmann, Michael Schwarz, diameter, hub height etc. and with the in- of future 10+MW wind turbines against Jannik Schottler, Lena Vorspel, creasing size comes the challenges of the both, high-fidelity CFD simulations and Ingrid Neunaber, Lars Kröger, Elias Modrakowski, Johannes Krauß, structural and aerodynamic scaling laws. dedicated measurements of the aerodynam- Joachim Peinke, Michael Hölling The so-called square-cube relationship be- ic coefficients of airfoil profiles. ForWind tween size and weight of the turbine rotors contributed to several numerical and experi- Project partners: requires the manufacturers and designers to mental tasks of the project. Energy Research Centre of the Nether- introduce thinner and thus more flexible ro- lands, ECN (Netherlands, coordinator) tor blades in order to be able to increase the The impact of turbulent inflow conditions Delft University of Technology, TU Delft rotor diameter of future wind turbines. This on the aerodynamic performance of an (Netherlands) holds in particular for offshore applications. DU00-W-212 airfoil profile was inves- Technical University of Denmark, DTU However, the design of lighter and more tigated based on data from wind tunnel (Denmark) flexible rotors introduces not only challeng- campaigns at 1 Million Reynolds number. Fraunhofer IWES (Germany) es to the material selection and the manu- These measurements with turbulent inflow University of Oldenburg, ForWind (Germany) facturing, but also challenges the design at comparably low Reynolds numbers were University of Stuttgart (Germany) tools used today. The aerodynamic design complemented by standard measurements National Renewable Energy Centre, process of wind turbines is heavily relying in laminar inflow conditions at the design CENER (Spain) on fast engineering tools based on blade Reynolds numbers of the airfoil as well as University of Glasgow (UK) (Until element momentum theory (BEM) or free high Reynolds number measurements in a September 1st 2015: University of vortex models. These so-called low-fidel- pressurized wind tunnel [3]. However, both Liverpool) ity models are preferred over high-fidelity of these campaigns were not able to account Centre for Renewable Energy Sources and computational fluid dynamics (CFD) codes, for the effects of turbulent inflow on the air- Saving, CRES (Greece) which resolve the local aerodynamics of foil aerodynamics. National Technical University of Athens, wind turbines, due to their significantly low- NTUA (Greece) er computational costs. However, these en- To investigate the impact of inflow turbu- Politecnico di Milano, Polimi (Italy) General Electric, GE (Germany) gineering models are validated for today’s lence on the aerodynamic performance, LM Wind Power (Denmark) wind turbines, but as future wind turbines representative flow patterns were selected grow in size, the models are used well out based on field measurements and scaled to of their validated range. match the dimensions of the wind tunnel Funding: European Union's Seventh model. These tailored inflow conditions Programme for research, technological Within the EU funded AVATAR project, the comprised periodic and extreme fluctua- development and demonstration consortium of 11 research institutes and two tions of the angle of attack as well as differ- industry partners focused on the improve- ent reproductions of measured inflow angle Ref.Nr. FP7-ENERGY-2013-1/no. 608396 ment and validation of aerodynamic model- time series. A sophisticated control pattern ling tools for large wind turbines, i.e. larger of the active grid was therefor used, allow- Duration: 01/11/2013 – 31/12/2017 than 10 MW. ing for the repeated reproduction of the tur- . bulent and unsteady inflow time series in multiple experimental cases [2].

High-resolution force measurements and pressure scans along the airfoil chord were taken for a total of 28 inflow cases. Along with the setup specifications, time series of the inflow and derived characteristic aerodynamic quantities, like lift, drag and

– 131 – 132ROTOR BLADES Annual Report 2015-2017

Figure 1: Upstream view of the wind tunnel test section with mounted active grid in the back and DU00-W-212 airfoil profile in the center

momentum coefficients, were provided to putational studies regarding the impact of of-the-art codes showed reasonable agree- validate the CFD simulations with inflow geometrical rotor parameters on the perfor- ment between the codes [5]. The effects of turbulence. Although, reasonable agree- mance and loads in the context of upscaling thicker airfoils were found to be in the same ment between the simulations and the ex- as well as the impact of turbulence intermit- order of magnitude like effects of increased periments were found in a lot of cases, some tency. For the first investigation, the AVA- Reynolds number, both of which are rele- deviations underline the challenges of both, TAR reference wind turbine (RWT) was vant for the upscaling of the current turbine the comprehensive measurement of flow subject to slight modifications of the blade designs towards 20MW. fields and aerodynamic quantities as well as thickness at the outer rotor regions. Airfoil the integration of the measured flow proper- thickness was increased by 2% and CFD Similar methodology was used to study the ties in computational methods [4]. Thus, a simulations of the altered profiles were used effect rotor solidity with the purpose to in- significant demand for further research on to derive the airfoil coefficients. These were vestigate the loading on upscaled wind tur- these methods was identified. included in power curved calculations and bines with different induction. Results sug- estimations of several characteristic turbine gest, that low induction rotors with slender Besides the validation data from wind tun- loads by means of engineering BEM tools. blades are preferred for these cases. nel experiments, ForWind contributed com- A comparison of results from several state-

– 132 – Annual Report 2015-2017 ROTOR BLADES Annual Report 2015-2017133

No exceptional or unexpected aero-elastic Summary behavior was found in both cases, but results should not be generalized and need The AVATAR project consortium of 11 sci- further validation. entific institutions and two industry partners made substantial efforts to compare and Further simulation activities were dedicat- validate different design tools for future ed to assessing the impact of the included large wind turbines. Validations were based wind fields on the turbine performance de- on wind tunnel measurements of airfoil rived from two different engineering tools characteristics in a wide range of Reynolds (BEM). While engineering tools usually numbers and in turbulent conditions. More- rely on wind fields with normal distributed over, comparisons to field measurements velocity changes, atmospheric measure- and high-fidelity CFD simulations provided ments typically show a higher probability valuable insights on the validity of state- for large wind speed changes – so-called of-the-art engineering tools for aero-elastic References intermittency. Both used BEM codes were wind turbine simulations as well as their found to process the generated normal- shortcomings. [1] www.eera-avatar.de distributed and intermittent wind fields in a [2] Heißelmann, H.; Peinke, J.; similar way, which is essential to allow for The validation data and test cases have been Hölling, M.: Experimental Airfoil a code-to-code comparison of actual wind made publicly available and detailed project Characterization under Tailored turbine loads. While rotation speeds of the results are published in scientific articles Turbulent Conditions, J. Phys.: Conf. turbine showed intermittent behavior in data and deliverable reports – all which are avail- Ser. 753 072020, (2016) [3] Ceyhan, O.; Pires, O.; from both tools, the assessment of fatigue able on the project website [1]. A summary Munduate, X.; Sørensen, N.N.; loads for a limited set of damage equivalent of some key project findings will be pre- Schaffarczyk, A.P.; Reichstein, T.; sented at the TORQUE conference 2018. load cases (DEL) returned mixed results. Diakakis, K.; Papadakis, G.; While on code indicated a high sensitivity Daniele, E.; Schwarz, M.; T. Lutz, of DEL to intermittent inflow, the other code T.; Prieto, R.: Summary of the Blind showed a lesser impact on the loads, which Test Campaign to predict the High emphasizes the need for further studies and Reynolds number performance of code-to-code comparisons in this regard DU00-W-210 [6]. However, it is apparent, that turbulence airfoil, 35th Wind Energy Symposium, intermittency can be a potent influence on AIAA SciTech Forum, (AIAA 2017-0915) the damage equivalent loads, which needs [4] Kim, Y.; Lutz, T.; Jost, E.; Gomez-Iradi, S.; Muñoz, A.; further investigation. Méndez, B.; Lampropoulos, N.; Stefanatos, N.; Sørensen, N.N.; Madsen, H. Aa.; van der Laan, P.; Heißelmann, H.; Voutsinas, S.; Papadakis, G.: AVATAR Deliverable 2.5: Effects of inflow turbulence on large wind turbines, project deliverable, (2016) [5] Schwarz, C. M.; Aparicio-Sanchez, M.; Astrain Juangarcía, D.; Gonzalez-Salcedo, A.; Ramos García, N.; Sessarego, M.; Sørensen, N.N.; Schramm, M.; Thomas, P.; Chasapogiannis, P.; Chaviaropoulos, T.; Manolesos, M.; Sieros,G.: D4.10: Effect of blade solidity and thickness on performance and loads, project deliverable, (2017) [6] Schwarz, C. M.; Ehrich, S.; Thomas, P.: D 4.7 Turbulence Intermit- tency, project deliverable, (2016)

– 133 – 134ROTOR BLADES Annual Report 2015-2017

Hybrid Laminates and Nanoparticle Reinforced Materials for Improved Rotor Blade Structures (LENAH)

Leibniz Universität Hannover ing their performance limits when it comes to use nanoparticle-modified materials in Institute of Structural Analysis to achieving a longer service life and better the whole rotor blade and to substitute the exploitation of the lightweight construction usual resin system. This can improve the Raimund Rolfes, Behrouz Arash, potential. One of the main problems con- mechanical properties of fiber-reinforced Christian Gerendt, Robin Unger cerning offshore rotor blades is the blade laminates and adhesives, which in turn im- stiffness, which needs to be increased with proves the strength and behavior of the ro- Funding: Federal Ministry of Education and Research increasing length. A large blade of inap- tor blade under different wind conditions. propriately low stiffness may contact the Nanomodified polymers are created by mix- Ref. Nr. 03SF0529 tower under extreme loading conditions. A ing nanoparticles with an average size of total loss of the turbine would be the result. about 20 nm (20 ∙ 10-9 m) and conventional Duration: 09/2015 - 07/2019 Ensuring sufficient stiffness of future blades polymers. The interaction between the poly- with today’s established materials would mer molecules and the nanoparticles, which either lead to higher weights in case of have a huge cumulative surface (‘nano ef- glass fiber reinforced polymers (GFRP) or fect’), lead to improved mechanical proper- to inacceptable high costs in case of carbon ties of the material. This beneficial effect can Introduction fiber reinforced polymers (CFRP). An ad- be applied for fiber reinforced polymers, as ditional issue is the high fatigue resistance well as for pure adhesive polymers. Within In order to perform the energy transforma- of the applied material, which is required the LENAH project, nanoparticle reinforced tion both successfully and economically, the to withstand up to 107 (100 million) load polymers are investigated by experimental efficiency of energy generation from renew- cycles during the wind turbine’s lifetime of tests and multi-scale finite element simula- able sources must be further increased. With about 20 years. If the fatigue resistance of tions. First of all, atomistic simulations on regard to wind energy, this means increas- the materials applied is too low, cracks may the nano scale are performed to derive the ing the energy yield per turbine, which is grow within the blade, yielding its total loss unknown material properties, followed by usually achieved by an increase of the ro- at worst. The fact that future blades with micro-scale simulations as well as compo- tor diameter. However, the rotor blades are lengths over 90 m will have to withstand nent simulations on the macro scale (see exposed to several loads due to wind and fatigue loadings of even higher amplitudes Fig. 1). The major aim is to understand and turbulences, the inertia of the blade dur- than today’s largest blades motivates the ur- characterize the influence of nanomodifica- ing rotation and others, leading to gradual gent need for materials with a high fatigue tions on the material properties. With the fatigue of the blade material. With higher resistance. The application of traditional knowledge gained, materials can be de- material strengths, less material is needed materials like GFRP or CFRP would again signed that merge the positive characteris- to make the blade resistant to a given load, yield too heavy and too cost-intensive con- tics of the base materials and allow a more or to push boundaries further when it comes structions. Within the LENAH project, in- efficient design of wind turbine rotor blades. to the design of lightweight blades with novative material concepts are to be inves- simultaneous increase of the blade length. tigated, which offer such improved fatigue In contrast, hybrid laminates are used lo- The aim of the LENAH research project is characteristics. The scientists are working cally in particularly stressed areas in order therefore to develop material systems that on improvements throughout the entire to provide a sufficient fatigue resistance. lead to a longer service life and improved process chain – from material development Hybrid materials, which combine different lightweight construction for wind turbines. to component design. To do so, the project materials to profit from their individual ben- The scientists are working on improvements takes into account two important fields efits, are playing an increasingly important throughout the entire process chain – from of material research: The development of role in industrial applications. Within the material development to component design. nanoparticle-modified plastics to increase LENAH project, hybrid laminates made service life and hybrid materials to optimize from GFRP and stainless steel as well as lightweight construction. GFRP and CFRP will be investigated con- Project Description cerning their applicability in highly stressed For the research topic "nanoparticle modifi- areas of wind turbine rotor blades. Here, Although today’s rotor blades of wind tur- cation", researchers are specifically chang- not only the optimization of the resulting bines can hold lengths larger than 85m, con- ing the material properties of materials material properties is considered, but also ventional materials are increasingly reach- through the use of nanoparticles. The aim is the manufacturability and the profitability.

– 134 – Annual Report 2015-2017 ROTOR BLADES Annual Report 2015-2017135

GFRP-steel laminates are to be used within stiffness of the blade’s supporting structure be developed for the production of coupon the load introduction area, where the blade in areas of high loading and increased light- and small component elements to basically is bolted to the hub of the turbine. Due to weight requirements (e.g. midsection of the characterize the material’s benefits. After- the bolt holes, local stress concentrations blade). The CFRP/GFRP hybrid laminates wards, promising materials will be further occur, which can damage the rather brittle can be used either directly to stiffen the investigated at the structural level using GFRP significantly. Local implementation blade or are implicitly necessary, when a more complex component tests. In parallel, of ductile steel foils in between the GFRP pure GFRP area has to transformed into a computational tools will be developed to plies yields a smooth introduction of the pure CFRP area (and vice versa) using well- predict the (failure) behavior of the devel- loads from the bolts into the laminate on designed ply substitution techniques. The oped material systems. The project follows the one hand, and increases the effective investigation of hybrid laminates includes a holistic approach that does not only focus joint strength under static and cyclic load- the development of computational design on improved characteristic material proper- ing conditions. Once a certain distance from tools and experimental investigations. ties, but rather aims at an understanding of the introduction area is ensured, the metal the mechanisms of action and the generally foils are substituted by regular GFRP layers. Both technology approaches will range from valid interrelationships. In a final step, the The application of GFRP-CFRP laminates basic material developments to component project partners collect, summarize and cat- is an efficient method to locally increase the testing. In a first step, material systems will egorize all the results with regard to their usability, cost-effectiveness and the requirements of the wind energy industry.

Summary

The aim of the project LENAH (Hybrid laminates and nanoparticle reinforced materials for improved rotor blade structures) is to develop innovative material systems for rotor blade lightweight construction and to understand their mechanisms of action. The project focuses on two different material systems: nanoparticle modified epoxy res- ins and hybrid laminates. Figure 1: Schematic diagram of the multi-scale finite element approach that is followed in the research topic "Nanoparticle Modification". The project findings will not only contribute to an implementation of further material improvements in the future, but the knowledge gained can also be used as a basis for the development of other material systems. Improved material properties through nanoparticle modification, modified design through hybrid materials, and lightweight construction optimization are the focus not only in wind energy technology but also in other sectors, such as aerospace technology.

Figure 2: Schematic illustration explaining the blade root strengthening via Fiber Metal Laminates investigated in the LENAH project.

– 135 – 136ROTOR BLADES Annual Report 2015-2017

Wind Turbine Load Control under Realistic Turbulent In-Flow Conditions

Carl von Ossietzky Universität Oldenburg Project Description PIV measurements periodic structures along Institute of Physics the shafts of the active grid (shown in fig. 2) Research Group: Turbulence, Wind Energy The PAK780 Project, or "Wind Turbine were identified. The structures can directly and Stochastics (TWiSt) Load Control under Realistic Turbulent In- be linked to the shape of the active grid Flow Conditions", is aimed to investigate flaps. To guarantee a homogeneous inflow Tom Wester, Michael Hölling, Joachim Peinke passive smart devices which are abled to along the spanwise direction of the airfoil it minimize extreme load events and fatigue was necessary to revise the grid. This led to Partners: loads of WEC by influencing the aerody- the new 2D active grid, which is shown in Technical University Berlin namics locally at the rotor blade. fig. 3. RTWH Aachen University of Stuttgart To investigate those occurring load events The diamond shaped flaps were removed Technical University Darmstadt it is necessary to generate repeatable in- as well as all horizontal shafts leading to a flow conditions. Therefore, an active grid grid with nine vertical shafts. Each shaft is Funding: German Research Foundation, (shown in fig. 1) should be used. During again controlled individually by a stepper (DFG)

Ref.Nr. PE 478/15-1.

Duration: 12/2015 – 04/2019

Introduction

The atmospheric boundary layer (ABL), in which wind energy converters (WEC) operate, contains a broad spectrum of inflow fluctuations. Mainly manifesting as gusts, such inflow fluctuations cause heavy load variations for the turbine rotor. To expand the lifetime of a WEC, the load fluctuations need to be reduced. To be able to do so it Figure 1: Active Grid which was used to generate repeatable turbulent inflows is essential to understand the aerodynamic reaction of rotor blades under tailored turbulent inflow conditions. Besides generating a 2D tailored inflow, the project aim is to identify aerodynamic extreme events and link them to the aerodynamic response of an airfoil. The airfoil is equipped with an adaptive camber mechanism for load mitigation. This mechanism was already presented for static measurements in [1] with promising results. Those results will be extended to dynamic and turbulent inflows.

Figure 2: Generated angle of attack variations by the active grid due to flap shape

– 136 – Annual Report 2015-2017 ROTOR BLADES Annual Report 2015-2017137

Figure 3: New designed 2D active grid with nine Figure 4: Periodic inflow at different spanwise positions of the vertical, individual movable shaft new 2D active grid measured with hot wire

motor. With this a purely 2D inflow can be The setup includes besides the active grid reduction of lift fluctuation and additional produced in the spanwise direction of the to generate the inflow also force balances to an increase in mean lift generated by the air- airfoil. This is shown in fig. 4, where the measure the lift, drag and torque and a high- foil. This is shown in figure 6, where the lift flow is measured at two different spanwise speed stereoscopic PIV system to measure coefficient for both airfoils is plotted for a positions. Both measurements are perfectly the flow around the airfoil. This enables the sinusoidal inflow of 8Hz at a mean angle of aligned and the phase shift which was meas- simultaneous measurement of the aerody- attack of 10°. The red curve represents the ured beforehand cannot be found anymore. namic response of the airfoil and the result- baseline airfoil and blue the ACP. From the ing forces. By this the ACP can be directly plot a reduction of minimum and maximum After improving the inflow for airfoil exper- compared to the baseline case and the dif- for the ACP can be observed, leading to a iments, the next step was the further investi- ferences under turbulent inflow for both air- decrease of lift fluctuations by 29%. - Fur gation of the adaptive camber profile (ACP) foils can be identified. thermore, the mean lift could be increased under dynamic inflow. The used setup is by 12% compared to the baseline Clark-Y shown in fig. 5. First measurements of the ACP under pe- airfoil. riodic sinusoidal inflow show a significant

Figure 6: Comparison of generated lift of the baseline airfoil in red and the ACP in blue. The data are phase averages, Figure 5: Wind tunnel setup to measure the flow field around the where the error bars represent the standard deviation. The ACP under 2D inflow conditions black vertical lines represent the time of the PIV Snapshots

– 137 – 138ROTOR BLADES Annual Report 2015-2017

One of the reasons for this can be seen in on the chord. By this the abrupt loss of lift, Summary figure 7. The figure shows PIV snapshots of which is seen for the baseline airfoil, is re- the flow around both airfoils. The upper line duced for the ACP. The second line shows Within the project “Wind Turbine Load shows the response of the airfoils to a big compared to this a rather small AoA. In this Control under Realistic Turbulent In-Flow angle of attack. Whereas the baseline airfoil case the ACP cambers, leading to a higher Conditions”, a method to generate defined shows a rather large separation with an early flow velocity around the airfoil and by this turbulent 2D inflow fields in a wind tunnel separation point, the ACP reduces the sepa- a higher lift is generated. This leads to an was developed. With given periodic inflow ration area due to the upwards deflection increase of the minimal lift compared to the fields it was possible to connect these with of the trailing edge flap. Also, the point of not cambering baseline airfoil. the aerodynamic response of the airfoil. For separation is mitigated further downstream this purpose, the loads had been measured simultaneously with the flow field, which was measured by means of high-speed stereoscopic Particle Image Velocimetry. By comparing the baseline airfoil with the adaptive camber airfoil, it was possible to identify significant differences in the aerodynamics of both airfoil types.

Figure 7: PIV measurements of the baseline airfoil (left) and the ACP (right). The upper line represents the time of the first black line, the lower one the second.

References

[1] Lambie B.: Aeroelastic investigation of a wind turbine airfoil with self- adaptive camber, Ph.D. thesis, Technische Universität Darmstadt (2011)

– 138 – Annual Report 2015-2017 CROSS139 DISCIPLINARY TOPICS Annual Report 2015-2017139

Smart Blades – Development and Design of Intelligent Rotor Blades

Carl von Ossietzky Universität Oldenburg the Fraunhofer Institute for Wind Energy dividual Pitch controllers (IPC). The results Institute of Physics and Energy System Technology (IWES), showed a significant reduction of the fatigue Research Group: Wind Energy Systems ForWind has participated with its wide va- loads in the rotor blades as well as the sup- Research Group: Turbulence, Wind Energy riety of expertise in each part of the project, port structure. The best results were found and Stochastics which is organized in four technologies. for the bend-twist coupled blades combined with MIMO IPC. Hauke Beck, Martin Bitter, Bastian Dose, Agnieszka Hölling, Tim Homeyer, Michael Hölling, Martin Kraft, Martin Project Description ForWind Hannover investigated the load-re- Kühn, Joachim Peinke, Francesco Perrone, duction potential of geometrical bend-twist Hamid Rahimi, Marijn van Dooren, Technology 1 – Passive Smart Blades coupling, i. e. the inclusion of a pre-sweep Lena Vorspel, Róbert Ungurán, Mehdi Vali with bend-twist coupling in the rotor plane towards the trailing edge. The focus of Technology 1 was the inves- The impact of the shape of the sweep func- Leibniz Universität Hannover tigation and design of wind turbine blades tion on the reduction of fatigue and extreme Institute of Structural Analysis, which feature bend-twist coupling as a pas- loads with respect to the reference turbine Institute of Turbomachinery and Fluid sive load control mechanism to reduce the was the target outcome. A detailed param- Dynamics, aerodynamic loads. eter study for the optimization of the sweep Institute for Wind Energy Systems function has been carried out [3] revealing ForWind Oldenburg was in charge of tasks a substantial reduction of the fluctuation of Majeed Bishara, Mehdi Garmabi, Eelco that deal with the development of new nu- Jansen, Raimund Rolfes, Carl Robert aerodynamic loads [4]. This led to a signifi- Brand, Benedikt Ernst, Jörg Seume, merical simulation tools and controller cant reduction of fatigue and extreme loads Torben Wolff, Claudio Balzani, Moritz concepts. The simulations included an auto- in the rotor blades and the support structure. Bätge, Johannes Ende, Daria Horte, mated optimization of complete rotor blades The flutter limit of the swept blade was ana- Jaione Ortega Gómez, Andreas Reuter, with respect to their aerodynamics [1]. The lyzed in the frequency and the time domain Alper Sevinc resulting code allows for variable imple- using HAWCStab2 and HAWC2. Analyti- mentation of bend-twist coupling at differ- cal models were developed and parameter Funding: Federal Ministry for Economic ent regions of the blade and the separation and sensitivity studies executed. The flutter Affairs and Energy (BMWi) from pure bending, pure twisting and bend- limit increased for swept blades compared twist coupled blades. A tool chain has been with the reference blade. A demonstration Ref.Nr. 0325601C (Hannover) developed that allows automated grid gen- rotor blade with a length of 20m with geo- 0325601D (Oldenburg) eration for 2D profiles and the complete ro- metrical bend-twist coupling has been de- tor [2]. With this a detailed investigation for signed, see Fig. 1, and provided to DLR for Duration: 12/2012 – 02/2016 wind turbines under the condition of yaw- manufacturing and field testing in a follow- misalignment has been performed which led up project. In order to support the analytical to an implementation of a correction in the and numerical studies, material and compo- BEM based tool FAST. Additionally, a new nent tests have been carried out in strong co- solver for fluid structure interaction (FSI) operation with Fraunhofer IWES. Material Introduction has been developed to account for changes tests under multi-axial fatigue loading and in the resulting aerodynamics while flexible structural tests on thick laminates and scaled Modern wind energy converters are increas- blades move in the flow due to acting forc- blade root bolt connections were carried ing in size. At the same time, loads acting es. After verifying of the new solver it has out. A deeper understanding of the structural on the rotor blades and the subsequent com- been applied to the reference blade of this performance of passive smart blades could ponents become more and more problem- project. In the field of controller concepts be extracted from the experimental work. atic, for both blade design and operation. a model of the reference and the bend-twist Within the Smart Blades project, different coupled blades have been used to find op- Technology 2 – Smart Blades with technologies have been considered that aim timal control strategies with state-of-the-art active trailing edge for a reduction of aerodynamic loads and Collective Pitch Controller (CPC) as well Technology 2 dealt with the development of their fluctuations. In close cooperation with as with Single Input Single Output (SISO) smart rotor blades with trailing edge flaps the German Aerospace Center (DLR) and and Multi Input Multi Output (MIMO) In- and actively deformable trailing edges.

– 139 – 140CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Figure 1: Blade design of the 20 m demonstration blade. (a) Primary blade structure with central shear web; (b) Blade structure with sandwich panels (pictures from [5]).

These are typically applied in the outer part tified which can measure the wind speed in Various combinations of active trailing of the rotor blade to reduce the loads acting front of the rotor. A continuous-wave (cw) edge flaps with individual pitch control to on the wind turbine. Laser Detection and Ranging System (Li- reduce the 1P or 2P Loads on the blade root DAR) has been purchased and used to char- were examined with help of the aeroelastic Focus of ForWind Oldenburg was in the re- acterise the incoming wind field in tempo- simulation tool HAWC2. The most suitable search of optimal control strategies with the ral and spatial resolution with approx. 400 strategy for a robust multi-variable control- aim to reduce particularly higher harmonic measures points/sec with ±30° inclination ler (MIMO) was the H-infinity controller loads. First relevant sensors have been iden- angle and a focus length of 10 to 150 m. (H∞). Furthermore, an “anti-wind-up” system was implemented around the saturation in the actuator balancing loop. It has been shown, that the maximum lifetime fatigue load reduction could be achieved with a combination of individual pitch and trailing edge flap control, see Fig. 2. The reduction of the 2P flapwise bending moments via the trailing edge flaps lead to load reduction on other components of the turbine as well.

The suitability of the tool chain that has been developed for Technology 1 has been proven for active trailing edge flaps as well. This includes automated meshing and 2D- CFD simulations, translation of 2D-polar in FAST-Simulation environment as well as run-time evaluation of 3D-CFD rotor simulations. Needed correction models for the inclined flow for FAST simulations have been developed as well.

ForWind Hannover was in charge of tasks considering structural, aerodynamic, and aeroacoustic aspects. A representative finite element discretization using shell elements was chosen as a basis for fatigue, strength, and vibration analyses at blade level. To facilitate the structural modeling and analysis activities, an input generator for the finite Figure 2: Power spectral density of the flapwise blade root bending moments at 13 m/s element program ABAQUS was developed. for different control combinations (pictures from [6]) Specific strength analyses have been carried

– 140 – Annual Report 2015-2017 CROSS141 DISCIPLINARY TOPICS Annual Report 2015-2017141

out for the flexible trailing edge. Moreo- sive measurement campaign was conducted ver, in the context of strength modeling, in the aero-acoustic wind tunnel at the DLR a novel hybrid micro-meso modeling ap- in Braunschweig, see [7]. CFD simulations proach has been developed for the simula- were compared with the experimental data tion of progressive failure considering the so that a model for the noise emission as a interaction between fiber kinking, split- function of the angle of attack and the flap ting, matrix cracking, and delamination. angle could be extracted. For aerodynamic considerations, a BEM- based unsteady aerodynamic model was Technology 3 – Active slat implemented in Matlab that captures the Along with the design of the slat done by dynamics of pitch and flap motions. Pitch, the DLR, wind tunnel tests have been per- flap and combined pitch/flap motions have formed in the wind tunnel in Oldenburg [9]. Figure 4: Lift coefficient for different been simulated. Unsteady 2D CFD simu- For the characterization of the effect of dif- mean angle of attacks of the profile lations have been carried out as well. The ferent positions (Hub) of the trailing edge of over trailing edge positions (Hub) of the BEM- and CFD-based simulation results the active slat, standard polar measurements active slat. were compared and showed generally good have been performed with combinations of agreement. From the CFD results, unsteady mean angle of attack of the profile and hub characteristics could be extracted, which position of the slat. Fig. 4 influence of slat. have been used for the derivation of a novel shows that the lift coefficient changes for empirical aerodynamic model. Critical load different settings. This was used to design and simulation cases have been extracted an experiment, where an active grid was for fixed flap angles, i. e. flap actuator fault used to generate angle of attack variations situations, in order to prepare activities on over time at the leading edge of the profile. the determination of flutter limits in a fol- This inflow resulted in a sinusoidal change low-up project. In order to characterize the in the angle of attack of 5Hz and about 4° aero-acoustic behavior of trailing edge flaps, amplitude. The Hub movement of the ac- wind tunnel experiments have been carried tive slat was also operated at 5Hz with at the out. After designing and manufacturing of a maximum Hub amplitude. Triggered meas- Fig. 5: Standard deviation of the measured test specimen preliminary studies were exe- urement allowed for different time shifts lift coefficient for different phase shifts cuted in Hannover for the development and between the start of the active grid motion between the motion of the slat and the incoming flow with changing angles of design of aerodynamic and acoustic meas- and the slat motion. Fig. 5 active slat effect attack. The profile was set to a mean urement systems and the respective evalu- hows the measured standard deviation of angle of attack of 10°. The red dotted line ation algorithms (beamforming). An exten- the lift coefficient over the resulting phase represent the measured standard deviation for identical inflow conditions but the fixed slat.

shift between the active grid motion and the slat motion. The red dotted line represents the standard deviation for the slat in a fixed position. It can be seen that the active slat technology is able to change the aerodynamic response of the complete profile even at a frequency of 5Hz. The final result of the experimental investigation shows that the active slat technology is also able to reduce the standard deviation of the lift coefficient and with this the changing loads acting on the profile.

Technology 4 - Cross-Technology Topics Figure 3: CFD simulation showing a vortex at the side edge of the trailing edge flap Activities that cannot be linked to one single resulting in noise (view on the suction side; flap deflection of 15°; angle of attack of 6°; of the investigated technologies were inte- picture from [8]) grated in the cross-technology topics. The

– 141 – 142CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

core goal was the evaluation of the different structural, aerodynamic, and aeroacoustic Furthermore, ForWind developed a holistic smart blades technologies. investigations regarding the concept of an evaluation model including technological active trailing edge. On the structural side, and economical aspects as well as reliabil- Fraunhofer IWES together with ForWind shell-type (“3D”) finite element models ity analyses for the turbine concepts under Hannover finalized the 3D design of a refer- have been used to perform vibration, stabil- investigation. ence blade as well as the specification of the ity, strength and fatigue analyses. Wind tun- reference turbine model, which both serve nel experiments including High-Speed PIV as a basis for the comparative evaluation of under turbulent inflow conditions have been the smart blades technologies. the basis for a better understanding of the effectiveness of slats in Technology 3. In the To estimate the risk potential for Smart cross-technology topics ForWind participat- Blades, ForWind Oldenburg created fault ed in developing a 7.5 MW reference wind models for the reference system, Technol- turbine model for loads simulation purposes ogy 1 and Technology 2. For Technology and corresponding reference rotor blades 3, a comparative evaluation of reliability including detailed lay-up specifications. was performed. The consequences of possible failures were investigated by means of simulation models. From this, design requirements and parameters in the form of sensitivity analyses for passive smart blades regarding shape and dimensional tolerance were derived. For active systems robust- References ness and design of critical components have [1] Vorspel, L.; Schramm, M.; Ref. No. 0325601C, Leibniz University been analysed as well. Necessary changes to Stoevesandt, B.; Brunold, L.; Brünner, Hannover, Institute for Wind Energy the relevant directives and standards were M.: A benchmark study on the Systems (2016) identified. There is only a small need for efficiency of unconstrained optimi- [6] Ungurán, R.; Kühn, M.: Combined changes in the regulations. However, due to zation algorithms in 2D-aerodynamic Individual Pitch and Trailing Edge Flap not yet performed validation, an increased shape design, Cogent Engineering, Control for Structural Load Alleviation of partial safety factor for model uncertainties vol. 4, no. 1, p. 1354509 (2017) Wind Turbines, 2016 American Control seems necessary. [2] Rahimi, H.; Daniele, E.; Conference (ACC) 6-8 July 2016 [Pisca- Stoevesandt, B.; Peinke, J.: taway, NJ] : IEEE (2016), ForWind Hannover finalized the formula- Development and application of a pp. 2307-231 tion of a techno-economical evaluation grid generation tool for aerodynamic [7] Brand, C. R.; Seume, J. R.: Flaps for wind turbine applications: Results of an model. Expert interviews were prepared, simulation of wind turbine, Wind Engineering, April 2016 vol. 40 no. 2 acoustic study, Proceedings of the EWEA conducted and evaluated in order to guar- pp. 148-172 conference, Paris, France (2015) antee an objective technology evaluation [3] Sevinc, A.; Bleich, O.; Balzani, C.; [8] Brand, C. R.: Numerische on a solid basis. The power curves were ex- Reuter, A.: Parameterized analysis of Simulation instationärer Strömung an tracted from the loads simulations provided swept blades regarding bend-twist Klappe, Deliverable No. 2.2.8.2, coordi- by Fraunhofer IWES. The corresponding coupling, Proceedings of DEWEK nated research project “Smart Blades”, annual energy yield was calculated and in- 2015, Bremen, Deutschland, 19.- funded by the German Federal Ministry cluded in the determination of the levelized 20.07.2015. for Economic Affairs and Energy, Ref. cost of energy for the different technologies. [4] Sevinc, A.; Bleich, O.; Balzani, No. 0325601C, Leibniz Universität Han- C.; Reuter, A.: Aerodynamic and nover, Institute of Turbomachinery and structural aspects of swept blades Fluid Dynamics (2015) Summary in the context of wind turbine load [9] Manso Jaume, A.; Wild, J.; reduction, Proceedings of the EWEA Homeyer, T.; Hölling, M.; Peinke, J.: conference, Paris, France (2015) Design and wind tunnel testing of a Within the Smart Blades project, ForWind [5] Bätge, M.: Design of a 20m leading edge slat for a wind turbine was involved in the development of new blade for the demonstration turbine, airfoil, DEWEK 2015, Proceedings numerical simulation tools and controller Technical Report, Deliverable No. and blade design concepts for Technology 1.2.6.1/1.2.6.3, coordinated research 1. Amongst others, fluid-structure interac- project “Smart Blades”, funded by tion was added to the tool chain that allows the German Federal Ministry for aerodynamic analyses for flexible blades. Economic Affairs and Energy, In Technology 2, ForWind participated in

– 142 – Annual Report 2015-2017 CROSS143 DISCIPLINARY TOPICS Annual Report 2015-2017143

GigaWind life – Life time Research on Support Structures in the Offshore Test Site alpha ventus

Leibniz Universität Hannover technology installed in the alpha ventus test Subproject 1: Institute of Structural Analysis field on the support structures using moni- Data Management and Analysis Institute of Geotechnical Engineering toring methods developed to date, enabling In this subproject, various monitoring, sim- Institute of Steel Construction significant scientific findings to be gained ulation and laboratory test data from OWT Institute for Building Materials Science from previous investments. Furthermore, were merged into one user database. An Ludwig-Franzius-Institute for Hydraulic, validated methods and structural models additional goal was to develop optimized Estuarine and Coastal Engineering access and storage strategies for such data Institute for Concrete Construction were developed, both for local investigations as well as for the holistic design of the and a modular framework for the effective Martin Achmus, Johannes Albiker, supporting structure of planned OWTs. evaluation of large amounts of data. Jan Dubois, Rasmus Eichstädt, Jan Häfele, Joshua Henneberg, Subproject 2: Ralf Herrmann, Arndt Hildebrandt, Project Description Monitoring and Inspection Mahmoud Jahjouh, Sebastian Kelma, The task of this subproject was the moni- Ludger Lohaus, Nikolai Penner, - Objectives toring of long-term measurements in alpha Alexander Raba, Raimund Rolfes, The objective of this collaboration is to ventus as well as of model experiments in Patrick Rzeczkowski, Tobias Schack, validate methods and structural models to order to gain in-depth knowledge regard- Dominik Schäfer, Peter Schaumann, enable a holistic and economical design of ing the degradation of materials and cor- Alexander Schendel, Christoph Tomann OWT supporting structures. To achieve this, rosion protection systems and to validate Funding: Federal Ministry for Economic a variety of research areas in the context of previously developed SHM methods. Other Affairs and Energy the GIGAWIND life subprojects, shown in important aspects were the identification Fig. 1, were investigated. of nonlinear effects (scour, hydrodynamic Ref.Nr. 0325575A masses, damping) as well as the determina- - Presentation of Subprojects tion of stresses and partial safety factors in Duration: 02/2013 – 01/2018 As part of the project "Validated Methods the context of new design concepts. and Structural Models for an Integral and Economic Design of OWT Support Struc- Subproject 3: tures" different tasks and objectives were Degradation Models Introduction sought within the individual subprojects. In the context of this subproject, degrada- These are briefly outlined below. tion models were developed, which enable As a continuation of the project GIGAWIND alpha ventus, the GIGAWIND life project aims to expand the new economic design concept for supporting structures of Off- shore Wind Turbines (OWT) developed in GIGAWIND alpha ventus with special emphasis on the operation of such converters, degradation mechanisms and interaction of supporting structures with its surrounding environment in the form of material damage to the support structure and the joints, fatigue, damage of the corrosion protection systems, and degradation of the pile bearing behavior as well as the determination of acting loads from waves and marine growth are considered.

The required damage and stress hyster- esis measured by the comprehensive sensor Figure 1: Structure of GIGAWIND life

– 143 – 144CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

the deduction of the influences of loads, - Reuss Bridge in Wassen, Switzerland, tion monitoring of support structures (see environmental conditions and environmen- 2007 [10], [11], [12], [13]). Investigations in [10] tal factors on the lifetime of the supporting further summarize fundamental findings of structures, providing statements about the At OWTs, monitoring systems on machine the SHM from various disciplines into axi- remaining useful life of the OWT. components represent currently the state of oms. However, the use of SHM in OWTs the art (see [4], [5]). On other components, needs some challenges to be tackled. This Subproject 4: such as the support structures, systems are includes, for example, the consideration of Adapted model experiments increasingly used for condition monitor- varying operating and environmental condi- Model experiments were the focus of this ing, but mainly on prototypes. Accordingly, tions as well as a corresponding further de- subproject. These should provide insights current data management and data analysis velopment or adaptation of the methods and into commonly “assumed” phenomena. In systems represent object-specific solutions. the hardware (see [14], [15], [16]). Current addition, scour effects, structural damage An approach to automated data evaluation efforts focus on the computation of state pa- and degradation of pile behavior were in- exists only in the form of alarm functions, rameters over longer time periods to better vestigated experimentally. In addition, ap- which send a notification when certain understand their variations and dependen- proaches to mineral corrosion protection are thresholds are exceeded, or in the form of cies and translate them into probabilistic optimized. SHM systems that record measured data descriptions. only for the duration of such events. Subproject 5: So far, there monitoring system enabling Method and Model Integration Subproject 2: continuous condition monitoring of the Within the framework of subproject 5, the Monitoring and Inspection corrosion protection system haven’t been integrated overall simulation using new Previous experience in the production and developed yet. Corrosion monitoring repre- program modules is used to provide the op- load bearing behavior of grouted joints for sents an innovative, technological concept portunity to map degradation processes un- OWTs is based predominantly on the results for all technical areas, in which, despite der real load situations in order to be able of small- and large-scale trials [6] and their limited accessibility, it is necessary to detect to model the time-varying behavior of an numerical models [7]. The description of a failure of the corrosion protection system OWT over its entire service life. stiffness degradation of nodes under real at an early stage and to initiate corrective loads cannot, however, be done due to scale measures in order to avoid greater damage - State of the Art effects. With the development of a proto- and thus higher costs. Subproject 1: type measuring unit in GIGAWIND alpha Data Management and Analysis ventus, real displacements in the grouted Subproject 3: A wide range of measurement options for joint could be determined for the first time. Degradation Models monitoring the condition of the structure The continuous measurement of the real This subproject deals with the different deg- of the structure (including optic fibers and displacement and the associated detection radation phenomena of the supporting struc- acoustic sensors) exist. With the emergence of possible stiffness changes require further tures of offshore wind turbines, which have of social networks and other data-intensive development. to be considered during the design process. applications, there has been a significant These are the marine growth, scouring, cor- development in the area of database sys- Previous laboratory tests showed very good rosion and protection, fatigue strength of the tems in order to be able to realize intensive performance for the mineral corrosion pro- steel components and changes in the struc- research processes. The rapid discovery of tection system in terms of durability (see tural behavior of soil. semi-structured data enabled new database [8]). However, these are not sufficient to es- systems such as Apache Cassandra [1] and timate the material degradation. The influ- The design of an offshore wind energy de- frameworks such as Hadoop [2]. However, ence of defects, e.g. in the form of cracks vice considering mentioned degradation existing tools for handling large amounts of in the mineral corrosion protection layer and phenomena is based on state-of-the-art data aren’t readily available to Structural their effects on the corrosion protection are technology utilizing the BSH standard 7005 Health Monitoring (SHM) applications and not yet known for this particular applica- [17], which describes the minimum require- necessary adaptations should be made. tion. Also, the tests according to ISO 20340 ments for the construction-related and struc- Examples of implemented object-specific [9] for the characterization of conventional tural components of offshore structures for isolated solutions for data analysis in bridge organic coatings cannot be transferred to the use of offshore wind energy. In addition, structures are (see [3]): the mineral system due to the experimental DNVGL-ST-0126 [18] regulates the neces- setup. sary design specifications for the supporting - Arsenal Bridge, Rock Island, IL, USA, structures of offshore wind turbines. Within 2009 Currently, major efforts are invested the DNVGL-ST-0126, reference is made to - Kishwaukee River Bridge, Illinois, USA, throughout the world and across industries further DNVGL offshore standards and rec- 2001 to develop effective systems for the condi- ommendations, which mainly deal with

– 144 – Annual Report 2015-2017 CROSS145 DISCIPLINARY TOPICS Annual Report 2015-2017145

the design of components of the foundation depth is investigated in this subproject and ware is referred to as the DataSet format structure. more accurate results about the reliability of (Fig. 3). The data format describes chrono- lifetime predictions [21] will be obtained. logically ordered data objects, the so-called Subproject 4: DataSets (DS). The sequence of several Adapted model experiments - Subprojects: DataSets is called DataSetList and contains Jian et al. [19] investigated the influence of the measured data from a measuring chan- currents with varying angles of attack and Subproject 1.1: nel or the evaluation data from an evalua- velocities in combination with waves. The Data transfer and download tion in sorted order. relevance and influence of currents on the Based on extensive measured data, which wave frequency, the wave rise and the forc- were generated during the first years of Subproject 1.4: es were taken into consideration. In addi- operation of the offshore wind farm alpha Modular Basis for data evaluation tion, marine growth leads to increased flow ventus, validation and optimization of the The software SMMEXS, which is the basis resistance due to greater surface roughness structural design is carried out in the overall of further development within the frame- [20] and has an effect on the hydrodynamic project. Subproject 1 evaluates the neces- work of the GIGAWIND life project, can be mass of water, which influences the dynam- sary data sources and creates interfaces for used both for individual operation as well as ics of the structure. In this context, detailed these data sources. In bilateral coordination for use in a client-server environment. Con- insights into the effects of short-sea swell, with the institutes involved, it was examined ceptually, the management of the measuring marine growth and their impact over the which databases are relevant for the project points and the configuration of the evalua- lifetime of OWT are still unknown. realization and which evaluation types and tion tasks is done by the local software SM- algorithms are of interest. The resulting MEXS. The data storage of the measure- Previous studies show high resistance of a process of data management and analysis is ment and evaluation data is performed by mineral corrosion protection against frost, shown in Fig. 2. the software LAMA (LArge Measurements penetration of water and chlorides [8]. In the Accessible) of the project partner Fraun- process, however, no overlays and previous Subproject 1.3: User database hofer Gesellschaft e.V. [22]. damage of the individual test specimens are The internal data format used for data stor- taken into account, thus are investigated age and data analysis in the SMMEXS soft- within the scope of this project.

The validation of degradation models is performed based on measurement data from real turbines or experiments. Exact planning and optimization using virtual experiments is used to check the requirements and reduce costs.

Subproject 5: Method and Model Integration Significant developments in numerical models and tools for the holistic simulation Figure 2: Data management and analysis process of OWTs took place in recent years. The focus was on modeling the global behavior of the overall system. Significant detail improvements were achieved in the field of wind field, rotor aerodynamics and plant control. However, the application of more precise forecasting models has been restricted to load bearing capacity or serviceability of structural elements. With the help of numerically efficient implementations of detailed models into the overall simulation, it is possible to take account of model details and degradation effects relevant for the impact side, thus significantly improving load predictions. A new level of modeling Figure 3: Data format used in SMMEXS

– 145 – 146CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Subproject 2.1: Monitoring of scour and marine growth In this subproject, the long-term wind and sea measurements on the FINO1 platform are analyzed. In addition to a description of the wind and wave characteristics, the investigations carried out on this database include in particular the correlation of wind and wave directions, which can have a major impact on loads, especially in dynami- Figure 4: Concept of data management and analysis system cally reacting OWTs. A sample resulting scatter diagram and correlation is shown in Fig. 5.

An important conclusion of this subproject is the fact that the direction and combination of wind and wave effects with recurrence periods of several decades are location dependent. At the location FINO1, the effective directions of the wind velocity and the significant wave height differ by 45 ° - 60 °. At the location FINO3, 140 km northeast of the FINO1 location, the wind speed and the significant wave height Hs50 have almost the same effective directions.

Subproject 2.2: Identification of condition parameters This section identifies parameters that describe the condition of a wind turbine. A state refers primarily to the state of damage of the entire wind turbine. Parameters are characteristic quantities of time series, Figure 5: Scatter diagram of wind speeds and significant wave height for example, statistical quantities, natural frequencies or residuals. Changes in the condition parameters, however, do not ex- clusively result from structural damage. They can also result from plant operation, environmental conditions, structural boundary conditions and measurement errors. Ac- cordingly, it is examined which influences are recognizable in the state parameters and whether these influences allow damage detection. In addition, the condition parameters are determined using measured data from the alpha ventus wind farm (Fig. 6), thus providing a realistic assessments.

Identification of such parameters is done via Artificial Neural Networks, and show that both the Frequency Domain Decomposition and the Artificial Neural Networks are applicable in real offshore wind turbines and Figure 6: Overview of data points and sensor positions. Left: AV07 Adwen OWT on a tripod. complement each other. The natural Right: R04 Senvion OWT on a jacket

– 146 – Annual Report 2015-2017 CROSS147 DISCIPLINARY TOPICS Annual Report 2015-2017147

frequencies and eigenmodes provide a first visible effect on the second and third natural Results obtained from SSI and FDD show overview of the overall dynamic behavior frequencies. that damage has a visible influence on all of the system and the influences of environ- natural frequencies, with the highest influ- mental and operating conditions. The sin- The difference in influence is due to the fact ence on the first natural frequency. Further- gular value decomposition curves indicates that the marine growth was located where more, damage greater than the wall thick- excitation frequencies, noise influences and the eigenmodes of higher natural frequen- ness, leads to the ingress of water into the signal energies. cies have larger shifts. structure and produces a jump in all natural frequencies. Subproject 2.3: The modal damping increases significantly Monitoring experiments to estimate for all eigenmodes, as more friction is in- Subproject 2.7: hydrodynamic masses troduced by the marine growth due to its Inspection intervals for tower and sub- The effect of hyrdrodynamic masses by ma- roughness. structure from degradation rine growth is investigated based on varying The investigations of grout joints (SP 3.6), modal parameters. To determine these ef- Subproject 2.4: welded components (SP 3.7) and screws fects, both "Datadriven Stochastic Subspace Monitoring experiments with damaged (SP 3.8) have shown that during the service Identification" (SSI) (see [23], [24]) and the structures life of the OWT, no degradation effects are Frequency Domain Decomposition (FDD) Since both marine growth damage has an expected on these structural components of are used. influence on the modal parameters ofa the support structure. However, the bedding structure, it is of paramount importance to conditions change due to scouring. Further- Marine growth was applied via synthetic determine differences between the influ- more, marine growth affects the loads on sheets, shown in Fig. 7, having a scaled sur- ence of marine growth and damage. As in the structure. Both effects lead to a chang- face roughness and density similar to ma- subproject 2.3, the "Datadriven Stochastic ing structural stiffness of the OWT and also rine growth. Subspace Identification" (SSI) (see [23], to changed loads. [24]) and the "Frequency Domain Decom- Results obtained from SSI and FDD show position" (FDD) are also used to determine During the design of OWT, assumptions for that the degree of marine growth has little these effects. such degradation phenomena are made ac- influence on the first natural frequency, but a cording to the relevant standards. However, Various degrees of damage were applied ongoing OWT measurement and monitor- symmetrically at the pile structure, which ing can provide detailed information on the was scaled in the sub-projects 4.3 and 4.4. current status of degradation, allowing more Damage was implemented via saw cuts accurate estimation of OWT exposure and having following depths: 2 mm, 7 mm, 12.5 remaining service life. mm and 25 mm. The sawn pile is shown in Fig. 8. In all tests of the damaged structure a Marine growth: marine growth was applied. Several investigations took place in the area

Figure 7: Application of synthetic sheets as Figure 8: Application of damage to experimental pile marine growth

– 147 – 148CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

of alpha ventus, focusing on marine growth. The growth fauna on the research platform FINO1 was recorded between 2005 and 2007 in spring, summer and autumn [25]. In the alpha ventus offshore wind farm, measurements were taken from 2009 to 2012 on up to four supporting structures (see [26], [27], [28], [29]). The supporting structures investigated here are mainly populated by mussels up to a water depth of 5 m. The results of the investigations listed are shown in Fig. 9 for both the FINO1 research platform and the OWT M7, M12, R1 and R6 offshore wind farm alpha ventus for water depths of 1, 5 and 10 meters. The thickness of the marine growth results from the wet Figure 9: (a) thickness of marine growth at FINO1 and in the offshore wind farm alpha ventus; mass per area determined in the investiga- (b) derived models of marine fouling for different months and the approach according to [72]. tions and a density of 1325 kg / m³. The effect of marine growth on natural frequencies is shown in Fig. 10.

Scour: Scouring has a direct impact on the structural response of OWT. Investigations of scour formation in monopile substructures show that the structural dynamic properties of OWT change significantly with increasing scouring depth [30]. Furthermore, it leads to the increase in fatigue loads [31]. The effect of scour in natural frequencies is shown in Fig. 11.

Fatigue Loads: Figure 10: Relation between natural frequencies of an OWT with jacket support structure With increasing marine growth, both the and different models of the marine (EF). cross section and the surface change, thus increasing the wave-induced loads on the jacket support structure. When using the Morrison equation (see description in SP 5.3), higher hydrodynamic drag coefficients are used, the values of which are taken from, for example, [32]. Furthermore, the investigations carried out show that the fatigue loads of an OWT with jacket support structure are increased with increasing scour. Therefore, the monitoring of scour development is necessary to be able to estimate the fatigue load during operation.

Subproject 2.8: Degradation models for steel elements from measurements In order to validate the developed degradation models with realistic loads, numerical Figure 11: Relation between natural frequencies of an OWT with jacket support structure holistic simulations of OWT are carried out and different scour depths.

– 148 – Annual Report 2015-2017 CROSS149 DISCIPLINARY TOPICS Annual Report 2015-2017149

on the basis of measured environmental parameters. The holistic numerical simulations were performed for the normal, undisturbed operation of the OWT according to the design load case DLC 1.2 (DLC, design load case) according to IEC 61400 3 [33]. Both long-term and short-term statistics are considered. The long-term statistics include the distribution and correlation of the measured environmental parameters. The short-term statistics take into account the stochastics of the generated time series. The long-term statistics of environmental parameters are shown in Fig. 12. Figure 12: Long-term statistics of the environmental parameters on the research platform FINO1: (left) Frequency distribution of wind direction and wind speed at hub height in 2010, (to the right) to the wind speeds associated averaged sea state To determine the structural stresses accord- parameters ing to DLC 1.2 [33], two load time-series with a length of ten minutes are generated for each load combination, as shown in Fig. 12 (right). For this purpose, the finite element software Poseidon was used, which is coupled with the program Flex5 to determine the wind-induced rotor-nacelle loads [34]. The sea states, whose induced loads are applied to the support structure within the Poseidon software, are modeled as a JONSWAP spectrum, as explained in [34] Figure 13: Tripod structure (left), CAD model of the prototype measuring unit made or [35]. The fatigue loads are determined of GIGAWIND alpha ventus (center) and salvaged prototype measuring unit after 22 from the time series of the structural stress- months of use (right) es generated in the OWT's holistic simulations using the rainflow counting algorithm, as described in [36]. The determined fatigue loads are used for the investigations carried out in subproject 3 on Grout joints (SP 3.6), welded components (SP 3.7) and screws (SP 3.8).

Subproject 2.8: Developments of the measurement unit In GIGAWIND alpha ventus project, a mobile measuring unit was developed at the Institute for Building Materials. This was installed and tested under maritime conditions on the support structure of a fully operational offshore wind turbine. After a 22-month measurement on a tripod structure in the North Sea, it was returned (see Fig. 13). The aim of the project is to analyze the salvaged prototype measurement unit, to identify its weak points and to increase their performance and reliability in a further development.

Figure 14: Assembly procedure of measuring unit II.

– 149 – 150CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

The resulting new measurement unit avoids all identified weakspots and has a higher performance and reliability. The assembly procedure of the new measuring unit is shown in Fig. 14.

Subproject 2.10: Mineral corrosion protection: tests and repairs To estimate the anticorrosive effect of the mineral corrosion protection system under real conditions, retrieval experiments on an offshore wind turbine in the alpha ven- Figure 15: Multi-stage manufacturing process of the plate-shaped test coupons tus test field are carried out. The aim of this investigation is to detect and analyze the corrosion protection effect of the specially developed mineral corrosion protection system. In addition to the analysis of the corrosion-protecting effect of an intact mineral corrosion protection layer, the analysis of a damaged (cracked) protective layer is also the focus of the investigations. For this purpose, special specimens, the so-called test coupons, are designed and manufactured for a period of ~ 1.5 years. The manufacturing process of the test coupon is shown in Fig. 15.

The resulting new measurement unit avoids all identified weakspots and has a higher performance and reliability. The assembly procedure of the new measuring unit is shown in Fig. 14. Figure 16: Schematic representation of a possible spiral wound Due to the difficult environmental condi- plastic formwork with internal textile reinforcement for appli- tions that exist in offshore wind turbines, cation of the mineral corrosion there are currently almost no repair concepts protection system or principles for these structures. Thus, the corrosion protection for these structures is usually measured and designed for the entire service life, so that a repair of the sys- and the existing water is displaced. The Subproject 2.11: tems during this time is not required. In the placement of a suitable formwork poses a Realistic stresses from alpha ventus case of damage, however, systems of this great challenge in the case of a pile of an measurement data type can hardly be repaired using conven- offshore wind energy installation due to the The measurement of stresses on the support tional methods. Here, a problem-free appli- geometry. One possible variant could be a structure of offshore wind turbines allows cation both above and below the waterline water-impermeable plastic formwork with realistic load assumptions for frequently oc- is possible. an inner textile reinforcement wound spi- curring events as well as for rare extreme rally around the pile (see Fig. 16). The tex- events. Prerequisite for reliable statements A comparable repair procedure has already tile reinforcement serves to limit the crack on rare loads are measurement series that been used on a trial basis on foundation width in the mortar layer and as a spacer be- span several years. For this purpose, the piles of a naval structure [37]. In the pro- tween the component surface and the plastic strains on the anchoring piles of the jacket cess, a space is created between the com- surface. The resulting gap can be pumped of the R4 turbine of the offshore wind farm ponent surface and a formwork (grid), in with mortar from below and the water can alpha ventus are examined. Due to numer- which mortar is introduced under pressure be displaced upwards. ous measurement failures, only short meas

– 150 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017151

urement time series of individual strain gauges are evaluated and related to the exposure parameters from subproject 2.1.

The distribution of daily extreme values of pile strains are shown in Fig. 17. It is well described by the Weibull III max distribution function. This applies in particular to the upper quantile range.

An example of the empirical distribution of the week extreme values of the strains at the R4_D-E1_1 stake in the observation period 19.10.2010 to 03.05.2013 (see [38]) is shown in Fig. 18.

The Weibull III min distribution is not suitable for describing the empirical distribution of strains in this database. In addition to Figure 17: Empirical distribution of the day extremes of strains at the R4_D-E1_1 stake and various distribution functions the maximum likelihood method, the least square method is also used to estimate the distribution parameters. However, the estimation methods used have only a very small influence on the course of the distribution functions. Also, the illustrated empirical distribution of the weekly strain extremes can be well described by the Weibull III max distribution.

With the statistical description of the pile strains, linear and elastic material behavior can be used to infer maximum and minimum strains of longer recurrence periods.

Subproject 3.1: Effect of marine growth on load coefficients Basics of wave and current loads on monopiles were extensively studied in recent years, especially in wave channels, which generate unidirectional waves. However, loads are caused by wave spectra, which scatter in their direction. Using the directional spectra, natural seaways are modeled much more accurately. Investigations of this kind are carried out in wave pools in which multidirectional sea state can be generated. Extensive tests were carried out on combined wave and current loads for load analysis and scour development, which are described in detail in the subproject 4. In the following, the findings of the experiments Figure 18: Empirical distribution of the week extreme values of strains at the R4_D-E1_1 are used to show the influences of marine stake and various distribution functions growth and interacting waves and currents.

– 151 – 152CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Figure 19: Position of measuring sensors Figure 20: History of maximum scouring depths due to tidal flow 1-6 with reference to the water level. Position of measuring segments 1-6 with reference to the water level

The installed force sensors on smooth and rosion protection is intended to protect the Subproject 3.6: Grout material degrada- marine grown cylinders are shown in Fig. 19. steel tower of an offshore wind turbine tion under fatigue loads and multi-axial The measured forces are compared to those against corrosion over its service life, es- stresses determined by the Morrison equation. pecially in the water slushing zone (see Offshore wind turbines in the research wind Fig. 21). farm alpha ventus are installed on the sub- Subproject 3.2: structures shown in Fig. 22 (left). Six of Time-dependent scour development and The protection system is based on the pas- the twelve installed systems are based on seasonal effects sivating effect of cement-bound building tripods and six on jackets. The anchoring In addition to a reliable estimate of ex- materials, which is well-known from classic of the foundation structures in the seabed pected maximum scour depths, an accu- reinforced concrete construction. is realized via rammed steel pipes. For the rate prediction of the time-dependent scour transfer of force between foundation struc- development is required, in particular for tures and ground pile, a Grout joint is used, the determination of suitable times for the installation or maintenance of offshore structures. Based on the adapted model experiments described in subproject 4.4, approaches for the description of the time- dependent scour development as a result of tidal flow and multidirectional seaway load are presented in this subproject. A comparison between the measured and predicted scour development is shown in Fig. 20.

Subproject 3.3: Durability of mineral corrosion protection Efficient corrosion protection systems are essential for the durability of an offshore wind turbine. In addition to the conventional corrosion protection systems, a mineral corrosion protection system is investigated within the scope of this project. The cor- Figure 21: Schematic representation of the corrosion rate of an OWT [112].

– 152 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017153

as shown schematically in Fig. 22 (center, right). In this connection, a small steel pipe is inserted into a large steel sleeve, and the resulting cavity is filled with a high strength mortar. In order to ensure a force transfer between steel and grout, the surfaces are profiled with build-up welds (shear ribs).

Due to the cyclic stresses of the Grout connection as a result of wind, waves and system operation, fatigue occurs which lead to a degradation of the connection. The objective of subproject 3.6 is to identify these degradation phenomena and to consider them in simulation models. An example of the numerical simulations of the grouted joints is shown in Fig. 23.

Based on the determined realistic loads, nu- Figure 22: Substructures of the alpha ventus wind turbines (left), external view of the merical simulations were performed to de- Grout connection between foundation structure and ground pile (center) and detailed termine the influence of the nonlinear grout view of the grout gap with a schematic representation of the structural behavior (right) material and stress state. The expected degradation of the grout material for the alpha ventus jacket was estimated, suggesting a characteristic damage after about 40 years of operation. In order to be able to better track the real degradation state, the application of measurement technology to detect the relative displacement between pile and sleeve is essential.

Subproject 3.8: Screw preload forces in non-cyclic loading To connect the segments of a steel tower for offshore wind turbines, screwed ring flange connections are used (see Fig. 24). These are also increasingly used instead of Grout connections for connecting transition pieces and monopiles. Due to the high dynamic Figure 23: Results of numerical simulations loads and large lever arms HV screw sets with very large diameters M64 or M72 are used. This is also the case in alpha ventus. At middle flanges between the tower sections and at the nacelle connection, slightly smaller diameters in the range M36-M48 are usually used. Nevertheless, codes of practice and regulations, which are used for the design of the support structure, are usually validated only for diameters ≤ M36. Therefore, the influence of the diameter on the fatigue strength of HV sets in [39] was experimentally investigated. Figure 24: Positions of ring flange connections in a support structure with jacket supporting structure (left) and HV bolt sets with large diameters (right)

– 153 – 154CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Subproject 3.9: tion of non-linear modulus methods, so- The experiments were carried out in the Method development for the estimation of called p-y methods (see [41], [18]), which wave basin of the Ludwig-Franzius-Institut accumulated pile head deformations describe the soil reaction due to constant (Fig. 26). The pool has a size of 24 m * 40 Monopiles are a frequently chosen sub- loading. In fact, however, the fatigue loads m with a usable length of 30 m and a width structure for offshore wind turbines. The are characterized by the founding reaction of 15 m at a maximum water depth of 1 diameters of these piles vary between about due to the continuous unloading and re- m. The installed pump capacity of 5 m³/s 5.0m and 8.0m. Under cyclic loads, the piles loading rather than the reaction to the sin- can generate currents of up to 0.3 m/s at 1 show permanent displacements and twists. gle constant load. m water depth and correspondingly higher Since offshore wind turbines are sensitive to speeds at shallower water depths. Both permanent deformations, especially against This subproject is intended to investigate regular unidirectional and irregular multi- misalignments, it is important to be able to the possibility of applying increased foun- directional waves as well as stationary flow determine them as accurately as possible. dation stiffness due to consideration of un- fields can be generated. The flow can be -or Some approaches have recently been de- loading and reloading stiffness. thogonal as well as oblique to the wave di- veloped [40], but so far none of these ap- rection. In order to produce the waves with proaches have been established. In the fol- Subproject 4.1: different angles of attack between θ = 5° lowing the so-called Stiffness Degradation Model experiments on sea spectrums and θ = 175°, the plant has a multi-element Method (SDM) is presented, which was for combined wave and current loads on wave machine with 72 single-wave paddle. developed and used in extensive parameter smooth and rough piles Each paddle is 40 cm wide and 1.8 m high studies to determine the pile behavior under In this subproject, experiments are conduct- and has a maximum stroke of 1.2 m (± 0.6 cyclic horizontal loads. A selection of these ed to investigate the interaction between m) and a maximum speed of 3 m/s. studies and the results obtained from them waves and current. are the core of this subproject. An example The direct comparison of the graphs, of a finite element model utilized in such In the superposition of waves and currents, shown in Fig. 27, shows that the forces in methods is shown in Fig. 25. the waves deform according to the Dop- the x-direction for the wave load without pler effect. If the flow moves in the same flow effect are 10-15% greater than for the Subproject 3.10: direction as the waves, the wavelength is forces in the wave direction with combined Cyclic stiffness of horizontally loaded piles stretched and the wave amplitude becomes orthogonal flow influence. Thus, the flow For the overall dynamic simulation of off- smaller. If they run in the opposite direc- (y-direction) influences the x-components shore wind turbines, foundation stiffness is tion, the wavelength is shortened and the of the wave forces and creates a vector tri- a key input parameter. It significantly influ- wave steeper (see [42], [43]). Currents af- angle with y-components from the x-forces ences the vibration behavior and thus the fect waves not only on the surface, but also and thus slightly reduces the original x- fatigue loads that occur (see SP 5). across the depth to the seabed [44]. The force of the wave. In addition, there is a interactions with regard to the formation of slight (in the graphs difficult to recognize) In design practice, lateral foundation stiff- swells and structural loads are investigated eccentricity of the forces, which show a nesses are usually applied by the applica- experimentally. shift to the flow-facing side.

Figure 25: 3D FE model for the num. Simulation and calculation steps of the Stiffness Degradation Method

– 154 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017155

Figure 26: Wave basin

Figure 27: x-forces (left) and y-forces (right) along the circumference of the cylinder with modeled marine growth at the time of the wave crest (blue) and the crest (red) with H = 6.8 cm, T = 1.2 s and vertical acting Flow at 0.32 m/s

Figure 28: Experimental setup in top and side view [205]. Dimensions in centimeters, not to scale

– 155 – 156CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Figure 29: left: scour development after approx. 8 h with a unidi- Figure 30: Transparent model pile with installed camera and auto- rectional flow load of u = 0.2 m/s; right: scour development after mated travel system, according to [109] approx. 8 h as a result of tidal flow with a maximum flow velocity (peak) of u = 0.25 m/s

Subproject 4.2: channel is driven by a pump system con- Subproject 4.4 and 4.5: Sea spectrum and scouring sisting of four individually controllable Model tests for hydrodynamic masses and In this subproject, experiments are con- pipe pumps with a total flow capacity of damaged structures ducted to investigate the generation of scour max. 0.5 m/s. The setup is shown in Fig. 28. For the experimental investigation of the ef- resulting from sea states similar to the north Resulting scouring is shown in Fig. 29. fects of marine growth and damage on the sea. modal behavior of offshore wind turbines, Scouring tests were also performed at the a scaled model of a supporting structure The model experiments take place in the Wave Basin. Resulting scouring is shown was used (see Fig. 31) based on the mono- circulation channel of the Ludwig-Franzius- in Fig. 30 for multi-directional waves and pile defined in the Offshore Code Compari- Institut. The scale is 1:40. The circulation currents. son Collaboration (OC3) [45]. A variety of sensors were also placed on the structure, which are shown in Fig. 32.

Figure 32: Measurement setup of the test Figure 31: Dimensions of the test structure structure

– 156 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017157

Subproject 4.7: Model tests for mineral corrosion protection As described in subproject 3.3, cracks occur within the thin mineral corrosion protection layer as a result of shrinkage or mechanical stress, which favor the chloride entry and increase the risk of corrosion of the steel support structure. The determination of the properties in the non-cracked mortar system Figure 33: Implementation of cracks in specimens. is therefore insufficient for an estimate of the effectiveness. With previous test methods, such as the Rapid Chloride Migration Test (RCM) [46], it is only possible to determine the chloride penetration behavior in the non- cracked state. It is therefore also necessary to determine the effects of a self-healing on the chloride penetration behavior.

Cracks are deliberately generated as shown in Fig. 33, which are then installed into migration cells and tested as shown in Fig. 34.

Subproject 5.1: Simulation and design package In order to standardize all structural models Figure 34: Installation of specimens used, the DeSiOLite framework (see Fig. 35) was derived from the previously developed DeSiO framework. It produces input files for predefined structural types for all the simulation tools used in GIGAWIND life. This ensures an improved verification process since all models are consistent.

In order to ensure and improve the networking between the subprojects, interfaces to the monitoring and data management software SMMEXS used in subproject 1 were included. Various import and export filters guarantee the simple migration of simulation data to the central data storage. This facilitates the validation process using measurement data from the offshore test field alpha ventus.

Subproject 5.2: Modular implementation of degradation models Based on SP 3.6, this subproject describes the implementation of the degradation model of predominantly axially stressed Grout connections in holistic numerical simulations with the help of the software interface POSEIDON-Flex 5. This type of im- Figure 35: Graphical user interface of DeSiOLite

– 157 – 158CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

plementation is comparable to the damage description of welded pipe nodes and differs significantly in the damage hypothesis. The basic flowchart is shown in Fig. 36.

Subproject 5.4: Dynamic and structural design using a holistic model When designing OWTs, different types of support structures are considered. The choice of the support structure is often determined by the location-related water depth and the size of the system. Different types of support structures used for OWT in the German North Sea are shown in Fig. 37. Jackets and tripods are well-known as support structure types with pile joints in the offshore wind farm alpha ventus. Support Figure 36: Overview of the implementation of degradation models in the framework structures of floating OWTs, whose use is only an economic option from water depths of 50 m, are not listed here.

The research work performed in this subproject aims at further developing the parametric monopile support structure in [47] to include other sub structures such as jackets, shown in Fig. 38.

Subproject 5.5: Validation of simulation tools using measurements from operating wind turbines In subproject 5.6, new approaches for the Figure 37: Supporting structures of offshore wind turbines modeling of jacket structures were verified. Based on this modeling some selected load cases were examined. Here, the jacket modeled using the classic bar models coupled with a super element approach. To validate the super element approach, the fatigue stresses in the nodal area (a double Y-node on the jacket head and a double K-node) were compared with measurements.

The super element approach provides significantly more realistic results for the maximum node damage, as shown in Fig. 39. The simpler calculations with beam elements are almost always on the safe side in all investigated time. For structural optimization or more realistic calculation of a residual lifetime, the use of the super element approach is recommended.

Subproject 5.6: Figure 38: Modeling of jacket substructures. Left: Ansys, Right: Poseidon

– 158 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017159

Verification of the superelement ap- Summary Subproject 3 includes degradation models, proach and developed modules: a holistic developing currently existing methods in approach The project focused on the validation of alpha ventus and introducing new models. The developed modules as well as the sug- methods and structural models for an in- gested superelement approach is validated tegral and cost-effective design of OWEA Subproject 4 carried out model experiments in this subproject. Comparisons with AN- support structures. In particular, the follow- to gain insights to phenomena that cannot be SYS, Poseidon and measurement data are ing topics were investigated: derived from alpha ventus data. performed. In subproject 1, a user database for merging Subproject 5 represented the merging of the system for laboratory and simulation data findings from the first four subprojects by was developed. using this knowledge in overall simulations to provide forecasts for the overall system In subproject 2, monitoring aspects were in- behavior. vestigated based on measurement data from alpha ventus and laboratory data.

Figure 39: Position and degradation of the considered jacket nodes

– 159 – 160CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

References

[1] Capriolo, E.: Cassandra High [13] Farrar, C.; Lieven, N.: Damage Conference, St. Louis, USA, 2006. Performance Cookbook, Birmingham, prognosis: the future of structural health [24] van Overschee, P.; De Moor, B.: Vereinigtes Königreich: Packt Public, monitoring, Phylosopical Transactions of Continuous-time frequency domain 2011. the Royal Society A: Mathematical, subspace system identification, Signal [2] White, T.: Hadoop: the definitive Physica and Engineering Sciences Processing, 52(2), pp. 179-194, 1996. guide, Sebastopol, CA, USA: O’Reilly 365(1851), pp. 623-632, 2007. [25] Krone, R.; Gutow, L.; Joschko T. J.; Media Inc., 2012. [14] Kraemer, P.: Schadensdiagnosever- Schröder, A.; Epifauna dynamics at an [3] Dong, Y.; Song, R.; Liu, H.: Bridges fahren für die Zustandsüberwachung von offshore foundation – Implications of Structural Health Monitoring and De- Offshore-Windenergieanlagen, future wind power farming in the North tection – Synthesis of Knowledge and Dissertation, Universität Siegen, 2011. Sea, Marine Environmental Research Technology, Fairbanks, AL, USA, 2010. [15] Kraemer, P.; Fritzen, C.-P.: Aspects of 85, pp. 1-12, 2013. [4] Link, M.; Petryna, Y.; Künzel, A.: Operational Modal Analysis for Structures [26] Kazmierczak, F.; Laudien, J.; Modellbildung und Parameteridentifi- of Offshore Wind Energy Plants, Structural Fürst, R.; Bönsch, R.; Benthosbio- kation für Grouted Joints an Offshore- Dynamics and Renewable Energy 1, logische Erhebungen während der Windenergieanlagen, in 4. VDI-Tagung pp. 145-152, 2011. Bauphase des Offshore-Windparks Baudynamik, Kassel, 2012, pp. 259-269 [16] Häckell, M.; Rolfes, R.: alpha ventus, Institut für Angewandte [5] Germanischer Lloyd: Guideline for Monitoring a 5MW Offshore Wind Energy Ökosystemforschung GmbH, the certification of wind turbines, 2010. Converter – Condition Parameters and Neu Broderstorf, 2010. [6] Anders, S.: Betontechnologische Triangulation Based Extraction of Modal [27] Kazmierczak, F.; Preuß, S.; Breyer, Einflüsse auf das Tragverhalten von Properties, Mechanical Systems and Signal S.; Kern, K.; Fürst, R.; Bönsch, R.: Fach- Grouted Joints, Institut für Baustoffe, Processing 40(1), pp. 322-343, 2013. gutachten "Benthos" zum Offshore- Leibniz Universität Hannover, 2008. [17] Bundesamt für Seeschifffahrt und Windpark alpha ventus – [7] Keindorf, C.: Tragverhalten und Hydrographie: Mindestanforderungen an 1. Betriebsjahr, Institut für Ermüdungsfestigkeit von Sandwichtür- die konstruktive Ausführung von Offshore- Angewandte Ökosystemforschung men für Windenergieanlagen, Leibniz Bauwerken in der aus-schließlichen GmbH, Neu Broderstorf, 2011. Universität Hannover, Dissertation, Wirtschaftszone (AWZ), BSH Standard [28] Preuß, S.; Breyer, S.; Fürst, R.; 2010. Konstruktion, 2015. Bönsch, R.: Fachgutachten "Benthos" [8] Lohaus, L.; Weicken, H.: Polymer- [18] DNV GL: Support Structures for Wind zum Offshore-Windpark alpha ventus – modified mortars for corrosion Turbines, Offshore Standard DNV-ST-0126, 2. Betriebsjahr, Institut für Angewandte protection at offshore wind energy 2016. Ökosystemforschung GmbH, converters, in 13th International [19] Jian, Y.; Zahn, J.; Zhu, Q.: Short rested Neu Broderstorf, 2012. Congress on Polymers and Concrete wave-current forces around a large [29] Preuß, S.; von Allwörden, S.; – ICPIC 2010, Funchal, Portugal, 2010, vertical circular cylinder, European Journal Nestler, S.; Kazmierczak, F.; Fuchs, A.; pp. 151-157. of Mechanics B/Fluids 27, pp. 346-360, Gloede, F.; Breyer, S.; Fürst, R.; [9] International Organization for 2008. Bönsch, R.: Fachgutachten "Benthos" Standardization, Performance require- [20] Forteath, G.; Picken, G.; Ralph, R.; zum Offshore-Windpark alpha ventus – ments for protective paint system for Williams, J.: Marine Growth Studies on the Bericht über das 3. Betriebsjahr, Institut offshore and related structures, ISO North Sea Oil Platform Monrose Alpha, für Angewandte Ökosystemforschung 20340, 2009. Marine Ecology 8, pp. 61-68, 1982. GmbH, Neu Broderstorf, 2013. [10] Farrar, C.; Worden, K.: [21] Dong, W.; Moan, T.; Goa, Z.: Fatigue [30] Prendergast, L. J.; Gavin, K.; An introduction to structural health reliability analysis of the jacket support Doherty, P.; An investigation into the monitoring, Philosphical Transactions structure for offshore wind turbine effect of scour on the natural frequency of the Royal Society A: Mathematical, considering the effect of corrosion and of an offshore wind turbine, Ocean Physical and Engineering Sciences inspection, Reliability Engineering and Engineering 101, pp. 1-11, 2015. (365)1851, pp. 303-315, 2007. System Safety 106, pp. 11-27, 2012. [31] van der Tempel, J.; Zaaijer, M. B.; [11] Farrar, C.; Jauregui, D.: [22] Herrmann, R.; Rabe, J.; Bolle, G.; Subroto, H.: The effects of Scour on the Comparative study of damage Marx, S.: Konzepte für Datenqualität und design of Offshore Wind Turbines, in identification algorithms applied to a Datenablage bei Entwurf und Umsetzung 3rd International Conference on Marine bridge: I. Experiment, Smart Materials von Monitoringsystemen, Bauingenieur Renewable Energy, Blyth, and Structures 7, pp. 704-719, 1998. 92, pp. 537-545, 2017. Großbritannien, 2004. [12] Chang, P.; Flatau, A.; Lui, S.: Re- [23] Brincker, R.; Andersen, P.: view Paper: Health Monitoring of Civil Understanding stochastic subspace [32] DNV GL, Loads and site conditions Infastructure, Structural Health identification, in Proceedings of the 24th for wind turbine, Standard ST-0437, Monitoring 2, pp. 341-358, 2003. International Modal Analysis 2016.

– 160 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017161

[33] International Electrotechnical for IEA task 23 offshore wind Commission, Wind turbines part 3: technology and deployment, 2010. Design requirements for offshore wind [46] Bundesanstalt für Wasserbau, turbines, IEC 61400-3, 2009. Chlorideindringwiderstand von Beton [34] Böker, C.: Load simulation and (MCL), Karlsruhe: BAW-Merkblatt, 2012. local dynamics of support structures for [47] Dubois, J.; Thieken, K.; offshore wind turbines, Shaker-Verlag, Schaumann, P.; Achmus, M.: Influence Aachen, 2010. of soil resistance approach on overall [35] Mittendorf, K.; Nguyen, B.; structural loading of large diameter Zielke, W.: Seegang und Seegangs- monopiles for offshore wind turbines, belastung, in 2. GIGAWIND Symposium, in Proceedings of the 1st International Hannover, 2002. Wind Energy Conference, Hannover, [36] Radaj, D.; Sonsino, C.: 2014. Ermüdungsfestigkeit von Schweißver- bindungen nach lokalen Konzepten, Düsseldorf: Verlag für Schweißen und verwandte Verfahren DVS Verlag GmbH, 2000. [37] Westendarp, A.: Instandsetzung von Wasserbauwerken heute und morgen, in: Bundesanstalt für Wasser- bau (BAW) (Hrsg.): Erhaltung von Wasserbauwerken und Brücken, Karlsruhe, 2010. [38] Germanischer Lloyd, RAVE - Sensor Description AV 04, 2011. [39] Schaumann, P.; Eichstädt, R.: Ermüdung sehr großer HV-Schrauben- garnituren, in Stahlbau 85(9), Ernst & Sohn, 2016, pp. 604-611. [40] Achmus, M.; tom Wörden, F.: Geotechnische Aspekte der Bemessung der Gründungskonstruktionen von Offshore-Windenergieanlagen, Bauingenieur 8, pp. 331-341, 2012. [41] American Petroleum Institute, Recommended Practice for Planning, Designing and Constructing Fixed Offs- hore Platforms – Working Stress Design, Recommended Practice API RP 2A-WSD, 21. Auflage, 2000. [42] Chakrabarti, S.: Handbook of offshore engineering (Volume 1), Elsevier, 2005. [43] Sarpkaya, T.; Isaacson, M.: Mechanics of wave forces on offshore structures, Van Nostrand Reinhold, New York, 1981. [44] Schneider, R.: Örtliche Bewertung der Schwingfestigkeit von Gewindever- bindungen, Dissertation, Technische Institut für Darmstadt, 2011. [45] Jonkman, J.; Musial, W.:Offshore code comparison collaboration (OC3)

– 161 – 162CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Innovative Wind Conversion Systems (10-20 MW) for Offshore Applications (INNWIND.EU)

Carl von Ossietzky Universität Oldenburg produced integrated tower and substructure WP4 addressed the challenges for the sup- Institute of Physics, (WP4). WP1 integrated the innovations in port structure in the 10-20MW wind turbine Research Group: Energy Meteorology, conceptual designs and focused as well on class. ForWind – University of Oldenburg, Research Group: Wind Energy Systems, external conditions, wind turbine control namely the Wind Energy Systems group, Research Group: Turbulence, Wind Energy and economical assessment. WP5 and 6 co-lead the Work Package and coordinated and Stochastics embedded the project by being responsible Task 4.1. The combination of large water Martin Kühn, Bernd Kuhnle, for the management and the dissemination depth and tall hub height due to the large Rasoul Shirzadeh, Martin Kraft, and exploitation. rotor, as well as large thrust forces and dy- Frederik Berger, Detlev Heinemann, namic loading results in strongly increased Joachim Peinke ForWind – University of Oldenburg par- demands for the support structure. Further- ticipated in the Work Packages 1, 2, 4, more, the increased rotor diameter yields in Funding: European Union’s Seventh lead Task 1.1. "External conditions", Task low rotational speeds and therefore possible Programme for research, 4.1. "Innovations on component level for excitations of the natural frequency of the Technological development and bottom-fixed support structures" and co- relatively stiff support structure by the rotor demonstration, Community’s Seventh coordinated WP4. or blade passing frequency. ForWind – Uni- Framework Programme versity of Oldenburg was engaged in load FP7-ENERGY-2012-1-2STAGE mitigation concepts both by sophisticated Ref.Nr. 308974 Project Description control of the rotor-nacelle-assembly with respect to the support structure as well as Duration: 11/2013 – 10/2017 In WP1, ForWind – University of Old- structural control concepts such as active, enburg, namely the Energy Meteorology semi-active and passive damper devices. group, lead the Innwind Task 1.1. on model- ForWind – University of Hanover analysed ling the wind conditions up to 300 m height. the feasibility of hybrid jackets comprising These wind conditions are crucially impor- classical steel tubes as well as sandwich el- tant for the design of the foreseen 10-20 ements to improve the ultimate limit state Introduction MW offshore turbines with rotor diameters strength of bottom-fixed multi member sup- of up to 250 m and maximum tip heights port structures. For this purpose an investi- INNWIND.EU is the follow-up of the fin- of up to 300 m. Over such a height range, gation of sandwich elements on component ished project "UpWind" addressing the massive changes in wind speed, wind direc- level was performed in Task 4.1. The con- vision of a 10-20MW wind turbine. This tion and turbulence can occur, which induce ceptual design of a hybrid jacket is made in project identifies further development and high loads in the blades and the whole tur- Task 4.3. innovation needs on system, subsystem, and bine. To model these conditions correctly, component level to achieve the objective of it is of utmost importance to simulate the a high-performance wind turbine beyond thermal stratification of the marine bound- Summary the state-of-the-art. It facilitates the realiza- ary layer in detail and then to compute the tion of 10-20MW wind turbines by inno- respective heat and momentum fluxes. In The first step in WP1 and especially Task vations on component level and decreases WP2, ForWind – University of Oldenburg, 1.1. was to compile and analyze existing their time to market for under the premise of namely the Turbulence, Wind Energy and measurements of the offshore wind resourc- cost effectiveness by necessary cost reduc- Stochastics group, is involved in the dem- es at higher atmospheres. Now the Energy tions of around 20% for the overall system. onstration and validation of new control Meteorology group has set up various nu- The project had a 5 years duration and 27 concepts by experiments at a medium scale merical modeling schemes for the analysis European partners are involved. The work turbine reproducing atmospheric turbulence of different weather situations, including packages can be divided in integrating, in a wind tunnel. The task here is to design extreme events. Especially, the results with embedding and innovative subcomponent- and built an active grid for the open jet fa- the meso-scale model WRF using the MYJ- based packages. The three subsystem work cility of the Delft University of Technol- scheme for the Planetary Boundary Layer packages focused on the light weight ro- ogy. This wind tunnel has a cross section (PBL) proved to be the best technique, com- tor (WP2), innovative, low-weight, direct of 3 times 3 m² and will be used for some pared to other models and PBL-schemes. drive generator (WP3), and standard mass- of the experiments in the work package. The next step will be the implementation of

– 162 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017163

these results in simpler 2-dimensional and trol (SPRC), a data-driven control technique dampers showed a high potential in lower- engineering models. In WP2, an active grid developed by the Technical University of ing the fatigues loads in sideway direction. is used to generate customized turbulence Delft. In Fig. 1 the setup in the wind tunnel Only small advantages in the fore-aft direc- for wind tunnel experiments. The used ac- is shown. tion could be seen by implementing these tive grid divides the cross section by 80 technologies. The stochastic wind input is horizontal and vertical elements, resulting In WP4, load mitigation strategies have here the governing factor. Due to higher in a mesh width of about 0,14 m. Each of its been analysed with respect to fatigue loads masses for the 20MW turbine design pas- axes is connected to a servomotor in such a on the support structure. These have been sive damper technologies showed less per- manner that each can be controlled individ- separated in control concepts and damp- formance due to space and mass restrictions ually by a real time system. Mounted on the ing devices. Firstly, control related load for such extra equipment. Active damper rods are square flaps which are, depending reduction technologies like speed exclusion technologies, like the magnetorheological on the orientation with respect to the inflow, zone and peak shaver have been included (MR) damper showed a higher potential for blocking and deflecting the wind. Dynamic to reduce oscillations in the low wind op- fatigue load reduction. Further investiga- changes of the angle of attack of the flaps to erational mode as well as thrust reduction tions are needed to improve the results and the flow are in the following described as near rated wind speed. Secondly, concepts to show the applicability on real structures. an excitation protocol. The excitation pro- for tower dampers have been analysed. tocol defines the dynamics of the generated Passive Tuned Vibration Absorbers (TVA), turbulence. As part of the project different Tuned Mass Dampers (TMD) and Viscous excitation protocols were developed and the Fluid Dampers (VFD) have been imple- used active grid were built to test the con- mented in the aeroelastic models of the 10 cept of Subspace Predictive Repetitive Con- MW reference turbine design. Tuned mass

References

[1] Kuhnle, B.; Kühn, M.: Unfavourable trends of rotor speed and systems dynamics for very large offshore wind turbines - Analysis of the 10MW INNWIND.EU reference turbine," in EWEA Offshore, Copenha- gen, 2015. [2] Shirzadeh, R.; Kühn, M.: Applica- tion of two passive strategies on the load mitigation of large offshore wind turbines, Journal of Physics: Confe- rence Series, vol. 749, no. 1, p. 012011, 2016. [3] Shirzadeh, R.: Support structure load mitigation of a large offshore wind turbine using a semi-active magnetorheological damper, in EERA DeepWind, Trondheim, Norway, p. 15 (2017) [4] Jensen, P. H. et al.: LCOE reduction for the next generation offshore wind turbines - Outcomes from the INN- WIND.EU project, October 2017 Figure 1: Model wind turbine of the Delft University of Technology in front of 3 m x 3 m active grid in the WindLab wind tunnel

– 163 – 164CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Integrated Research Programme on Wind Energy (IRPWIND)

Leibniz Universität Hannover (LUH) Project Description the static maximum tension bearing capac- Institute of Structural Analysis, ity. Some results of these tests are shown in Institute of Geotechnical Engineering, IRPWIND with more than ten WP and Fig. 2. Institute of Steel Construction, about 50 tasks covers a broad range of top- Institute for Risk and Reliability, ics in the wind energy sector. That is why In WP 7.4, the results of these tests were Institute of Electric Power Systems only some WP in which LUH was involved analyzed. First, existing theoretical methods for the determination of the maximum ten- Clemens Hübler, Tim Berthold, Tanja can be presented here. sion bearing capacity of statically, axially Grießmann, Andreas Ehrmann, Kirill A. Schmoor, Karsten Schröder, Matthias WP 7.2 is intended to improve and validate loaded piles (jacket piles) were validated or Bode the structural reliability of offshore wind rather their model error was determined us- turbine support structures. One task in this ing the data of the pile load tests. Second, Funding: European Union’s Seventh WP is the conduction of large-scale geo- the data of the dynamic tests was used to Programme for research, technological technical tests with model structures. This validate existing approaches for the lateral development and demonstration means that pile foundations were tested in soil stiffness (p-y curves), if these methods the large geotechnical test pit (10 m depth, are applied in a dynamic context. For the Ref.Nr. Grant agreement no 609795 14 m large and 9 m wide) in the new Test first aspect, one exemplary result is that the standard approach in the API standard [1] is Duration: 03/2014 – 02/2018 Center for Support Structures in Hannover (see Fig. 1). much less accurate than CPT-based methods (ICP, UWA, Fugro, and NGI), as large The large-scale tests are intended to deter- model errors occur (see Table 1). mine soil-structure interaction effects in order to support probabilistic calculations of Regarding the dynamic results, it can be the reliability of offshore wind turbine sup- seen that existing p-y approaches are not Introduction port structures. The calculations itself were always suitable for dynamic applications. performed in WP 7.4 of the IRPWIND pro- In Fig. 3, the deviations of measured (using IRPWIND is a large EU project in the wind ject. Six piles (lengths of 5-7m and diam- impact hammer tests of WP 7.2) and simu- energy sector with 24 international part- eters of around 0.3m) were tested statically lated eigenfrequencies (using different p-y ners. Its aim is to foster better integration and dynamically to determine, inter alia, curves) are shown. No theoretical of European wind energy research activities with the objective of accelerating the transition towards a low-carbon economy and of maintaining and increasing European Table 1: Derived model errors on basis the performed pile load tests in WP 7.2 competitiveness. In this context, IRPWIND includes several non-research aspects (work Pile load tests Related pile capacity Qcalculation/Qmeasurement packages WP 1-5) that are intended to improve the European collaboration. An ex- API ICP UWA Fugro NGI ample is the mobility call for experienced researchers (WP 5) to transfer knowledge. Pile 1 0.42 0.84 0.78 0.96 1.91 In research, IRPWIND focusses on three Pile 2 0.70 1.21 1.12 1.16 2.86 main aspects. The first one is the optimiza- Pile 3 0.47 1.00 0.93 1.23 2.09 tion of wind farms through the validation of Pile 4 0.45 0.85 0.79 0.89 1.83 integrated design models (WP 6). The sec- Pile 5 0.50 1.07 1.02 1.30 2.28 ond one is the reduction of the uncertainty Pile 6 0.52 0.98 0.91 1.02 2.10 in order to increase efficiency and reliability of future wind turbines (WP 7). The last one Mean model error 0.51 0.99 0.93 1.09 2.18 is the transformation of the energy supply Standard deviation 0.10 0.14 0.13 0.16 0.37 system (WP 8). LUH is involved in all three Coefficient of variation 0.19 0.14 0.14 0.15 0.17 research aspects.

– 164 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017165

Figure 2: Raw data of the load-displacement curves for all the tensile loading tests

Figure 1: Photo of the test pit and the installed piles

model can predict the correct eigenfrequencies for all tests. Although this can easily be explained, as p-y curves have originally Figure 3: Normalized first bending eigenfrequencies of the piles for been designed for static loads, it is an im- different embedded lengths calculated using different soil models [1-6] portant result, since p-y curves are widely utilized in transient, dynamic wind turbine simulations. References [3] Kallehave, D.; LeBlanc Thilsted, C.; Liingaard, M.: Modification of the API Summary [1] American Petroleum Institute, 2002. p-y formulation of initial stiffness of Recommended Practice for Planning, sand (2012) IRPWIND successfully demonstrated the Designing and Constructing Fixed [4] Wiemann, J.; Lesny, K.; Richwien, W.: potential of European collaborations for Offshore Platforms – Working Stress Evaluation of the Pile Diameter Effects the wind energy research. For example, the Design. Recommended Practice RP on Soil-Pile Stiffness (2004). 2A-WSD [5] Sørensen, S.: Soil-structure inter- geotechnical tests in Hannover were jointly [2] Sørensen, S. P. H.; Ibsen, L. B.; action for non-slender, large-diameter planned, designed, and conducted by LUH Augustesen, A. H.: Effects of diameter offshore monopiles. PhD Thesis, and the IWES Hannover and in alignment on initial stiffness of py curves for Aalborg University Denmark (2012) with the needs of Aalborg University. So, it large-diameter piles in sand. Numerical [6] Thieken, K.; Achmus, M.; Lemke, K.: was possible to use the same experiments Methods in Geotechnical Engineering A new static p-y-approach for piles for various subsequent numerical studies. (2010). with arbitrary dimensions in sand. This is just one example of successful Euro- Geotechnik 38 (2015). pean collaboration in IRPWIND.

– 165 – 166CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Extreme Ocean Gravity Waves: Understanding and Predicting Breathers with Wave Breaking and in Coastal Waters

Carl von Ossietzky Universität Oldenburg rogue waves is analyzed by a scale-depend- the according conditional probability is es- Institute for Physics ent stochastic approach, which interlinks pecially high. To quantitatively evaluate the Research Group: Turbulence, Wind Energy fluctuations of waves for different spacings. prediction quality of our stochastic recon- and Stochastics (TWiSt) With this approach we are able to determine struction, we apply in fig. 2 the receiver op- a stochastic cascade process, which pro- erating characteristic curve (ROC) [11, 12]. Ali Hadjihosseini, Christian Behnken, vides information of the general multipoint The ROC compares the rate of true predic- Pedro Lind, Matthias Wächter, statistics. Furthermore the evolution of sin- tions with the rate of false alarms for events Joachim Peinke gle trajectories in scale, which characterize of a given magnitude. The most quantitative Partners: wave height fluctuations in the surroundings index describing a ROC curve is the area Norbert Hoffmann, Technical University of a chosen location, can be determined, and under it, which is known as accuracy. In all Hamburg-Harburg, (Coordinator) time series of according wave heights can three cases in fig. 2 we have an accuracy of Efim Pelinovsky, Russian Academy of be reconstructed. greater than 80% which indicates that our Science, Nizhny Novgorod multi-point procedure is a proper method Nail Akhmediev, Australian National for time series reconstruction and can be University, Canberra Project Description used for short time prediction purposes. Funding: VolkswagenStiftung We use a stochastic multi-point approach Ref.Nr. 8482 which is explained and discussed in [9]. For Summary details we refer the reader to that publica- Duration: 09/2014 – 05/2019 tion. As a short summary, we describe the In this project we have presented a new ap- evolution of time series of wave heights proach for a comprehensive analysis of the h(t) as a stochastic cascade process which complexity of ocean wave dynamics. We evolves in the time scale τ=t* - t. Here, t* have been able to show for the first time denotes the latest time instant at which the that by our stochastic approach not only can Introduction sea level height h* = h(t*) is to be deter- the joint N-point statistics be grasped, but mined. It turns out that this cascade process also extreme events, rogue waves, can be Oceanic rogue waves are usually defined as presents Markov properties which allow for captured statistically. We have also shown extremely large waves that occur suddenly the derivation of a closed expression of the how for each instant in time the conditional and unexpectedly, even in situations where probability of a certain height h* to occur, probability of the next wave height can be the ocean appears relatively calm and quiet namely [9]. determined. As the height profile of waves [1]. While there are numerous reports from changes from moment to moment, also the sailors claiming to have observed a rogue probability of the next value of the wave wave in the open ocean, rogue waves are very rare, which makes researching or fore- height is changing dynamically. These casting difficult [2]. Because of their size changes may thus clearly give rise to meas- rogue waves can be extremely dangerous, ures indicating the risk of the appearance even to the large ocean liners, appearing of rogue waves ahead of their actual emer- in different forms of rare large-amplitude gence. Most interestingly, this was possible,

events [3,4]. As a prototypical example Here we use the short notation ζ j := h* - hj although in the measured data only one of extreme events emerging in a stochas- for height increments. Equation (1) espe- event of a rogue wave was recorded. From tic “background”, rogue waves have been cially allows for the reconstruction of wave our analysis of the occurrence probabilities investigated from various perspectives, time series, incorporating correct non-linear it becomes clear that the rogue wave for e.g., using tools from nonlinear waves and multi-point statistics. these wave conditions is an integral part of soliton theory [5–7]. the entire complex statistics. An example of such reconstruction is pre- In this project, based on data from the Sea of sented in fig. 1. Note the occurrence of an Japan and the North Sea, the occurrence of extreme wave height in a situation when

– 166 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017167

References

[1] Birkholz, S.; Brée, C.; Demircan, A.; Steinmeyer G.: Phys. Rev. Lett., 114 (2015) 213901 [2] Soto-Crespo, J.; Devine, N.; Akhmediev, N.: Phys. Rev. Lett., 116 (2016) 103901 [3] Kharif, C.; Pelinovsky, E.; Slunyaev, A.: Rogue Waves in the Ocean, in Springer Series in Advances in Geophysical and Environmental Mechanics and Mathematics (Springer, Berlin, Heidelberg) 2010 [4] Onorato, M.; Residori, S.; Bortolozzo, U.; Montina, A.; Arecchi, F.: Phys. Rep., 528 (2013) 47 [5] Remoissenet, M.; Waves Called Solitons: Concepts and Experiments (Springer) 1999 [6] Chabchoub, A.; Hoffmann, N. P.; Akhmediev, N.: Phys. Rev. Lett., 106 (2011) 204502 [7] Birkholz, S.; Brée, C.; Veselić, I.; Figure 1: Reconstructed time series (a) after equation (1). Two time windows are marked Demircan, A.; Steinmeyer, G.: Sci. by (b) and (c) for which the corresponding multi-conditioned PDFs are given in (d) and Rep., 6 (2016) 35207 (e). To show the changing volatility of the multi-conditioned PDFs (black), the unconditio- [8] Hadjihosseini, A.; Peinke, J.; nal PDF (red) estimations from all data are shown too. Note the obvious changes of the Hoffmann, N.: Stochastic analysis of likelihood of large wave amplitudes. ocean wave states with and without rogue waves, New J. Phys. 16, 053037 (2014) [9] Hadjihosseini, A.; Wächter, M.; Hoffmann, N.; Peinke, J.: Capturing rogue waves by multi-point statistics, New J. Phys. 18, 013017 (2016) [10] Hadjihoseini, A.; Lind, PG.; Mori, N.; Hoffmann, NP.; Peinke, J.: Rogue waves and entropy consumption, Europhys. Lett. 120, 30008 (2017) [11] Hanley, J. A.; McNeil, B. J.: The meaning and use of the area under a receiver operating characteristic (ROC) curve, Radiology 143, pp. 29- 36 (1982) [12] Kantz, H.; Altmann, E. G.; Hallerberg, S.; Holstein, D.; Riegert, A.: Dynamical interpretation of extreme events: Predictability and predictions Extreme Events, in: Nature and Society: The Frontiers Collection, eds. S Albeverio, V Jentsch and H Kantz, Springer, pp. 69-93 (2006) Figure 2: ROC curve for three different estimations of Pextreme , for hr = 5.2 m (solid black

line), hr = 3.5 m (dashed blue line) and hr = 2.5 m (dotted red line).

– 167 – 168CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

SAMS Save Automation of Maritime Systems

Carl von Ossietzky Universität Oldenburg, Introduction Project Description Computing Science, Business Engineering Shipping is always a joint undertaking of In order to facilitate research in the domain Axel Hahn humans and their technology. The increas- of integrated and save maritime supervi- ing usage of the oceans for transportation, sion, control and assistance technology, the Funding: Federal State of Lower Saxony fishing, exploration and even recreation in- PhD program "Save Automation of Mari- time Systems (SAMS)" was approved and Ref.Nr. 0325492B crease threats for humans, the environment and even the economy in case of disrupted initiated. SAMS is a joint PhD program of Duration: 10/2014 – 09/2018 supply chains. Consequently, the challenge the University of Oldenburg, the Jade Uni- faced by marine engineers and the scientific versity of Applied Science and the research community is designing save and environ- institute OFFIS. It is funded by the Federal ment friendly maritime systems. State of Lower Saxony, which provided

Figure 1: SAMS Kick-Off Meeting at the OFFIS in February 2015

– 168 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017169

funds for the granting of 15 "Georg-Chris- uate vocational training. Especially since Summary toph-Lichtenberg Scholarships" for the pro- maritime education in Germany is not car- gram in the funding period from 1 October ried out at university level, this cooperation SAMS is a joint PhD program for safety of 2014 to 30 September 2018. From approxi- makes it possible to obtain scientific doctor- maritime systems. SAMS offers 15 three mately 200 applicants, 15 were selected on ates for nautical science as well. year scholarships within the period between the basis of their application documents and 2014 until 2018. The objective of the pro- subsequent interviews and the scholarships gram is the joint transfer of research from were awarded. design and analysis technologies of safety critical systems to maritime sociotechni- The cooperation between University of cal systems with their specific dynamic, Oldenburg, the Jade University of Applied geographical and hydromechanical char- Science and the research institute OFFIS acteristics, organizational forms, technolo- enables the combination of basic research gies, environmental conditions and safety and application perspectives with postgrad- requirements.

Figure 2a, 2b: First SAMS Autumn School at the University of Oldenburg in October 2015

– 169 – 170CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Ventus Efficiens – AP 3

Leibniz Universität Hannover Introduction on a numerical model of wind turbine and grid and the Institute for Drive Systems Institute for Drive Systems and Power Investigations of the cumulative output and Power Electronics (IAL) researches Electronics coupling between the converter’s control Institute of Electric Power Systems power of a wind park showed great power fluctuations, significantly outside the ex- systems. Farshid Goudarzi, Lutz Hofmann, pected variance [1]. These results lead to the Daniel Herwig, Axel Mertens assumption, that the fluctuations of the single wind turbine’s output powers within the Project Description Funding: Lower Saxony Ministry for park are correlated to each other. The root Science and Culture (MWK) cause of this phenomenon is not yet known. I) Hardware-in-the-Loop (IAL) The topic is investigated in cooperation with At the IAL a test bench for detailed investi- Ref.Nr. VWZN3024 the University of Oldenburg, work group gations of the converters controls is imple- Turbulence, Wind Energy and Stochastics mented. This Hardware-in-the-Loop system Duration: 12/2014 – 12/2019 (TWiSt). allows the real-time emulation of the converter controls in Matlab/Simlink. Aerody- Within the project, the problem is ap- namic coupling effects have to be simulated proached from three different points of in condensed models. view, with different couple-mechanisms in focus. The TWiSt investigates coupling- The test bench enables the modelling of real mechanics between wind turbines based grid configurations for a scaled output pow- on aerodynamics. The Institute of Electric er at different voltage levels. This includes Power Systems (IfES) addresses coupling the collection grid within a wind farm. For between the wind farm and the grid based this purpose, an overall network consisting

Figure 1: Example of a grid configuration with three wind turbines, a grid emulator and passive loads

– 170 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017171

Figure 2: Microgrid test bench

of several converters in the range of some The converters are able to create Matlab/ tion generator) based wind turbines is ten kilowatts has been set up in the labora- Simulink models in real time mode. Apart simulated in time-domain in Matlab envi- tory. The aim is to analyse the interactions from the electrical system, modelling of fur- ronment. Fig. 3 illustrates the structure of between the individual power electronic ther components in the power flow chain is overall control system of the modelled wind systems as well as between the passive grid also possible. For example, the mechanical turbines which consists the following con- elements. Moreover, the test bench can be drive train of a wind turbine can be emu- trol concepts: used to investigate alternative grid configu- lated and analysed in a simplified form for a rations and control methods for future dis- known wind flow at the rotor blades. - Pitch angle and speed control tributed grids. An overview of the test bench - Machine-side converter control (the inde- is shown in Fig. 1. A decentralized, synchronized measure- pendent control of the generator active and ment system captures the measured data reactive power) The passive components between the sys- with a sample rate of 10 kHz, the data be- - Line-side converter control (the DC volt- tems are simulated by modular power line ing merged and depicted in a virtual system age control and the reactive current sup- models. The initial configuration consists of control centre. Three-phase current and port) ten line models (NA2XSY 3x1x120mm²) voltage measurement is done for each meas- with an equivalent length of 2.1 km and 20 uring point. In this model, the dynamic behavior of the kV scaled output power. The models are generators and the control systems of the also applicable to a low-voltage grid using As a next step, the test bench will be ex- converters is represented through differen- NAYY 4x150mm² models with an equiva- tended to include a 200 kVA grid-emulator. tial equation systems. According to the time lent length of 300 m. Future test bench The system is driven by a real-time proces- frame of interest (seconds range) for the enhancements will also include linear and sor system, capable of emulating further proposed investigation the passive electri- non-linear loads. grid components behind the PCC in Matlab/ cal network is modelled by its steady-state Simulink. model which are represented through a set The point of common coupling (classical of algebraic equations. Consequently, the public mains) is simulated by a converter II) Numerical modeling (IfES) fundamental oscillation of voltages and system emulating the grid. This enables a In order to analyse the power fluctuations currents is considered through their RMS decoupled setting of the grid parameters, mentioned above, a dynamic wind farm values (Root Mean Square). The resulting e.g. frequency, harmonics or asymmetries. model contains DFIG (Doubly-fed induc- overall system equations can be solved for

– 171 – 172CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Figure 3: Structure of overall control system

each time step as an algebraic-differential equation system using an appropriate numerical method.

Summary

Within a wind park, a strong coupling of the wind turbines output power has been observed [1], that cannot be explained with coupling mechanisms known today. As the root cause is unknown, the effect is investigated with three possible origins of the coupling in focus – Aerodynamics, the power References grid and the wind turbines control systems. [1] Anvari, M.; Wächter, M.; Peinke, Each study has a very detailed model of the J.: Phase locking of wind turbines aspect in focus and simplified models of the leads to intermittent power other two domains. The goal is to narrow production, EPL (Europhysics Letters), down the point of origin of the observed Volume 116, Number 6 (2016), doi: new type of coupling between wind tur- 10.1209/0295-5075/116/60009 bines.

– 172 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017173

Ventus Efficiens – Collaborate Research to increase the Efficiency of Wind Turbines in the Energy System

Carl von Ossietzky Universität Oldenburg Project Description fig. 1). Customized control protocols pre- Institute of Physics describe the motion of each motor and with Research Groups: Subproject I: Increase of the efficiency it the distribution of the blockage over the Energy Meteorology (EnMet), of energy conversion in wind energy cross-sectional area of the grid over time. Wind Energy Systems (WESys), systems This procedure allows for the generation of Turbulence, Wind Energy and Stochastics (TWiSt) turbulent wind fields behind the grid that Work package 1: Generation and feature e. g. vertical velocity profiles, dif- Detlev Heinemann, Gerald Steinfeld, modeling of atmospheric turbulence ferent turbulence intensities, wind gusts Jens Tambke, Lüder von Bremen, This WP aims at generating flow condi- and defined higher order statistics. To in- Björn Witha, Martin Kühn, Martin Kraft, tions that feature atmospheric like char- crease the basic understanding of the gen- Marijn Van Dooren, Matthias Wächter, acteristics over a long test section and a eral correlation between motions of the Luis Vera-Tudela, Agnieszka Hölling, large domain, respectively. This should shafts and the flow situations generated Lars Kröger, Jannik Schottler, be done for experimental investigations behind the grid, tests with other smaller Sebastian Ehrich, Lisa Rademacher, in the wind tunnel as well as for simula- active grids and reduced complexity have Michael Hölling, Joachim Peinke, tions. For the latter the Continuous Time been performed. The results show that the David Bastine Random Walk (CTRW) model has been basic approach for designing control pro- used in order to generate inflow conditions Funding: Ministry for Science and Culture tocols can be applied to all active grids in Lower Saxony (MWK) that show intermittency on various scales. independent of their size and number of Currently the work focuses on stabilizing motors, respectively. First tests with the Ref.Nr. ZN3024 the created properties and minimizing the big grid performed in late 2017 confirm decay of the generated turbulence. Experi- these results. The dimensions of the new Duration: 01/2015 – 12/2019 mentally, a so-called active grid has been wind tunnel also allow for using LiDAR designed and constructed for the big wind wind scanner systems in order to measure tunnel with an outlet of 3m by 3m in the the velocity with up to 400 points per sec- new WindLab building. This active grid ond. First test have already been performed Introduction consists of 80 motors that are connected in the wind tunnel in Milano that show the to shafts with flaps mounted to them (see proof of this concept based on the results. The planned research activities within the framework of this project concentrate on increasing efficiency in wind energy in general and its integration into the energy supply structure in particular. The subprojects thus deal with increasing efficiency along the entire impact chain, i. e. starting with energy conversion in wind energy systems (subproject I), increasing the efficiency of load-bearing structures (subproject II) as well as drive trains and grid connection (subproject III) up to the validation of efficiency increase in the overall system of research wind turbines (subproject IV). ForWind participates with its wide variety of expertise in the subprojects I, III and IV, where each of the subprojects is again divided into work packages (WPs). For the sake of visibility the contribution of ForWind Figure 1: Left: Technical drawing of the big active grid with 80 motors. The blue field Oldenburg to each WP will be described in- illustrates the generated turbulent wind. Right: Constructed active grid mounted to the dividually in this report. outlet of the wind tunnel with an outlet of 3m by 3m.

– 173 – 174CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Work package 2: Interaction of turbulence and wind energy systems Experimentally, the topic of this WP has been addressed with the help of wind turbine models. Two types of turbine models of different sizes have been developed, the MoWiTO 0.6 with a rotor diameter of about 0.58m and the MoWiTO 1.8 with a rotor diameter of 1.8m (see fig. 2). MoWi- TO stands for Model Wind Turbine Olden- burg. Due to the smaller size of the MoW- iTO 0.6, wind farms consisting of 2 and more models can be arranged in the wind tunnel in Oldenburg. Within this project, there will be a total of eleven MoWiTOs 0.6 available for experiments, all of which feature pitch and torque control [1]. First experiments focused on measurements in the wake of such models e. g. for differ- Figure 2: Left: Model wind turbine Oldenburg (MoWiTO 0.6) with a diameter of 0.58m. ent inflows [2] and in yaw conditions, re- Right: Model wind turbine Oldenburg (MoWiTO 1.8) with a diameter of 1.8m. Pictures are spectively [3]. The impact on the dynamic not in scale. response of the models was investigated by creating inflow conditions with Gauss- ian and non-Gaussian statistics [4]. All re- situations, the large-eddy simulation model el. This led to an improved reproduction of sults show, that even though the Reynolds should be driven by data from a mesoscale actually observed wind direction changes. numbers are not comparable to real wind simulation model. In the framework of the Moreover, aiming at a verification of wind turbines, general effects can be reproduced ventus efficiens project the results of the farm simulations with the large-eddy simu- in the wind tunnel under controlled and re- coupling between mesoscale models and lation model PALM, simulations of the producible boundary conditions. The larger the large-eddy simulation model PALM flow conditions within the German off- MoWiTO 1.8 on the other hand is designed was improved by taking information on the shore wind farm Riffgat have been carried in such a way, that the local Reynolds num- meso-scale advection of momentum into out. A simple actuator disc model has been ber with respect to the chord of the rotor account in the large-eddy simulation mod- used to parameterize the wind turbine ef blades is large enough to be aerodynamically comparable to real wind turbines. Therefore, the MoWiTO 1.8 can be used to investigate the flow around the blades or even near wake characteristics, where the aerodynamic has a significant impact on the results. In order to be able to identify highly dynamic effects, which are due to the interaction of turbulent wind fields with the turbine models, stochastic methods such as Langevin processes are used and further developed within this project.

Additionally to the experimental investigations, one objective in this work package is to gain a better knowledge on the superposition of wind turbine wakes in wind farms by the use of simulation tools. The methodology chosen to achieve this objective was turbulence-resolving simulations Figure 3: Line-of-sight-velocities at different positions within the wind farm Riffgat. The with a large-eddy simulation model. In black line shows results obtained from large-eddy simulations with PALM, while the co- order to allow for the reproduction of real lored lines show actual LiDAR measurements.

– 174 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017175

fect in these simulations. The wind farm verification of the results of the wind farm Power Systems from ForWind Hannover Riffgat that consists out of thirty wind tur- parameterization in WRF, also large-eddy possible coupling and synchronization ef- bines allows for the analysis of up to ten- simulations of the flow conditions around fects within wind farms, respectively. The fold wakes. First comparisons with data the cluster consisting of the wind farms question at stake is whether such effects from LIDAR measurements in the farm in- Riffgrund and alpha ventus using the mod- result from aerodynamic interaction e. g. creased the confidence in the results from el PALM have been carried out. through wakes or from electrodynamic in- the large-eddy simulations (see fig. 3). teraction over the local electrical network. Work package 4: Control strategies to ForWind Oldenburg will address this prob- Work package 3: Meteorological interac- compensate effects due to turbulence lem by building up a wind farm of up to tion of wind energy systems and feed-in From the experimental point of view, this nine MoWiTOs 0.6 in the wind tunnel. This work package aims at the promotion WP just started since the implementation The challenge up to now is the electrical of ensemble forecasts as standard tool for of advanced controller strategies comes af- connection between these models since the wind power forecast. Within this project ter the construction and characterization of the model turbines are equipped with DC it could be shown that different ensemble the model turbines in use. Like already re- motors and therefore show significantly systems need different calibration meth- ported in WP2 of this subproject, the mod- different behavior compared to double fed ods. Generally, calibration can improve the els have been built and it has been shown asynchronous generators used in most real quality of the ensemble forecast consider- that general, turbulence induced effects can modern wind turbines. ably [5]. Moreover, it was shown that the be observed with these models. Next steps so-called analog-ensemble method, which will focus on the further development of Subproject 4: Validation of efficiency is based on only one deterministic forecast the controller to e. g. mitigate stochastic increases in the overall system of run and historical data of observations and loads, minimize wake effects and optimize research wind turbines forecast runs, yield similar or even better wind farm output of two and more tur- forecast skills than conventional ensemble bines. Besides the experimental approach, In this subproject ForWind Oldenburg methods [6]. one objective of this work package is a bet- and Hannover are working together on ter integration of meteorological methods the preparation of future measurement Another objective of this work package and information in methods for the control campaigns at the planned research wind was to gain a better understanding of the of wind turbines and wind farms. For that energy plants (FWEA) to record the dy- reasons for the occurrence of wind fluctua- purpose a number of setups for the large- namic behaviour of the FWEA. In order tions. Data from a measurement campaign eddy simulation model PALM resulting to be able to evaluate influences from soil- and mesoscale meteorological simulations in a variety of different atmospheric flows structure interaction as well as interactions were used to investigate the impact of the have been developed in the framework of between the atmospheric flow and the op- atmospheric stability on wind fluctuations the ventus efficiens project. Moreover, the erating behaviour of the FWT, ForWind [7]. Moreover, the analysis of data from a wind turbine model PALM has been fur- institutes from the fields of wind physics large offshore wind farm in the North Sea ther developed by adding a yaw, pitch and (Oldenburg), civil engineering, electrical showed that small wind direction changes torque controller to the ADMR wind tur- engineering and mechanical engineering come along with large power fluctuations bine model, so that these parameters now (Hanover) are involved. The close inter- for certain wind directions. Data from Li- can be used to develop new control strate- disciplinary cooperation of the scientists DAR measurements were used to inves- gies for wind turbines and wind farms. The leads to an improved understanding of tigate power fluctuations of single wind large-eddy simulation model PALM can the system and enables the use of research turbines. thus be used as a numerical test environ- synergies. Subproject IV is divided into ment for the development of new control four work packages and is managed by the Finally, the work package aims also at strategies. research group WESys (Oldenburg). At mesoscale simulations of the wind condi- Leibniz University Hannover, the Institute tions around wind farms. Here, the region- Subproject 3: Increased efficiency of for Turbomachinery and Fluid Dynamics al impact of the planned exploitation of the the electric drive train and the grid coordinates the contributions there. wind resources in the German Bight has connection been studied by using the standard wind Work package 1: Coordination of farm parameterization in the mesoscale Work package 2: Investigation of electro- measurement campaigns and technology model WRF. The results showed that clus- fluid-mechanical coupling mechanisms in With the internal and external partners ters of offshore wind farms can have wakes wind farms involved, the measurement targets, the that reach several tens of kilometers down- In this WP ForWind Oldenburg wants to measurement campaigns and the necessary stream of the cluster. The wake direction investigate in close cooperation with the tools were coordinated, and synergies were is heavily influenced by the large-scale Institute for Drive Systems and Power developed. All measurement campaigns atmospheric flow. In order to prepare a Electronics and the Institute of Electric were examined regarding permissible

– 175 – 176CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

and intended operation of the plant. New teraction between the properties of the am- Low-level jet events are expected to come measurement technology and instrumenta- bient atmosphere and the aeroelastic loads along with large loads. Especially such tion concepts were developed for the wind felt by a wind turbine. For that purpose the situations shall be investigated with the turbine, which go far beyond standard in- coupling between the large-eddy simula- PALM-FAST-coupling in the framework strumentation. For testing and validation of tion model PALM and the aeroelastic code of the ventus efficiens project. PALM shall instrumentation several test sites have been FAST is used. The computational require- be driven by data from the mesoscale mod- developed. ments for coupled runs could be reduced el WRF in order to produce realistic low- considerably by implementing an actuator level jet events for the analysis. In a first Work package 2: Wind fields, aerodyna- sector method in PALM. This means that step using data from a measurement cam- mics and aeroelasticity in the modified version of the coupling paign at the FINO1 platform it has been The overall objective of this work package PALM has a much larger time step than investigated whether the mesoscale model is to gain a better understanding of the in- FAST. WRF can reproduce observed low-level

Figure 4: Wind speeds measured during a low-level jet event by a LiDAR system on the offshore platform FINO1 (top row). Wind speeds obtained from a simulation with the meso-scale model WRF for the same period of time and position as those of the LiDAR measurements (bottom row).

– 176 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017177

jet events. The result was that, generally, Work package 4: Overall system, Summary WRF can reproduce low-level jet events, simulation although the intensity, the position and In this work package models for describ- Within the ventus efficiens project -For the timing of the low-level jet event can ing complex wind turbine system are re- Wind Oldenburg participates with simu- slightly differ between the simulation and fined. Different approaches are pursued, lations, stochastic analysis as well as ex- the observation (see fig.4). which focus on the representation of the perimental investigation. In the field of vibrational behaviour with a numerical experiments much work has been done to A continuous-wave short-range LiDAR model, the monitoring of the dynamic per- expand the infrastructure with additional with high temporal and spatial resolution formance characteristics with a stochastic equipment and, more importantly, to gain a recorded measurements of the rotor inflow model and the monitoring of the fatigue detailed understanding of the possibilities of an eno114 turbine. For postprocessing load based on the transformation from opening up by characterizing their perfor- lower order modelling of turbine inflow standard operating data. All three priorities mance in detail. The different wind turbine has been developed [8]. are intended to support the validation of models in combination with the ability existing approaches on a real-world plant to generate user-defined turbulent inflow and can be understood as trailblazers for conditions with an active grid are the foun- the following research projects [9-12]. dation for many of the upcoming tasks in this project. A combination of this to real world behaviour is given via the measurement and analysis of real megawatt wind turbines in free field.

References

[1] Schottler, J., Hölling, A., Peinke, J., [5] Junk, C., Delle Monache, L., EERA DeepWind – 14th Deep Sea Hölling, M.: Design and implementa- Allessandrini, G., Cervone, G., Offshore Wind R&D Conference, tion of a controllable model wind tur- von Bremen, L.: Predictor-weighting Trondheim, Norway [Poster & Energy bine for experimental studies, Journal strategies for probabilistic wind power Procedia] of Physics: Conference Series, Vol. 753, forecasting with an analog ensemble, [10] Vera-Tudela L., Kraft M., Kühn M.: No.7, 2016 Meteorol. Z., 24, pp. 361-379, 2015 How do missing data affect your data- [2] Neunaber, I., Schottler, J., [6] Späth, S., von Bremen, L., Junk, C., driven model? Monitoring fatigue loads Peinke, J., Hölling, M.: Comparison of Heinemann, D.: Time-consistent calibra- in wind turbines with SCADA data. the development of wind turbine wake tion of short-term regional wind power Offshore Wind Energy 2017, London, UK under different inflow conditions, ensemble forecasts, Meteorol. Z., 24, pp. [Poster] Progress in turbulence VII – Springer 381-392, 2015 [11] Pedro G. Lind, Vera-Tudela, L., Proceedings in Physics 196, pp. 177-182 [7] Mehrens, A., von Bremen, L.: Wächter, M., Kühn, M., Peinke, J.: (2016) Comparison of methods for the Normal behaviour models for wind [3] Schottler, J., Hölling, A., Peinke, identification of mesoscale wind speed turbine vibrations: An alternative ap- J., Hölling, M.: Wind tunnel tests on fluctuations, Meteorol. Z., 26, proach, Energies 10, 1944, 2017 controllable model wind turbines in pp.333-342, 2017 [12] Bastine, D., With a, B., Wächter, M. yaw, 34th Wind Energy Symposium, [8] Kidambi Sekar A.P., Kühn, M.: Lower and Peinke, J.: Towards a Simplified Dy- Vol. 1523, 2016 Order Modeling of Wind Turbine inflow namic Wake Model Using POD Analysis, [4] Schottler, J., Reinke, N., Hölling, with Short Range Lidars, WESC 2017, Energies 8, pp. 895-920 (2015) A., Whale, J., Peinke, J., Hölling, M.: Lyngby, Denmark [Talk] On the impact of non-Gaussian wind [9] Seifert J.K., Vera-Tudela L., statistics on wind turbines – an experi- Kühn M.: Training requirements of a neu- mental approach, Wind Energ. Sci., 2, ral network used for fatigue load pp. 1-13, 2017 estimation of offshore wind turbines,

– 177 – 178CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

German Research Facility for Wind Energy (DFWind)

Leibniz Universität Hannover 27.8 % (as of 2014, BMWi 2014). In order MW) including an extensive recording of Institute of Structural Analysis to make a contribution to reducing green- the environmental conditions. The work in house gases and preventing further global this subproject should enable the research Raimund Rolfes, Christian Gebhardt, warming in line with climate policy goals, platform to test novel components and con- Tanja Grießmann, Dorian Pache, the Federal Government plans to increase cepts in the overall system on the one hand Benedikt Hofmeister, Stefan Wernitz the share of renewable energy sources in and to derive and improve validated indus- Funding: Federal Ministry for the electricity generation in Germany to 55% trial models for the description of large wind Environment, Nature Conservation, to 60% and 80% respectively by 2035 and turbines on the other hand. Building and Nuclear Safety (BMU) 2050 respectively. Wind energy, with a total share of 34.8 % of the electricity supply Ref.Nr. 0325492B from renewable energies, already plays an Project Description outstanding role today and is classified as a Duration: 03/2015 – 08/2017 leader in the medium to long term. The aim of the project is to create the basis for a wind energy research and develop- In order to ensure the continued competi- ment platform with which numerous topics tiveness of the European and German wind for their use on land as well as at sea can be energy industries and to establish the use of dealt with extensively and in a hitherto unat- Introduction wind energy as the cheapest form of energy, tained quality along the entire impact chain. supplementary research and development The focus will be on a holistic approach Due to the climate problem as well as envi- efforts in the field of wind energy are neces- in which the research focus will be on the ronmental and resource issues, a conversion sary - as in other branches of industry (e.g. interaction of the subsystems in the overall of energy systems to systems supplied sus- automotive, aviation). This will enable wind wind turbine system, also taking into ac- tainably with renewable energies is being energy to be expanded in line with plans to count the mutual influence of two separate pushed forward worldwide. increase the share of renewable energies in wind turbines up to the effect on the inter- Germany. connected grid. In addition to investigating Wind energy plays a key role in this process purely physical and technical aspects, the in Germany, Europe and numerous other In 2012, the first step towards the creation structure will also be able to contribute to countries on the American and Asian con- of a national research platform was taken answering socio-economic and ecological tinents. According to the International Ener- with the start of the project "PROWind - questions relating to wind energy. gy Agency (IEA), a total potential of 2,300 Setting up a research wind farm" initiated GW should be available by 2050, which by DLR, which will allow and advance a DFWind is divided into two phases for re- corresponds to a reduction in CO2 emissions wide range of applied and basic research search technical upgrading. Phase 1 pre- of 4.8 Gt per year. In recent years, the in- and developments in the wind energy sector. sented in this report serves to prepare the stalled capacity of wind turbines (WT) has The construction and research operation of installation and instrumentation of the increased significantly - from 17.4 GW in the platform is being promoted by DLR in research wind turbines as well as the first 2000 to 369.6 GW in 2014 (GWEC). The close cooperation with ForWind - Center for meteorological measurements at the undis- German wind turbine manufacturers oper- Wind Energy Research of the Universities turbed site. This includes the development ating in the quality segment as well as the of Oldenburg, Hannover and Bremen and of various additional instruments in order associated suppliers and engineering ser- the Fraunhofer Institute for Wind Energy to ensure rapid progress in the scientific up- vice providers are among the world's lead- Systems within the German Research Alli- grading of the research wind turbines after ing companies both in terms of technology ance for Wind Energy (FVWE). the installation of the turbines in phase 2. It standard and market share. According to is planned that this second phase will begin Handelsblatt, five of the 15 largest manu- The project German Research Facility for after conclusion of the contractual agree- facturers come from Europe, including all Wind Energy (DFWind) aims to prepare ment with the turbine manufacturer with German manufacturers. and implement a worldwide unique research an overlap of several months compared to infrastructure. This consists of two exten- phase 1 described here. In Germany, the share of renewable ener- sively instrumented research wind turbines gies for gross power generation is already of the current multi-MW class (2.5 to 3.5

– 178 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017179

Figure 1: DFWind research wind farm consisting of two wind turbines of the current multi-MW class and meteorological measurement masts

Global Objectives of DFWind of the research platform, consisting of 4. Networking and Transfer The overarching project objective is to several wind turbines and equipment for The cooperation of the FVWE part- answer research questions to promote the the measurement of wind fields, noise ners bundles the different competences expansion and reliable operation of wind and environmental conditions, aims at and focuses of a large research facility energy against the background of economic the elimination of considerable gaps in (DLR), a university network (ForWind) efficiency, security of supply and accept- knowledge in scientific questions. as well as an application and service ori- ance by the population, or to deal with ented research institute (IWES) and thus questions that were previously difficult or 3. Use of Synergy Effects in Interdiscipli- already represents an innovative coop- impossible to answer. This objective can be nary Research eration format. On this basis, an intensive divided into four categories: The interdisciplinary exchange is strong- networking of teaching, basic and applied ly promoted and the prospect of a suc- research, services and innovation man- 1. Economy and Industry cessful economic evaluation of results is agement takes place on a national and in- The research initiative will enable the further improved by the work of the pro- ternational level. The DFWind platform demonstration and testing of pre-com- ject partners with different technical ori- will stimulate close cooperation with petitive concepts, the implementation of entations in the research network on the other partners from science and industry, which is currently still associated with an jointly operated platform. The creation thus enable the establishment of new and unacceptably high economic risk for in- of an extensive measurement and model improved products and services in com- dustry in commercially used facilities. database will enable the further develop- petition, and close exploitation gaps in ment of methods and models and their line with the German government's high- 2. Answering Scientific Questions validation in a holistic context. tech strategy. The planned extensive instrumentation

– 179 – 180CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Scientific and Technical Objectives of the transient pressure fluctuations on the guarantees energy and impulse conserva- DFWind Phase 1 tower; sensors for recording the deformation. These properties are an important in- In accordance with the energy conversion tion of the blade adjustment mechanisms; novation for predicting the behavior of the chain and the technical design of wind tur- processing of measurement data synchro- complete wind turbines during operation bines, the partners identified various work- nous to the rotor position (WP 3). close to their operating limits in view of the ing packages (WP) for phase 1 of DFWind, ever larger, heavier and more flexible rotor which is presented more detailed here: - Preparation of the instrumentation of blades and tower structures of future wind the foundation structure (for foundation turbines. - Elaboration of the concrete research in- movements and soil structure interaction); strumentation of the wind turbines to- preparation of the measuring technique With regard to the overall project, the de- gether with the manufacturer towards the for monitoring the grouted joint between velopment should be carried out with design end of the project period (improved basic foundation and tower structure as well as data of the turbines in the DFWind/PROW- instrumentation, instrumentation for cer- for measuring the anchor forces; prepara- ind measuring field. As data for these- tur tification, line-guided measurement of tion of the instrumentation for recording bines is not available until the end of the electrical quantities of the line layout as the bolt forces at highly loaded locations; project phase 1, the data set of the NREL well as power electronics, improved elec- preparation of a measuring set-up and a 5MW reference turbine [1], which is widely tromagnetic compatibility and lightning concept for automated data processing used in the field of wind energy, was used protection; Specification of the experi- and identification of dynamic parameters instead. The parameterization of the models mental turbine control; development and (modal parameters, damage parameters) of was coordinated in such a way that an up- construction of a prototype of the data the foundation structure within the frame- date with the turbines actually installed in acquisition and data management system; work of long-term monitoring (WP 4). the measuring field is easy to implement. Development of parameterized numerical For this purpose, a basic support of the input component models as well as overall sys- In addition to planning the instrumenta- format of FAST [2] was implemented. tem models, which will later be validated tion of the wind turbines and the site, it is in phase 2 with the acquired measurement planned to procure and install initial instru- The developed models are mechanically data on rotor blade, drive train and sup- mentation in the fields of meteorology and simplified because comparatively long time port structure (WP 1). acoustics and in the fields of load-bearing periods have to be simulated efficiently in structure and geotechnics. The reason for the overall simulation. The models for rotor - Conception, procurement, coordination that is to record the yet undeveloped site in blade and tower described here are therefore and testing of the measuring instruments advance and thus prepare it for subsequent based on a beam theory, which is state of the for far-field measurements of wind,- tur research projects in science and industry. art in research and industry [3]. The models bulence, temperature and humidity, both for foundation, drive train, nacelle and hub for low-frequency profile and wind field Some Scientific Results in DFWind are based on condensed multi-body interac- measurements and for high-frequency Phase 1 tions. point measurements on meteorological Component Models measuring masts; planning, procurement, In this task models for the structural me- The geometrically exact beam theory has coordination and testing of the acoustic chanical simulation of wind turbine compo- already been successfully used in [4] for the measuring technology (WP 2). nents were generated. The theoretical basis simulation of rotor blades of wind turbines is a theory for flexible multibody mechan- with an angle-based formulation. For this - Preparation of a measurement system ics, which enables a strong coupling for task, a formulation of the geometrically ex- for monitoring the manufacture of rotor structural interactions and at the same time act beam theory within the framework of a blades, in particular with regard to overall structure tests; development of structural health monitoring systems (SHM) for damage detection on rotor blades; Prepa- ration of the aerodynamic measurement of rotor blades (pressure measurement technology, electrical and pneumatic connection concepts, protection against environmental influences, development of a mounting, mounting design for 5-hole and LIDIC probes as well as microphones on the rotor blade); development of a measurement technology for recording Figure 2: Illustration of a blade model with directors

– 180 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017181

director-based description of the geometry The overall model is to be validated with the Since no model data of the turbines to be was used. aid of measurement data from the DFWind installed in the research wind farm are research wind turbines. For the validation, available yet, the models were parameter- This formulation is particularly suitable in high-resolution meteorological measure- ized with the data of the NREL 5MW refer- the context of large rotations as they occur ment data are available in DFWind phase 2, ence turbine [1]. This data set enables the in wind energy rotors. A total Lagrangian so that the aerodynamic load of the turbine construction and calculation of a realistic representation guarantees path independ- is known precisely. High-resolution data model and the subsequent verification of the ence in the transient simulation. The time from the turbine control system (SCADA implementation. integration method used has the property data) enable the modelling of the aerody- of impulse and energy conservation. This namic and electrical control variables. The Design data of the turbines in the DFWind means that in this formulation no energy extensive instrumentation with accelerom- test field can be easily entered into the pa- disappears due to the numerical method and eters and strain gauges to determine the rameterized models in DFWind phase 2 as rigid body movements are preserved. dynamic structural response of the wind soon as they become available. Based on the turbine thus enables the comparison of real data set of the NREL 5MW reference tur- The aforementioned techniques, which are operation and aeroelastic simulation. A vali- bine, a large number of publications exist, implemented in an in-house simulation dation based on a direct comparison of the which can be used to verify the simulation framework of the Institute of Structural time data is aimed at, supplemented by sta- code. Analysis (ISD, FW-Hannover), result in tistical evaluations and analysis of the struc- advantages in the calculation accuracy with tural eigenmodes and eigenfrequencies. Operational Modal Analysis long time scales and large changes of posi- It is of great need to identify modal proper- tion of the structural components, as they The verification of the overall parameter- ties like natural frequencies, mode shapes occur with wind turbines. ized model was carried out with respect to and damping values to understand the [2]. Further, compatibility to the input data structural dynamical behavior of WTs. In Within the scope of this task, the multibody format of the FAST turbine simulation code phase 2 of DFWind those features are to be model and the servo components were kin- was established. This enables a model im- identified and monitored for the WT sup- ematically adapted to the formulation of the port from this code, which is widely used port structure and rotor blades, respectively. beam elements, so that a consistent, flexible in research. For the calculation of the struc- These parameters form the basis for the multibody model results when joining the tural dynamic response in the range of the adaptation of numerical simulation models components. nominal speed, in addition to the mechani- and at the same time represent an important cal design, the exact representation of the component of structural health monitoring. Overall System Models system controller is decisive. The controller It is well known that modal parameters vary A parameterized overall model for a wind from [1] can be used for verification. as a function of changing environmental turbine was developed on the basis of a consistent flexible multibody approach. The developed overall model forms the basis for the transient simulation of wind turbines in an aero-servo-elastic context. On the basis of the simulation results, a cost-effective and efficient dimensioning of the mechanical plant components can be carried out.

The component models developed in the previously mentioned task are combined in the overall model. The structural components are linked according to the kinematics of wind turbines by means of constraints. In addition, electrical components were developed for the overall model. A generator model provides realistic dissipation of the mechanical shaft power, a turbine controller drives the turbine to the nominal operating point and controls the angle of attack of the rotor blades. Figure 3: First mode shape of support structure (left) and first mode shape of rotor blades (right)

– 181 – 182CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

and operational conditions (EOC). How- dure is based on the assumption that the lo- ever, varying modal parameters within the cal structural changes solely slightly effect same operating conditions can be damage those properties. indicators. In DFWind operational modal analysis (OMA) is carried out using acceleration measurements. Therefore, preliminary Structural Health Monitoring modal testing was performed on a swept The condition monitoring of structures can pre-stressed concrete support structure in a be divided into four or five general phases laboratory setup. [7], [8]: 1. damage detection, 2. damage localization, 3. damage classification, 4. dam- The tower consists of twelve ring-shaped age quantification and 5. lifetime prognosis. reinforced high-strength concrete segments Purely measurement data and vibration- stacked on top of each other, each 5 cm based approaches that do not use structural thick, 60 cm in outer diameter and 0.5 m models (e.g. FE models) can only cover high. The total height is approximately 6.67 the first two phases of SHM. The decision m together with the tensioning structure at whether and where the monitored structure the top of the tower. The tower has a cylin- is damaged is usually made by evaluating drical structure with a cylindrical profile. damage sensitive features, also called con- The stability of the tower is provided by dition parameters (CP) [9]. All quantities a pre-stressed single tendon, which is an- that can be derived directly (e.g. maximum chored in the foundation and runs centrally values) or indirectly (e.g. natural frequen- through the tower up to a 460 kg head con- cies) from the measurement data are de- struction. scribed as such. Absolute and relative CPs exist. The former can be determined by As a system identification technique the looking at a single data set. These include Covariance-Driven Stochastic Subspace natural frequencies or damping values. Rel- Identification (SSI-Cov) was considered ative CPs in turn result from the application [5]. The resulting eigensolutions were Figure 4: Pre-stressed concrete support of a comparison metric to two or more data evaluated or with other words identified structure equipped with accelerometers in series. The feature tested describes the aver- as modal properties using the triangula- a laboratory setup age power of a synthetic difference pro tion based extraction of modal parameters method (TEMP) [6]. Instead of simulating the change of modal parameters due to environmental variability, different damage scenarios were considered. Those were realized by performing saw cuts between the concrete segments at different positions and depth. In a monitoring scenario, the number of identified modes varies across the data sets. This makes it difficult to monitor the modal parameters of different system states (under different EOCs) and/or different measurements. The reasons is that it is not possible to compare the ith eigensolution in the reference state or in the one data set with the ith eigensolution in the next system state or other data set. In addition, the eigenfrequencies change their values as the degree of damage increases due to the reduction of stiffness. In case of the laboratory test, an at- tempt was made to compute the MAC value and frequency deviation in order to trace the course of the natural frequencies over the individual structural states. This proce- Figure 5: Monitoring of identified natural frequencies and mode shapes (dashed line)

– 182 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017183

Figure 6: Experimental setup and damage positions (circled) of the benchmark test on a three-story frame structure

cess (relative quantity) and can be used for References damage detection and localization as well. [8] Farrar, C. R.; Worden, K: Structural Prof. Armin Lenzen and Dr. Max Vollmer- [1] Jonkman,J.; Butterfield, S.; Musial, W.; health monitoring: a machine learning ing from the University of Applied Sciences Scott, G.: Definiton of a 5-MW Reference perspective, Chichester, West Sussex, in Leipzig defined the examined CP based Wind Turbine for Offshore System Deve- U.K. Hoboken, N.J: Wiley, 2013. on H estimation and state projections [10], lopment, Golden, Colorado: NREL, 2009. [9] Häckell, M. W.: A Holistic Evaluation ∞ Concept for Long-Term Structural Health [11]. It was applied in DFWind phase 1 us- [2] Jonkman, J.; Buhl, L. M.: FAST User's Monitoring, Mitteilungen des Instituts ing benchmark test data provided by the Guide, NREL, 2005. [3] Wang, Q.;Jonkman, J.; Sprague, M.; für Statik und Dynamik der Leibniz Los Alamos National Laboratories [12]. Jonkman, B.:BeamDyn User's Guide and Universität Hannover 25/2015, 2015. The measurements were performed on a Theory Manual, NREL, 2016. [10] Lenzen, A.; Vollmering, M.: An out- three-story frame structure under laboratory [4 Wang,Q.; Sprague, M.;Jonkman, J.;N. put-only damage identification method conditions. Johnson, N.; Jonkman, B.: BeamDyn: a based on H∞ theory and state projec- high-fidelity wind turbine blade solver in tion estimation error (SP2E), Structural Apart from measurements in the reference the FAST modular framework, Wind Ener- Control and Health Monitoring, 2017. state (undamaged), five different damage gy, vol. 20, pp. 1439-1462, 2017. [11] Lenzen, A.; Vollmering, M: On scenarios were realized on two different [5] Katayama, T.: Subspace methods for experimental damage localization by positions and data acquired. Bolts between systen identification, Springer, 2005. SP2E: Application of H ∞ estimation and oblique projections, Mechanical brackets and plates were loosened (scenario [6] Häckell M. W.; Rolfes, R.: Monitoring Systems and Signal Processing, D05, D10, DHT) and removed (scenario a 5MW offshore wind energy converter - condition parameters and triangulation vol. 104, pp. 648-662, 2018. DB0). The removal of entire brackets was based extraction of modal parameters, [12] Wernitz, S.; Pache, D.; the most severe damage state (scenario Mechanical Systems and Signal Proces- Grießmann, T.; Rolfes, R.: Performance DBB). The damage analysis showed that all sing, vol. 40, no. 1, pp. 322-343, 2013. of an H_infinity estimation based dama- structural modification were detected and [7] Rytter, A.: Vibration based inspection ge localization approach in the context most scenarios located. Therefore, the proof civil engineering structures, Aalborg: of automated Structural Health posed condition parameter showed a prom- Dept. of Building Technology and Structu- Monitoring, 9th European Workshop ising potential for future SHM and damage ral Engineering, Aalborg University, 1993. on Structural Health Monitoring, 2018. assessment in DFWind phase 2.

– 183 – 184CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Smart Blades 2 – Construction, Testing and Further Development of Intelligent Rotor Blades

Carl von Ossietzky Universität Oldenburg affecting the energy yield. Different smart achieved by using a continuous-wave short- Institute of Physics blades technologies are being investigated, range LiDAR system. The LiDAR meas- Research Group: Wind Energy Systems components are being designed and tested ures on a conical area around the nacelle Research Group: Turbulence, Wind Energy and simulation models are in course of top at a distance of 60 m in front of the ro- and Stochastics validation at different scales, including field tor. Besides the testing activities, strategies tests. In addition to the partners from the to include structural bend-twist coupling Lars Neuhaus, Piyush Singh, forerunner project “Smart Blades” com- in addition to those worked on in the fore- Agnieszka Hölling, Michael Hölling, Martin Kraft, Martin Kühn, prising the ForWind research institutes, runner project are being studied at an 80 m Joachim Peinke, Hamid Rahimi, the German Aerospace Center (DLR) and scale. A parametric study on helical stringer Marijn van Dooren, Lena Vorspel, the Fraunhofer Institute for Wind Energy configurations has started, aiming to find a Bastian Dose, Frederik Berger, Systems (IWES), several partners from the weight-specific optimum solution for the in- Vlaho Petrovic, Mohsen Forghani, wind energy industry joined the consortium, tegration of structural bend-twist coupling. Matthias Wächter, Stephan Voß, namely General Electric Deutschland Hold- Róbert Ungurán ing GmbH, Henkel AG & Co. KG, Nordex Technology 2 – Rotor blades with Energy GmbH, SSB Wind Systems GmbH active trailing edge Leibniz Universität Hannover, & Co. KG, Senvion GmbH, Suzlon Energy The main goal of ForWind Oldenburg in Institute of Structural Analysis, Limited and WRD Wobben Research and Technology 2 is to build up a new setup for Institute of Turbomachinery and Fluid Development GmbH. ForWind is partici- validating the aerodynamic and servody- Dynamics, namic performance of a profile with an ac- Institute for Wind Energy Systems pating with its wide variety of expertise in all project activities, which are divided into tive trailing edge. The idea is to expose such Mehdi Garmabi, Ayan Haldar, four technologies. a profile to varying inflow conditions com- Eelco Jansen, Helge Jauken, ing from pre-defined time changing angles Hinesh Madhusoodanan, Raimund Rolfes, of attack at the leading edge of the profile. Christoph Jätz, Jörg Seume, Torben Wolff, Project Description The resulting forces will be measured and Claudio Balzani, Tobias Holst, Sina Lotfi- controlled using the active trailing edge. omran, Pablo Noever Castelos, Technology 1 – Rotor blades with enburg has been modified in order be able to Jelmer Derk Polman, Andreas Reuter, bend-twist coupling meet the requirements for the needed inflow Michael Wentingmann In this technology, 20 m rotor blades with conditions. First tests have been done and a passive bend-twist coupling were manu- these inflow conditions have been charac- Funding: Federal Ministry for Economic factured by DLR for a full-scale blade test terized. Affairs and Energy (BMWi) at Fraunhofer IWES and for field tests on Ref.Nr. 0324032C (ForWind Hannover) the CART 3 research wind turbine at the The activities of ForWind Hannover are 0324032D (ForWind Oldenburg) National Renewable Energy Laboratory subdivided into structural, aeroelastic, aer- (NREL) in Boulder, Colorado, USA. These oacoustic investigations and model devel- Duration: 09/2016 – 11/2019 tests should provide the consortium with the opment. In the structural part, the fatigue necessary data to validate codes, tools and behavior of segments with an active trailing models developed in the previous Smart edge that are cyclically tested at DLR, are Blades project. For this purpose ForWind simulated using a fatigue model imple- was involved in the definition of sensor mented in ABAQUS. For that purpose the Introduction positions in the full-scale blade test and the related finite element models were created, specification of test cases that must be cov- see Fig. 1. The aim is to apply and validate As wind turbines continue to grow in size, ered in the field tests. The developed simu- within SmartBlades2 the fatigue damage the slenderness of wind turbine rotor blades lation tools have been fed with the CART model that has recently been extended to increases. Hence, the loads acting on wind 3 geometry. CFD grids as well as Finite 3D stress states. In addition, an active trail- turbines and their components become in- Element models for the 20m blades have ing edge based on the concept of multi- creasingly critical. The SmartBlades2 pro- been prepared. Additionally, ForWind is re- stable structures is investigated. Therein, a ject aims to reduce the aerodynamic loads sponsible for measuring the CART 3 wind bi-stable structure was designed to switch and their fluctuations without significantly turbine’s incoming wind field, which is between the required maximum flap de

– 184 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017185

designing the pre-defined complex inflow situations has been measured. In close cooperation with DLR the specifications for the kinematics for the passive adaptive slat have been defined. In the field of the rigid slat, ForWind Oldenburg is mainly interested in the effect of a rigid slat on the dynamic response of a wind turbine. Stochastic analysis using Langevin processes will be used to investigate e. g. the dynamics of the power output of a wind turbine with and without a rigid slat under realistic inflow conditions. This will be achieved with data from a real multi MW wind turbine from the project partner Suzlon, which will be equipped with Figure 1: Fatigue simulation for flexible trailing edge under representative variable a rigid slat. Additionally, in this project For- amplitude fatigue loading. Wind Oldenburg designed and built a model turbine for wind tunnel experiments. Fig. 2 shows the Model Wind Turbine Oldenburg, which has a diameter of 1.8 m (MoWiTO flections with low energy consumption via that appropriately resolves the different 1.8). The size of the blades is large enough piezo actuators. Finite element simulations scales at a reduced computational cost com- so that aerodynamically functional slats can have been carried out in order to show the pared with a pure LES approach. be designed and constructed for it. The tur- feasibility of the concept [1]. bine is equipped with sensors to measure, Technology 3 – Rotor blades with among other things, torque, power output, In the context of aeroelasticity, ForWind adaptive slat rotational frequency, thrust and bending Hannover implemented an empirical aero- The encouraging result in the previous moments of the blades. The system features dynamic model based on URANS simula- Smart Blades project regarding Technology a torque control and three motors allow for tions. The model was developed with the 3 led to an extended consideration in Smart- individual pitch control. This turbine will aero-servo-elastic simulation tool FAST Blades2. This includes the active slat, a rigid be tested under various inflow conditions in the forerunner project. It is planned to version of the slat as well as a so-called pas- in the turbulent wind tunnel in Oldenburg. validate this model by means of 2D wind sive adaptive slat, equipped with a kinematic This wind tunnel is equipped with an ac- tunnel experiments within the project at a that allows the slat to move in a predefined tive grid (see fig. 2) that allows generating later stage. Besides, the development of a manner based on local inflow conditions. In flexible inflow conditions featuring e. g. high-fidelity aero-servo-elastic model with this technology ForWind Oldenburg is in- different turbulence intensities and higher two-way coupling between the structure volved in wind tunnel test with the active order statistics. The inflow situations for and aerodynamics has started. A shell-type and the passive adaptive slat. For the active the wind tunnel tests have been designed (“3D) finite element model represents the slat an open as well as a closed loop control and the active grid performance has been structure, the aerodynamic model is based strategy are being tested and compared with characterized. Additionally, ForWind is also on BEM, and the controller for torque-speed respect to the ability to mitigate force fluctu- involved in the acoustic analysis of the slat and pitch control relates the rotor behavior ations. In the closed loop control the move- technology. Currently, the validation of the to the wind inflow. ment of the slat is based on the acting forces numerical models regarding the acoustic on the profile whereas the open loop control simulations is being performed. Moreover, ForWind Hannover aims to per- uses the previously measured information form high-fidelity CFD/CAA aeroacoustic of the evolution of the inflow conditions. In Technology 4 – Cross-Technology Topics simulations. For that purpose, a parameter- contrast to the Smart Blades project where Activities at ForWind Hannover that can- ized CAD model of the specimen geometry just sinusoidal variations of the inflow angle not be linked directly to a single investi- that has been tested in the forerunner pro- have been tested, more complex flow situa- gated technology but to all are integrated ject has been created. The meshing of the tions are planned for this project. So far dif- in the cross-technology topics. These com- fluid volume with CCM+ has started, so that ferent controller concepts have been inves- prise the modeling and simulation of adhe- remeshing for variable flap angle configu- tigated as well as the communication set-up sive joints, the investigation of aero-elastic rations will be possible in the near future. with the active slat electronics and mechan- stability (flutter) and the further evaluation A literature review resulted in the decision ics. The active grid has been characterized of techno-economical aspects of each tech- to use DDES, a hybrid RANS/LES method in detail and a transfer matrix that allows for nology.

– 185 – 186CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Figure 2: Model Wind Turbine Oldenburg with a diameter of 1.8m (MoWiTO 1.8) in front of the active grid in the big wind tunnel in Oldenburg

In the work package focused on adhesive solid elements in order to extract full 3D measurements of the thermal and chemical joints, trailing edge adhesive joints have stress states, see fig. 3 (c). Global stress cal- shrinkage have started. been added to the finite element modeling culations show clear 3D stress states in the process developed in the in-house tool adhesive, highlighting that the commonly For the investigations of flutter limits of dif- MoCA (Model Creation and Analysis tool) used one-dimensional shear stress proof is ferent smart blades technologies, the setup and including local mesh refinement, see physically not meaningful [2-4]. The mod- of FAST models of the Smart Blades refer- fig. 3 (a)-(b). The adhesive is modeled with eling of the curing kinetics is finished and ence turbine has started. The structural

a b c

Figure 3: Finite element model of the trailing edge adhesive joint. (a) Slice of the blade, (b) detail of the trailing edge with local mesh refinement, (c) hybrid shell/solid element model at the trailing edge with the adhesive solid elements highlighted in red.

– 186 – Annual Report 2015-2017 CROSS DISCIPLINARY TOPICS Annual Report 2015-2017187

properties for the beam model have been recalculated on the basis of the 3D blade model from the forerunner project. Pre-test simulations for a code-to-code comparison with project partners have jointly been specified. These will start in the near future.

The technology evaluation model has been reviewed and improvement potential has been identified. The technology evaluation will be executed in regular time intervals in order to identify development needs throughout the project. A final evaluation aims for the comparison of the effectiveness of the different technologies. References

Summary [1] Haldar, A.; Jansen, E.; Rolfes, R.: Snap-through of Multistable Variable Stiffness Composites using The project is running and the partners are Piezoelectric Actuators, in: Ferreira, A. working on fulfilling their ambitious goals. J. M.; Larbi, W.; Deu, J.-F.; Tornabene, ForWind is active in the development of all F.; Fantuzzi, N. (Edts.): Proc. of ICCS20 the project’s technologies, including struc- – 20th International Conference on tural investigations on bend-twist coupled Composite Structures, Sept. 04-07, blades, wind field and wind tunnel meas- 2017, doi: 10.15651/978-88-938-5041-4 urements, structural, aeroelastic, and aeroa- [2] Noever Castelos, P.; Voosen, M.; coustic modeling as well as the techno-eco- Balzani, C.: Non-proportional nomic technology evaluation. First results multi-axial stress states and their are already available, but the major results influence on the fatigue life of trailing edge adhesive joints in wind turbine are expected towards the end of the project rotor blades, Wind Energy Science at the end of 2019. Conference, Technical University of Denmark, Lyngby Campus, Copenhagen, Denmark, June 26-29, 2017 [3] Noever Castelos, P.; Wentingmann, M.; Balzani, C.: Fatigue analysis of trailing edge adhesive joints in wind turbine rotor blades accounting for non-proportional multiaxial stress histories, in: Ferreira, A. J. M.; Larbi, W.; Deu, J.-F.; Tornabene, F.; Fantuzzi, N. (Edts.): Proc. of ICCS20 – 20th International Conference on Composite Structures, Sept. 04-07, 2017, doi: 10.15651/978-88-938-5041-4 [4] Noever Castelos, P.; Wentingmann, M.; Balzani, C.: Comparative study of finite-element-based fatigue analysis concepts for adhesive joints in wind turbine rotor blades, Proceedings of the 7th GACM Colloquium on Computational Mechanics for Young Scientists from Academia and Industry, Stuttgart, Germany, Oct. 11-13, 2017

– 187 – 188CROSS DISCIPLINARY TOPICS Annual Report 2015-2017

Interdisciplinary Research Center on Critical Systems Engineering for Socio-Technical Systems II

Carl von Ossietzky Universität Oldenburg, Project Description along both spatial-temporal and cognitive Research Center Safety Critical Systems / dimensions. It calls for robust and adaptable Computing Science, Business Engineering The Project "Critical Systems Engineering designs, seamlessly catering with adverse for Socio-Technical Systems II" (CSE-2) and changing environmental conditions. It Axel Hahn is funded by the 'Volkswagenstiftung' and calls for executable and composable models the MWK from January 2017 to October of socio-technical systems, both human and Funding: Federal State of Lower Saxony 2018. Building on the successful work of technical, allowing to adaptively, as it were, and Volkswagenstiftung the predecessor project CSE-1, the project’s "zoom" into detailed levels when reaching Duration: 01/2017 – 10/2018 objective combined goals on a technical re- critical states to provide fine-grained views search level as well as on a strategic level, of the actual interactions, as well as the need summarized as 'enhancing the state of the to aggregate to coarse views in order to cope art of human-centered critical systems engi- with the sheer complexity of such models. neering and increasing the visibility of the Introduction state of Lower Saxony as one of the main To aid the strategic impact of the project, actors on national and European level in this CSE-2 has been structured into two main Critical Systems, i.e., systems whose failure interdisciplinary research area with high so- research areas, namely: either endanger human life or cause drastic cial relevance'. economic losses, form the technological A. foundational research of the princi- backbone of today's society and are an in- CSE-2 focuses on instances of such socio- ples of Human-Machine Cooperation in tegral part in such vastly diverse industrial technical systems in the transportation do- safety critical real time systems, paving sectors as automotive, aerospace, maritime, main, where the overarching objectives are the way for a submission of a proposal automation, energy, health care, banking, to achieve safe and green mobility, through for a collaborative research center at the and others. cooperative semi-autonomous guidance of German Science Foundation vehicles with humans in the loop, such as in B. establishing a research infrastructure their roles as drivers, operators, navigation which is unique in Europe for experi- officers, flight controllers, etc., and consider ments and experimental evaluation in the two industrial sectors key in Lower Saxony, scope of CSE. the automotive and maritime domains. Such systems are safety critical – human errors, technical failures and malicious manipula- Summary tion of information can cause catastrophic events leading to loss of life. Creating suf- The Interdisciplinary Research Center on ficiently precise real-time mental or digital Critical Systems Engineering for Socio- images of real-world situations and assur- technical Systems addresses critical sys- ing their coherence among all involved ac- tems, which rely on synergistically blending tors (both humans and technical systems) human skills with IT-enabled capabilities as a basis for coordinated action is a major of technical systems to jointly achieve the challenge in socio-technical system design. overarching objectives of the system-of- This calls for constructive approaches in- systems. volving intuitive and scalable patterns of cooperation, between humans and technical systems, seeking for a balanced sharing of tasks best matching both the abilities of humans and technical systems, or between technical systems. It calls for insights in understanding humans in their interaction with technical systems. It calls for layered approaches in aggregating information

– 188 – Annual Report 2015-2017 QUALIFICATION AND KOORDINATION Annual Report 2015-2017189

Continuing Studies Programme Wind Energy Technology and Management (“Windstudium”)

Moses Kärn and Christoph Schwarzer Project Description experts and satisfy the highest of academic didactic standards. Partners: Wind energy is a highly innovative and in- Wind Energy Agency WAB, terdisciplinary industry sector whose further The interdisciplinary approach is the central City of Oldenburg successful development is closely tied to the theme that the program is based upon and Funding: Participants’ fees and co- availability of highly skilled and educated is represented by the following special char- sponsoring by Norddeutsche Landesbank experts from various backgrounds. Because acteristics. (NORD/LB), GE Wind Energy GmbH, of the complex nature of the field part-time UKA Umweltgerechte Kraftanlagen GmbH education is an important instrument to Group Dynamics: The selection process & Co. KG provide professionals with high-level sys- will be based on bringing together a group tematic know-how and interdisciplinary of people with a wide range of experience understanding of wind energy technology, in their academic and occupational areas, projects, and economic issues. be it technological, planning management, administration, or law. Thus, the student The Continuing Studies Programme Wind group’s line-up will reflect the heterogene- Energy Technology and Management is ous profiles, which are often found within especially designed to support companies a company’s typical departmental work Introduction of the wind energy sector and is directed group. to professionals in the field as well as to In the years 2015, 2016, and 2017 the Con- those who wish to enter this field. It offers Project Work: For the duration of the pro- tinuing Studies Programme Wind Energy comprehensive systematic understand- gramme two to three groups of students will Technology and Management (“Windstudi- ing of wind energy projects from scientific team-up as virtual companies to develop a um”) started with its 10th, 11th, and 12th grounds to technical, legal and economic wind energy project from the green field to one-year-long courses. The Windstudium realization, as well as skills in planning and the selling of the wind farm. This project has maintained its unique position as part- project management. work provides a lot of practical experience, time qualification for professionals despite which reflects what the students learn in the fact that the wind energy businesses have The programme offers a mix of learning classes, and copies ‘real’ experience in com- been facing uncertain regulatory standards methods and is therefore designed to fit the munication with experts from a variety of and policies in the year 2013 and following requirements of professionals: self-study of disciplines. (e. g. debates about the “Strompreisbremse” reading materials, a two-day seminar once and replacing the fixed feed-in-tariffs by an every month, and project work in teams. Teaching Material: The reading material is auction process with the EEG 2017). The The total duration of the programme is elev- divided into basic and speciality sections. Windstudium is still operated by the found- en months. The Diploma of Advanced Stud- The basic sections are obligatory for all ing institutions ForWind, WAB in close co- ies (DAS) “Certified Wind Energy Expert” the students. And, students must select half operation with the City of Oldenburg, and is issued by the University of Oldenburg of the specialized sections to be examined the companies GE Wind Energy GmbH, upon successfully passing the examinations. upon, therefore allowing them to focus on Bremer Landesbank (now NordLB), and their personal areas of expertise and interest. UKA Umweltgerechte Kraftanlagen GmbH The realization of wind energy projects & Co. KG. requires that experts from a variety of dif- Lecturers and Co-Lecturers: The seminars ferent disciplines work closely together. are lead by the authors of the teaching mate- The "Windstudium" addresses exactly such rial who are leading experts from universi- interdisciplinary challenges, and fosters a ties and industry. The main lecturer's semi- 'know-how-transfer' from acknowledged nars are complemented by co-lecturers and experts in the field and from Universities, guest speakers who give insight into their thus providing current and expert knowl- professional experience. edge. The program’s study materials were developed in partnership with the Univer- The administrators of the Continuing Stud- sity of Oldenburg’s continuing education ies Programme in Wind Energy Technology

– 189 – 190QUALIFICATION AND KOORDINATION Annual Report 2015-2017

and Management operate an active network Summary of alumni for the students to stay in touch with each other, to assist them in the ex- The Continuing Studies Programme Wind change of professional information, and to Energy Technology and Management is support further continuing education possi- especially designed to support companies bilities. Yearly alumni meetings take place and professionals of the wind energy sec- in Oldenburg or Bremerhaven. tor. After twelve years of successful operation it has consolidated its position as well-established programme of continuing academic education in the wind energy sector even if the numbers of applications and participants has been declining since the year 2013.

Link

www.windstudium.de

– 190 – Annual Report 2015-2017 QUALIFICATION AND KOORDINATION Annual Report 2015-2017191

European Wind Energy Master (EWEM)

Carl von Ossietzky Universität Oldenburg Introduction Project Description ForWind Head Office ForWind is co-founder of the European The EWEM- European Wind Energy Mas- Institute of Physics Wind Energy Master (EWEM) which is ter is an advanced two-year (120 ECTS) Research Group: Wind Energy Systems master course (MSc) with four specializa- Research Group: Energy Meteorology supported by the European Commission as tions (see also Fig. 1): Research Group: Turbulence, Wind Energy an Erasmus Mundus Master Course. A net- and Stochastics work of associated partners, composed of the main wind energy research institutions, • Wind Physics Martin Kühn, Joachim Peinke, industry and NGOs reinforces the consor- • Rotor Design Detlev Heinemann, Moses Kärn tium. During the five cohorts since 2012 • Electric Power Systems EWEM has educated 169 students from 43 • Offshore Engineering Partners: nationalities. Delft University of Technology (TU Delft), The EWEM prepares graduates for a career Technical University of Denmark (DTU), in research, both in industry and in aca- Norwegian University of Science and demia, and is closely linked to the research Technology (NTNU) taking place at the four universities and their partners. ForWind / University of Old- enburg (UOL) is teaming-up with DTU in the Wind Physics track. Graduates receive a double degree in M.Sc. Engineering Physics (UOL) and M.Sc. Wind Energy Engineering (DTU). From the 169 students in the first five cohorts 23 students have graduated in the Wind Physics track.

Figure 1: Structure of the European Wind Energy Master (EWEM). Abbreviations: Delft University of Technology (TU Delft), Technical University of Denmark (DTU), Norwegian Link University of Science and Technology (NTNU), ForWind Institute at the Carl von Ossietzky Universität Oldenburg (UOL). www.windenergymaster.eu (source: http://ewem.tudelft.nl/about-ewem/programme)

– 191 – 192EVENTS 2015 Annual Report 2015-2017

Smart Blades Symposium in Oldenburg

Carl von Ossietzky Universität Oldenburg ForWind Head Office Institute of Physics Research Group: Turbulence, Wind Energy and Stochastics

Joachim Peinke

Co-organizers: Oliver Paschereit (TU Berlin), Christian Willberg (DLR)

With the aim of discussing approaches, The organizers of the symposium and invited guest speakers (from left): methods and results of the two ongoing re- Flemming Rasmussen (DTU), Prof. Dr. Joachim Peinke (ForWind), search projects on intelligent rotor blades, Prof. Dr. Oliver Paschereit (TU Berlin), Dr.-Ing. Christian Willberg (DLR), researchers of the BMWi joint research Prof. Dr. Jan-Willem van Wingerden (TU Delft) project "Smart Blades - Development and Design of Intelligent Rotor Blades" and the DFG research project "Lastenkontrolle von Windturbinen unter realistischen tur- bulenten Anströmungen" ("Load Control of Wind Turbines under realistic turbulent flow") came together for a Symposium in Oldenburg in February 2015.

Approximately 60 scientists from For- Wind, DLR, Fraunhofer IWES Nord- west, the RWTH Aachen University, TU Welcome address by Prof. Dr. Joachim Peinke Lab tour with active grid Berlin, TU Darmstadt, the University of Oldenburg, and the University of Stutt- gart met for two days at the Institute of Physics at the University of Oldenburg. They discussed recent research results, exchanged ideas with each other and with invited international speakers and visited the wind tunnel and laboratories on site. The Symposium was organized by Prof. Dr. Joachim Peinke (ForWind), Prof. Dr. Oliver Paschereit (TU Berlin), and Dr.-Ing. Christian Willberg (DLR). .

Participants of the Smart Blades Symposium in Oldenburg

– 192 – Annual Report 2015-2017 EVENTS 2015 Annual Report 2015-2017193

Job and Education Fair “zukunftsenergien nordwest” 2015

Carl von Ossietzky Universität Oldenburg ForWind Head Office

Christoph Schwarzer

Partners: Windenergieagentur WAB e.V., Oldenburger Energiecluster OLEC e.V., Wirtschaftsförderung der Stadt Bremerhaven (BIS), WFB Wirtschaftsförderung Bremen GmbH, Wirtschaftsförderung Stadt Oldenburg, Hochschule Bremerhaven

Sponsored by: Bremer Landesbank, Enercon GmbH

The job and education fair for renewable and get informed about career opportuni- The "zukunftsenergien nordwest" offered energies and energy efficiency, the "zukun- ties in a personal conversation. The fields exhibitors a high profile public platform ftsenergien nordwest", opened its gates for include electrical engineering/electronics, where companies and service providers the sixth time on the 20th and 21st March mechanical engineering, industrial engi- meet together with an audience interested 2015 in Hall 4 of the Bremen Exhibition neering, economics, civil engineering and in their sector along with potential young Centre. 62 exhibitors were presenting energy technology. Beside the exhibition, applicants and those looking to enter from themselves on the job fair: companies, uni- visitors could take part in workshops and other fields of expertise. versities, institutes for further training and company presentations. The forums of- research institutes involved in renewable fered valuable tips on starting a career, The "zukunftsenergien nordwest" is sup- energies and energy efficiency. 2.500 visi- starting salaries or assessment centers. This posed to take place once a year alternat- tors met attractive employers with open job program was complemented by excursions ing in Oldenburg and Bremen. It started opportunities and traineeships. The major- to well known companies and regional 2010 in Oldenburg. The program and all ity of the exhibitors came from the wind facilities as well as application trainings. exhibitors with their details on personnel energy sector or are leaders in the field of Admission to the fair and the supporting requirements can be found online at www. energy efficiency. Visitors could talk to the program with excursions and lectures was zukunftsenergien-nordwest.de. personnel directly at the trade fair booth free of charge.

– 193 – 194EVENTS 2015 Annual Report 2015-2017

Hannover Messe 2015

ForWind Head Office

Committee on Economic Affairs, Employment and Transport of the Lower Saxony Parliament

In April 2015 ForWind participated again measuring device allows it to be mounted As in previous years, ForWind took part in the Hannover Messe – the world's big- directly on the wind turbine. It actively in the Tec2You student campaign and an- gest industrial fair – at the joint stand of supports the control of the turbines and swered the questions of several school Lower Saxony, introducing ForWind and helps to reduce the wind gust load. groups. its services in research, consultancy, education, and events. Another exhibit from the field of flow Besides the various visitors from industry, measurement was the sphere anemometer. research and the interested public, repre- Amongst others, ForWind presented the ForWind scientists improved and refined sentatives of ForWind were pleased about active grid, which makes it possible to pro- its technology continuously in the past discussions with the Minister for Science duce turbulent dynamic flows under con- years. It allows wind speed and direction and Culture, Dr. Gabriele Heinen-Kljajić, trolled conditions in the laboratory. The to be measured simultaneously and at high the Minister for Economic Affairs, Em- heart of the large wind tunnel currently frequency. Due to its particularly stable ployment and Transport, Olaf Lies, and the under construction in Oldenburg can re- shape and the absence of moving parts, the Minister for the Environment, Energy and alistically simulate all types of wind situ- device is ideally suited for offshore use. Climate Protection, Stefan Wenzel. ations with its rhombic aluminum wings, from the mild breeze to strongly turbulent Interested visitors of the fair could also in- For the first time, Maroš Šefčovič, Vice storms. This opens up new possibilities for form themselves about the further training President for Energy Union of the new Eu- the operation and further development of opportunities at ForWind: Two continuing ropean Commission, visited the event to wind turbines. study programs on onshore and offshore learn more about wind energy research and wind energy are aimed at specialists and the turbulent wind tunnel under construc- Furthermore, ForWind showed a newly managers in this field of industry. In its tion in Oldenburg. In addition, members of developed wind scanner. This is based on seminars, the ForWind Academy combines Lower Saxony's state parliament and the the remote sensing method Lidar (Light practice-related questions with current sci- European Parliament informed themselves detection and ranging), which measures entific findings. about current developments at ForWind. air flows with high temporal and spatial resolution. The new compact design of the . – 194 – Annual Report 2015-2017 EVENTS 2015 Annual Report 2015-2017195

Member of the European Parliament, Reinhard Bütikofer (Die Vice-President for Energy Union of the new Grünen/EFA) and Dr. Stephan Barth, Managing Director of European Commission, Maroš Šefčovič ForWind"

from left to right: Lower Saxony Minister for the Environment, Lower Saxony Minister for Science and Culture, Energy and Climate Protection, Stefan Wenzel; Dr. Gabriele Heinen-Kljajić, Michael Lindenthal, Head of Department 5 of the Ministry of and Dr. Stephan Barth, Managing Director of ForWind the Environment (Energy, Climate Protection), and Dr. Stephan Barth, Managing Director of ForWind

Lower Saxony Minister for Economic Affairs, Employment and ForWind was once again a tour point for many school groups as Transport, Olaf Lies part of the young talent initiative Tec2You.

– 195 – 196EVENTS 2015 Annual Report 2015-2017

Visit of the Vice-President for the Energy Union and Energy Minister

ForWind Head Office and Dr. Werner Brinker, CEO of EWE AG The speaker of the executive board of For- visited ForWind at the University of Old- Wind, Prof. Dr. Martin Kühn, explained enburg together with a delegation of repre- the spectrum of ForWind's research activi- The Vice-President of the European Com- sentatives from industry and government. ties and the planned investigations to the mission for the Energy Union, Maroš The core of the visit was a tour through delegation. Šefčovič and Wenzel were im- Šefčovič, the Lower Saxony Minister for the research laboratory for turbulence and pressed by the size and future experimen- the Environment, Energy and Climate wind energy systems and the large turbu- tal possibilities of the 20.5 million Euro Protection, Stefan Wenzel, the Regional lence wind tunnel currently under con- project. Following the visit, Šefčovič an- Ministers for Energy & Environment struction. nounced via his Twitter account: "We need Nienke Homan (Province of Groningen) spread the word of Lower Saxony world and Tjisse Stelpstra (Province of Drenthe), class research on Sustainable Energy”.

Figure 1: from left to right: Prof. Dr. Martin Kühn (ForWind), Maroš Šefčovič (Vice-President of the European Commission for the Energy Union), Dr. Detlev Heinemann (ForWind), Stefan Wenzel (Lower Saxony Minister for the Environment, Energy and Climate Protection), Prof. Dr. Joachim Peinke (ForWind), Nienke Homan (Minister for Energy & Environment Province of Groningen), Dr. Werner Brinker (CEO of EWE AG), Tjisse Stelpstra (Minister for Energy & Environment Province of Drenthe), and Dr. Stephan Barth (ForWind)

– 196 – Annual Report 2015-2017 EVENTS 2015 Annual Report 2015-2017197

– 197 – 198EVENTS 2015 Annual Report 2015-2017

Cooperation with Taiwanese University – first Summer School realized

Carl von Ossietzky Universität Oldenburg In close cooperation with the National Lectures and exercises given by ForWind Institute of Physics Cheng Kung University (NCKU) in researchers Prof. Dr. Joachim Peinke, Dr. Research Group: Energy Meteorology Tainan, Taiwan, ForWind has organized a Detlev Heinemann, Hendrik Heißelmann Research Group: Turbulence, Wind Energy Summer School. Funded by the German and Michael Schmidt as well as Jan-Chris- and Stochastics Academic Exchange Service (DAAD), toph Hinrichs from Aerodyn Engineering the Summer School "Aerodynamics and GmbH covered the basics of wind energy Joachim Peinke, Detlev Heinemann, Energy Meteorology for Renewable generation and added a focus on the re- Hendrik Heißelmann, Michael Schmidt Energy Conversion: Aspects of the gional typhoon risks and their mitigation Partner: Jan-Christoph Hinrichs, Aerodyn Physical, Statistical and Meteorological by typhoon-resistant turbines. Engineering GmbH Understanding of Wind, Turbulence and Turbines", took place from 7th to 14th Within the framework of the Summer Funding: German Academic Exchange September 2015 at the NCKU in Tainan. School, the Department for Aeronautics Service (DAAD) and Astronautics (DAA) at NCKU and More than 70 students and 16 representa- ForWind Oldenburg agreed to further tives of the Taiwanese industry participated extend their cooperation. in the 6-day course at the NCKU Depart- ment of Aeronautics and Aerodynamics.

Prof. Dr. Ta-Hui Lin, Director of the Depart- In front (from left to right): Prof. Dr. T.S. Leu, Prof. Dr. J.J. Miau (both NCKU ment for Aeronautics and Astronautics (DAA), Tainan), Prof. Dr. Cheng-Dar Yue (National Chiayi University, Taiwan), Hendrik handed over an award to Prof. Dr. Joachim Heißelmann, Dr. Detlev Heinemann, Prof. Dr. Joachim Peinke and Michael Schmidt Peinke for the successful cooperation at the (all ForWind). Summer School. Prof. Dr. Cheng-Dar Yue made his PhD at the University of Oldenburg.

– 198 – Annual Report 2015-2017 EVENTS 2016 Annual Report 2015-2017199

WindLab – New Research Building for Wind Energy Research

Carl von Ossietzky Universität Oldenburg ForWind Head Office

Research Group: Wind Energy Systems Research Group: Energy Meteorology Research Group: Turbulence, Wind Energy and Stochastics

Martin Kühn, Joachim Peinke, Detlev Heinemann

In March 2015 the carcass of the new research laboratory for turbulence and wind energy systems (WindLab) was completed and the topping-out ceremony took place on the Wechloy campus of the Univer- sity of Oldenburg. The heart of the 2,300 square meter new building is a turbulent On the photo (from left): Energy meteorologist Dr. Detlev Heinemann, turbulence researcher Prof. Dr. Joachim Peinke; Chief Building Director Wilhelm Wickbold; wind tunnel, which enables scientists to in- interim University President Prof. Dr. Katharina Al-Shamery; Finance Minister vestigate the interaction of turbulent flows Peter-Jürgen Schneider; ForWind Managing Director Dr. Stephan Barth; with wind turbines and their components, architect Veit Schäfer and complete wind farms. The aim is a better understanding of turbulent flows and to obtain precise data on the operating behavior of wind turbines and large offshore been successfully researched at the Univer- ogy (Bremerhaven) and the Max Planck wind farms. The new building was ap- sity of Oldenburg for many years. With the Institute for Dynamics and Self-Organi- proved by the German Council of Science WindLab, excellent conditions have been zation (Göttingen) will benefit from the and Humanities in 2012 and classified as created to further strengthen research in the new facilities. Scientists from Oldenburg, particularly worthy of promotion. Guests field of wind energy". led by wind energy expert Prof. Dr. Mar- at the topping-out ceremony included tin Kühn, turbulence researcher Prof. Dr. Lower Saxony's Finance Minister Peter- The WindLab offers space for more Joachim Peinke and energy meteorologist Jürgen Schneider and Wilhelm Wickbold, than 130 scientists, a wind tunnel with a Dr. Detlev Heinemann, were responsible Chief Construction Director of the State 30-meter-long measuring section and wind for the contents of the application for the Construction Management Lüneburger speeds of up to 150 kilometers per hour. new research building. Heide. Minister Schneider was impressed "With the WindLab and the associated by the progress of the construction and ex- turbulence wind tunnel, we have a unique In comparison to wind tunnels, such as pressed his thanks to all those involved in research infrastructure for wind energy," those used in aviation, realistic wind fields the construction so far. emphasized University President Prof. Dr. can be simulated in the turbulent Olden- Dr. Hans Michael Piper at the opening cer- burg wind tunnel as they occur in nature. After completion of the carcass in 2015, the emony. Both the technical equipment and The investigations should help to increase entire building was completed at the end of the synergies resulting from the coopera- the efficiency of wind farms and avoid 2016. It was officially inaugurated on Janu- tion of the interdisciplinary team of experts technical and financial risks. "Our great ary 26th, 2017, in the presence of Lower are outstanding. vision is to achieve a new quality in wind Saxony's Minister for Science and Culture, energy research through the combination Gabriele Heinen-Kljajić. "The sustainable Wind energy scientists from the Universi- of measurements in the open field, numeri- energy supply from renewable energies is ties of Oldenburg and Hanover, the Jade cal simulations and the new experimental a central challenge of our time," explained University, the Fraunhofer Institute for possibilities in the turbulent wind tunnel," the Minister. "Renewable energies have Wind Energy and Energy System Technol- explained Peinke.

– 199 – 200EVENTS 2016 Annual Report 2015-2017

The WindLab was designed by the architect's office HammesKrauseArchitekten.

The total costs of around 20 million euros are shared equally between the national government and the state of Lower Saxony.

Front of the new WindLab

The new WindLab from a bird's eye view in summer 2016

Symbolic key handover in the wind tunnel: Science Minister Gabriele Heinen-Kljajić (3rd from right) with University President Prof. Dr. Dr. Hans Michael Piper (2nd from right) and Chief Construction Director of the State Construction Management Lüneburger Heide, Michael Brassel. Also in the picture (from left): ForWind Managing Director Dr. Stephan Barth; wind energy expert Prof. Dr. Martin Kühn, turbulence researcher Prof. Dr. Joachim Peinke; State Secretary of the Lower Saxony Ministry for the Environment, Energy and Climate Protec- tion, Almut Kottwitz; Energy Programme Director at the German Aerospace Center (DLR), Bernhard Milow; Oldenburg Mayor Germaid Eilers-Dörfler

– 200 – Annual Report 2015-2017 EVENTS 2016 Annual Report 2015-2017201

GeCoLab – Test Bench for Generators and Converters Systems

Leibniz Universität Hannover converter/machine interactions is almost converters. The study of megawatt genera- Institute for Drive Systems and Power completed. GeCoLab stands for Generator- tors with or without the power electronic Electronics Converter-Laboratory. After the construc- is an ongoing important topic due to the tion phase in 2013, technical infrastructure very different phenomena and uncertain- Bernd Ponick, Axel Mertens and electrical components were supplied in ties of large electrical drive systems. The 2014. Power electronics, transformers and increasing complexity of electrical grids Funding: Federal Ministry for Economic Affairs and Energy (BMWi) power distribution followed by the end of and rapid emerging of frequency converter- 2014. With the hall nearly filled to capac- based generators encourage a much deeper Ref.Nr. 0325398 ity, the installation phase started in January study of the converter-generator interaction. 2015. Everything arrived in time and fitted GeCoLab has been created to deal with Duration: 01/2012 – 09/2016 into place. The last items – the electrical ma- these challenges. chines – arrived in early summer 2015, then we could start off with the operation step by The system diagram (see fig. 1) gives an step. Finally, first start-up tests under power impression on the universal possibilities of- were successfully made in autumn 2015. fered by the test bench facility, and this not Introduction only for the test machines and converters specifically equipped for testing, but also It is done! In the last two years, our vision Project Description for specimens, be it machines or converters, turned into reality: A new large-scale Uni- provided by customers. An impression of versal test bench GeCoLab to investigate The GeCoLab is a universal motor and the test bench shows the fig. 2. steady-state and dynamic properties of elec- generator test bench which enables a deep trical machines and converters including investigation of electrical machines and

Figure 1: Single line diagram

– 201 – 202EVENTS 2016 Annual Report 2015-2017

Equipment parameters currents rical short circuits. - rated voltage up to 690 V • Additional losses through harmonics - Carry out investigations on the converter- - permanent magnet synchronous machine • Winding defects and their diagnosis generator interactions and their influence with respective full converter on other system components, such as -1 [PN = 1.2 MW, nN = 375 min With the test bench it is possible bearings and gearboxes. (0 - 750 min-1)] - Test your motor or generator prototype - doubly-fed induction machine with with different types of converters for respective wind converter fault diagnostics, performance valida- Summary -1 [PN = 2.08 MW, nN = 1780 min ] tion, and analytical modelling and design - converter-based grid modelling method development GeCoLab offers an optimal solution for re-

[SN = 4.4 MVA; unbalance, freely adjust- - Carry out investigations on our genera- search on the electrical drive train of wind able frequency and harmonics] tion (PMSM and DFIG) or your com- turbines and Hydrogenerators. A quick - maximum component weight 20 t ponents using our sensor types, such as change of components is made possible by - base area of span 10 m x 4.3 m torque, position, speed, voltage, current, a tensioning field and on-site 20t crane. The flux, vibration, temperature, etc. machine foundation is decoupled from the Tests - Carry out investigations on both con- building by means of pneumatic spring ele- - steady-state and transient operating ventional and innovative converter and ments. In addition, a readily accessible ter-

conditions (Ψ, PV, M, U, I, ECD, ...) generator concepts including control minal box is installed to replace the inverter - temperature rise and loss distribution and filter design methods. This includes with a test object. Researchers on the men- - diagnostic methods investigations into dynamics and system tioned topics result in an improved valida- stability, stationary and transient thermal tion of analytical modelling, diagnostic pro- on electrical machines: loading, various methods of grid control cedures and advanced simulation model for • flux distribution, end-winding fields and the behaviour of grid faults, such as electrical and mechanical components for a • current displacement, circulating voltage dips, symmetrical and asymmet- better design of generator and converter.

Figure 2: Installed test bench References

[1] https://www.tth.uni-hannover. de/159.html

– 202 – Annual Report 2015-2017 EVENTS 2017 Annual Report 2015-2017203

Minister President of Lower Saxony Stephan Weil visits ForWind

ForWind Head Office ForWind at the University of Oldenburg. turbulence generators and scaled research On his summer trip, Weil took a look at wind turbines for this purpose, which Weil the new research infrastructure of the was impressed by. "Wind energy is one of Laboratory for Turbulence and Wind En- the supporting elements of the future en- "Lower Saxony is No. 1 of the Wind En- ergy Systems (WindLab), with its unique ergy supply," said the Minister President of ergy Lands and is also at the forefront of turbulence wind tunnel. The system makes Lower Saxony. "The University of Olden- wind energy research. I am delighted that it possible to reproduce and analyze the in- burg has been expanding research in this we have one of the top addresses for interaction between realistic turbulent wind field for many years and is making an ac- ternational research in this field here in flows on the one hand and wind turbines tive contribution to the success of the en- Oldenburg," said Minister President Ste- and wind farms on the other in the lab. ergy transition." phan Weil during his visit in June 2017 at The scientists have developed powerful

Minister President of Lower Saxony Stephan Weil (3rd from right) visits ForWind. Besides (from left to right) ForWind managing director Dr. Stephan Barth, wind energy expert Prof. Dr. Martin Kühn, Oldenburg’s Mayor Jürgen Krogmann, University President Prof. Dr. Dr. Hans Michael Piper and turbulence researcher Prof. Dr. Joachim Peinke.

Minister President Stephan Weil also inspected Oldenburg's Didn't fear the self-test: Minister President Weil, together with the new turbulence wind tunnel and the associated components. University President Prof. Dr. Dr. Hans Michael Piper and the wind energy expert Prof. Dr. Martin Kühn, stepped into the wind.

– 203 – 204EVENTS 2017 Annual Report 2015-2017

ForWind Symposia

Carl von Ossietzky Universität Oldenburg Research Group: Wind Energy Systems Martin Kühn, Joachim Peinke, ForWind Head Office Research Group: Energy Meteorology Detlev Heinemann, Stephan Barth Institute of Physics Research Group: Turbulence, Wind Energy and Stochastics

Symposium "Results from Wind Physics"

Wind turbines in wind farms and wind farm clusters influence each other through interaction with the atmospheric boundary layer. This has far-reaching consequences, in particular for the achievable energy yields and the turbine loads. Both the large offshore wind farms with nominal capacities of several 100 MW and the onshore wind farms have optimization potential in planning, monitoring and control. In order to give interested persons from industry and project partners an overview of current results and an insight in wind energy research projects, ForWind organized the Symposium “Results from Wind Phys- ics” ("Ergebnisse aus der Windphysik") on Participants of the ForWind Symposium "Results from Wind Physics" June 14th, 2017, in Oldenburg.

In cooperation with the partners involved, results from the following recently completed collaborative projects were presented:

• GW Wakes – Analysis of shadowing effects and wake turbulence characteristics of large Offshore-wind farms by comparison of alpha ventus and Riffgat • OWEA Loads – Probabilistic load description, monitoring and reduction of loads of future offshore wind energy converters • CompactWind – Increase of the Over- The current state of research was present- layer to plant loads, plant monitoring and all Energy Yield per Area at Sites Due ed and discussed on a cross-project basis. control strategies for individual plants and to the Utilization of Advanced Single The topics ranged from the interaction of wind parks. Turbine and Wind Farm Control Strat- wind parks with the atmospheric boundary egies

– 204 – Annual Report 2015-2017 EVENTS Annual Report 2015-2017205

Symposium "Perspectives on Turbulence and Wind Energy Research"

As part of the further development of the research concept of the Carl von Ossietzky University Oldenburg in the field of wind physics, ForWind organized a symposium with young scientists in the WindLab on December 1st, 2017. The focus was on the perspectives in turbulence and wind energy research.

The program covered almost the full range of topics from turbulence research to wind energy applications, from physical and meteorological to engineering questions as well as theoretical, numerical and experimental methods. Five longer lectures by young scientists were combined with four shorter in-house lectures. In addition, we were able to invite Prof. Marc Avila, Director of ZARM Bremen, to discuss the perspectives of further cooperation with the University of Bremen.

References

[1] https://symposium2017.forwind.de [2] https://symposium2017-2.forwind.de

– 205 – 206DOCUMENTATION Annual Report 2015-2017

Documentation

Publication Lists Bromm, M., Vollmer, L., Kühn, M.: Gebhardt, C. G.; Benedikt, H.; Hente, C.; Numerical investigation of wind Rolfes, R.: turbine wake development in Nonlinear dynamics of slender struc- Peer Reviewed Articles directionally sheared inflow tures: a new object-oriented frame- Wind Energy, 20:381–395 (2017), work published online 21 July 2016 Computational Mechanics, pp. 1-34 Albiker, J.; Achmus, M.; Frick, D.; (2018) Flindt, F.: Chattopadhyay, K.; Kies, A.; Lorenz, E.; 1g Model Tests on the Displacement von Bremen, L.; Heinemann, D.: Häfele, J.; Hübler, C.; Gebhardt, C. G.; Accumulation of Large-Diameter Piles The impact of different PV modu- Rolfes, R.: Under Cyclic Lateral Loading le configurations on storage and An improved two-step soil-structure Geotechnical Testing Journal 40 (2): additional balancing needs for a fully interaction modeling method for March 2017 renewable European power system dynamical analyses of offshore wind Renewable Energy, 113, pp. 176-189, turbines Andersen, S. J.; Witha, B.; Breton, S.-P.; doi:10.1016/j.renene.2017.05.069 (2017) Applied Ocean Research, 55, pp. 141-150 Sørensen, J. N.; Mikkelsen, R. F.; (2016) Ivanell, S.: Cordes, U.; Kampers, G.; Meißner, T.; Quantifying variability of Large Eddy Tropea, C.; Peinke, J.; Hölling, M.: Hartwig, S.; Marx, S.: Simulations of very large wind farms Note on the limitations of the Theo- Torsionstragverhalten extern vorge- J. Phys.: Conf. Ser., 625, 012027, doi: dorsen and Sears functions spannter Kreissegmente mit trockenen 10.1088/1742-6596/625/1/012027 (2015) Journal of Fluid Mechanics, vol. 811, 2017 Fugen Beton- und Stahlbetonbau 112, Anvari, M.; Werther, B.; Lohmann, G.; Dörenkämper, M.; Optis, M.; pp. 740-746 (2017) Wächter, M.; Peinke, J.; Beck, H.-P.: Monahan, A; Steinfeld, G.: Suppressing power output fluctuations On the offshore advection of bounda- Hartwig, S.; Marx, S.: of photovoltaic power plants ry-layer structures and the influence Torsionstragverhalten extern Solar Energy, vol. 157, pp. 735–743, 2017 on offshore wind conditions vorgespannter Kreissegmente mit Boundary-Layer Meteorol., 155, pp. trockenen Fugen Bastine, D.; Witha, B.; Wächter, M.; 459-482, doi:10.1007/s10546-015-0008-x Beton- und Stahlbetonbau 112, Peinke, J.: (2015) pp. 740-746 (2017) Towards a Simplified Dynamic Wake Model using POD Analysis Dörenkämper, M., Witha, B., Steinfeld, Haselsteiner, A. F.; Ohlendorf, J.-H.; Energies, 8, pp. 895-920, doi:10.3390/ G., Heinemann, D., Kühn, M.: Wosniok, W.; Thoben, K.-D. en8020895 (2015) The impact of stable atmospheric Deriving environmental contours from boundary layers on wind-turbine highest density regions Beck, H., Schmidt, M., Kühn, M.: wakes within offshore wind farms In: Coastal Engineering; Volume 123; Dynamic data filtering of long-range Journal of Wind Engineering and pp. 42-51; May 2017 Lidar wind speed measurements Industrial Aerodynamics, 144, Remote Sens., 9(6), p. 561 (2017) pp. 146-153 (2015) Hellsten, A.; Luukkonen, S. M.; Steinfeld, G.; Kanani-Sühring, F.; Markkanen, T.; Brandt, S.; Broggi, M.; Häfele, J.; Gauterin, E., Kammerer, P., Kühn, M., Järvi, L.; Lento, J.; Vesala, T.; Raasch, S: Gebhardt, C. G.; Rolfes, R.; Beer, M.: Schulte, H.: Footprint Evaluation for Flux and Meta-models for fatigue damage Effective wind speed estimation: Com- Concentration Measurements for estimation of offshore wind turbines parison between Kalman Filter and an Urban-Like Canopy with Coupled jacket substructures Takagi–Sugeno observer techniques Lagrangian Stochastic and Large-Eddy Procedia Engineering, Band 199, Seite(n): ISA Transaction 62, pp. 60-72 (2016) Simulation Models 1158-1163. DOI: 10.1016/j. Boundary-Layer Meteorol., 157(2), pp. proeng.2017.09.292 (2017) 191-217. https://doi.org/10.1007/s10546- 015-0062-4 (2015)

– 206 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017207

Herrmann, R.; Rabe, J.; Bolle, G.; Kies, A.; Schyska, B.U.; von Bremen, L.: Luhur, M. R., Peinke, J., Wächter, M., Marx, S.: Curtailment in a Highly Renewable Kühn, M.: Konzepte für Datenqualität und Daten- Power System and Its Effect on Stochastic Model for Aerodynamic abfrage bei Entwurf und Umsetzung Capacity Factors Force Dynamics on Wind Turbine von Monitoringsystemen Energies, 9(7), 510, doi:10.3390/ Blades in Unsteady Wind Inflow Bauingenieur 92, pp. 537-545 (2017) en9070510 (2016) Journal of Computational and Nonlinear Dynamics, Volume 10, Issue 4 (2015) Herrmann, R.; Rabe, J.; Bolle, G.; Kies, A.; Schyska, B.U.; von Bremen, L.: Marx, S.: The Effect of Hydro Power on the Mehrens, A. R.; von Bremen, L.: Konzepte für Datenqualität und Optimal Distribution of Wind and Solar On the correlation of spatial wind Datenablage bei Entwurf und Generation Facilities in a Simplified speed and solar irradiance variability Umsetzung von Monitoringsystemen Highly Renewable European Power above the North Sea Bauingenieur 92, pp. 537-545 (2017) System Advances in Science and Research, 13, Energy Procedia, 97, pp. 149-155 (2016) pp. 57-61, doi:10.5194/asr-13-57-2016 Hübler, C.; Gebhardt, C. G.; Rolfes, R.: (2016) Development of a comprehensive Kies, A.; Schyska, B.U.; von Bremen, L.: database of scattering environmental The optimal share of wave power in a Mehrens, A. R.; von Bremen, L.: conditions and simulation constraints highly renewable power system on the Comparison of methods for the iden- for offshore wind turbines Iberian Peninsula tification of mesoscale wind speed Wind Energy Science, 2, pp. 491-505 Energy reports, 2, pp. 221-228, fluctuations (2017) doi:10.1016/j.egyr.2016.09.002 (2016) Meteorol. Z., 26(3), pp. 333-342, doi:10.1127/metz/2017/0826 (2017) Hübler, C.; Gebhardt, C. G.; Rolfes, R.: Kies, A.; Schyska, B.U.; von Bremen, L.: Hierarchical four-step global The Demand Side Management Mittelmeier, N., Allin, J., Blodau, T., sensitivity analysis of offshore wind Potential to Balance a Highly Trabucchi, D., Steinfeld, G., Rott, A., turbines based on aerodynamic time- Renewable European Power System Kühn, M.: domain simulations Energies, 9(11), 955, doi:10.3390/ An analysis of offshore wind farm Renewable Energy, 111, pp. 878-891 en9110955 (2016) SCADA measurements to identify key (2017) parameters influencing the magnitude Knigge, C.; Raasch, S.: of wake effects Hümme, J.; von der Haar, C.; Lohaus, L.; Improvement and development of Wind Energ. Sci., 2, pp. 477-490 (2017) Marx, S.: one- and two-dimensional discrete Fatigue behaviour of a normal- gust models using a large-eddy Mittelmeier, N., Blodau, T. and Kühn, M.: strength concrete – number of cycles simulation model Monitoring offshore wind farm power to failure and strain development Journal of Wind Engineering and Industri- performance with SCADA data and Structural Concrete 17, pp. 637-645 al Aerodynamics, 153, pp. 46-59 (2016) advanced wake model (2016) Wind Energy Science 2, pp. 175-187 (2017) Lind, P. G.; Vera-Tudela, L.; Wächter, M.; Junk, C.; Späth, S.; von Bremen, L.; Kühn, M.; Peinke, J.: Oelker, S.; Lewandowski, M.; Alla, A. A.; Delle Monache, L.: Normal behaviour models for wind Ohlendorf, J. H.; Haselsteiner, A. F. Comparison and Combination of Regi- turbine vibrations: An alternative Logistikszenarien für die Errichtung onal and Global Ensemble Prediction approach von Offshore-Windparks – Systems for Probabilistic Predictions of arXiv:1710.08676 Herausforderungen der Hub-Height Wind Speed Wirtschaftlichkeitsbetrachtung neuer Weather and Forecasting, 30(5), pp. Lind, P., Vera-Tudela, L., Wächter, M., Logistikkonzepte 1234-1253, doi:10.1175/WAF-D-15-0021.1 Kühn, M., Peinke, J.: In: Industrie 4.0 Management; Volume 3; (2015) Normal behaviour models for wind No. 1; pp. 24-28; ISSN 2364-9208 (2017) turbine vibrations: Comparison of Junk, C.; Delle Monache, L.; neural networks and a stochastic Oelker, S.; Lewandowski, M.; Freitag, M.: Allessandrini, S.; Cervone, G.; von Bre- approach Konzept einer preagierenden Instand- men, L.: Energies 10; 12, p. 1944 (2017) haltungsstrategie – Effiziente Instand- Predictor-weighting strategies for haltung und automatisierte Logistik probabilistic wind power forecasting Lohaus, L.: in der Betriebsphase von Offshore- with an analog ensemble Untersuchungen zum Ermüdungsver- Windenergieanlagen Meteorol. Z., 24(4), pp. 361-379, halten eines höherfesten In: Industrie Management; Zeitschrift für doi:10.1127/metz/2015/0659 (2015) Normalbetons industrielle Geschäftsprozesse; Jahrgang Beton- und Stahlbetonbau 110, 31; Nr. 5; S. 40-44; ISSN 1434-1980 (2015) pp. 699-709 (2015)

– 207 – 208DOCUMENTATION Annual Report 2015-2017

Pauscher, L., Vasiljevic, N., Callies, D., Scholz, T.; Raischel, F.; Lopes, V. V.; Lehle, Tabrizi, A. B.; Whale, J.; Lyons, T.; Lea, G., Mann, J., Klaas, T., Hieronimus, B.; Wächter, M.; Peinke, J.; Lind, P. G.: Urmee, T.; Peinke, J.: J., Gottschall, J., Schwesig, A., Kühn, M., Parameter-free resolution of the Modelling the structural loading of a Courtney, M: superposition of stochastic signals small wind turbine at a highly An Inter-Comparison Study of Physics Letters A, vol. 381, no. 4, pp. turbulent site via modifications to the Multi- and DBS Lidar Measurements in 194–206, 2017 Kaimal turbulence spectra Complex Terrain Renewable Energy, vol. 105, pp. 288–300, Remote Sens., 8, 782 (2016) Schottler, J.; Hölling, A.; Peinke, J.; 2017 Hölling, M.: Perrone, F., Kühn M.: Brief communication: On the influence Thieken, K.; Achmus, M.; Lemke, K.: Offshore Wind Turbine Tower Fore-Aft of vertical wind shear on the combined A new static p-y approach for piles Fatigue Load Reduction by Coupling power output of two model wind with arbitrary dimensions in sand Control and Vibrational Analysis turbines in yaw geotechnik 38 (4), pp. 267-288 (2015) Journal of Ocean and Wind Energy, Vol. Wind Energy Science, vol. 2, no. 2, pp. 2, No. 3, pp. 168–175 (2015) 439–442, 2017 Thieken, K.; Achmus, M.; Lemke, K.; Terceros, M.: Reinke, N.; Homeyer, T.; Hölling, M.; Schottler, J.; Reinke, N.; Hölling, A.; Evaluation of p-y approaches for large Peinke, J.: Whale, J.; Peinke, J.; Hölling, M.: diameter monopiles in sand Flow modulation by an active grid On the impact of non-Gaussian wind International Journal of Offshore and arXiv:1703.00721, 2017 statistics on wind turbines-an Polar Engineering (IJOPE), Vol. 25, No. 2, experimental approach June 2015, pp. 134-144 Rockel, S.; Peinke, J.; Hölling, M.; Wind Energy Science, vol. 2, no. 1, p. 1, Bayoán Cal, R.: 2017 Trabucchi, D.; Vollmer, L.; Kühn, M.: Dynamic wake development of a Shear layer approximation of Navier- floating wind turbine in free pitch Schramm, M.; Rahimi, H.; Stokes steady equations for motion subjected to turbulent inflow Stoevesandt, B.; Tangager, K.: non-axisymmetric wind turbine wakes: generated with an active grid The Influence of Eroded Blades on description, verification and first Renewable Energy, vol. 112, pp. 1–16, Wind Turbine Performance Using application 2017 Numerical Simulations Journal of Physics: Conference Energies, vol. 10, no. 9, p. 1420, 2017 Series, 753, 032030, doi: 10.1088/1742- Santos-Alamillos, F. J.; Pozo-Vázquez, D.; 6596/753/3/032030 (2016) Ruiz-Arias, J.A.; von Bremen, L.; Schwack, F.; Stammler, M.; Poll, G.; Tovar-Pescador, J.: Reuter, A.: Trabucchi, D., Vollmer, L., Kühn, M.: Combining wind farms with Comparison of Life Calculations for 3-D shear-layer model for the concentrating solar plants to provide Oscillating Bearings Considering simulation of multiple wind turbine stable renewable power Individual Pitch Control in Wind wakes: description and first assess- Renewable Energy, 76, pp. 539-550, Turbines ment doi:10.1016/j.renene.2014.11.055 (2015) J. Phys.: Conf. Ser. 753 pp.112013 (2016) Wind Energy Science. 2, pp. 569-586. (2017) Schendel, A.; Hildebrandt, A.; Goseberg, Shimada, S.; Ohsawa, T.; Kogaki, T.; N.; Schlurmann, T.: Steinfeld, G.; Heinemann, D.: Traphan, D.; Wester, T. T. B.; Gülker, G.; Processes and evolution of scour Effects of Sea Surface Temperature Peinke, J.; Lind, P. G.: around a monopile by tidal currents Accuracy on Offshore Wind Resource Directed percolation in aerodynamics: Coastal Engineering, 139, pp. 65-84 Assessment Using a Mesoscale Model resolving laminar separation bubble (2018) Wind Energy, 18(10), pp. 1839-1854, on airfoils doi:10.1002/we.1796 (2015) DOI: 10.1103/PhysRevX.8.021015 Schmietendorf, K.; Peinke, J.; Kamps, O.: arXiv:1707.06832 The impact of turbulent renewable Späth, S.; von Bremen, L.; Junk, C.; energy production on power grid Heinemann, D.: Trujillo, J.J., Seifert, J.K., Würth, I., stability and quality Time-consistent calibration of short- Schlipf, D., Kühn, M.: The European Physical Journal B, vol. 90, term regional wind power ensemble Full-field assessment of wind turbine no. 11, p. 222, 2017 forecasts near-wake deviation in relation to yaw Meteorol. Z., 24(4), pp. 381-392, misalignment Schmoor, K.; Achmus, M.: doi:10.1127/metz/2015/0664 (2015) Wind Energ. Sci., 1, pp. 41-53 (2016) Optimum Geometry of Monopiles with Respect to the Geotechnical Design Journal of Ocean and Wind Energy Vol. 2, No. 1, February 2015, pp. 54-60

– 208 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017209

Tsiapoki, S.; Häckell, M.W.; Vollmer, L.; Lee, J. C.-Y.; Steinfeld, G.; von der Haar, C.; Marx, S.: Grießmann, T.; Rolfes, R.: Lundquist, J. K.: Ein additives Dehnungsmodell für Damage and ice detection on wind A wind turbine wake in changing ermüdungsbeanspruchten Beton turbine rotor blades using a structural atmospheric conditions: LES and lidar Beton- und Stahlbetonbau 112, pp. 31-40 health monitoring framework measurements (2017) DOI: 10.1177/1475921717732730. (2017) J. Phys.: Conf. Ser., 854, 012050, doi :10.1088/1742-6596/854/1/012050 (2017) Vorspel, L.; Schramm, M.; Stoevesandt, van Dooren, M., Trabucchi, D., Kühn, M.: B.; Brunold, L.; Bünner, M.: A Methodology for the Reconstruction Vollmer, L., Steinfeld, G., and Kühn, M.: A benchmark study on the efficiency of of 2D Horizontal Wind Fields of Wind Transient LES of an offshore wind unconstrained optimization algorithms Turbine Wakes Based on Dual-Doppler turbine in 2D-aerodynamic shape design Lidar Measurements Wind Energ. Sci., 2, pp. 603-614 (2017) Cogent Engineering, vol. 4, no. 1, p. Remote Sens. 8, 809 (2016) 1354509, 2017 von der Haar, C.; Hümme, J.; Marx, S.; van Dooren, M. F., Campagnolo, F., Lohaus, L.: Wahl, B.; Feudel, U.; Hlinka, J.; Wächter, Sjöholm, M., Angelou, N., Mikkelsen, T., Untersuchungen zum Ermüdungs- M.; Peinke, J.; Freund, J. A.: Kühn, M.: verhalten eines höherfesten Conditional Granger causality of diffu- Demonstration and uncertainty Normalbetons sion processes analysis of synchronised scanning lidar Beton- und Stahlbetonbau 110, pp. 699- The European Physical Journal B, vol. 90, measurements of 2-D velocity fields in 709 (2015) no. 10, p. 197, 2017 a boundary-layer wind tunnel Wind Energ. Sci., 2, pp. 329-341 (2017) von der Haar, C.; Marx, S.: Design Aspects of Concrete Towers for Vera-Tudela, L., Kühn, M.: Wind Turbines Analysing wind turbine fatigue load Journal of the South African Institution of prediction: The impact of wind farm Civil Engineering 57, pp. 30-37 (2015) flow conditions Renewable Energy Journal, 107, pp. von der Haar, C.; Marx, S.; Krompholz, R.: 352–360 (2017) Ultraschalluntersuchungen an statisch beanspruchten Betonproben Vollmer, L.; van Dooren, M.; Trabucchi, Beton- und Stahlbetonbau 110, D.; Schneemann, J.; Steinfeld, G.; pp. 759-766, (2015) Witha, B.; Trujillo, J.; Kühn, M.: First comparison of LES of an offshore von der Haar, C.; Marx, S.: wind turbine wake with dual-Doppler Untersuchungen zur Steifigkeit und lidar measurements in a German Ultraschallgeschwindigkeit dynamisch offshore wind farm beanspruchter Betonproben J. Phys.: Conf. Ser., 625, 012001, doi: Beton- und Stahlbetonbau 111, 10.1088/1742-6596/625/1/012001 (2015) pp. 141-148, (2016)

Vollmer, L., Steinfeld, G., Heinemann, D., von der Haar, C.; Marx, S.: Kühn, M.: Development of stiffness and ultra- Estimating the wake deflection sonic pulse velocity of fatigue loaded downstream of a wind turbine in concrete different atmospheric stabilities: An Structural Concrete 17, pp. 630-636, LES study (2016) Wind Energ. Sci., 1, pp. 129–141 (2016) von der Haar, C.; Hümme, J.; Marx, S.; Vollmer, L., Steinfeld, G., Heinemann, D., Knigge, C.; Raasch, S.: Kühn, M.: Improvement and development of The wake deflection downstream of a one- and two-dimensional discrete wind turbine in different atmospheric gust models using a large-eddy stabilities: An LES study simulation model Wind Energ. Sci., 1, 129-141 (2016) Journal of Wind Engineering and Industrial Aerodynamics, 153, pp. 46-59 (2016)

– 209 – 210DOCUMENTATION Annual Report 2015-2017

Articles, Conference Proceedings, Achmus, M.; Abdel-Rahman, K.; Schäfer, Bahrami, O.; Tsiapoki, S.; Kane, M.B.; Conference Contributions D.; Kuo, Y.-S.; Chung, C.-Y.; Tseng, Y.-H.: Lynch, J.P.; Rolfes, R.: Interaction Diagrams for Driven Steel Extraction of Environmental and Op- Piles under Cyclic Axial Loading erational Conditions of Wind Turbine 27th Int. Ocean and Polar Engineering from Tower Response Data for Im- Conference (ISOPE), San Francisco/USA, proved Structural Health Monitoring Abdel-Rahman, K.; Achmus, M.: June 25-30, 2017 the 11th International Workshop on Numerical Modelling of Pile Groups Structural Health Monitoring (IWSHM Composed of two open-ended Steel Achmus, M.; Abdel-Rahman, K.; Schaefer, 2017), Stanford University, California, Piles D.; Kuo, Y.-S.; Chung, C.-Y.; Tseng, Y.-H.: USA. (2017) GeoMEast – Sustainable Civil Capacity Degradation Method for Infrastructures: Innovative Infrastructure Driven Steel Piles under Cyclic Axial Bartschat, A.; Morisse, M.; Mertens, A.; Geotechnology, July 15-19, 2017, Loading. Wenske, J.: Sharm El-Sheikh, Egypt In Proceedings of the 27th International Analysis of Dynamic Interactions Ocean and Polas Engineering Conference between Different Drivetrain Achmus, M.; Lemke, K.; Abdel-Rahman, (pp. 362-368). San Francisco, Kalifornien, Components with a Detailed Wind K.; Kuo, Y.-S.: USA (2017). Turbine Model, The Science of Mak- Numerical Approach for the Derivation ing Torque from Wind (TORQUE 2016) of Interaction Diagrams for Piles under Achmus, M.; Terceros, M.; Thieken, K.: München,October 5-7, 2016 Cyclic Axial Loading Assessment of p-y Approaches for Proceedings of the Twenty-fifth (2015) Piles in Normally and Over- Basheer, N.; Abdel-Rahman, K.; International Ocean and Polar Consolidated Soft Clay Albiker, J.; Chakraborty, T.; Achmus, M.: Engineering Conference Kona, Big Island, 27th Int. Ocean and Polar Engineering Numerical Modelling of Offshore Rock Hawaii, USA, June 21-26, 2015 Conference (ISOPE), San Francisco/USA, Socketed Monopile Foundations June 25-30, 2017 15th Int. Conference on Structural and Achmus, M.; Thieken, K.; Lemke, K.: Geotechnical Engineering (ICSGE), Cairo/ Evaluation of a new p-y approach for Achmus, M.; Thieken, K.; tom Wörden, F.; Egypt, December 20-22, 2015 piles in sand with arbitrary dimensions Terceros, M.: 3rd International Symposium on Frontiers Assessment of pile length criteria for Bastine, D., Wächter, M., Peinke, J., in Geotechnics (ISFOG), Oslo, Norway, monopiles Trabbucchi, D., Kühn M.: June 10-12, 2015 8th Int. Offshore Site Investigation and Characterizing Wake Turbulence with Geotechnics Conference, Society of Staring Lidar Measurements Achmus, M.: Underwater Technology (SUT), September Wake Conference, Visby, Sweden. Journal Design of Foundation Elements for 12-14, 2017, London/UK of Physics: Conference Series 625, 012006 Offshore Wind Energy Converters (2015) Keynote lecture, 15. Int. Conference on Achmus, M.: Structural and Geotechnical Engineering Bearing behavior and design of Baust, S.: (ICSGE), Cairo/ Egypt, monopiles used as foundations for Global and local FE-modelling of rotor December 20-22, 2015 offshore wind energy converters bearings in Ansys and validation of Keynote lecture at the 3rd International results Achmus, M.; Gütz, P.; Gerlach, T.: Soil-Structure Interaction Symposium, 1st Young Tribological Researcher Numerical modeling of the behavior October 18-19, 2017, Izmir, Turkey Symposium, 2017, Hannover of bucket foundations in sand under transient tensile loading Albiker, J.; Achmus, M.; Thieken, K.: Beck, H., Trujillo, J.J., Kühn, M.: 6th Int. Conference on Structural Engi- Numerical Simulation of Cyclic Analysis of wake sweeping effects neering, Mechanics and Horizontally Loaded Piles Under based on load and long-range lidar Computation (SEMC), Cape Town/ South Special Loading Conditions measurements Africa, September 5-7, 2016 12th German Wind Energy Conference DEWEK 2015, Bremen, Germany, DEWEK 2015, May 19-20, 2015, Bremen Proceedings of the German Wind Energy Achmus, M.; Terceros, M.; Thieken, K.: Conference DEWEK (2015) Evaluation of p-y approaches for large Andersen, S. J.; Witha, B.; Breton, S.-P.; diameter monopiles in soft clay Sørensen, J. N.; Mikkelsen, R. F.; Benedikt, H.; Gebhardt, C. G.; Hente, C.; 26th Int. Ocean and Polar Engineering Ivanell, S.: Rolfes, R.: Conference (ISOPE), Rhodes/Greece, Quantifying Variability In LES Sparsity pattern extraction for June 26-July 2, 2016 Simulations of Very Large Wind Farms assembly of KKT-like matrices in Wake Conference 2015, Visby, Sweden, multibody dynamics June, 09-11, 2015 (oral) (2015) Proceedings in Applied Mathematics and Mechanics, 89th GAMM Annual Meeting (2018)

– 210 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017211

Berger, F., Kühn, M.: Bustamante, A., Vera-Tudela, L., Franke, J.; Richrath, M.; Thoben, K.-D. Investigating the aerodynamic Kühn, M.: Forscher automatisieren Rotorblatt- implications of slender wind turbine Evaluation of wind farm effects on fertigung blade design fatigue loads of an individual wind In: MM Maschinen Markt, DEWEK 2015, Bremen, Germany turbine CW Composites World; 35/2016; Proceedings of the German Wind Energy at the EnBW Baltic 1 offshore wind farm pp. 30-35; August 2016 Conference DEWEK (2015) Wake Conference, Visby, Sweden. Journal of Physics: Conference Series 625, 012020 Fuchs, A.; Reinke, N.; Guelker, G.; Boersma, S., Vali, M., Kühn, M., van (2015) Peinke, J.: Wingerden, J.-W.: Experimental investigations of flow Quasi Linear Parameter Varying Dörenkämper, M.; Junk, C.; von Bremen, fields generated by fractal and regular modeling for wind farm control using L.; Steinfeld, G.; Heinemann, D.; grids the 2D Navier-Stokes equations Kühn, M.: Progress in Turbulence VI, Proc. iTi ACC 2016; Boston, MA, USA 2016. Advances in monitoring, simulation Conference in Turbulence, Bertinoro, pp. 4409 – 4414 (2016) and short-term forecasting at the off- Springer, 2017 shore wind farm “EnBW Baltic 1” Böske, L.N.; Raasch, S.: DEWEK 2015, Bremen, Germany, 19 - 20 Gerendt, C.; Jansen, E.; Rolfes, R.: Large-eddy simulation as a tool for May 2015 (oral) (2015) A Fatigue Damage Model for Fiber onshore wind turbine site assessment Metal Laminates Including Fiber/ 4th International Education Forum Dubois, J.; Schaumann, P.; Thieken, K.; Matrix Degradation on Environment and Energy Science Achmus, M.: 6th ECCOMAS Thematic Conference on (ACEEES), Maui, Hawaii, USA Zum Ansatz der lastabhängigen the Mechanical Response of Composites 6-10 Dec. 2015 Gründungssteifigkeit in gesamt- (ECCOMAS COMPOSITES), dynamischen Simulationen von Eindhoven, Netherlands (2017) Böske, L.N.: Offshore Windenergieanlagen Large-eddy simulation as a tool for Seminar „Stahl im Wasserbau 2015“, Gütz, P.; Achmus, M.; Thieken, K.: site assessment in complex September 30-October 10, 2015, Model testing of suction bucket environment Braunschweig foundations under tensile loading in Wind Energy Science Conference, non-cohesive soil Kopenhagen, Danmark Dubois, J.; Thieken, K.; Terceros, M.; 8th Int. Offshore Site Investigation and June 26-29, 2017 Schaumann, P.; Achmus, M.: Geotechnics Conference, Society of Advanced Incorporation of Soil- Underwater Technology (SUT), Brandt, S.; Broggi, M.; Häfele, J.; Hübler, Structure Interaction into Integrated September 12th-14th 2017, London/UK C.; Gebhardt, C. G.; Beer, M.; Rolfes, R.: Load Simulation Meta-models for fatigue damage 26th Int. Ocean and Polar Engineering Gütz, P.; Achmus, M.; Schröder, C.: estimation of offshore wind turbines Conference (ISOPE), Rhodes/Greece, Comparison of methods for the jacket substructures. June 26-July 2, 2016 determination of the tensile bearing Procedia Engineering (199), pp. 1158-1163 behavior of suction bucket (2017) Ehrmann A.; Penner N.; Gebhardt C. G.; foundations in sand Rolfes R.: 3rd International Soil-Structure Interac- Bromm, M., Kühn, M.: Offshore Support Structures with tion Symposium, October 18th-19th 2017, Numerical investigation of wake Suction Buckets: Parameter Fitting of a Izmir, Turkey development in a stable atmospheric Simplified Foundation Model boundary layer in: Proceedings of the 26th International Häfele, J.; Rolfes, R.: DEWEK 2015, Bremen, Germany, Ocean and Polar Engineering Conference, Approaching the ideal design of jacket Proceedings of the German Wind Energy ISOPE 2016, Rhodes (Greece) substructures for offshore wind Conference DEWEK (2015) turbines with a Particle Swarm Ehrmann A.; Gebhardt C. G.; Rolfes R.: Optimization algorithm Buhr, R.; Aposporidis, A.; Munk, J.; A Hierarchical Clustering Approach for In Proceedings of the 26th International Steinfeld, G.; Heinemann, D.: Jacket Substructures for Offshore Wind Ocean and Polar Engineering Conference RANS simulations of flows above Farms according to Location- pp. 156-163. Rhodos, Griechenland (2016) forest canopies of different topologies Dependent Environmental Conditions Wind Energy Science Conference 2017, 1st Wind Energy Science Conference, Lyngby, Denmark, June 26-29, 2017 (oral) WESC 2017, Lyngby (Denmark) (2017)

– 211 – 212DOCUMENTATION Annual Report 2015-2017

Häfele, J.; Hübler, C.; Gebhardt, C. G.; Hübler, C.; Müller, F.; Gebhardt, C. G.; Kies, A. et al.: Rolfes, R.: Rolfes, R.: Trade-off between reduced storage An improved two-step soil-structure Global Sensitivity Analysis of Offshore and extended inter-country interaction modeling method for Wind Turbine Substructures transmission needs in an European dynamical analyses of offshore wind 15th International Probabilistic Workshop energy system dominated by turbines (2017) renewables Applied Ocean Research (55), pp. 141-150 12th German Wind Energy Conference (2016) Ivanell, S.; Arnqvist, J.; Witha, B.: (DEWEK 2015), Bremen, Germany, Ryningsnäs Benchmark – A forested May 19 – 20, 2015 (oral) (2015) Häfele, J.; Hübler, C.; Gebhardt, C. G.; site modelled with different microscale Rolfes, R.: model approaches Kies, A. et al.: Efficient Fatigue Limit State Design Wind Energy Science Conference 2017, The temporal development of storage Load Sets for Jacket Substructures Lyngby, Denmark, June 26 - 29, 2017 needs in the European energy Considering Probability Distributions (oral) (2017) transition of Environmental States DPG Spring Meeting, Berlin, Germany, In Proceedings of the 27th International Jacobsen, S., Lehner, S., Hieronimus, J., March 15 – 20, 2015 (oral) (2015) Ocean and Polar Engineering Conference Schneemann, J., Kühn, M.: pp. 167-173. San Francisco, Kalifornien, Joint offshore wind field monitoring Krüger, S.; Witha, B.; Heinemann, D.: USA (2017) with spaceborne SAR and platform- Investigation of the Development of based LiDAR measurements Internal Boundary Layers by Means of Häfele, J.; Hübler, C.; Gebhardt, C. G.; ISRSE36, Berlin, Germany (2015) Large-Eddy Simulation Rolfes, R.: EMS 2017, Dublin, Ireland, Reconsidering fatigue limit state load Kies, A.; Schyska, B.U.; von Bremen, L.: September 4 – 8, 2017 (oral) (2017) sets for jacket substructures utilizing How Much Can Demand Side Manage- probability distributions of ment Contribute to Balance a Fully Krutova, M.; Kies, A.; Schyska, B.U.; von environmental states Renewable European Power System? Bremen, L.: Proceeding of the Twenty-seventh EMS 2016, Trieste, Italy, The Smooting Effect in an Afro- International Ocean and Polar Engineer- September, 12 – 16, 2016 (oral) (2016) Eurasian Renewable Power Grid ing Conference San Francisco, CA, USA, EMS 2016, Trieste, Italy, June 25-30, 2017; ISBN: 978-1-880653-97-5 Kies, A. et al.: September 12 – 16, 2016 (oral) (2016) (2017) Backup, Storage and Transmission Esti- mates of a Supra-European Electricity Kuhnle, B., Kühn, K.: Hahmann, A.; Montornés, A.; Casso, P.; Grid with High Shares of Renewables Strukturelle Dämpfer – Ein alter Hut Volker, P.; Witha, B.: 14th Wind Integration Workshop, Brus- oder Innovation für große Bigger is better. Is it really? sels, Belgium, October 20 – 22, 2015 Windenergieanlagen? Wind Energy Science Conference 2017, (oral) (2015) 6. VDI - Fachtagung Schwingungen in Lyngby, Denmark, 26 - 29 June 2017 (oral) Windenergieanlagen 2015, Bremen, (2017) Kies, A. et al.: Germany (2015) Transmission, Storage and Backup Hansen, M.; Schmidt, B.; Kelma, Estimates for a Global Electricity Grid Kuhnle, B., Kühn, M.: S.; Schmoor, K.; Goretzka, J.: with High Shares of Renewables Unfavourable trends of rotor speed Probabilistic Assessment of the 11th Eawe PhD Seminar on Wind Energy and systems dynamics for very large Foundation of Offshore Wind Turbines in Europe, Stuttgart, Germany, offshore wind turbines –Analysis of Proceedings of the IABSE Conference - September 23 - 25, 2015 (oral) (2015) the 10MW INNWIND.EU reference Elegance in Structures, Nara, Japan, turbine May 13th-15th 2015 Kies, A. et al.: EWEA Offshore 2015, Copenhagen, Effects of Scandinavian hydro power Denmark (2015) Hübler, C.; Häfele, J.; Ehrmann, A.; on storage needs in a fully renewable Rolfes, R.: European power system for various Lieske, M.; Schendel, A.; Schulz, N.; Effective consideration of soil transmission capacity scenarios Stahlmann, A.; Hildebrandt, A.; characteristics in time domain EGU General Assembly 2015, Vienna, Schlurmann, T.: simulations of bottom fixed offshore Austria, April 12 – 17, 2015 (oral) (2015) Untersuchungen zur Hydro- und wind turbines Morphodynamik bei Wellen- In Proceedings of the 26th International Strömungs-Interaktion anhand Ocean and Polar Engineering Conference physikalischer Modellversuche (3D- (pp. 127-134). Rhodos, Griechenland Wellen-Strömungsbecken) (2016) In Tagungsband zum Kongress der Hafen- technischen Gesellschaft e.V. (pp. 163- 172). Bremen: Eigenverlag HTG (2015)

– 212 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017213

Lohaus, L.; Tomann, C.; Weicken, H.: Morisse,M.; Bartschat, A.; Wenske, J. ; Penner N.; Rolfes R.: Combined formwork and textile Mertens, A.: Soil Stiffness Evaluation and Bearing reinforcement system for a mineral Dependency of the Lifetime Estimation Behavior of a Suction Bucket Jacket corrosion protection layer for offshore of Power Modules in Fully Rated Wind Prototype application Turbine Converters on the Modelling Presentation at the International Modal In 11th International Symposiom of Fer- Depth of the Overall System Analysis Conference, IMAC 2017, Garden rocement and 3rd ICTRC International 18th European Conference on Power Elec- Grove (USA) Conference on Textile Reinforced Con- tronics and Applications (EPE'16-ECCE crete, RILEM Proceedings PRO98 Europe), Karlsruhe, Germany, Raasch, S.: (pp. 287-294). Aachen (2015) September 5.-9. 2016 Scalability of atmospheric turbulence simulations on next generation hybrid Mehrens, A. R.: Morisse,M.; Bartschat, A.; Wenske, J. ; computer architectures – Challenges Korrelation der Windgeschwindigkeit Mertens, A.: and perspectives über der Nord- und Ostsee Converter Lifetime Assessment for Heraeus-Seminar “Science Application 4. Fachtagung Energiemeteorologie 2016, Doubly-Fed Induction Generators for Exascale Computing - Exploring New Bremerhaven, Germany, April 20 – 22 Considering Derating Control Avenues towards Scalability and Fault- (oral) (2016) Strategies at Low Rotor Frequencies Tolerance”, Bad Honnef, Germany TORQUE 2016, München, Germany, September 9, 2015 Mehrens, A. R.; von Bremen, L.; Oktober 5.-7, 2016 Heinemann, D.: Raasch, S.: Characterization of Mesoscale Wind Neunaber, I.; Schottler, J.; Peinke, J.; Simulation of atmospheric turbulence Fluctuations in space and time Hölling, M.: 10th Spring School on Transition and 12th German Wind Energy Conference Comparison of the development of Turbulence (EPTT 2016), São José dos (DEWEK 2015), Bremen, Germany, a wind turbine wake under different Campos, Brazil May 19 – 20, 2015 (oral) (2015) inflow conditions September 22, 2016 in Progress in Turbulence VII, Springer Mehrens, A. R.; von Bremen, L.; Proceedings in Physics 196, pp. 177-182, Raasch, S.: Heinemann, D.: 2017 New trends in high resolution large- Mesoscale wind fluctuations and their eddy simulation of the atmospheric correlation with solar variability Ohlendorf, J.-H.; Franke, J.; Schmohl, T.; boundary layer: From fundamental to 15th EMS Annual Meeting, Sofia, Thoben, K.-D. applied research Bulgaria, September 07 – 11, 2015 (oral) Automation strategies in rotor blade Center for Mechanics, Uncertainty and (2015) production: preforming vs. direct Simulation in ENgineering – MUSEN, TU textile layup Braunschweig, Germany Mittelmeier, N., Blodau, T., Kühn, K.: In: JEC Composite Magazin; Volume 99; June 22, 2017 Offshore Wake Model Validation pp. 35-37; ISSN 1639-965X; 2015 – Methodology for Linking Model Raasch, S.: Results and Operational Data Ohlendorf, J.-H.; Franke, J.; Rolbiecki, M.; High resolution atmospheric turbu- DEWEK 2015, Bremen, Germany, Schmohl, T.; Thoben, K.-D.; lence simulations for applied problems Proceed-ings of the German Wind Energy Ischtschuk, L. : SIAM conference on mathematical and Conference DEWEK (2015) Preforming in großer Dimension – computational issues in the geosciences, innovativer Ansatz in der Erlangen, Germany Mittelmeier, N., Blodau, T., Steinfeld, G., Rotorblattfertigung September 13, 2017 Rott, A., Kühn, M.: In: Konstruktion; Zeitschrift für Produk- An analysis of offshore wind farm tentwicklung und Ingenieur-Werkstoffe; Richrath, M.; Franke, J.; Ohlendorf, J.-H.; SCADA measurements to identify key Volume 68; No. 3; pp. 75-79; ISSN 0720- Thoben, K.-D. parameters influencing the magnitude 5953; 2016 Effektor für die automatisierte of wake effects Direktablage von Textilien in der Journal of Physics: Conference Series. Ohsawa, T.; Orita, T.; Kozai, K.; Rotorblattfertigung 2016; 753(3):032052 (2016) Shimada, S.; Steinfeld, G.; In: Lightweight Design; Volume 10; No. 5; Heinemann, D.: pp. 48-53; ISSN 1865-4819 (Print) / 2192- Monner, H. P.; Huxdorf, O.; Pohl, M.; Accuracy estimation of hub-height 8738 (Online); 2017 Riemenschneider, J.; Homeyer, T.; wind speeds estimated from scatter- Hölling, M.: ometer and mesoscale model Richrath, M.; Franke, J.; Ohlendorf, J.-H.; Smart structures for wind energy EWEA Offshore 2015, Copenhagen, Den- Thoben, K.-D. turbines mark, 10-12 Mar 2015 (poster) (2015) Effector for automated direct textile 25th AIAA/AHS Adaptive Structures placement in rotor blade production Conference, AIAA SciTech Forum, Grape- In: Lightweight Design worldwide; Vol- vine, Texas, January 9-13, 2017 ume 10; No. 5; pp. 42-47; 2017

– 213 – 214DOCUMENTATION Annual Report 2015-2017

Rott, A., Boersma S., van Wingerden, Schneemann, J., Bastine, D., v. Dooren, Schyska, B.U.; Kies, A.; von Bremen, L.: J.-W., Kühn, M.: M., Schmidt, J., Steinfeld, G., Trabucchi, Deriving an optimal Grid Design for Dynamic Flow Model for Real-Time D., Trujillo, J.J., Vollmer, L., Kühn, M.: integrating Offshore Wind Energy Application in Wind Farm Control “GW WAKES”: measurements of wake successfully Journal of Physics: Conference Series, effects in »alpha ventus« with syn- Integration of Sustainable Energy Confer- Volume 854 (2017) chronised longe-range LiDAR winds- ence, Nuremberg, Germany, July 11 – 12, canners 2016 (oral) (2016) Schendel, A.; Hildebrandt, A.; DEWEK, Bremen, Germany, Proceedings Schlurmann, T.: of the German Wind Energy Conference Schyska, B.U.; von Bremen, L.: Experimental study on the progression DEWEK (2015) Optimizing Power System Integration of scour around a monopile in based on the Energy Ratio unidirectional and tidal currents Schneemann, J., Hieronimus, J., EAWE PhD Seminar, Stuttgart, Germany, In Proceedings of the 6th International Jacobsen, S., Lehner, S., Kühn, M.: September 23 – 25, 2015 (oral) (2015) Conference on the Application of Physical Offshore wind farm flow measured by Modelling in Coastal and Port Engineer- complementary remote sensing Schyska, B.U.; Kies, A.; von Bremen, L.: ing and Sciene (Coastlab16). Ottawa, techniques: radar satellite TerraSAR-X Curtailment Affected Capacity Factors Kanada (2016) and lidar windscanners – The Role of Storages Wake Conference, Visby, Sweden. Journal International Renewable Energy and Stor- Schmidt, B.; Hansen, M.; Marx, S.: of Physics: Conference Series 625, 012015 ages Conference, Düsseldorf, Germany, Directional Dependence of Extreme (2015) March 15 – 17, 2016 (Poster) (2016) Load Parameters for Offshore Wind Turbines Schottler, J.; Mühle, F.; Bartl, J.; Peinke, Schyska, B. et al.: Proceedings of the 25th International J.; Adaramola, M. S.; Sætran, L.; Facing Europe: Revised wind power Offshore and Polar Engineering Hölling, M.: upscaling algorithms Conference (ISOPE), Kona, Big Island, Comparative study on the wake DPG Spring Meeting, Berlin, Germany, Hawaii, USA, 2015 deflection behind yawed wind turbine March 15 – 20 , 2015 (2015) models Schmidt, B.; Marx, S.; Hansen, M.: Journal of Physics: Conference Series, Schyska, B.; von Bremen, L.: Measurement Based Investigation of Volume 854, conference 1, 2017 Conviviality in Energy: Designing a Re- Directional Dependence of Extreme newable Energy System in a Degrowth Load Parameters for Offshore Wind Schröder, K.; Gebhardt, C.G.; Rolfes, R: Society Turbines Damage Localization at Wind Turbine 11th biennial conference of the European Proceedings of the 12th German Wind Support Structures Using Sequential Society for Ecological Economics, Leeds, Energy Conference (DEWEK), Bremen, Quadratic Programming for Model UK, June 30 – July 03, 2015 (oral) (2015) Germany, May 19th-20th 2015 Updating Proceedings of the 8th European Work- Seifert, J., Vera-Tudela, L., Kühn, M.: Schmidt, M., Trujillo, J.J., Kühn, M.: shop on Structural Health Monitoring Data analysis via clustering Orientation correction of wind (2016) exemplified by wind turbine wake direction measurements by means of identification staring lidar Schwack, F.; Stammler, M., Flory, H.; 12th EAWE PhD Seminar on Wind Energy J. Phys.: Conf. Ser. 749 012005 (2016) Poll, G.: in Europe, May 25-27, 2016, Lyngby Free Contact Angles in Pitch Bearings Denmark Schmoor, K.; Achmus, M.; Foglia, A.; and their Impact on Contact and Stress Wefer, M.: Conditions Seifert, J., Vera-Tudela, L., Kühn, M.: Reliability of design approaches for WindEurope Conference, Hamburg (2016) Training requirements of a neural net- axially loaded offshore piles and its work used for fatigue load estimation consequences with respect to the Schyska, B.U.; Couto, A.; von Bremen, L.; of offshore wind turbines North Sea Estqueiro, A.: EERA DeepWind – 14th Deep Sea Off- 15th International Conference of the Weather Dependent Upscaling Algo- shore Wind R&D Conference, Trondheim, International Association for Computer rithms for European Wind Power Norway. Energy Procedia 137; 315-322 Methods and Advances in EMS 2016, Trieste, Italy, September 12 – (2017) Geomechanics (15th IACMAG), October 16, 2016 (oral) (2016) 19-23, 2017, Wuhan, China

– 214 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017215

Shimada, S.; Ohsawa, T.; Takeyama, Y.; Stammler, M.; Wenske, J.: Thieken, K.: Kogaki, T.; Steinfeld, G.; Heinemann, D.: Integration von Schadensmechanimus- Entwicklung neuer p-y- Ansätze für Comparison of wind speeds simu- analyse und Blattlagertests in den Pfähle beliebiger Abmessungen lated with WRF using seven planetary Entwicklungsprozess von WKA In Baugrundtagung – Beiträge zur Spezi- boundary layer schemes at two off- VDI-Fachtagung Gleit- und Wälzlagerun- alsitzung – Forum für junge Geotechnik- shore met masts in the North Sea gen, 5/7/2015, Schweinfurt Ingenieure. Berlin (2016). EWEA Offshore 2015, Copenhagen, Denmark, March 10 – 12, 2015 (poster) Stammler, M.; Reuter, A.: Thieken, K.; Achmus, M.; Lemke, K.: (2015) Blade bearings: damage mechanisms Bewertung statischer p-y Ansätze für and test strategies horizontal belastete Pfähle beliebiger Shirzadeh, R. Kühn, M.: Conference for Wind Power Drives, Durchmesser in nichtbindigem Boden Application of two passive strate- 3/4/2015 - 3/4/2015, Aachen Pfahlsymposium 2015, Braunschweig, gies on the load mitigation of large February 19-20, 2015 offshore wind turbines Steinfeld, G.; Schmidt, M.; Journal of Physics: Conference Series 749; Heinemann, D.: Thieken, K.; Achmus, M.; Terceros, M.; 012011 (2016) WASA – Wind- und Stabilitätsatlas für Dubois, J.; Gerlach, T.: die südliche Nordsee Describing Six-Degree-of-Freedom Shirzadeh, R., Kühn, M.: 4. Fachtagung Energiemeteorologie, Response of Foundations Supporting Support structure load mitigation of Bremerhaven, Germany, OWEC Jacket Structures a large offshore wind turbine using April 20 - 22, 2016 (oral) (2016) 26th Int. Ocean and Polar Engineering a semi-active magnetorheological Conference (ISOPE), Rhodes/Greece, damper Stümpel, M.; Marx, S.: pp.745-753, June 26-July 2, 2016 EERA DeepWind – 14th Deep Sea Off- An Innovative Hybrid Substructure shore Wind R&D Conference, Trondheim, Made of High-Strength Concrete and Tomann, C.; Schack, T.; & Lohaus, L.: Norway. Energy Procedia (2017) Ductile Cast Iron for Offshore Wind Real Offshore Exposure Tests of a Min- Turbines eral Corrosion Protection System Shrestha, B., Kühn, M.: Proceedings of the 13th German Wind In Proceedings of the 27th International Adaptation of Controller Concepts for Energy Conference (DEWEK), Bremen, Offshore and Polar Engineering Support Structure Load Mitigation of Germany, Oct. 17-18, 2017 Conference. San Francisco, Kalifornien, Offshore Wind Turbines USA (2017) 13th Deep Sea Offshore Wind R&D Stümpel, M.; Marx, S.; Schaumann, Conference, EERA DeepWind’ 2016, P. Seidl, G. Göhlmann, J.: Trabucchi, D., Steinfeld, G., Bastine, D., Trondheim, Norway, Energy Procedia An Innovative Hybrid Substructure for Trujillo, J.J., Schneemann, Kühn, M.: (2016) Offshore Wind Turbines Study of wake meandering by means Proceedings of the 27th International of fixed point lidar measurements: Shrestha, B., Kühn, M.: Offshore and Polar Engineering Confer- Spectral analysis of line-of-sight wind Tailoring the Employment of Offshore ence (ISOPE), San Francisco, USA, component Wind Turbine Support Structure Load June 25-30, 2017 Wake Conference, Visby, Sweden. Journal Mitigation Controllers of Physics: Conference Series 625, 012016 Journal of Physics: Conference Series; Terceros, M.; Achmus, M.; Thieken, K.: (2015) 753(5):052032 (2016) Evaluation of p-y approaches for piles in soft clay Trabucchi, D., Vollmer, D., Kühn, M.: Späth, S.; von Bremen, L.; Junk, C.: 8th Int. Offshore Site Investigation and Shear layer approximation of Navier- Time-consistent calibration of regional Geotechnics Conference, Society of Stokes steady equations for non- wind power ensemble forecasts based Underwater Technology (SUT), axisymmetric wind turbine wakes: on COSMO-DE EPS September 12-14, 2017, London/UK Description, verification and first COSMO/CLM/ART User Seminar 2015, application Offenbach, Germany, March 06 - 08, 2014 Thieken, K.; Achmus, M.; Lemke, K.: Journal of Physics: Conference Series; (oral) (2015) Evaluation of a new p-y approach for 753(3):032030 (2016) piles in sand with arbitrary dimensions Späth, S. et al.: In Proceedings of the 3rd International Trabucchi, D., Trujillo, J. J., Ritter, K., Ways to generate probabilistic wind Symposium in Frontiers in Offshore Geo- Steiner, J., Kühn, M.: power forecasts from ensemble technics (pp. 741-746). Oslo, Norwegen Nacelle Based Lidar Measurements for weather forecasts (2015) the Characterisation of the Wake of an Wind Integration Workshop 2015, Offshore Wind Turbine under Different Offenbach, Germany, Atmospheric Conditions March 06 - 08, 2015 (oral) (2015) EERA DeepWind – 14th Deep Sea Off- shore Wind R&D Conference, Trondheim, Norway, Energy Procedia 137; pp. 77-88 (2017) – 215 – 216DOCUMENTATION Annual Report 2015-2017

Trujillo, J. J., Beck, H., Müller, K., Unguran, R., Kühn, M.: van Dooren, M., Trabucchi, D., Beck, H., Cheng, P. W., Kühn, M.: Combined individual pitch and trailing Schneemann, J., Friedrichs, W., Kühn, M.: A test case of meandering wake edge flap control for structural load Assessment of the global wind and the simulation with the Extended-Disk alleviation of wind turbines local wake direction at alpha ventus Particle model at alpha ventus ACC 2016; Boston, MA, USA; using long range scanning LiDAR EERA DeepWind – 14th Deep Sea Off- pp. 2307 – 2313 (2016) EWEA Offshore 2015, PO.ID 183 shore Wind R&D Conference, Trondheim, Copenhagen, Denmark (2015) Norway. Energy Procedia (2017) Vali; M., Wingerden, J.-W. van, Boersma, S., Petrovic, V., Kühn, M.: Vera-Tudela, L., Kühn, M.: Tsiapoki, S.; Häckell, M.W.; Rolfes, R.: A predictive control framework for Evaluation of a wind turbine fatigue Monitoring of a 35 m Wind Turbine optimal energy extraction of wind load monitoring system based on Rotor Blade During a Fatigue Test by a farms standard SCADA signals in different Modular SHM-Scheme Journal of Physics: Conference Series wind farm flow conditions The 10th International Workshop on 753; 052013 (2016) DEWEK 2015, Bremen, Germany, Structural Health Monitoring (IWSHM Proceedings of the German Wind Energy 2015), Stanford University, California, Vali; M., Wingerden, J.-W. van, Kühn, M.: Conference DEWEK (2015) USA. DOI: 10.12783/SHM2015/347 (2015) Optimal Multivariable Individual Pitch Control for Load Reduction of Large Vollmer, L., van Dooren, M., Trabucchi, Tsiapoki, S.; Krause, T.; Häckell, M.W.; Wind Turbines D., Schneemann, J., Steinfeld, G., Ostermann, J.; Rolfes, R.: ACC 2016; Boston, MA, USA; Witha, B., Trujillo, J., Kühn, M.: Combining a vibration-based SHM- pp. 3163 – 3169 (2016) First comparison of LES of an offshore scheme and an airborne sound wind turbine wake with dual-Doppler approach for damage detection on Vali, M, Petrovic, V., Boersma, S., lidar measurement in a German off- wind turbine rotor blades Wingerden, J.-W., Kühn, M.: shore wind farm In the 8th European Workshop on Adjoint-based model predictive Wake Conference, Visby, Sweden. Journal Structural Health Monitoring (EWSHM control of wind farms: Beyond the of Physics: Conference Series 625, 012001 2016), Bilbao, Spain. (2016) quasi steady-state power (2015) maximization Tsiapoki, S.; Lynch, J.P.; Häckell, M.W.; IFAC PapersOnLine 50-1; 4510-4515 (2017) Vollmer, L.; van Dooren, M.; Rolfes, R.: Trabucchi, D.; Steinfeld, G.; Kühn, M.: Improvement of the damage detection Valldecabres, L., Friedrichs, W., von Comparison of dual-Doppler lidar performance of a SHM framework Bremen, L., Kühn, M.: measurements and Large Eddy using AdaBoost: Validation on an Analysis of wind field fluctuations on Simulations of an offshore wind operating wind turbine wind farm power output turbine wake 11th International Workshop on 12th EAWE PhD Seminar on Wind Energy RAVE Offshore Wind R&D 2015, Bremer- Structural Health Monitoring (IWSHM in Europe, 25-27 May 2016, Lyngby haven, Germany, October 13 - 15, 2015 2017), Stanford University, California, Denmark (oral) (2015) USA. (2017) Valldecabres, L., Friedrichs, W., von von Bremen, L. et al.: Unger, R.; Gerendt, C.; Daum, B.; Bremen, L., Kühn, M.: Advances in monitoring, simulation Jansen, E.; Rolfes, R.; Ueing, C.; Spatial-temporal analysis of coherent and short-term forecasting at the off- Englisch, N.; Sayer, F.; Exner, W.; offshore wind field structures shore wind farm “EnBW Baltic 1” Mahrholz, T.; Lorsch, P.; Wierach, P.: measured by scanning Doppler-lidar 12th German Wind Energy Conference Hybrid Laminates and Nanoparticle Journal of Physics: Conference Series; (DEWEK 2015), Bremen, Germany, Reinforced Materials for Improved 753(7):072028. (2016) May 19 - 20, 2015 (oral) (2015) Rotor Blade Structures Advances in Rotor Blades for Wind van Dooren, M., Kühn, M., Petrovic, V., von Bremen, L.; Junk, C.; Späth, S.; Turbines, Bremen, April 5-7 2017, Bremen, Bromm, M., Botasso, C. L., Camapgnolo, Heinemann, D.: Germany F., Sjöholm, M., Angelou, N., Mikkelsen, Statistical methods for the develop- T., Croce, A., Zasso, A.: ment of skillful probabilistic wind Unger, R.; Daum, B.; Braun, U.; Rolfes, R.: Demonstration of synchronised scan- power forecasts Experimentally calibrated modeling ning lidar measurements of 2D velocity 3rd International Conference Energy & technique for the cross linking mecha- fields in a boundary-layer wind tunnel Meteorology, Boulder, Colorado, USA, nism of epoxy resin and its influence Journal of Physics: Conference Series; June 22 - 26, 2015 (oral) (2015) on mechanical properties, 753(7):072032. (2016) 20th International Conference on Com- posite Structures (ICCS), Paris, France (2017)

– 216 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017217

von Bremen, L.; Kies, A.; Schyska, B.; Weber, S.; Schlüter, M.; Borowski, D.; Witha, B.; Steinfeld, G.; Drüke, S.; Heinemann, D.: Mertens, A.: Günther, R.; Schmidt, M.; Heinemann, D.: Analysis of 100% Renewable Power Simple analog detection of turn-off WASA – A Wind and Stability Atlas Systems: Deployment Strategies and delay time for IGBT junction Providing Information on Long-Term efficient Feed-in Monitoring temperature estimation Atmospheric Conditions in the 3rd International Conference Energy & 2016 IEEE Energy Conversion Congress Southern North Sea Meteorology, Boulder, Colorado, USA, and Exposition (ECCE), Milwaukee, WI, 3rd International Conference Energy & June 22 -26, 2015 (oral) (2015) USA, September 18.-22. 2016 Meteorology, Boulder, Colorado, USA, June 22 - 26, 2015 (oral) (2015) von der Haar, C.; Marx, S.: Wernitz, S.; Pache, D.; Grießmann, T.; Strain Development of plain High Rolfes, R.: Wurps, H.; Steinfeld, G.; Heinemann, D.: Strength Concrete under Fatigue Performance of an H∞ estimation Optimization of the Mellor-Yamada Loading based damage localization approach scheme in the meso-scale model WRF Proceedings of the 4th International in the context of automated Structural for offshore wind conditions based on Symposium on Ultra-High Performance Health Monitoring measurements and large eddy simula- Concrete and High Performance Proceedings of the 9th European tions Construction Materials (HiPerMat), Workshop on Structural Health EMS 2016, Trieste, Italy, Kassel, Germany, March 9-11, 2016 Monitoring (2018) September 12 - 16, 2016 (oral) (2016)

von der Haar, C.; Wedel, F.; Marx, S.: Wester, T.; Traphan, D.; Gülker, G.; Wurps, H.; Tambke, J.; Bye, J. A. T.; Wolff, Numerical and Experimental Peinke, J.: J.-O.; Steinfeld, G.: Investigations of the Warming of Percolation: Statistical description of Necessary Accuracy of Offshore Wind Fatigue-Loaded Concrete a spatial and temporal highly resolved and Turbulence Measurements for the Proceedings of the fib Symposium 2016, boundary layer transition Improvement of Micrometeorological Cape Town, South Africa, Progress in Turbulence VI, Proc. iTi Models November 21-23, 2016 Conference in Turbulence, Bertinoro, 3rd International Conference Energy & Springer, 2017 Meteorology, Boulder, Colorado, USA, Wagner, D.; Witha, B.; Wurps, H.; June 22 -26, 2015 (oral) Flügge, M.; Reuder, J.: Witha, B.; Vollmer, L.; Steinfeld, G.: Evolution and propoerties of Low Level The PALM wind turbine model: An LES Wurps, H.; Steinfeld, G.; Tambke, J.: Jet events over the southern North Sea tool for modelling wind turbine wakes Influence of PBL schemes and rea- Wind Energy Science Conference 2017, in the atmospheric boundary layer nalyses data on wind fields above the Lyngby, Denmark, June 26 - 29, 2017 Wind Energy Science Conference 2017, North Sea simulated with the meso- (oral) (2017) Lyngby, Denmark, June 26 - 29, 2017 scale model WRF (oral) (2017) COSMO/CLM/ART User Seminar 2015, Wagner, D.; Steinfeld, G.; Witha, B.; Offenbach, Germany, March 06 - 08, 2015 Wurps, H.; Krüger, S.; Reuder, J.: Witha, B.; Dörenkämper, M.; (poster) (2015) Frequency and evolution of Low Level Heinemann, D.: Jet events over the southern North Sea The effect of heterogeneous land-use analysed from WRF simulations and on the land-sea transition and power LiDAR measurements production of offshore wind farms EMS 2017, Dublin, Ireland, EUROMECH Colloquium 576: Wind farms September 04 - 08, 2017 (oral) (2017) in complex terrains, Stockholm, Sweden, June 08 - 10, 2016 (oral) (2016) Wagner, D.; Steinfeld, G.; Witha, B.; Flügge, M; Reuder, J.: Witha, B.; Steinfeld, G.; Heinemann, D.: Low Level Jets over FINO1 - Observa- Turbulent Wake Flow Characteristics tions and Modelling Investigated with LES IRP Wind Conference 2017, Amsterdam, 3rd International Conference Energy & Netherlands, September 25 - 26, 2017 Meteorology, Boulder, Colorado, USA, (oral) (2017) June 22 - 26, 2015 (oral) (2015)

– 217 – 218DOCUMENTATION Annual Report 2015-2017

Project Reports Ehrich, S.; Lind, P.; Shrestha, B.; Steinfeld, Kühn, M., Cheng, P.W. (Hrsg.): G.; Tambke, J.; Trujillo, J.J.; Wächter, M.; Regelung von Offshore-Windparks Vera Tudela, L.; Wurps, H.; Zadorozhnyy, durch lokale Leistungsprognosen A.; Kühn, M. (Eds.): sowie Monitoring der Leistungs- und Final report of the research project Belastungscharakteristik (Baltic I) Achmus, M.; Abdel-Rahman, K.: RAVE "Probabilistische Last- Final report of the BMWi research project Ermittlung des Pfahlgruppeneffektes beschreibung, Monitoring und 0325215A, Universität Oldenburg (2015) einer Zweiergruppe offener Stahl- Reduktion der Lasten zukünftiger rohrpfähle Offshore-Windenergieanlagen“ Marx, S.; Hansen, M.; von der Haar, C.; Project report for Tennet TSO GmbH, OWEA Loads (2017) Diederley, J.; Lohaus, L.; Hümme, J.: November 01, 2016 Entwicklung und versuchstechnische Foglia, A.; Gintautas, T.; Hübler, C.; Erprobung von ermüdungsfesten Aman, T. Z.; Bastine, D.; Beck, H.; Schmoor, K.; Sørensen, J.D.: Gründungskonstruktionen aus Beton Dörenkämper, M.; van Dooren, M. F.; Report on geotechnical tests with für Offshore-Windenergieanlagen Friedrichs, W.; Funcke, P. G.; model structures: Work Package 7.2 - Final report of research project ProBeton Heinemann, D.; Hieronimus, J.; Kühn, D72.2 (BMWi-FKZ 0325670), p. 119 M.; Peinke, J.; Reuter, R.; Schmidt, J.; EU project IRPWIND, pp. 1-43 Schmidt, M.; Schneemann, J.; Meyer Forsting, A.R.; Verelst, D.R.; Steinfeld, G.; Stoevesandt, B.; Grießmann, T.; Hansen, K.S.; Merz, K.; Kvittem, M.; Kelma, S.: Trabucchi, D.; Trujillo, J. J.; Vollmer, L.; Karimirad, M.; Schröder, K.; Tande, J.O.: Roadmap for offshore design codes: Voß, S.; Wächter, M. & Witha, B. Kühn, Upload of the formatted data (floating Work Package 6.2 – Deliverable D62.8 M. & Schneemann, J. (Eds.): WT) to the database: Work Package 6.1 EU project IRPWIND, pp. 1-30 Final report of the research project – Deliverable D61.3 "Analyse der Abschattungsverluste EU project IRPWIND, pp. 1-10 Natarajan, A.; Ehrmann, A; Hübler, C.:, und Nachlaufturbulenzcharakteristika Sørensen, J.D.; Schmoor, K; Dimitrov, N.; großer Offshore-Windparks durch Ver- Häfele, J.; Gebhardt, C.G.; Rolfes, R.: Pfaffel, S.; Faulstich, S.; Gintautas, T.; gleich von alpha ventus und Riffgat“ Innovative design of a hybrid-type Wandji, W.N.: GW Wakes (2017) jacket for 10MW turbines Probabilistic: Work Package – D7 4.2 Hannover, 2017 EU project IRPWIND, pp. 1-111 Bedon, G.; Grießmann, T.; Hansen, K.S., Schröder, K.; Tu, Y.: Hansen, M.; Schmidt, B.: Ohlendorf, J.-H.; Rolbiecki, M.; Schmohl, Upload of the formatted data (fixed Safety of Offshore Wind Turbines T.; Franke, J.; Ischtschuk, L.: WT) to the database: Work Package 6.1 (WP 1) mapretec – ein Verfahren zur – Deliverable D61.4, Marx, S. (ed.): Probabilistic Safety preform-Herstellung durch ebene EU project IRPWIND, pp. 1-40 Assessment of Offshore Wind Turbines. Ablage für ein räumliches Bauteil als Final report of research project PSA- Basis einer automatisierten Bockholt, S., Voß, S., Bromm, M., OWEA, pp. 15-28 Prozesskette zur Rotorblattfertigung Kühn, M., Botasso, C., Campagnolo, F., Universität Bremen, Institute of Heinemann, D.: Hübler, C., Hofmeister, B.: integrated Product Development (BIK); Erhöhung des Flächenenergieertrags in Turbine Modeling SAERTEX GmbH & Co. KG., 2015 Windparks durch avancierte Anlagen- 14th eawe PhD seminar, Vrije Universiteit und Parkregelung Brussel, Belgium, Schmidt, B.; Hansen, M.; Schaumann, P.; ("CompactWind") September 18-20, 2018 Rolfes, R.; Goretzka, J.: Final report of the research project (2017) Foundation and Support Structure Kühn, T.; Altmikus, A.: (WP 4) Bromm, M.; Heinemann, D.; Kühn, M.; AssiSt – BMWi Schlussbericht gemäß Marx, S. (ed.): Probabilistic Safety Petrović, V.; Rott, A.; Steinfeld, G.; van NKBF98 Assessment of Offshore Wind Turbines. Dooren, M.; Vollmer, L.; Voß, S. (Eds.): Förderkennzeichen: 0325719 Final report of research project PSA- Final report of the research project OWEA, pp. 56-84 "Erhöhung des Flächenenergieertrags Kühn, M.: in Windparks durch avancierte Anla- IEA Task 32: Wind Lidar Systems for Schröder, K.; Bode, M.: gen- und Parkregelung“ Wind Energy Deployment (LIDAR) Design of offshore wind farms – data CompactWind (2017) Final report (2015) assimilation: Work Package 6.1 – Deliverable number 61.1 Drüke, S.; Steinfeld, G.; Schmidt, M.: (Definition of conventions and first data Numerical wind and stability atlas for structures), EU project IRPWIND, pp. 1-29 the North Sea Deliverable D1.3 in EU Project ClusterDesign, pp. 1-148

– 218 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017219

Schröder, K.; Berthold, T.: Design of offshore wind farms – Data assimilation: Work Package 6.1 – Deliverable number 61.2 (Installation of a database), EU project IRPWIND, pp. 1-27 Steinfeld, G.; Özdemir, H.; Mittelmeier, N.: Validated wake models (ForWind and ECN) Deliverable D6.3 in EU Project ClusterDesign, pp. 1-88

Tessmer, J. (Hrsg.): Smart Blades – Entwicklung und Konstruktion intelligenter Rotor- blätter: Teilvorhaben D Windphysik Final report by DLR, Fraunhofer IWES, ForWind – Universität Hannover, and ForWind – Universität Oldenburg (FKZ 0325601A/B/C/D)(2016)

Tsiapoki, S.; Krause, T.; Rolfes, R.; Oster- mann, J.: Erweiterung und Erprobung eines Schadensfrüherkennungs- und Eisde- tektionssystems für Rotorblätter von Windenergieanlagen Final report of the research project SHM. Rotorblatt by Institute of Structural Analysis (ISD) and Institut für Theoretische Nachrichtentechnik und Informationsverarbeitung (tnt), Leibniz Universität Hannover. (2016)

von der Haar, C.; Marx, S.: Auslegung von Suction Buckets in Betonbauweise Achmus, M. (ed.): Suction Bucket Gründungen als innovatives und montageschallreduziertes Konzept für Offshore-Windenergieanlagen. Final report of research project WindBucket (BMWi-FKZ 0325406B), pp. 151-172

– 219 – 220DOCUMENTATION Annual Report 2015-2017

Public Lectures, Articles in Barth, S.: Poppinga, T., Behrendt, T., Schmidt, A. H.: Magazines and Other Publications Windenergie – zum Stand der Technik Kollaborative Entwicklung eines und Entwicklungen Masterprogramms und Ringvorlesung Umwelt-Technik, verschiedener Zertifikatskurse im Hildesheim, January 12, 2017 modularen Baukastensystem. In Entwicklung von wissenschaftlichen Achmus, M.: Barth, S.: Weiterbildungs-programmen im MINT- Design of Foundation Elements for Presentation "ForWind" Bereich Offshore Wind Energy Converters Sommerreise Ministerpräsident Stephan Arnold, M., et al. (Hrsg.), Waxmann- Keynote lecture, 15. Int. Conference on Weil, Oldenburg, June 22, 2017 Verlag, Münster, 2017, ISBN: 978-3-8309- Structural and Geotechnical Engineering 3694-7 (ICSGE), Cairo/ Egypt, Dec. 20, 2015 Barth, S.: From turbulence to torque – why wind Schmidt, A. H.: energy needs pre-competitive research Achmus, M.: Das Modell des Lehrbaustein-Gitters and large research infrastructures Bearing behavior and design of mono- für MINT-Fächer. Eine neue Art der Invited talk, 6th international conference piles used as foundations for offshore Erfassung von Lerninhalten. In on Renewable Power Generation (RPG), wind energy converters Entwicklung von wissenschaftlichen Wuhan, China, October 19, 2017 Keynote lecture at the 3rd International Weiterbildungsprogrammen im MINT- Soil-Structure Interaction Symposium, Bereich Dörenkämper, M.: Izmir, Turkey, Oct. 18, 2017 Arnold, M., et al. (Hrsg.), Waxmann- An investigation of the atmospheric Verlag, Münster, 2017, ISBN: 978-3-8309- influence on spatial and temporal Barth, S.: 3694-7 power fluctuations in offshore wind Windenergie – zum Stand der Technik farms und Entwicklungen Stammler, M.: Dr. Hut Verlag, 168 pages (2015) Ringvorlesung Umwelt-Technik, Verfahren zum Ermitteln einer Hildesheim, October 21, 2015 Notwendigkeit, einen Schmierzustand Kärn, M., Schwarzer, C.: eines Rotorblattlagers einer Barth, S.: Praxisbeispiel Windstudium: Windkraftanlage zu verbessern, Excellent wind power research for Lernprozesse in der Weiterbildung als Steuervorrichtung zum Durchführen industrial projects soziale Prozesse gestalten. In: eines solchen Verfahrens und Invited talk, The Niedersachsen Event at Entwicklung von wissenschaftlichen Windkraftanlage mit einer solchen EWEA Offshore 2015, Copenhagen, Weiterbildungsprogrammen im MINT- Steuervorrichtung March 10, 2015 Bereich DE 10 2015 218 659 A1 (2017) Arnold, M., et al. (Hrsg.), Waxmann- Barth, S.: Verlag, Münster, 2017, ISBN: 978-3-8309- Unger, R.; Gerendt, C.; Daum, B.; Jansen, TV-Interview 3694-7 E.; Rolfes, R.; Ueing, C.; Englisch, N.; Radio Bremen Buten & Binnen Sayer, F.; Exner, W.; Mahrholz, T.; March 20, 2015 Kühn, M., Gasch, R., Sundermann, S.: Lorsch, P.; Wierach, P.: Kap. 8 Strukturdynamik Lebensdauererhöhung und Barth, S.: In: Gasch, R., Twele, J., Windkraftanla- Leichtbauoptimierung durch TV-Interview gen – Grundlagen und Entwurf, 9. Aufl., nanomodifizierte und hybride NDR Hallo Niedersachsen Springer (2016) Werkstoffsysteme im Rotorblatt September 15, 2015 Forum Industrial Supply – Hannover Kühn, M.: Messe, 2017, Hannover, Germany Barth, S.: Kap. 16 Offshore-Windparks Von vorwettbewerblicher In: Gasch, R., Twele, J., Windkraftanla- Unterstutzung bis zum Betrieb von gen – Grundlagen und Entwurf, 9. Aufl., Prufstanden fur die Industrie Springer (2016) Invited talk, 12th Austrian Wind Energy Symposium, Vienna Örlü, R.; Talamelli, A.; Oberlack, M.; March 10, 2016 Peinke, J. (Eds.): Progress in Turbulence VII Barth, S.: Proceedings of the iTi Conference in TV-Interview Turbulence 2016 NDR Hallo Niedersachsen Springer, 2017, DOI: 10.1007/978-3-319- October 25, 2016 57934-4

– 220 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017221

Lectures Kühn, M.: Kühn, M.: Wind Energy Utilisation Wind Physics Student's Lab Achmus, M.: Carl von Ossietzky Universität Carl von Ossietzky Universität Soil Investigation and Foundations for Summer term 2016 Winter term 2017/2018 Wind Turbines Haus der Technik (HdT) Essen, Kühn, M.: Kühn, M.: June 22, 2015 Wind Physics Measurement Project Advanced Metrology Carl von Ossietzky Universität Carl von Ossietzky Universität Achmus, M.: Summer term 2017 Winter term 2017/2018 Open questions regarding the bearing behavior of suction buckets Kühn, M.: Raasch, S.: DNV GL Cross-Industry Symposium, Design of Wind Energy Systems Turbulence II Hamburg, Carl von Ossietzky Universität Leibniz Universitat Hannover, 28 Sept. 2015 Summer term 2017 Winter term 2015/2016, 2016/2017, 2017/2018 Achmus, M.: Kühn, M.: On the design of suction-installed Aeroelastic Simulation of Wind Raasch, S.: foundations Turbines Introduction to simulating flows with Seminar Research on Offshore Founda- Carl von Ossietzky Universität LES models tions, Imperial College, London, Summer term 2017 Leibniz Universitat Hannover, 16 March 2016 Summer term 2015, 2016, 2017 Kühn, M.: Hansen, M.; Bode, M.; Diederley, J.: Introduction to Engineering Physics Thoben, K.-D.: Sonderkonstruktionen im Massivbau Carl von Ossietzky Universität Technisches Zeichnen KL I-1 / Engineer- Leibniz Universität t Hannover, Summer term 2017 ing Drawing KL I-1 Winter term 2015/2016 Universität Bremen, Kühn, M.: Winterterm 2015/16 & Hansen, M.; Bode, M.; Diederley, J.: Wind Physics Student Lab Winterterm 2016/17 Sonderkonstruktionen im Massivbau Carl von Ossietzky Universität Leibniz Universität t Hannover, Summer term 2017 Thoben, K.-D.: Winter term 2016/2017 Anwendung eines 3D-CAD-Systems / Kühn, M.: Application of 3D-CAD-Systems Hansen, M.; Bode, M; Birkner, D.: Wind Energy Utilisation Universität Bremen Sonderkonstruktionen im Massivbau Carl von Ossietzky Universität Winterterm 2015/16 & Leibniz Universität Hannover, Summer term 2017 Winterterm 2016/17 Winter term 2017/2018 Kühn, M.: Thoben, K.-D.: Kühn, M.: Mechanics Extended Products / Extended Wind Physics Measurement Project Carl von Ossietzky Universität Products Carl von Ossietzky Universität Winter term 2017/2018 Universität Bremen Summer term 2016 Winterterm 2015/16 & Kühn, M.: Winterterm 2016/17 Kühn, M.: Introduction to Renewable Energies Design of Wind Energy Systems Carl von Ossietzky Universität Thoben, K.-D.: Carl von Ossietzky Universität Winter term 2017/2018 Konstruktionssystematik / Produktent- Summer term 2016 wicklung / Design Methods and Tools Kühn, M.: Universität Bremen Kühn, M.: Laboratory Project II Winterterm 2015/16 & Aeroelastic Simulation of Wind Tur- Carl von Ossietzky Universität Winterterm 2016/17 bines Winter term 2017/2018 Carl von Ossietzky Universität Thoben, K.-D.: Summer term 2016 Kühn, M.: Einführung in die Maschinenelemente Wind Turbine Design Project KL I-2 / Introduction into Machine Ele- Kühn, M.: Carl von Ossietzky Universität ments KL I-2 Introduction to Engineering Physics Winter term 2017/2018 Universität Bremen. Carl von Ossietzky Universität Sommerterm 2015 & Sommerterm 2016 & Summer term 2016 Sommerterm 2017

– 221 – 222DOCUMENTATION Annual Report 2015-2017

Thoben, K.-D.: PhD Theses Späth, S.: Auslegung von Maschinenelementen / Statistische Korrektur von Ensemble- Konstruktionsentwurf KL II-1 / Di- vorhersagen der regional aggregierten mensioning of Machine Elements / Dörenkämper, M.: Windleistung Engineering Design KL II-1 An investigation of the atmospheric Ph.D. Thesis Universität Bremen influence on spatial and temporal Carl von Ossietzky Universität Oldenburg Sommerterm 2015 & Sommerterm 2016 & power fluctuations in offshore wind (2015) Sommerterm 2017 farms Ph.D. Thesis Schmidt, B.: Thoben, K.-D.: Carl von Ossietzky Universität Oldenburg Kombinierte extreme Einwirkungen für Anwendung und Vergleich von Kreativ- (2015) Offshore- Windenergieanlagen itätstechniken / Applying and Compar- PhD Thesis, ing Creativity Techniques Herrmann, R.: Leibniz Universität Hannover (2017) Universität BremenSommerterm 2015 & Simulation und Regelung von Sommerterm 2016 & Sommerterm 2017 Resonanzversuchsständen zur Schmoor, K.A.: Untersuchung der Bauteilermüdung Probabilistische Analyse zum Sicher- Thoben, K.-D.: PhD Thesis, heitsniveau von Offshore-Gründung- CAD-Management und virtuelle Leibniz Universität Hannover (2017) spfählen Produktentwicklung / CAD Manage- PhD Thesis, ment and Virtual Product Development Junk, C.: Leibniz Universität Hannover (2017) Universität BremenSommerterm 2015 & Statistical methods for probalististic Sommerterm 2016 & Sommerterm 2017 wind and wind power forecasting Thieken, K.: Ph.D. Thesis Geotechnical Design Aspects of Thoben, K.-D.: Carl von Ossietzky Universität Oldenburg Foundations for Offshore Wind Energy Anwendung von Konstruktionsmeth- (2015) Converters oden / Application of Design Methods PhD Thesis, Universität BremenSommerterm 2015 & Kapp, S.: Leibniz Universität Hannover (2015) Sommerterm 2016 & Sommerterm 2017 Lidar-based Reconstruction of Wind Fields and Application for Wind Tur- Vollmer, L.: bine Control Influence of atmospheric stability on Ph.D. Thesis wind farm control Carl von. Ossietzky Universität Oldenburg PhD Thesis, (2017) Carl von Ossietzky Universitat Oldenburg (2017) Kies, A.: Modelling the influence of demand von der Haar, C.: side management, storage and curtail- Ein mechanisch basiertes ment on a highly renewable European Dehnungsmodell für ermüdungs- power system beanspruchten Beton PhD Thesis, PhD Thesis Carl von Ossietzky Universitat Oldenburg Leibniz Universität Hannover (2016) (2016)

Lewandowski, M.: Entwicklung eines Integrationsmodells zur Zustandserfassung und Optimierung der Instandhaltung komplexer technischer Systeme PhD Thesis, Universität Bremen (2016)

Maier, M.: Entwicklung einer systematischen Vorgehensweise für bionischen Leichtbau PhD Thesis, Universität Bremen (2015)

– 222 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017223

Diploma, Master, Bachelor, and Assessment of engineering wake mod- Bock, H.: other Theses els available in WindPro for prediction Numerische Untersuchung eines Stahl- of energy yield at onshore wind farms betonsegmentturms mit alternativer Master thesis Fugenausbildung Ahrens, T.: Carl von Ossietzky Universität (2015) Master Thesis, Vergleich von Teilsicherheitskonzepten Leibniz Universität Hannover (2017) für die Suction-Bucket-Installation Barak, E.: Diploma Thesis, Konstruktion eines Leichtbau-Effek- Bock, H.: Leibniz Universitat Hannover (2016) torrahmens unterstützt durch die Bemessungsmodelle für Segmentfugen Anwendung von FE-Methoden von Windenergieanlagen Albe, R.: Master Thesis Bachelor Thesis, Validierung von Berechnungsmethoden Universität Bremen (2016) Leibniz Universität Hannover (2017) zur Bestimmung der axialen Tragfähig- keit von Offshore-Pfahlgründungen Barnhoorn, J.: Böhm, M.: Master Thesis, Turbulence Estimation from a Continu- Studie zum Extremlastverhalten von Leibniz Universitat Hannover (2016) ous-Wave Scanning Lidar Jacket-Tragstrukturen für Offshore Wind- Master thesis energieanlagen Alemans, L.: Carl von Ossietzky Universität (2016) Bachelor Thesis, Reliability-based model for steel Leibniz Universität Hannover (2016) structures inspection planning Baust, S. Master thesis Global and local FE-Modelling of rotor Borgelt, J.: Institute of Physics – Wind Energy blade pitch bearings in Ansys and Numerischen Untersuchungen an einer Systems, Carl von Ossietzky Universität validation of results Muffenverbindung (2015) Bachelor Thesis, Bachelor Thesis, (2016) Leibniz Universität Hannover (2017) Al-Madi, N.: Brandt, L.; Gründer, J.; Projahn, S.; Zur Anwendung von Kombinations- Bertram, L.: Wittfoth, S.: verfahren für Einwirkungen des Untersuchungen zur Anwendbarkeit Intelligentes Lager- und Logistiksys- Hochbaus auf Offshore-spezifische der p-y Methode für Pfähle großer tem am Beispiel von Textilrollen zur Einwirkungen Durchmesser in geschichteten, automatisierten Herstellung von Faser- Seminar Thesis, nichtbindigen Baugrundverhältnissen Kunststoff-Verbundbauteilen Leibniz Universität Hannover (2016) Diploma Thesis, Project Thesis Leibniz Universitat Hannover (2015) Universität Bremen (2015) Al-Madi, N.: Entwurf und Bemessung der Tragkon- Besecke, G.: Brandt, S.: struktion einer Windenergieanlage Korrelationsuntersuchungen an Optimierung von Jacket-Tragstrukturen Master Thesis, Messzeitreihen von Wind- und für Offshore-Windenergieanlagen Leibniz Universität Hannover (2017) Seegangsereignissen sowie Bachelor Thesis, Dehnungszeitreihen erfasst an einer Leibniz Universität Hannover (2015) Altenhöner, R.: Jacket-Tragstruktur Untersuchung zur Anwendbarkeit Seminar Thesis, Braun, F.: der p-y Methode für Pfähle großer Leibniz Universität Hannover (2016) Untersuchungen zur erforderlichen Durchmesser Einbindetiefe von Monopiles als Diploma Thesis, Bill, M.: Gründungselement für Offshorewind- Leibniz Universitat Hannover (2015) Reibwertermittlung zwischen energieanlagen Betonfertigteilen Master Thesis, Assam, E. M.: Project Thesis, Leibniz Universitat Hannover (2017) Feed Forward Control of a 7,5 MW Leibniz Universität Hannover (2017) Smart Blades Wind Turbine using a Budvytyte, E.: Linear Kalman Filter Birkner, D.: Untersuchungen zum progressiven Master thesis Numerische Untersuchung von Rammversagen bei offenen Stahlrohr- Carl von Ossietzky Universität (2017) Spannungsumlagerungen an einem rammpfählen Schwerkraftfundament infolge Master Thesis, Astudillo, F.: Ermüdungsbelastung Leibniz Universitat Hannover (2015) Master Thesis, Leibniz Universität Hannover (2016)

– 223 – 224DOCUMENTATION Annual Report 2015-2017

Buß, M.: Giron, C.: Iywer, S.: Untersuchung zur axialen Pfahltrag- Optimization of the Wind Energy Pro- Assessing and Minimizing the Sam- fähigkeit von Offshore Windenergiean- duction Through Efficient Portfolios pling Errors in Dual-Doppler Remote lagen in der Nordsee Master thesis Sensing Data Bachelor Thesis, Carl von Ossietzky Universität (2017) Master thesis Leibniz Universitat Hannover (2015) Carl von Ossietzky Universität (2016) Goldau, N.: Carreño, D.: Analyse von Strömungsnetzberech- Junge, M.: Optimization of Preventive nungen für Suction Buckets bei ge- Schadenslokalisation gestützt auf Maintenance for Onshore Wind Energy schichtetem und geneigtem Baugrund model updating im Zeitbereich am Converters Master Thesis, Beispiel eines vorgespannten Beton- Master thesis Leibniz Universitat Hannover (2017) turmes Carl von Ossietzky Universität (2015) Master Thesis, Grove, S.: Leibniz Universitat Hannover (2017) Dietrich, J.: Schadenslokalisation auf Basis au- Extremlastverhalten von Tragstruk- tomatisierter Modellanpassung am Klein, F.: turen für Offshore-Windenergieanla- Beispiel eines Rotorblatts für Kollisionssicherheit einer hybriden gen unter seismischer Anregung Windenergieanlagen Substruktur für Offshore Windener- Bachelor Thesis, Master Thesis, gieanlagen Leibniz Universitat Hannover (2017) Leibniz Universitat Hannover (2017) Master Thesis, Leibniz Universität Hannover (2015) Dörpmund, L.: Hamza, M. I.: Parameterstudie zur Untersuchung der Assessment of different wake tracking Klein, F.: maßgebenden Einflussgröße zur Ab- methods Skalierbarkeit einer hybriden schätzung der kritischen Einbindetiefe Bachelor thesis Substruktur für von Monopiles Carl von Ossietzky Universität (2017) Offshore-Windenergieanlagen Bachelor Thesis, Seminar Thesis, Leibniz Universitat Hannover (2017) Hempen, C.: Leibniz Universität Hannover (2015) Shut Down Procedures for Wind Tur- Fligg, A. M.: bines with Smart Blades Krehl, S.: Validation of remotely sensed tem- Master thesis Zum Tragverhalten von Sphäroguss perature and humidity profiles against Carl von Ossietzky Universität (2016) und Beton radiosoundings Master Thesis, Master Thesis, Henkel, M.: Leibniz Universität Hannover (2016) Carl von Ossietzky Universität Oldenburg Application of virtual sensing to (2017) offshore wind energy converters with Krüger, S.: jacket substructures Investigation of the Development of Frouzakis, N.: Master Thesis, Internal Boundary Layers by Means of Implementation of virtual LIDARs in Leibniz Universität Hannover (2016) Large-Eddy Simulation LES for wake analysis Master Thesis, Master thesis Herrera Rodríguez, O. R.: Carl von Ossietzky Universität Oldenburg Carl von Ossietzky Universität (2017) Analysis of the impact of atmospheric (2017) stability on wake meandering Sensitiv- Garr, J.: ity of wind resource estimates with Külpmann, A.: Entwicklung einer aktiven Vorrichtung the mesoscale model WRF on land Case Study of mesoscale wind für definierte Bedingungen bei Scher- surface data fluctuations during cold air outbreaks prüfungen von technischen Textilien Master Thesis, Master Thesis, Bachelor Thesis Carl von Ossietzky Universität Oldenburg Carl von Ossietzky Universität Oldenburg Universität Bremen (2016) (2015) (2017)

Gharbaoui, J.: Hieronimus, J.: Korthals, F. A.: Optimierung der Genauigkeit von Wind speed measurements in an off- Numerische Untersuchungen zum Roboter-Systemen durch integrierte shore wind farm by remote sensing. A Verbundverhalten von Sphäroguss und Sensorik comparison of radar satellite TerraSAR- Beton Master Thesis X and ground-based LiDAR systems Master Thesis, Universität Bremen (2015) Master thesis Leibniz Universität Hannover (2017) Carl von Ossietzky Universität (2015)

– 224 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017225

Kurka, M.: Miesner, E.: Plut, M.: Konstruktion eines aktiven Adapters Untersuchung eines Schmelzklebers als A comparison of two approaches for zur Anbindung einer Handhabungsein- Bindersystem in der automatisierten the representation of wind farms in heit an ein Roboterportalsystem Rotorblattfertigung WRF during stable atmospheric Bachelor Thesis Master Thesis conditions Universität Bremen (2015) Universität Bremen (2016 Master thesis Carl von Ossietzky Universität (2016) Li, C. S.: Miryeganeh, A.: Design of experiment methods for the Off-shore Wind Turbine Tower Con- Prakash, G.: damage evaluation of floating wind trol by Applying an Optimal Control Investigation of a Condition-Monitor- turbines Technique ing approach for wind turbine power Bachelor thesis Master thesis converters based on insulation resist- Carl von Ossietzky Universität (2017) Carl von Ossietzky Universität (2015) ance measurements Master thesis Lilje, D.: Müller, F.: Carl von Ossietzky Universität (2015) Einfluss einer heterogenen Verteilung Vergleich von Sensitivitätsanalysen zur von Baugrundeigenschaften auf den Bestimmung signifikanter Parameter Recalde, R. F. V.: axialen und lateralen Pfahlwiderstand bei komplexen Modellen von Development of an analysis procedure Master Thesis, Windenergieanalgen for evaluating dynamic behavior of Leibniz Universitat Hannover (2017) Master Thesis, operational wind turbines using me- Leibniz Universitat Hannover (2017) chanical load measurement data Martínez, R. A.: Master thesis Design and Verification of Risk Niehaus, J.: Carl von Ossietzky Universität (2017) Management Methodologies for its Entwicklung einer Vorrichtung zur Application in the Wind Industry teilautomatisierten Bestimmung Reckweg, L.: Master thesis der Biegesteifigkeit von technischen Extraction of Modal Parameters of a Carl von Ossietzky Universität (2015) Textilien Rotor Blade Bachelor Thesis Bachelor thesis Menke, R.: Universität Bremen (2017) Carl von Ossietzky Universität (2017) Investigation of the Wake Propagation of a Wind Turbine in Complex Terrain Orjuela Velasco, J.: Reepschläger, L.: Blades on Lidar Measurements of the Study on the Low Level Jet events oc- Systematische Analyse technischer Perdigao Experiment curring at heights relevant for offshore Risiken im Betrieb von Offshore-Wind- Master Thesis wind energy converters in the German energieanlagen als Grundlage für die Carl von Ossietzky Universität (2016 Bight Bewertung des Bachelor Thesis, Instandhaltungsprozesses Meyer, J.: Carl von Ossietzky Universität Oldenburg Bachelor thesis Implementierung eines Modells zur (2017) Carl von Ossietzky Universität (2015) Abschätzung des Zugtragverhaltens von Suction Bucket-Gründungen in Pache, D.: Richrath, M.: nichtbindigen Böden Anwendung eines vorwärtsgerichteten Konstruktion komplexer Komponenten Bachelor Thesis, neuronalen Netzes zur Erweiterung einer Handhabungseinheit zur Leibniz Universitat Hannover (2016) eines SHM-Schemas automatisierten Ablage technischer Master Thesis, Textilien Michaely, M.: Leibniz Universitat Hannover (2016) Master Thesis Validierung des Ansatzes zur Bestim- Universität Bremen (2015) mung der elastischen Verformung von Plitzner, S.: Monopiles Analyse von Drapierwerkzeug- Roß, A.: Diploma Thesis, konzepten für die Ablage von Glas- Situational activation of Load Leibniz Universitat Hannover (2015) fasergelegen in eine Rotorblatt-Haupt- Mitigation Concepts for a Commercial form Offshore Wind Turbine Middendorf, J.: Master Thesis Master thesis Schubspannungsverteilung in Universität Bremen (2015) Carl von Ossietzky Universität (2017) geschliffenen Segmentfugen Seminar Thesis, Rowlinson, B.: Leibniz Universität Hannover (2017) Wind Turbine Standstill Instability Analysis Master thesis Carl von Ossietzky Universität (2015)

– 225 – 226DOCUMENTATION Annual Report 2015-2017

Rudloff, S.; Schnabel: Seifert, J. K.: Sved, D.: Entwicklung einer Einheit zur Investigation of the Effect of Wind Konstruktiver Entwurf einer Leichtbau- automatisierten Aufwicklung Direction Dynamics and Measurement Handhabungseinheit zur automatisi- konfektionierter Fasergelegebahnen Errors on Active Wake erten Ablage technischer Textilien Bachelor Thesis Master thesis Master Thesis Universität Bremen (2016) Carl von Ossietzky Universität (2015) Universität Bremen (2017)

Salim Dar, A.: Selle, J.: Syedykh, D.: Wind turbine wake interaction in Analyse des Adapterbereiches einer Numerische Untersuchungen zum complex terrain using Large-Eddy Hybrid-Windenergieanlage Einfluss des Installationsvorgangs auf Simulation Bachelor Thesis, das Zugtragverhalten von offenen Master Thesis, Leibniz Universität Hannover (2016) Verdrängungspfählen Carl von Ossietzky Universität Oldenburg, Master Thesis, Danish Technical University (2018) Sommer, K.: Leibniz Universitat Hannover (2016) Konstruktion eines modularen Sambahamphe, B. S.: Ablagesystems für Roboterwerkzeuge Theuer, F.: Geographical Information System (GIS) Bachelor Thesis Development of a Pressure Model in Analysis of Global Wind Atlas Data Universität Bremen (2016) the Wake of Wind Turbines Master thesis Bachelor thesis Carl von Ossietzky Universität (2017) Soylu, T.: Carl von Ossietzky Universität (2016) Numerische Schädigungsuntersuchung Sangster, M.: eines Schwerkraftfundamentes infolge Troya, J.: Wind Farm Layout Optimisation for Ermüdungsbelastung Sensitivity of wind resource estimates both Loads and Production using Bachelor Thesis, with the mesoscale model WRF on land Surrogate Wake Models and Automat- Leibniz Universität Hannover (2016) surface data ed Optimisation Methods Bachelor Thesis, Master thesis Steiner, J.: Carl von Ossietzky Universität Oldenburg Carl von Ossietzky Universität (2017) Analysis of the Effect of Wind Turbine (2016) Yaw Misalignment on the Turbine Schillebeckx, D.: Wake Path by Lidar Wake Tracking Tun, L.: Sensivity Testing and Technical Master thesis Entwurf eines 145 m hohen Implementation of Mesoscale Carl von Ossietzky Universität (2016) Spannbetonturmes für eine Generalization Procedure for the WRF Windenergieanlage der 3-MW-Klasse Model Stell, L.: Seminar Thesis, Master thesis Monitoring der Rotorblätter von Leibniz Universität Hannover (2017) Carl von Ossietzky Universität (2016) Windenergieanlagen im Betrieb Master Thesis, Vali, M.: Schödler, J.: Leibniz Universitat Hannove (2017) Controller design for a wind turbine Identification of wind turbine with bend-twist coupled blades production losses due to icing using Stelljes, P.: Master thesis neural networks Untersuchung zu großmaßstäblichen Carl von Ossietzky Universität (2015) Master thesis Modellversuchen zum axialen Carl von Ossietzky Universität (2017) Tragverhalten Vernekar, V.: Master Thesis, Grid codes variant and implementa- Schröder, L.: Leibniz Universitat Hannover (2015) tion strategy for country specific Sceening potential areas and requirements techno-economic assessment of wind Stindt, D.: Master thesis farms for the planning of Danish Analyse von Werkzeugkonzepten Carl von Ossietzky Universität (2015) offshore wind energy zur automatisierten Drapierung von Master thesis Glasfaserge-legen in der Rotor- Vikili Dastjerd, B.: Carl von Ossietzky Universität (2016) blattfertigung Entwicklung eines Funktionstest- Bachelor Thesis programms für das Rotorblattlager der Schwabe, R.: Universität Bremen (2015) Referenzanlage IWT 7.5-164 und Aerodynamic Analysis and Optimisa- Erstellung eines Algorithmus zur Er- tion of a Reference Rotor for a 10 MW zeugung eines beschleunigten Offshore Wind Turbine with respect to Dautertestprogramms. the Support Structure Master Thesis, Master Thesis Universität Kassel (2016) Carl von Ossietzky Universität (2016)

– 226 – Annual Report 2015-2017 DOCUMENTATION Annual Report 2015-2017227

Von Hindte, C.: Tragverhalten zyklisch axialbelasteter Stahlrohrpfähle Master Thesis, Leibniz Universitat Hannover (2017)

Wagner, D.: Evolution and Properties of Low Level Jet Events above the Southern North Sea Master Thesis, Carl von Ossietzky Universität Oldenburg (2017)

Wegmann, A.: Optmimierungsmöglichkeiten durch Augmented Reality in der Instandhal- tung und im technischen Service Bachelor Thesis Universität Bremen (2017)

Wlodarczyk-Zimny, A.: Konzeptioneller Entwurf einer Bewegungskompensationsplattform für Windener-gieanlagenkomponenten auf Transportschiffen Bachelor Thesis Universität Bremen (2017)

Zain, J.: Degradation of pile loading behaviour under cyclic axial load Master Thesis, Leibniz Universitat Hannover (2016)

Zindler, J.: Numerische Untersuchungen zum Verbundverhalten von Sphäroguss und Beton Seminar Thesis, Leibniz Universität Hannover (2017)

– 227 – 228DOCUMENTATION / STAFF Annual Report 2015-2017

List of ForWind Staff Ambroise, Hugues, M.Sc. Behnken, Christian, M.Sc. Universität Oldenburg Universität Oldenburg Members Institute of Physics Institute of Physics AG Energy Meteorology (EnMet) AG Turbulenz, Windenergie and +49 441 798 5078 Stochastics (TWiSt) [email protected] +49 441 798-5052 Abdel-Rahman, Khalid, Dr.-Ing. M.Sc. [email protected] Universität Hannover Balzani, Claudio, Dr.-Ing. Institute for Geotechnical Engineering Universität Hannover Behrendt, Werner, Dipl.-Ing. (IGtH) Institute of Wind Energy Systems (IWES) Universität Bremen +49 511 762 2273 +49 511 762 5612 Bremen Institute for Metrology, [email protected] [email protected] Automation and Quality Science (BIMAQ) +49 0421 218 64636 Abulrazek, Abdulkarim, M.Sc. Barth, Stephan, Dr. [email protected] Universität Oldenburg ForWind Head Office Institute of Physics +49 441 798 5091 Beltrán, Raúl, M.Sc. AG Turbulence, Wind Energy and [email protected] Universität Hannover Stochastics (TWiSt) Institute for Concrete Construction (IfM) +49 441 798 5023 Basaldella, Marco, M.Sc. +49 511 762 2189 [email protected] Universität Hannover [email protected] Institute for Building Materials Science Achmus, Martin, Prof. Dr.-Ing. (IfB) Berger, Frederik, M.Sc. Universität Hannover +49 511 762 3477 Universität Oldenburg Institute for Geotechnical Engineering [email protected] Institute of Physics (IGtH) AG Wind Energy Systems (WESys) +49 511 762 4155 Bartschat, Arne +49 441 798 5065 [email protected] Fraunhofer Institute for Wind Energy [email protected] Systems (IWES) Adler, Johannes, Dipl.-Ing. +49 471 14290 520 Berthold, Tim, M.Sc. Universität Bremen [email protected] Universität Hannover Institute for Electrical Drives, Power Institute for Risk and Reliability (IRZ) Electronics and Devices (IALB) Bastine, David +49 511 762 17883 +49 421 218-62675 Universität Oldenburg [email protected] [email protected] Institute of Physics AG Turbulence, Wind Betcke, Jethro, M.Sc. Akbarzadeh Lalkami, Siamak M.Sc. Energy and Stochastics (TWiSt) Universität Oldenburg Universität Oldenburg +49 441 798 5054 Institute of Physics Institute of Physics [email protected] AG Energy Meteorology (EnMet) AG Energy Meteorology (EnMet) +49 441 798 3927 +49 441 798 5076 Battermann, Dagmar [email protected] [email protected] ForWind Head Office +49 441 798 5060 Birkner, Dennis, M.Sc. Albensoeder, Stefan, Dr. [email protected] Universität Hannover Universität Oldenburg Institute for Concrete Construction (IfM) Institute of Physics Beck, Hauke, M.Sc. +49 511 762 3354 AG Energy Meteorology (EnMet) Universität Oldenburg [email protected] +49 441 798 5075 Institute of Physics [email protected] AG Wind Energy Systems (WESys) Bode, Dennis, M.Sc. +49 441 798 5064 Universität Bremen Albiker, Johannes, Dipl.-Ing. [email protected] Institute for Integrated Product Universität Hannover Development (BIK) Institute for Geotechnical Engineering Behnke, Karsten +49 421 218-64879 (IgtH) Fraunhofer Institute for Wind Energy [email protected] +49 511 762 4155 Systems (IWES) [email protected] +49 471 14290 521 [email protected]

– 228 – Annual Report 2015-2017 DOCUMENTATION / STAFF Annual Report 2015-2017229

Bode, Matthias, M.Sc. Chattopadhyay, Kabitri, M.Sc. Dose, Bastian, M.Sc. Universität Hannover Universität Oldenburg Universität Oldenburg Institute for Risk and Reliability (IRZ) Institute of Physics Institute of Physics [email protected] AG Energy Meteorology (EnMet) AG Turbulence, Wind Energy and +49 441 798 5074 Stochastics (TWiSt) Böske, Lennart, M.Sc. +49 441 798 5056 Universität Hannover Dapperheld, Jörg [email protected] Institute of Meteorology and Climatology Universität Oldenburg (IMUK) Institute of Physics Dubois, Jan, M. Sc. +49 511 762-3232 AG Turbulence, Wind Energy and Universität Hannover [email protected] Stochastics (TWiSt) Institute for Steel Construction +49 441 798 3535 +49 511 762-2492 Bohne, Tobias, M.Sc. [email protected] [email protected] Universität Hannover Institute of Structural Analysis (ISD) Daum, Benedikt, Dipl.-Ing. Dr. Ehrich, Sebastian, Dipl. Phys. +49 551 762-2859 Universität Hannover Universität Oldenburg [email protected] Institute of Structural Analysis (ISD) Institute of Physics +49-511-762- 5696 AG Turbulence, Wind Energy and Brink, Michael, M.Sc. [email protected] Stochastics (TWiSt) Universität Bremen +49 441 798 5057 Institute for Integrated Product Dean, Aamir, Dr.-Ing. [email protected] Development (BIK) Universität Hannover +49 421 218-64873 Institute of Structural Analysis (ISD) Ehrmann, Andreas, Dipl.-Ing. [email protected] +49-511-762-4519 Universität Hannover [email protected] Institute of Structural Analysis (ISD) Bromm, Marc, M.Sc. +49 551 762-17213 Universität Oldenburg Decker, André, Dr.-Ing. [email protected] Institute of Physics Universität Bremen AG Wind Energy Systems (WESys) Institute for Integrated Product Eichstädt, Rasmus, Dipl.-Ing. +49 441 798 5063 Development (BIK) Universität Hannover [email protected] +49 421 218-64874 Institute for Steel Construction (IfS) [email protected] +49 511 762-5949 Brudler, Evelyn, M.Sc. [email protected] Universität Oldenburg Dibbern, Christoph, M.Sc. Institute of Physics Department of Computing Science Ernst, Alexander, Dipl.-Ing AG Wind Energy Systems (WESys) +49 441 798-4482 Universität Bremen +49 441 798 5081 [email protected] Institute for Electrical Drives, Power [email protected] Electronics and Devices (IALB) Dollinger, Christoph, M.Sc. +49 421 218-62691 Bruns, Marlene, M.Sc. Universität Bremen [email protected] Universität Hannover Bremen Institute for Metrology, Institute of Structural Analysis (ISD) Automation and Quality Science Fleischer, Simon, M.Sc. +49-511-762-3083 (BIMAQ) Universität Bremen [email protected] +49 421 218 64628 Institute of Automation (IAT) [email protected] Brunko, Alexander, Dipl.-Ing. Forghani, S. Mohsen, M.Sc. Universität Hannover Dörenkämper. Martin, M.Sc. Universität Oldenburg Institute for Electrical Drives, Power Universität Oldenburg Institute of Physics Electronics and Devices (IALB) Institute of Physics AG Wind Energy Systems (WESys) +49 421-218 62772 AG Energy Meteorology (EnMet) +49 441 798 5068 [email protected] +49 441 798 5077 [email protected] [email protected] Buhr, Renko, M.Sc. Franke, Jan, Dipl.-Ing. Universität Oldenburg Universität Bremen Institute of Physics Institute for Integrated Product AG Energy Meteorology (EnMet) Development (BIK) +49 441 798 5076 +49 421 218 64876 [email protected] [email protected]

– 229 – 230DOCUMENTATION / STAFF Annual Report 2015-2017

Friedrichs, Wilm, Dr.-Ing. Gottlieb, Gerrit, M.Sc. Hahn, Axel, Prof. Dr.-Ing. Universität Oldenburg Universität Hannover Universität Oldenburg Institute of Physics Institute of Structural Analysis (ISD) Computing Science, Business Engineering AG Wind Energy Systems (WESys) +49-511-762-3108 +49 441 798 4480 +49 441 798 5066 [email protected] [email protected] [email protected] Grießmann, Tanja, Dr.-Ing. Hammer, Annette, Dr. Fuchs, André, M.Sc. Universität Hannover Universität Oldenburg Universität Oldenburg Institute of Structural Analysis (ISD) Institute of Physics Institute of Physics +49 511 762 2247 AG Energy Meteorology (EnMet) AG Turbulence, Wind Energy and [email protected] +49 441 798 3545 Stochastics (TWiSt) [email protected] +49 441 798 5023 Groß, Günter, Prof. Dr. [email protected] Universität Hannover Hanf, Michael, Dipl.-Ing. Institute of Meteorology and Climatology Universität Bremen Gebhardt, Cristian, Dr.-Ing. (IMUK) Institute for Electrical Drives, Power Universität Hannover +49 511 762 5408 Electronics and Devices (IALB) Institute of Structural Analysis (ISD) [email protected] +49 421 218 62795 +49 511 762 2859 [email protected] [email protected] Gülker, Gerd, Dr. Universität Oldenburg Hansen, Michael, Dr.-Ing. Gerasch, Wolf-Jürgen, Dipl.-Ing. Institute of Physics Universität Hannover Universität Hannover AG Turbulence, Wind Energy and Institute for Concrete Construction (IfM) Institute of Structural Analysis (ISD) Stochastics (TWiSt) +49 511 762 3353 +49 511 762 2247 +49 441 798 3511 [email protected] [email protected] [email protected] Haselsteiner, Andreas F., M.Sc. Gerendt, Christian, M.Sc. Gutierrez, Monica Almonacid, M.Sc. Universität Bremen Universität Hannover Universität Oldenburg Institute for Integrated Product Institute of Structural Analysis (ISD) Institute of Physics Development (BIK) +49-511-762-2266 AG Wind Energy Systems (WESys) +49 421 218-64878 [email protected] +49 441 798 5060 [email protected] [email protected] Gerlach, Tim, Dipl.-Ing. Haunhorst, Frauke, Dipl. Kffr. Universität Hannover Hadjihosseini, Ali, M.Sc. ForWind Head Office Institute for Geotechnical Engineering Universität Oldenburg +49 441 798 5090 (IGtH) Institute of Physics [email protected] +49 511 762 3529 AG Turbulence, Wind Energy and [email protected] Stochastics (TWiSt) Heinemann, Detlev, Dr. +49 441 798 5043 Universität Oldenburg Gharibi, Zeinab, .M.Sc. [email protected] Institute of Physics Universität Oldenburg AG Energy Meteorology (EnMet) Institute of Physics Häfele, Jan, Dipl.-Ing. +49 441 798 3543 AG Turbulence, Wind Energy and Universität Hannover [email protected] Stochastics (TWiSt) Institute of Structural Analysis (ISD) +49 441 798 5055 +49 551 762-4208 Heißelmann, Hendrik, Dipl.-Phys. [email protected] [email protected] Universität Oldenburg Institute of Physics, Groke, Holger, Dr.-Ing. Hähne, Hauke, M.Sc. AG Turbulence, Wind Energy and Stocha- Universität Bremen Universität Oldenburg stics (TWiSt) Institute for Electrical Drives, Power Institute of Physics +49 441 798 3643 Electronics and Devices (IALB) AG Turbulence, Wind Energy and [email protected] +49 421 218-62677 Stochastics (TWiSt) [email protected] +49 441 798 5054 Henneberg, Joshua, M.Sc. [email protected] Universität Hannover Institute for Steel Construction +49 511 762-4995 [email protected]

– 230 – Annual Report 2015-2017 DOCUMENTATION / STAFF Annual Report 2015-2017231

Herraez Hernandez, Ivan, M.Sc. Hörmeyer, Jasmin, M.Sc. Jahjouh, Mahmoud, Dr.-Ing. Universität Oldenburg Universität Hannover Universität Hannover Institute of Physics Institute of Structural Analysis (ISD) Institute of Structural Analysis (ISD) AG Turbulence, Wind Energy and +49 441 798 4703 +49 511 762-4939 Stochastics (TWiSt) j.hö[email protected] [email protected] +49 441 798 5055 [email protected] Hoffmann, Felix, Dipl.-Ing. Jansen, Eelco, Dr. Universität Bremen Universität Hannover Herrmann, Ralf, Dr.-Ing. Institute for Electrical Drives, Power Institute of Structural Analysis (ISD) Universität Hannover Electronics and Devices (IALB) +49-511-762-17425 Institute for Concrete Construction (IfM) +49 421 218 62664 [email protected] +49-030-8104-3888 [email protected] [email protected] Jauken, Helge, M.Sc. Hofmann, Lutz, Prof. Dr.-Ing. habil. Universität Hannover Hesseling, Christina, Dipl.-Phys. Universität Hannover Institute of Structural Analysis (ISD) Universität Oldenburg Institute for Electric Power Systems (IEE) +49-511-762-2885 Institute of Physics +49 511 762 2801 [email protected] AG Turbulence, Wind Energy and [email protected] Stochastics (TWiSt) Junk, Constantin, M.Sc. +49 441 798 3514 Hofmeister, Benedikt, M.Eng. Universität Oldenburg [email protected] Universität Hannover Institute of Physics Institute of Structural Analysis (ISD) AG Energy Meteorology Hieronimus, Julian, M.Sc. +49 511 762 17213 +49 441 798 5079 Universität Oldenburg [email protected] [email protected] Institute of Physics AG Wind Energy Systems (WESys) Hogreve, Sebastian, Dipl.-Ing. Kärn, Moses +49 441 798 5060 Universität Bremen ForWind Head Office [email protected] Bremen Institute for Mechanical +49 441 798 5082 Engineering (bime) [email protected] Hildebrandt, Arndt, Dipl.-Ing. +49 421 218 64837 Universität Hannover [email protected] Kaminski, Nando, Prof. Dr.-Ing. Franzius-Institute for Hydraulic, Institute for Electrical Drives, Waterways and Coastal Engineering Hohmann, Claas, Dipl.-Ing. Power Electronics and Devices (IALB) +49 511 762 2582 Universität Bremen +49 421 218 62660 [email protected] Institute of Automation (IAT) [email protected]

Hillers, Bernd, Dipl.-Ing. Holzke, Wilfried, Dipl.-Ing. Kampers, Gerrit Malte, M.Sc. ForWind Head Office Universität Bremen Universität Oldenburg Coordinator Institute for Electrical Drives, Power Institute of Physics +49 441 798 5090 Electronics and Devices (IALB) AG Turbulence, Wind Energy and [email protected] +49 421 218-62675 Stochastics (TWiSt) [email protected] +49 441 798 3455 Hölling, Agnieszka, Dipl.-Ing. [email protected] Universität Oldenburg Homeyer, Tim, Dipl.-Phys. Institute of Physics Universität Oldenburg Kanani, Farah, Dipl.-Met. AG Turbulence, Wind Energy and Institute of Physics Universität Hannover Stochastics (TWiSt) AG Turbulence, Wind Energy and Institute of Meteorology and Climatology +49 441 798 3535 Stochastics (TWiSt) (IMUK) [email protected] +49 441 798 3514 +49 511 762-3232 [email protected] [email protected] Hölling, Michael, Dr. Universität Oldenburg Hübler, Clemens, M.Sc. Kelma, Sebastian, Dipl.-Ing. Institute of Physics Universität Hannover Universität Hannover AG Turbulence, Wind Energy and Institute of Structural Analysis (ISD) Institute for Steel Construction (IfS) Stochastics (TWiSt) +49 511 762 2885 +49 511 762-4139 +49 441 798 3951 [email protected] [email protected] [email protected]

– 231 – 232DOCUMENTATION / STAFF Annual Report 2015-2017

Kidambi Sekar, Anantha Padmanabhan, Kühn, Martin, Prof. Dr. Dipl.-Ing. Maas, Oliver M.Sc. Universität Oldenburg Universität Hannover Universität Oldenburg Institute of Physics, Institute of Meteorology and Climatology Institute of Physics AG Wind Energy Systems (WESys) (IMUK) AG Wind Energy Systems (WESys) +49 441 798 5061 [email protected] +49 441 798 5065 [email protected] [email protected] Marx, Steffen, Prof. Dr.-Ing. Kuhfuss, Bernd, Prof. Dr.-Ing. Universität Hannover Kies, Alexander, M.Sc. Universität Bremen Institute for Concrete Construction (IfM) Universität Oldenburg Bremen Institute for Mechanical +49 511 762 3352 Institute of Physics Engineering (bime) [email protected] AG Energy Meteorology (EnMet) +49 421 218-64800 +49 441 798-5078 [email protected] Martens, Susanne, M.Sc. [email protected] Universität Hannover Kuhnle, Bernd, Dipl.-Ing. Institute of Structural Analysis Vibrations Knigge, Christoph, Dr. Universität Oldenburg Department, AG Acoustics Universität Hannover Institute of Physics, +49 441 798 4703 Institute of Meteorology and Climatology AG Wind Energy Systems (WESys) [email protected] (IMUK) +49 441 798 5064 +49 511 762 3232 [email protected] Matzat, Sebastian, M.Sc. [email protected] Universität Bremen Lang, Walter, Prof. Dr.-Ing. Institute of Automation (IAT) Kraft, Martin, Dipl.-Ing. Universität Bremen Universität Oldenburg Institute for Microsensors, - actuators Mehrens, Anna Rieke, M.Sc. Institute of Physics, and -Systems (IMSAS) Universität Oldenburg AG Wind Energy Systems (WESys) +49 421 218-62602 Institute of Physics +49 441 798-5065 [email protected] AG Energy Meteorology (EnMet) [email protected] +49 441 798 5079 Lind, Pedro, M.Sc. [email protected] Krause, Thomas, Dipl.-Ing. Universität Oldenburg Universität Hannover Institute of Physics Meints, Irene Institute for Information Processing (tnt) AG Turbulence, Wind Energy and Universität Oldenburg +49 511 762 5052 Stochastics (TWiSt) Institute of Physics kraus @tnt.uni-hannover.de +49 441 798 5090 AG Wind Energy Systems (WESys) [email protected] +49 441 798-5060 Krishnappa, Likith, M.Sc. [email protected] Universität Bremen Löw, Kathrin, Dipl.-Ing. Institute for Integrated Product Universität Hannover Menck, Oliver Development (BIK) Institute for Steel Construction (IfS) Fraunhofer Institute for Wind Energy +49-421 218 64861 Tel.: +49 511 762 3540 Systems (IWES) [email protected] [email protected] +49 471 14290 523 [email protected] Kröger, Lars, M.Sc. Lohaus, Ludger, Univ.-Prof. Dr.-Ing. Universität Oldenburg Uniiversität Hannover Mertens, Axel, Prof. Dr.-Ing. Institute of Physics, Institute for Building Materials Science Universität Hannover AG Turbulence, Wind Energy and (IfB) Institute for Drive Systems and Power Stochastics (TWiSt) +49 511 762 3722, 2990 Electronics (IAL) +49 441 798 3535 [email protected] +49 511 762 2471 [email protected] [email protected] Lohmann, Gerold, Dipl. Hydrol. Krüger, Sonja, M.Sc. Universität Oldenburg Michels, Kai, Prof. Dr.-Ing. Universität Oldenburg Institute of Physics Universität Bremen Institute of Physics +49 441 798-3577 Institute of Automation (IAT) AG Energy Meteorology (EnMet) [email protected] +49 421 218 62500 +49 441 798 5074 [email protected] sonja.krü[email protected]

– 232 – Annual Report 2015-2017 DOCUMENTATION / STAFF Annual Report 2015-2017233

Mörz, Tobias, Prof. Dr. Oneschkow, Nadja, Dr.-Ing. +49 441-798-5068 Universität Bremen Universität Hannover [email protected] Center for Marine Environmental Sciences Institute for Building Materials Science (MARUM) (IfB) Pierl, Dennis, Dipl.-Ing. +49 421 218 65840 +49 511 762 3105 Universität Bremen [email protected] [email protected] Institute of Automation (IAT)

Moriße, Marcel Orlik, Bernd, Prof. Dr.-Ing. Poland, Kathleen, Dipl.-Eng. Universität Hannover Universität Bremen Universität Oldenburg Institute for Drive Systems and Power Institute for Electrical Drives, Power Institute of Physics Electronics (IAL) Electronics and Devices (IALB) AG Wind Energy Systems (WESys) +49 421 218 62680 +49 441 798 5065 Munk, Johannes, M.Sc. [email protected] [email protected] Universität Oldenburg Institute of Physics Pache, Dorian, M.Sc. Poll, Gerhard, Prof. Dr.-Ing. AG Energy Meteorology (EnMet) Universität Hannover Universität Hannover +49 441 798 5074 Institute of Structural Analysis (ISD) Institute for Machine Design and [email protected] +49 511 762 3083 Tribology (IMKT) [email protected] +49 511 762 2416 Neuhaus, Lars Kristian, M.Sc. [email protected] Universität Oldenburg Pawletta, Peter Institute of Physics ForWind Head Office Ponick, Bernd, Prof. Dr.-Ing. AG Turbulence, Wind Energy and +49 441 798 5087 Universität Hannover Stochastics (TWiSt) [email protected] Institute for Drive Systems and Power +49 441 798 5024 Electronics (IAL) [email protected] Peinke, Joachim, Prof. Dr. +49 511 762 2571 Universität Oldenburg [email protected] Neunaber, Ingrid, M.Sc. Institute of Physics Universität Oldenburg AG Turbulence, Wind Energy and Preihs, Stefan, Dr.-Ing. Institute of Physics Stochastics (TWiSt) Universität Hannover AG Turbulenz, Windenergie and +49 441 798 3536 Institute of Communications Technology Stochastics (TWiSt) [email protected] (ikt) +49 441-798-3455 +49 511 762 2819 [email protected] Penner, Nikolai, M.Sc. [email protected] Universität Hannover Neuweiler, Insa, Prof. Dr. Institute of Structural Analysis (ISD) Prestel, Isabelle, M.Sc. Universität Hannover +49 551 762-4273 Universität Oldenburg Institute of Fluid Mechanics and Environ- [email protected] Institute of Physics mental Physics in Civil Engineering AG Energy Meteorology (EnMet) +49 511 762 3567 Perrone, Francesco, M.Sc. +49 441-798-5079 [email protected] Universität Oldenburg [email protected] Institute of Physics Oetting, Heike AG Wind Energy Systems (WESys) Puczylowski, Jaroslaw, M.Sc. Universität Oldenburg +49 441 798 5063 Universität Oldenburg Institute of Physics [email protected] Institute of Physics AG Turbulence, Wind Energy and AG Turbulence, Wind Energy and Stochastics (TWiSt) Peters, Jan-Hendrik, MSc. Stochastics (TWiSt) +49 441-798-3455 Universität Bremen +49 441 798 3643 [email protected] Institute for Electrical Drives, Power [email protected] Electronics and Devices (IALB) Ohlendorf, Jan-Hendrik, Dipl.-Ing. +49 421 218 62674 Pyei Phyo, Lin, M.Sc. Universität Bremen [email protected] Universität Oldenburg Institute for Integrated Product Institute of Physics Development (BIK) Petrovic, Vlaho, Dr. AG Turbulence, Wind Energy and +49 421 218 64876 Universität Oldenburg Stochastics (TWiSt) [email protected] Institute of Physics +49 441 798 3455 AG Wind Energy Systems (WESys) [email protected]

– 233 – 234DOCUMENTATION / STAFF Annual Report 2015-2017

Raasch, Siegfried, apl. Prof. Dr. Reinke, Nico, Dipl.-Phys. Rott, Andreas, Dipl.-Math. Universität Hannover Universität Oldenburg Universität Oldenburg Institute of Meteorology and Climatology Institute of Physics Institute of Physics (IMUK) AG Turbulence, Wind Energy and AG Wind Energy Systems (WESys) +49 511 762 3253 Stochastics (TWiSt) +49 441 798-5064 [email protected] +49 441 798 3455 [email protected] [email protected] Raba, Alexander, Dipl.-Ing. Rzeczkowski, Patrick, Dipl.-Ing. Universität Hannover Reuter, Andreas, Prof. Dr.-Ing. Universität Hannover Institute for Steel Construction (IfS) Universität Hannover Institute for Building Materials Science +49 511 762-4975 Institute of Wind Energy Systems (IfB) [email protected] +49 511 762 4712 +49 511 762 4206 [email protected] [email protected] Rademacher, Lisa, M.Sc. Universität Oldenburg Reuter, Rainer, Dr. Saathoff, Jann-Eike, M.Sc. Institute of Physics Universität Oldenburg Universität Hannover AG Turbulenz, Windenergie and Institute of Physics Institute for Geotechnical Engineering Stochastics (TWiSt) AG Marine Physics (IGtH) +49 441-798-3455 +49 441 798 3522 +49 511 762 3808 [email protected] [email protected] [email protected]

Radulovic, Luka, M.Sc. Richrath, Marvin, M.Sc. Sander, Aljoscha, M.Sc. Universität Hannover Universität Bremen Universität Bremen Institute for Steel Construction (IfS) Institute for Integrated Product Institute for Integrated Product +49 511 762 17212 Development (BIK) Development (BIK) [email protected] +49 421 218-64875 +49 421 218-64867 [email protected] [email protected] Rahimi, Hamid, M.Sc. Universität Oldenburg Rinn, Philip, Dipl.-Phys. Schack, Tobias, M.Sc. Institute of Physics Universität Oldenburg Universität Hannover AG Turbulence, Wind Energy and Institute of Physics Institute for Building Materials Science Stochastics (TWiSt) AG Turbulence, Wind Energy and (IfB) +49 441 798 5016 Stochastics (TWiSt) +49 511 762 5243 [email protected] +49 441 798 5052 [email protected] [email protected] Reck, Martin, M.Sc. Schäfer, Dominik, Dipl.-Ing. Universität Oldenburg Rockel, Stanislav, Dipl.-Phys. Universität Hannover Institute of Physics Universität Oldenburg Institute for Geotechnical Engineering AG Wind Energy Systems (WESys) Institute of Physics +49 511 762 3736 +49 441 798 3560 AG Turbulence, Wind Energy and [email protected] [email protected] Stochastics (TWiSt) +49 441 798 3455 Schaumann, Peter, Prof. Dr.-Ing. Redecker, Marc Allan, Dipl.-Ing. [email protected] Universität Hannover Universität Bremen Institute for Steel Construction (IfS) Institute for Integrated Product Rolfes, Raimund, Prof. Dr.-Ing. habil. +49 511 762 3781 Development (BIK) Universität Hannover [email protected] +49 421 218-64877 Institute of Structural Analysis (ISD) [email protected] +49 511 762 3867 Schendel, Alexander, Dipl.-Ing. [email protected] Universität Hannover Reichel, Juliane, Dr. Ludwig-Franzius-Institute (LuFI) ForWind Head Office Roschen, Julian, M.Sc. +49 511 762 2592 +49 441 798 5085 Universität Bremen [email protected] [email protected] Institute of Automation (IAT) Schleich, Florian Fraunhofer Institute for Wind Energy Systems (IWES) +49 471 14290 524 [email protected]

– 234 – Annual Report 2015-2017 DOCUMENTATION / STAFF Annual Report 2015-2017235

Schlurmann, Torsten, Prof. Dr.-Ing. Scholle, Niklaas, Dipl.-Ing. Seidel, Elke Universität Hannover Universität Hannover ForWind Head Office Franzius-Institute for Hydraulic, Institute for Building Materials Science +49 441 798 5092 Waterways and Coastal Engineering (IfB) [email protected] +49 511 762 19021 +49 511 762 3258 [email protected] [email protected] Seifert, Janna, M.Sc. Universität Oldenburg Schmidt, Andreas, Dipl.-Ing. Schottler, Jannik, M.Sc. Institute of Physics Universität Oldenburg Universität Oldenburg AG Wind Energy Systems (WESys) Institute of Physics Institute of Physics +49 441-798-5036 AG Wind Energy Systems (WESys) AG Turbulenz, Windenergie and [email protected] +49 441 798 5086 Stochastics (TWiSt) [email protected] +49 441 798-5090 Segelken, Katharina, M.A. [email protected] ForWind Head Office Schmidt, Boso, Dipl.-Ing. +49 441 798 5088 Universität Hannover Schramm, Matthias, M.Sc. [email protected] Institute for Concrete Construction (IfM) Universität Oldenburg +49 511 762 3355 Institute of Physics Segelken, Lars [email protected] AG Turbulenz, Windenergie and Universität Oldenburg Stochastics (TWiSt) Institute of Physics Schmidt, Michael, Dipl.-Phys. +49 441 798-5090 +49 441 798-5074 Universität Oldenburg [email protected] [email protected] Institute of Physics AG Wind Energy Systems (WESys) Schröder, Christian, Dipl.-Ing. Seume, Jörg, Prof. Dr.-Ing. +49 441 798 5086 Universität Hannover Universität Hannover AG Energy Meteorology (EnMet) Institute for Geotechnical Engineering Institute of Turbomachinery and Fluid +49 441 798 5074 (IGtH) Dynamics (TFD) +49 511 762 4152 +49 511 762 2733 Schmietendorf, Katrin, M.Sc. [email protected] [email protected] Universität Oldenburg Institute of Physics Schröder, Karsten, Dr. Shah, Syed Muzaher Hussain, M.Sc. AG Turbulenz, Windenergie and Universität Hannover Universität Oldenburg Stochastics (TWiSt) Institute of Structural Analysis (ISD) Institute of Physics +49 441 798-5054 +49 511 762 2885 AG Turbulenz, Windenergie and [email protected] [email protected] Stochastics (TWiSt) +49 441-798-5067 Schmoor, Kirill, Dipl.-Ing. Schüttler, Jochen, Dr.-Ing. [email protected] Universität Hannover Universität Bremen Institute for Geotechnical Engineering Institute of Automation Shirzadeh, Rasoul, M.Sc. (IGtH) Universität Oldenburg +49 511 762 5106 Schwack, Fabian Institute of Physics [email protected] Universität Hannover AG Wind Energy Systems (WESys) Institute for Machine Design and +49 441-798-5060 Schneemann, Jörge, Dipl.-Phys. Tribology (IMKT) [email protected] Universität Oldenburg +49 471 14290 524 Institute of Physics [email protected] Shrestha, Binita, M.Sc. AG Wind Energy Systems (WESys) Universität Oldenburg +49 441 798 5062 Schwarzer, Christoph, Dipl.-Oek. Institute of Physics [email protected] ForWind Head Office AG Wind Energy Systems (WESys) +49 441 798 5084 +49 441-798-5063 Schneider, Sebastian, Dipl.-Ing. [email protected] [email protected] Universität Hannover Institute for Concrete Construction (IfM) Schyska, Bruno, M.Sc. Singh, Piyush M.Sc. +49 511 762 3359 Universität Oldenburg Universität Oldenburg [email protected] Institute of Physics Institute of Physics AG Energy Meteorology (EnMet) AG Turbulenz, Windenergie and +49 441 798 5079 Stochastics (TWiSt) [email protected] +49 441-798-5024 [email protected]

– 235 – 236DOCUMENTATION / STAFF Annual Report 2015-2017

Sonnenschein, Michael, Prof. Dr. Thoben, Klaus-Dieter, Prof. Dr.-Ing. habil. Ungurán, Robert. Dipl.-Ing. Univversität Oldenburg Universität Bremen Universität Oldenburg Department of Computing Science Institute for Integrated Product Institute of Physics +49 441 798 2750 Development (BIK) AG Wind Energy Systems (WESys) [email protected] +49 421 218 50005 +49 441 798 5064 [email protected] [email protected] Sorg, Michael, Dipl.-Ing. Universität Bremen Tomann, Christoph, Dipl.-Ing. Vali, Mehdi, M.Sc. Bremen Institute for Metrology, Universität Hannover Universität Oldenburg Automation and Quality Science (BIMAQ) Institute for Building Materials Science Institute of Physics +49 421 218 64620 (IfB) AG Wind Energy Systems (WESys) [email protected] +49 511 762 3719 +49 441 798-5063 [email protected] [email protected] Späth, Stephan, Dipl.-Met. Universität Oldenburg Trabucchi, Davide, M.Sc. van Dooren, Marijn, M.Sc. Institute of Physics Universität Oldenburg Universität Oldenburg AG Energy Meteorology (EnMet) Institute of Physics Institute of Physics +49 441 798 5079 AG Wind Energy Systems (WESys) AG Wind Energy Systems (WESys) [email protected] +49 441 798 5062 +49 441 798 5064 [email protected] [email protected] Stammler, Matthias Fraunhofer Institute for Wind Energy Tracht, Kirsten, Prof. Dr.-Ing. Valldecabres Sanmartin, Laura, M.Sc. Systems (IWES) Universität Bremen Universität Oldenburg +49 471 14290 522 Bremen Institute for Mechanical Enginee- Institute of Physics [email protected] ring (bime) AG Wind Energy Systems (WESys) +49 421 218 64840 and AG Energy Meteorology (EnMet) Steinfeld, Gerald, Dr. [email protected] +49 441-798-5064 Universität Oldenburg [email protected] Institute of Physics Traphan, Dominik, Dipl.-Phys. AG Energy Meteorology (EnMet) & Universität Oldenburg Varasteh, Kamaloddin, M.Sc. AG Turbulence, Wind Energy and Institute of Physics, AG Turbulence, Wind Universität Bremen Stochastics (TWiSt) Energy and Stochastics (TWiSt) Institute for Integrated Product +49 441 798 5073 +49 441 7983 514 Development (BIK) [email protected] [email protected] +49 421 218-64872 [email protected] Tambke, Jens, Dipl.-Phys. Trei, Wilke, M.Sc. Universität Oldenburg Universität Oldenburg Vera-Tudela Carreno, Luis Enrique, M.Sc. Institute of Physics Institute of Physics Universität Oldenburg AG Energy Meteorology (EnMet) AG Energy Meteorology (EnMet) Institute of Physics +49 441 798 5072 +49 441-798-5077 AG Wind Energy Systems (WESys) [email protected] [email protected] +49 441 798 5063 [email protected] Terceros, Mauricio, M.Sc. Trujillo, Juan José, M.Sc. Universität Hannover Universität Oldenburg Vollmer, Lukas, M.Sc. Institute for Geotechnical Engineering Institute of Physics Universität Oldenburg (IGtH) AG Wind Energy Systems (WESys) Institute of Physics +49 511 762 4153 +49 441 798 5062 AG Energy Meteorology (EnMet) [email protected] [email protected] +49 441 798 5079 [email protected] Thieken, Klaus, Dipl.-Ing. Tsiapoki, Stavroula, Dipl.-Ing. Universität Hannover Universität Hannover von Bremen, Lueder, Dr. Institute for Geotechnical Engineering Institute of Structural Analysis Universität Oldenburg (IGtH) +49 551 762-8063 Institute of Physics +49 511 762 4153 [email protected] AG Energy Meteorology (EnMet) [email protected] +49 441 798 5071 [email protected]

– 236 – Annual Report 2015-2017 DOCUMENTATION / STAFF Annual Report 2015-2017237

Vorspel, Lena, M.Sc. Wolff, Miriam, Dipl.-Inform. Universität Oldenburg Universität Oldenburg Institute of Physics, Institute of Physics AG Turbulence, Wind AG Energy Meteorology (EnMet) Energy and Stochastics (TWiSt) +49 441 798 5092 +49 441 798 5056 [email protected] [email protected] Wurps, Hauke, M.Sc. Voss, Stephan M.Sc. Universität Oldenburg Universität Oldenburg Institute of Physics Institute of Physics AG Energy Meteorology (EnMet) AG Wind Energy Systems (WESys) +49 441 798 5092 +49 441 798 5065 [email protected] [email protected] Yassin, Khaled, M.Sc. Wächter, Matthias, Dr. Universität Oldenburg Universität Oldenburg Institute of Physics Institute of Physics AG Turbulenz, Windenergie and AG Turbulence, Wind Energy and Stocha- Stochastics (TWiSt) stics (TWiSt) +49 441 798-5058 +49 441 798 5051 [email protected] [email protected] Zack, Charis Wahl, Benjamin, M.Sc. ForWind Head Office Universität Oldenburg +49 441 798 5081 Institute of Physics [email protected] AG Turbulenz, Windenergie and Stochastics (TWiSt) Zhao, Shumian, M.Sc. +49 441-798-5090 Universität Oldenburg [email protected] Institute of Physics AG Turbulenz, Windenergie and Weber, Simon Stochastics (TWiSt) Universität Hannover +49 441 798-5055 Institute for Drive Systems and Power [email protected] Electronics (IAL) Zorn, Christian, Dipl.-Ing. Wernitz, Stefan, M.Sc. Universität Bremen Universität Hannover Institute for Electrical Drives, Power Institute of Structural Analysis (ISD) Electronics and Devices (IALB) +49 511 762 4273 +49 421 218 62669 [email protected] [email protected]

Wester,Tom, M.Sc. Universität Oldenburg Institute of Physics AG Turbulenz, Windenergie and Stochastics (TWiSt) +49 441 798-5023 [email protected]

Witha, Björn, Dipl.-Met. Universität Oldenburg Institute of Physics AG Energy Meteorology (EnMet) +49 441 798 5075 [email protected]

– 237 – 238 Annual Report 2015-2017

Imprint

Publisher ForWind – Center for Wind Energy Research f the Universities of Oldenburg, Hannover and Bremen

Executive Board Prof. Dr. Martin Kühn Prof. Dr.-Ing. Bernd Orlik Prof. Dr. Joachim Peinke Prof. Dr.-Ing. habil. Raimund Rolfes Prof. Dr.-Ing. Peter Schaumann Prof. Dr.-Ing. Klaus Dieter Thoben

Address ForWind Head Office Carl von Ossietzky University of Oldenburg W33 - WindLab Küpkersweg 70, 26129 Oldenburg +49 441 798-5090 [email protected] www.forwind.de

Editorial Coordination Dr. Stephan Barth Jana Stone

Photos ForWind

Layout Dagmar Weinreich-Brunner

– 238 –