ET -AFM - 9901 v______' MASTER
Database on Wind Characteristics L.
'• K.S. Hansen The Technical University of Denmark (editor) & M.S. Courtney 1 Ris0 National Laboratory, Denmark s (editor) Q 1 % a* ET - ARM - 9901
Database on Wind Characteristics
K.S. Hansen1 & M.S.Courtney 2
1The Technical University of Denmark
2Ris0 National Laboratory, Denmark (editors)
Contract JOR3-CT95-0061 Final report
Resarch funded in part by THE EUROPEAN COMMISSION in the framework of the Non Nuclear Energy Programme JOULE III
Department of Energy. Engineering Technical University of Denmark August 1999 DISCLAIMER
Portions of this document may be illegible in electronic image products. Images are produced from the best available original document. Abstract: This report describes the work and results of the project: Database on Wind Characteristics which was sponsered partly by the European Commision within the framework of JOULE III program under contract JOR3-CT95-0061
The organisations that participated in the project consists of five research organisations: MlUU(Sweden), ECN(The Netherlands), CRES(Greece), DTU(Denmark), Rise (Denmark) and one wind turbine manufacturer: Vestas Wind System AZS(Denmark).
The overall goal was to build a database consisting of a large number of wind speed time series and create tools for efficiently searching through the data to select interesting data.
The project resulted in a database located at DTU, Denmark with online access through the Internet. The database contains more than 50.000 hours of measured wind speed measurements.
A wide range of wind climates and terrain types are represented with significant amounts of time series. Data have been chosen selectively with a deliberate over-representation of high wind and complex terrain cases. This makes the database ideal for wind turbine design needs but completely unsuitable for resource studies. Diversity has also been an important aim and this is realised with data from a large range of terrain types; everything from offshore to mountain, from Norway to Greece.
Keywords: wind, turbulence, windloads, extreme wind conditions, gusts, wind shear, flow measurements, databases.
Authors: Kurt S. Hansen, Department of Energy Engineering, Technical University of Denmark Mail: [email protected]
Michael S. Courtney MetSupport ApS, Denmark Mail: [email protected] Contents
Executive summary ...... 1
1. Introduction ...... 2 1.1 Background ...... 2 1.2 Participants...... 3
2. Overview of the database system...... 6 2.1 Introduction ...... 6 2.2 Server organisation ...... 6 2.3 Data handling, input ...... 7 2.4 Data handling, output - the web interface...... 9
3. Archiving interesting data...... 10 3.1 Selecting interesting data...... 10 3.2 Converting to a standard format ...... 10 3.3 Data Arrival Procedure ...... 15 3.4 Checking the quality ...... 18 3.5 Characterising the data...... 19 3.6 Storing the details - the SQL database...... 21 3.7 Technical details...... 26
4. Finding and visualising time series...... 32 4.1 Downloading time series using FTP...... 32 4.2 Searching for data using a web browser ...... 33 4.3 Advanced queries ...... 37 4.4 Visualising time series...... 38
5. Contents of the database...... 40 5.1 Site information ...... 42
6. Extremes...... 66 6.1 Mean wind speed and turbulence intensity...... 66 6.2 Wind speed probability density distribution ...... 67 6.3 Maximum wind speed...... 68 6.4 Maximum gusts ...... 68 6.5 Wind direction change ...... 69 6.6 Gust directional index...... 70
7. Conclusion ...... 71
Annex A1: Definitions ...... A2 Annex B1: Project descriptions ...... A16 Annex B2: Site descriptions ...... A20 Annex B3: Master Sensor file...... A27 Annex B4: Template for common file format ...... A36 Executive summary
Vast amounts of wind data have been measured in many different locations. Of the thousands of hours of time series collected, only a tiny proportion are available for use by the wind turbine and wind engineering communities. This report describes the project of collecting a small but representative portion of these data and making them available on the World Wide Web.
In order to implement a suitable search system we have constructed a database for the detailed registration of field measurements, ranging in scope from the administrative level down to the mounting details of individual sensors. Wind data are quality checked according to a number of different criteria such as presence of spikes, noise and trends. Subsequently data are indexed using a variety of parameters, including conventional statistics and extremes, turbulence intensity, gust, acceleration and wind shear.
A wide variety of wind climates and terrain types are represented together with significant amounts of data measured in and close to wind farms. Data have a typical temporal resolution of 1-20 Hz and as such are intended for design and simulation studies (not wind resource applications). Emphasis has been gived to ensure a high level of documentation of the measurement setups which are included in the database. Furthermore, a search and data selection system has been developed that fully utilizes interactive nature of the World Wide Web.
After quality control and indexing, the actual wind data are copied in a standard format to CD- ROM’s. All CD-ROM’s reside in a juke box containing up to 150 disks, giving a capacity of around 100 GB (with present CD technology). The juke box is accessible via ftp which also allows direct downloading of data files from a web browser.
Data are provided by all the participants of the project, covering most of the countries of the European Union. By spring 1999 we have gathered and included approximately 50.000 hours of wind speed measurements representing more than 20 sites.
As an indication of the applicability of the database, we can refer to two projects that are allready benefiting from this rich resource, The Joule NEWGust project, coordinated by Delft University of Technology, uses wind data from the database to verify and calibrate theoretical models for gust shape. A Danish funded project uses wind time series to examine the precise form of the high sped tails of the wind speed distribution and determine how much these deviate from Gaussian form. This has particular relevance to fatigue loading of the wind turbines.
In conclusion it should be remembered that the database coordinators welcome offers of new data. We consider the Database on Wind Characteristics to be the natural repository for many of the numerous existing and future wind datasets.
1 1. Introduction.
It is now possible for designers of wind turbines to generate simulated time series of turbulence that to a good approximation describe ’normal ’ conditions in situations with reasonably homogeneous terrain.
When it comes to terrain inhomogeneities and windfarms we have some knowledge, but it is very difficult to cover the wide range of conditions that are of interest in a general manner. Therefore it is necessary to have access to time series of real data, covering a wide range of conditions, also those that fall beyond the most common ones. The main objective of this project is to provide wind turbine manufacturers with easy access to such data by collecting much of the large mass of existing wind time series and making them available on the Internet. This will enable the designers of wind turbines to optimize their designs under a variety of conditions without requiring them to collect new data.
Real data from rare events of violent, inhomogeneous, instationary conditions such as those found during frontal passages, in squall lines and in very rough mountainous terrain are not generally available to wind turbine designers or people defining standards. But real data like these do exist, they have been measured as part of wind turbine evaluation programs all over Europe. With this project we will provide general access to this huge existing mass of data, thereby greatly increasing the usefulness of this data to the wind energy community.
Apart from wind turbine designers, the database will also be useful to designers of buildings, bridges and other structures subject to the forces of the wind. Whilst the primary audience is the wind energy industry, follow-up projects is planned to provide access to the database to the broader wind engineering community. Subsequently, it is also the intension to include data from countries outside Europe, thereby extending the coverage to all areas of relevance to the European wind turbine industry.
1.1 Background.
This project was initiated as a consequence of an IEA Expert meeting on wind conditions for wind turbine design [1] which was held in Hamburg, 1994. A primary conclusion of this meeting was the need to collect many of the existing wind time series and make them generally available to the wind energy community.
The project has been proposed and coordinated by Rise (Coordinator) and DTU (Contractor) and the two institutions have in cooperation performed the following tasks:
- Prepared proposal for the DGXII, Joule programme. - Defined workplan and selected partners. - Developed all database definitions and templates for data reporting. - Implemented web and database interface. - Prepared and added data to the database.
2 4
The contact to the data providers has been coordinated and handled by the Contractor. Table 1 -1 contains the list of dataprovidersvjho have contributed time series to the database during the project period.
Table 1-1: Data providers
associated contractors
CRES/GR 3 sites 1556 hours
ECN/NL 2 sites 98 hours
MIUU/SV 6 sites 30000 hours
Riso/DK 3 sites 3468 hours
DTU/DK 1 site 64 hours
subcontractors
NTNU/N 3 sites 22887 hours
DEWI/D 2 sites 1125 hours
Ciemat/E 1 site 280 hours
ENEL/I 1 site 542 hours
ESTB 1 site 242 hours
VUB/B 1 site >100 hours
ENETI/PT 1 site 5 hours
DWD/D 1 site 8 hours
GH/UK 1 site > 600 hours
1.2 Participants.
The partners that participated in the Database on Wind Characteristics project, listed in the table below, consists of two universities, 3 research organizations and a wind turbine manufacturer. Rise National Labotory and Technical University of Denmark did the main work while the role of the other partners was to deliver measured wind data and give input to a proper and usable data handling interface. Furthemore a number of data suppliers has been paid for data delivery.
3 Table 1-2: Overview of consortium Organisation name Role Function Country 1 Rise COO Coordination Denmark 2 DTU C Data handling Denmark associated contractors 3 MIUU AC Data supply Sweden 4 ECN AC Data supply The Netherlands 5 CRES AC Data supply Greece 6 VESTAS AC Discussion Denmark subcontractors
7 DEWI SC Data supply Germany 8 CSTB SC Data supply France 9 ENEL SC Data supply Italy 10 Ciemat SC Data supply Spain 11 NTNU SC Data supply Norway 12 DWD SC Data supply Germany 13 INETI SC Data supply Portugal 14 GarradHassan SC Data supply UK 15 VUB SC Data supply Belgium
The following persons participated or contributed to the project.:
• Rise National Laboratory (RIS0): M.S.Courtney, J.Hojstrup, H.E.Jorgensen, E.L.Petersen. • Technical University of Denmark (DTU): K.S. Hansen,D.N.Sorensen. • Department of Meteorology Uppsala University: A-S.Smedman, M.Magnusson • Netherlands Energy Research Foundation (ECN): J. Dekker • Center of Renewable Energy Sources (GRES): G.GIinou • Vestas Wind Systems A/S: H.K.Jergensen, M.Friedrich • DEWI: A.AIbers • CSTB: C.Sacre • ENEL: M.Cavaliere • Ciemat: J.Navarro • NTNU: J.Lovseth • DWD: B. Hennemuth • INETI: A.Estanqueiro • GarradHassan & Partners LtD: D.Quarton, M.Johnston • VUB: LDeWilde
4 V
1
i
-!
,1
. !
.1 i
■1 i I
i ■'i i
!
1 • I 2. Overview of the database system.
2.1 Introduction.
From the initial phase of the project [spring 1996] is was obvious that the time series database should be implemented as a world wide web (www) system. In this way any user with an Internet connection and a web browser, regardless of computer or operating system type, can obtain access to
- All raw time series. - Signal quality, indexed values and statistics. - Background documentation. - Interactive, online queries. - Interactive, immediate download of interesting time series. - Online graphical view of results. - Online graphical view of time series.
Due to the fast development in internet-tools, much consideration was given to ensure an advanced, yet stable, platform for development and maintenance of the database. By the time both the development software and the developers abilities have matured, the technology has moved on. Now the database operator must use significant resources in keeping the old technology running with a constant stream of new and not always compatible browser versions.
2.2 Server organisation.
The objective of the project is to organise raw, measured time series together with an appropriate background documentation. Organising the data also includes the process of establishing an easy access to specific signals and time series - in other words establishing an useful index database for all the time series.
From the user side, the database system is organised in three elements: web server1, database server and FTP-server as illustrated in figure 2-1. These three servers can be physically present on a single machine. At present the FTP server and the database server reside on one computer, the web server on a separate machine. The function of the three servers are described below:
FTP server gives access to the raw time series which are stored on CD-ROM’s located in a 150 disk jukebox (approximately 100 Gbytes total capacity).
Web server is used as a gateway and provides browser access to the ftp server, input definition documents, background documentation and issues queries to the database server by use of SQL2 protocol.
'Server is used in the terms of giving access to the hosted data or documents. 2 Standard Query Language used for database queries.
6 ftp server JUKEBOX containing
Web browser Web server Internet • documentation
Database server • alts Information* • Instrument Information • screening results • bselc statistics • Indexed values
Figure 2-1: Structure of database servers.
Database server: contains large tables with project, site and instrumentation specific information. The main contents of the database server is the basic statistics for all time series, together with indexed values based on a reference period of 10 minutes.
2.3 Data handling, input.
The data handling process for entering data into the database consists of several steps as illustrated in figure 2-2. All the actions reflect the internal database structure, which was defined before the criteria for data handling could be settled.
The preliminary step is to identify the potential time series suitable for the database in terms of terrain, frequency and number of sensors.
The required time series are converted from "local" format to the common file format. This work is performed by the data provider, who has knowledge of the original data format and details concerning the measurement setup. In addition documentation of the measuring site and instrumentation setup are prepared in a common descriptive format. When complete, the data files and documentation are forwarded to the Contractor together with a data arrival message containing a list of time series attached, this formally completes the data transfer.
The second step consists of data format checking and registration. The site_code is created, and all the documentation are included in the database. All time series are checked for correct file structure and format and is corrected where necessary. Afterwards the time series are copied to consecutive numbered CD-ROMs, included in the jukebox and made publicly accessible through the FTP server. Finally, each time series is registered in the database, including information on time, location and basic statistics.
7 The third step consists of a data quality screening according to some predefined screening criteria prepared in accordance with the database structure. The results from this quality screening are included in the database. Afterwards, the time series are indexed at a ten minute level according to indexing criteria, defined in accordance with the internal database structure. The result of the indexing process is finally included in the database. Both the data screening and indexing process are performed with a dedicated client/server software package developed explicitly for this purpose.
Common Selection format criteria specification
/ User X. Raw Conversion Documentation Input data programme, Common format
Database structure
DATABASE Screening criteria Documentation / Datax Registration screening Quality Perams. programme Statistics Index Parents. Indexing criteria / Data xi indexing Directory programme structure Data input
WEB SERVER SERVER Jukebox
Query systems Browse docum. Project status Query Input defs. parents
User activity web browser - ftp download Data output
Figure 2-2: Structure of data handling.
8 2.4 Data handling, output - the web interface.
User access to the database server is established through a web browser (e.g. Netscape Communicator or Internet Explorer) at the address http://www.winddata.com . Through a browser a user can inspect or download background information, definitions and documentation. Most importantly, the user has access to the search system which allows identification of time series having the properties of interest. Searching is performed by filling out the desired parameters on pre-defined forms. The connection between the query forms and the database server is based on SQL and is invisible for the user.
Access to the data included in the database is performed with two different query types:
Simple queries enables quick and simple access and are based on nominaf 3 values and basic statistics1*. Here, the idea is to reduce the volume of statistics as much as possible by representing the time series by one mean speed {the mean of all the speed sensors), one direction (the mean of all the direction sensors) and one turbulence intensity (the mean turbulence intensity from all the speed sensors) only. Obviously, only a general impression of the time series remains but sufficies to satisfy perhaps 80% of user queries. Being relatively compact, searching in this database is fast and straightforward. The results are presented either in tables, plots or as a list of files referring directly to raw time series located in the jukebox.
Advanced queries are based on 10 minutes statistics and indexed values. The statistics and the indexed values can be used in a large number of combinations. Here, the statistics contain values for each channel for each ten minutes of the run (in contrast to basic (=nominal) statistics which contain only one speed, one direction and one turbulence intensity per run). Due to the level of detail, searching in these data is slower than for a simple query. The output from the advanced queries consists of a list of ten minute periods fulfilling the specific query parameters. Furthermore the results contain links to time series located in the jukebox.
Each run is characterized with only nominal speed, one nominal direction and one nominal turbulence intensity. The exact definition of the nominals are given in annex A1. Basic statistics represent periods varying between 600 and 3600 seconds.
9 3. Archiving interesting data.
This chapter describes the time series selection criteria and the process of entering data into the database. An overview of the database structure is given together with a short technical description of the database server.
3.1 Selecting interesting data.
The objective of the project is to create a database with large amounts of time series of wind data representing a range of terrain types, instrumentation and operational wind turbine conditions. Establishing selection criteria for time series suitable for the database was the preliminary goal of the project. The minimum requirements for acceptable data include:
At least one wind speed measurement. At least one wind direction measurement. Time series saved at a suitable temporal resolution (1 - 25 Hz). At least 50 hours of measurements. Site or terrain type should be wind energy or wind engineering relevant. Well documented site and instrumentation.
Many experiments fulfill these criteria. To further narrow the choice it is necessary to consider the use of the database. Desirable features in discriminating amongst time series include:
High wind speeds. Rare extreme events (storms). Sonic anemometer measurements (3D turbulence). Long periods of uninterrupted data. Good spatial (horizontal and vertical) resolution. Rare terrain type (for example offshore). Measurements from wind farms including wake conditions. Climatological parameters (temperature, radiation, pressure, rain) available. Long term wind climatology available in addition to time series data.
3.2 Converting to a standard format.
There are as many different data formats as there are experimenters. The second major goal of the project was to define a standard format both for the background information and for the time series themselves.
3.2.1 Background information.
In order to facilitate description of the background information for inclusion in the database, it was necessary to define reporting templates. These templates ( empty forms ) reflect the underlying database structure and comprise 3 parts.:
10 1. Project description file This description contains information about the project, institutions and the contact address for people working with the project together with list of references related to the project. A detailed description of the input format is listed in Annex B1. Projects usually "own" one site but for an experiment carried out at a regional or national level (for example a regional resource survey) it might be appropriate to define one governing project and a number of “child" sites.
2. Site description file This file documents descriptive features of the site: the location, terrain type, orography type; the mast (or masts) and any nearby wind turbines. Drawings, photos, maps and wasp description files may be included in the description. A detailed definition of the format is given in Annex B2. As indicated, there may well be several masts at one site. When the mast separation (spacing) approaches several kilometres, it may be more appropriate to define separate sites. A guiding rule here could be whether there would be any reasoning in using the time series data from the separated masts together in the same analysis.
3. Master sensor list. The master sensor list defines all the instruments and channels entered in the database together with mounting information. Here it is important to distinguish between models, instruments and signals.
An instruments a physical sensor, defined by its mocfe/type and serial number. All its salient physical features (e.g. dimensions and weight) are defined in the model type. The serial number specifies the precise physical manifestation of the given model type. Unique instrument characteristics are for example purchase date, service and calibration records.
A given model has one or more output signals, the discrete physical properties measured by the model. For example a cup anemometer has one signal (wind speed) whilst a sonic anemometer has four signals (3D wind velocity vector and virtual temperature).
The internal database structure further defines mountings, a complete definition of where an instrument is mounted and channels, a given signal from a given instrument at a given mounting. For the sake of simplicity, the master sensor list merges signal and mounting definitions together so that in practice a signal definition defines a channel in the database. A detailed definition of the master sensor list format is given in Annex B3.
3.2.2 Time series data files - common file format.
Several different options for data formats were discussed and it was finally decided to use a simple, flexible ASCII-file format, which could be adapted to the commonly used platforms (UNIX, MS-DOS, MS-Win95, MS-WinNT, Apple/Mac). This format is rather space consuming, but since the data are always stored in compressed form (using pkzip), this is not considered to be a problem. File compression and decompression software are readily available on all platforms. A further choice concerns how to handle multiple channels and multiple frequencies. Whilst data handling is simplified with only one channel per file, for many sites this would result in a large number of files per run. One file only containing all channels with possible differing frequencies would be unwieldy both for input and output. One file per discrete frequency results in few files per run (1 or 2) and has a simple internal file format. This was the solution adopted. The file name is derived from the date, time and frequency as follows:
tttt_fff.dat where tttt is the start time (hours and minutes) fff is the sampling frequency in tenths of Hz.
The common file format consists of a header with general information and basic statistics, followed by the scaled data (one row per scans / one column per channel). A definition of the subheader contents are given below and in detail in Annex B4. The last part named [Additional Statistics] is optional:
[Common File Header] 1. Unique site code 2. Date and time of recording, 3. Name of project information file, 4. Name of site information file, 5. Name of master sensor file 6. Sequence number 7. Available frequencies covering this period 8. Name of files covering the same period 9. Duration 10. Sensor configuration number 11. Runname based on recording time 12. Site name which could be equal to site code
[File Header] 13. Name and location data file (file structure) with extension "dat". 14. Number of scans 15. Number of sensors 16. Frequency of the current file
[Sensor Statistics] 17. Channel type, quality, height, wake, name, mean, stdv., min, max, unit
[Additional Statistics] 18. Channel type, quality, height, wake, name, mean, stdv., min, max, unit
[Data field]
The major fields are described below.: [Common File Header] contains basic information referring to the site and is repeated in all
12 the files (other frequencies) associated to this run.
[File Header] defines the size of the current data file, unique to this frequency.
[Sensor Statistics] contains the basic statistics for each of the channels present; one row per sensor. From left to right, the fields are:
sensor type (s=speed, d=direction etc.) quality index (1=good, -1=bad) sensor height wind turbine wake (0=no wake in signal, 1=wake in signal) sensor name mean standard deviation minimum maximum units
[Additional Statistics] is optional and contains statistics for channels which are not present in the data file as time series. The additional statistics section is used to include other relevant information which are available such as climatology and wind turbine power outputs. The fields are defined as for the [sensor statistics], described above.
[Data Field] section contains all the scaled data one scan per row, meaning that all channels listed in the same row are recorded at the same time. All data are stored in physical units [m/s], [deg] or [degC]. For time series delivered for inclusion, sonic data should be stored as raw, unaligned time series. Data alignment is performed at DTU, ensuring a uniform alignment procedure and the aligned data (together with the raw, unaligned data) are included in the public version of the time series.
To conserve space, the data files do not contain any time signal. This time signal has to be derived based on starting time and the frequency, i.e. t, = tsum + (i-1)/frequency [sec]. An example of the data format is shown in Jbox 3-1 /and further details are given in Annex B4.
13 Box 3-1: Example of data file in common file format.
3.2.3 File structure.
In order to keep track of the vast amount of time series, the following directory structure has been defined. The directory structure is based on three key parameters: site_code, year and day of recording. Time of recording is given by the first four digits of the file name as described above. This is illustrated in figure 3-1.
14 /------— File structure in Jukebox I sit* I vnr dtv I file site codel Drive:— site_code2 emden 1988 sited code4 1989 1990 1991 -1992 day001 1993 day002 hhmmjftop day033 1738_250jUp f FILENAME CONVENTION ^ 1838_250.zlp hhmm_fff.dat (uncompressed data) day134 hhmm_fff.zip (compressed data) 2040_010.zip hh=hour 2040_250.zip mm=minute fff=10 % scan frequency [Hz] J
Figure 3-1: File structure.
3.2.4 Data Arrival Message.
The data provider prepares the background information and the time series in the common file format according to the description given in section 3.1.1. All the background information and the time series are forwarded to the Contractor either by CD-ROM or uploaded to the FTP server. The data is enclosed with a data arrival message (DAM) which contains information about the data structure and amount of delivered data which all together are forwarded to the contractor.
The contents of the DAM should as a minimum contain: Site name List of delivered directories + no. of files in each directory including project, site description and master sensor f iles. Description of data: Year & day_names of measuring period. Total number of files: Total number of Megabytes
Data should be packed using PKZIP (but not as self-unpacking files).
3.3 Data Arrival Procedure.
After data arrival the post processing are performed according to the flow chart shown in figure 3-2. Data arrival, validation, screening & indexing.
FTP
CD-ROM zip file and Mssala. integrity data M/O-Disk, d*ta test structure test isesi Other
final ABgnmentot zip Create sonic data integrity production ft/or test CD_ROM recalculation ot statistics
Data Data screening indtodng
Figure 3-2: Data handling process.
The following steps are performed as part of the data arrival procedure.:
1. All received material is checked for attached virus according the current state of the antivirus software package.
2.. All time series are checked for correct zip-file integrity which involve a checksum control of the compressed file.
3. All the time series are checked for correct directory structure according to the definition given in section 3.2. The directory structure is rearranged if the structure does not agree with time and date information from the [Common File Header].
4. All time series are checked for correct internal file structure as listed in section 3.2. The file format is fixed in case of errors or deviations before further use.
5. Sensor statistics given in the file header are checked by recalculating the statistics from the time series data. Here a rudimental quality control is applied with two possible sanctions; i) adjustment of the data quality parameter in the [sensor statistics] header to indicate reduced signal quality or ii) amputation of channels with severe problems.
6. Sonic anemometer data are aligned and covariances calculated as described below. The aligned data are added to the time series files and the covariances are added as additional statistics.
16 7. The time series are stored on a new CD-ROM with consecutive numbers [prod_001 , prod_002,..]
8 . The new CD-ROM is tested for integrity (checksum).
9. The new CD-ROM is mounted in the jukebox and is now publicaliy accessible.
10. Details of the time series and the sensor run statistics are entered in the database.
Sonic anemometer signal coordinate transformations
The signals from sonic anemometers consists of 3 unaligned wind speed components (X,Y,Z) reffering to an orthogonal coordinate system. A fourth signal, temperature (derived from the instantaneous speed of sound) is usually also available. The standard Rise alignment procedure for sonic signals is used. This process transforms the unaligned x,y,z coordinate system to a second orthogonal system u,v,w where the u vector points in the direction of the mean flow and hence the mean speeds in the v and w directions are zero. Since there are many different preferences regarding sonic alignment, both the unaligned (x,y,z) and the aligned (u,v,w) sonic signals are present in the published time series file. In addition, a sonic speed signal and sonic direction signal derived from the transformed u,v,w components are written to the data file. Due to ambiguity with respect to sonic axis definitions and often incomplete knowledge of the physical orientation of the sonic, it was decided not to attempt to treat the sonic direction as an absolute quantity. Instead the mean sonic direction is always zero and only the standard deviation is available as a statistic. Covariances, rotation angles and friction velocity are written to the time series header as additional statistics. These are listed in /Box 3-2/:
17 Box 3-2: Output from sonic signal alignment.
3.4 Checking the quality.
Documented data quality control is a major feature of this project. In the following section we describe the various stages in the screening and how the results of this process are stored in the database.
3.4.1 Data screening process.
The data screening process consists of 7 steps which are performed before the data are entered into the database as listed in /box 3-3/.
jlBMggllMBllll Box 3-3: Data screening process
18 The main features of the screening items are to ensure a reasonable and documented data quality. The results of the screening are included as binary information (yes/no= -1/1) plus some of the scalars. i is used to ensure that only active signals are registrered and used in the database. ii is used to ensure that extreme ranges correspond to something like a normal distribution -failure here indicates possible spikes. iii range checking of normalised 4th moment iv range checking of the normalised 6th moment v range checking the signal according to instrument specification vi possible noise vii possible spikes
3.5 Characterising the data.
The purpose of the data indexing is to generate suitable descriptive statistics of the time series which can subsequently be used as selection parameters, in order to pick out relevant time series. Obviously the choice of indexing parameters strongly reflects the intended use of the wind data - wind turbine design.
Run and "nominal" statistics Basic statistics (mean, standard deviation, minimum, maximum) are generated for all channels both at the run (600..3600 seconds) and at the 10 minute level. A further reduction is made by generating “run nominal" values for speed, direction and turbulence intensity by taking a simple mean of all the channels of the relevant signal type, regardless of sensor height. “Nominal speed" for example is the mean of all the speed channels for a given run, “nominal direction ” is the mean (with correct 360-0 handling) of all the direction channels and “nominal turbulence intensity" is the mean turbulence intensity for all the speed channels. These 3 nominal values give a very concise picture of the entire time series and are appropriate for simple queries.
Ten minute indexing Indexing of data are performed in a number of different steps depending of the actual signal type. The basic period for all indexing purposes is 600 seconds as opposed to the actual run duration (from 600 to 3600 seconds) for run and “nominal ” statistics.The results from this data indexing process are included in the search databases and used in the advanced queries. Only time series with nominal wind speed above 3.0 m/s are indexed since low winds are not of interest in terms of wind induced structural loadings. For gust and acceleration parameters, time constants used are At= 2, 5, 10 and 30 seconds. For time series scanned with low frequency (< 2 Hz) only At=5,10 and 30 are used.
Basic statistics The calculated main statistics referring to 600 seconds are: i Mean value. ii Standard deviation.
19 iii Minimum value. iv Maximum value. v Turbulence intensity (for U > 0). vi Stationarity factor (trend variance contribution) (for U > 0), for a calculated linear trend h, is h2/12. vii Trend corrected turbulence intensity (for U > 0) is given by 100/U‘sqiKo 2^, - hz/12) [%].
Determination of gusts The maximum gust size VG is determined for 4 different periods, At=2,5,10 and 30 seconds: The gust shape is illustrated on Figure A1-1, page A9.
i positive gust size +VG is detected as max[St+At -SJ ii negative gust size -VG is detected as min[St+A, - SJ
Determination of wind speed acceleration The maximum speed acceleration VA is determined for 4 different periods, At = 2, 5,10 and 30 seconds. For time series scanned with low frequency (z 2 Hz), only At=5,10 and 30 are used.
i maximum positive acceleration, +VA is determind as max([St+At -SJ/At) during the period At ii minimum negative acceleration, -VA is determind as min([SUA, -SJ/At) during the period At
Determination of wind direction changes The maximum change in wind direction AD, given by
AD =(IDUAf Dtl),
is determined for 4 different periods, At = 2, 5,10 and 30 seconds. For time series scanned with low frequency (s 2 Hz) only At=5,10 and 30 are used. For each At, the maximum value of AD found in each 600 seconds is recorded in the database.
Similarly, the so-called wind directional “acceleration", DA given by
DA = (ID,+6t-D,l/At)
is also determined for the same periods as AD and the maximum value for each 600 seconds of signal is recorded in the database. Note that DA is really an angular velocity but is refered to as directional acceleration because it is analogous to the gust acceleration.
Determination of gust directional Index The maximum value of the Gust Directional Index, GDI is determined for 4 different periods, At = 2,5,10 and 30 seconds. The GDI factor expresses the contemporarily of the maximum gust and the maximum directional change.
20 qqI = ( abs(U(t+Af)-U{f)) _ abs(Dir(t+Af)-Dir(f)) { max(abs((U(t+Af)-U(f)))) max(abs({Dir{t+At)-Dir(t))))J
Maximum value of the GDI index lies within the interval [1,2], A value of 1 corresponds to no correlation whilst 2 implies full correlation between gust and directional change.
Determination of linear wind shear The maximum value for each of the time intervals At = 2, 5, 10 and 30 seconds, of the positive and negative linear wind shear, Sh given by
Sh = (Vh2-Vh1)/Ah is determined for all pairs of speed sensors satisfying the following criteria;
sensors on the same meteorological mast, lower height (hi) > 15m, upper height (h2) < 100m, sensor height difference Ah > 20m.
Determination of shear factor and exponent The variation of wind speed with height can be expressed as
Vh = factor xhexponenl .
The shear factor and exponent are recorded as indexing parameters in cases where the following conditions are fulfilled;
speeds at 3 or more heights on the same mast available, wind speeds increase monotonically with height.
If these conditions are fulfilled, the shear factor and exponent are derived from linear interpolation of the qualifying 600 second mean speeds plotted log-log.
3.6 Storing the details - the SQL database.
3.6.1 Introduction.
Every single item of information, with the exception of the time series themselves, is held in the central SQL database. In many ways the database can be considered as the heart of the entire system. The structure of the database is a reflection of the logical entities and processes that have been identified and implemented. Designing the database has been an iterative process deeply linked with the design of the other major components of the system - the input format specifications, data registration, screening, indexing and the web interface. In the following sections, the basic structure of the database is presented and a brief technical description is given.
21 3.6.2 Structure of the database.
The task of the database is to hold all the information needed for selecting time series together with all the available documentation. Both search flexibility and performance are crucially dependent on the underlying database structure. We have chosen to split the database into five classes, each containing a number of tables. These classes are (Figure 3- 3):
Projects. Sites. Instrument. Runs. Ten_mins.
The first three classes deal with documentation, the last two statistics and indexing parameters. In principle a search for data may contain elements from any of these five, in practice only the last three or four are used. The classes are described below.
3.6.3 Projects, Figure 3-4.
This class contains the institutional background of the project:
Why the project was undertaken. Who was in charge. Who else was involved. Who funded the work. Publications. Documentation relating to the whole project (maps and graphs).
A project typically "owns" one but possibly several sites, for example where one experiment is conducted over a region with a number of separate sites.
3.6.4 Sites, Figure 3-5.
This class contains the description of one site, including:
Location. Terrain type. Orography type. Graphical documentation relating to the site (maps, graphs, photos, drawings). Mast descriptions. Wind turbine descriptions.
A site "owns" one or more masts and possibly a number of wind turbines. For each mast, the following is described:
Location.
22 Height. Description. Roughness data for 12 sectors. Wind turbine wake presence for 12 sectors. Mast graphical documentation (photos, drawings).
Many wind measurements stem from wind turbine tests or wind park measurements. Wind turbine wakes may be a desired or an undesired feature of the wind data but in any case require documentation. Any wind turbines in the vicinity of the measurement site that can influence the data are described:
Location. Description. Diameter. Hub height. Rated power. Rated wind speed.
3.6.5 Instruments, Figure 3-6.
Here we must set out some clear definitions:
Model: a specific type of instrument from a specific manufacturer. Instrument: one particular example of a given model. Signal: one parameter (of possibly several) from a given model. Mounting: where and how an instrument is mounted. Channel: a given signal from a given instrument at a given mounting.
An example may help to clarify these definitions: - channel SY_43m consists of the Y axis speed (signal) from a Gill Solent R3 Sonic Anemometer (model), serial number 234 (instrument) mounted at 43m on mast 3 (mounting).
Once these definitions and distinctions are understood, a logical form for documentation can be devised. Starting with a model, these has the following essential properties:
Manufacturer. Model specification. Type (cup, vane, sonic, ...). Physical properties (weight, dimensions). Graphical documentation (drawings, photos, wiring diagrams). Signals (output parameters).
Signals are defined in there own table with the following basic properties:
Type (speed, direction, sonic_x, sonic_y .....). Time or length constant.
23 Range. Accuracy.
One physical manifestation of a given model is defined as an instrument. It’s physical properties and output parameters are already defined in the models and signals tables. What is important here is the properties specific to this particular instrument:
Serial number. Date purchased. Date last calibrated. Service record.
Now that we have a complete definition of a particular instrument and its output signals we need to define precisely where it was used. The mountings table performs this function with the following essential properties:
Mast. Height. Boom or top mounted. Boom properties (length, depth, form, direction). Mast dimensions at boom height. Sensor orientation.
A channel of data in a file of time series stems from a particular signal of a given instrument mounted as given by the corresponding mountings table entry. Thus channels have the following essential properties -
Name (as appears in a time series file). Instrument. Signal. Mounting.
3.6.6 Runs, Figure 3-7.
Two levels of indexing and statistics have been adopted, the run level with simple statistics of a complete run and ten_mins - detailed statistics and index parameters (gust, direction change and shear) for each ten minute period of a run. In the database structure these two indexing levels are clearly separated Here we describe the run level.
A time series is registered in the runs table, where the run is given a unique internal index and the following salient features are recorded:
Project and site. Run name. Start date and time. Run duration. Position in time series sequence.
24 Each time series file (one per frequency per run) is registered in the run_files table, with the following features:
Frequency. Number of scans. File name, size (packed and unpacked) and volume name.
As described in the previous section, statistics at the run level comprise run statistics; basic statistics of all channels for the complete run and run nominals: speed, direction and turbulence intensity statistics condensed from all the sensors of the appropriate type down to one value for each parameter. These two statistics types are contained in the run_statistics and run_nominals tables. A third table, addstat contains the additional statistics data that may optionally be added to the time series file header to document climatology, wind turbine power output or other relevant, non-wind parameters.
Screening performed at the run level is documented in the run_screening table.
3.6.7 Ten_mins, Figure 3-8.
All the ten minute indexing is documented in the ten_mins level of the database. Here the central table is ten_min_index where each ten minute period of each run is registered and given a unique, internal index. Indexing results are stored in the following tables:
Ten_min_stats. Ten_min_speeds. Ten_min_dirs. Ten_min_gust. Ten_min_accels. Ten_min_dir_change. Ten_min_shears. Ten_min_shear_fit. Ten_min_gdi.
Note that ten_min_speeds and ten_min_dirs contain speed and direction statistics respectively, duplicated from ten_min_stats in order to improve query speed (somewhat at the expense of strict database practice). Ten minute screening is documented in the ten_min_screening table.
25 3.7 Technical details
The database is implemented using Inprise (formerly Borland) Interbase SQL server, version 5.1, running under Microsoft NT server 4.0. This product is a compact, yet reasonably well performing database server with close integration to other Borland products used in the project (notably Delphi and IntraBuilder).
Since database performance is closely related to the available physical memory, the Interbase server runs on a computer with 256MB main memory, of which approximately half is dedicated to the database server. The database indexing files (close to 2GB) are stored on a RAID array of SCSI drives.
raBlipffiljftSSESp
I PROJECTS T SITES INSTRUMENTS)#! RUNS 111 TEN MIN
•facts' .. • PObUcaHoni ;
. * • .graphs" •, J>hij3tos$; e._diaiwtng8 : • ..graphs • MdcieEw&ing iShearsTi...
• MaisVphotds r;e:CfcaHheisv::.
• Wind.turblnes:
Example. :class .-project table ;_maps = pmject_maps
Figure 3-3 - Definition of classes
26 ppips
IIP ' ■PPIR V'! PUBLICATIONS • reference • description • article_bin #B#0 PllF
PROJECT MAPS 1: • description • map_bin . -p.:= PP: %!4.PPr': S#:N PPpjYP.!
PROJECT GRAPHS • description • graph_bin
" ■::; -:'i:' SPpp fip'::'/
Figure 3-4: Contents of project class.
27 u- , SIT 1 1" - "... MAST SECTC
OROGRAPHY TYPES 1
s |
fS^5 iV^'SS" • photo_bin 1
' ST l
Figure 3-5: Contents of site class.
28 :^ristrum&hPinfdrm^i^
INSTRUMENTS MODEL PHOTOS 1 E;, • seriaLno • description • purchased • photo_bin • lasLcaSbreled • lasLserviced
MODEL WIRING • description INSTRUMENT TYPES |m$; •types • wiringjDln • description
"■ .' ,z:» SIGNAL TYPES • types • description • derived_type • units
KT MOUNTINGS SIGNALS • sensor.dlrection • time or length constant • top_nx>unted • mirwnax_values • boom_helght • heighLoffset • boom_dir • accuracy • boomjength • accuracyjxt • boomjdimenston • boom„shape • masLdimension • effective heloht
, : - - T
MODELS CHANNELS • manufacturer • name • model specification • sensor_configu ration • description • sensor.offset • weight • sensor height
I . . MODEL DRAWINGS hr-r ADDITIONAL CHANNELS • description • drawing_pln l • • height • description L-~
Figure 3-6: Contents of instrument class. RUNS • rurrame column_no • start_date • mean, min, max and stdv • year_no • range • week_no • turbjnf • day_no • stationjactor • statLhour • frend_corrected_ti • start_minute • quality index • start_second • wake • duration • sequence number & total • sensor configuration
RUN PROC VERSION BUN SCREENING • registered_date • stdv_good • inchecked date • range_good • aligned date • ln_signal_range • statsed date • four_mom_good • screened date • six_mom_good • indexedjdate • no_spikes • under_10_spikes • under_100_spikes • no_nolse • no_fower_wake BUN.AOMIN ALS • number_of_spikes • no_of_speeds • kurtosis • no_ol_dics • nomlnaLspeed • stdv_of_speeds • nominal_dir • stdv_of_dirs • nomlnalji • stdv_ol_tis
BUN-BILES ADDSIAI »frequency • mean, min, max and stdv. • scans • qualityJndex > fi(e_name • wake • volume_name • packed_slze »unpacked_size
Figure 3-7: Contents of run class.
30 j
Figure 3-8: Contents often-min class. 4. Finding and visualising time series.
This chapter describes how to locate, download and visualise interesting time series. Physically, the time series data reside on cd-roms loaded in a so-called "jukebox". This is a storage system comprising 4 cd-rom drives, racks for 150 cd-roms and a robot system for automatically loading and changing cd-roms. With the help of control software running on an NT server, access to the data on the cd-roms is completely transparent, appearing as part of the local file system for local users.
For remote users, the jukebox data is accessible from the ftp server at address 130.225.71.50 (nt50.afm.dtu.dk). This is the first and most direct method of access to the data. All the time series data files are visible but none of the background information, statistics or indexing parameters. For a user requiring data from a specific site at a specific time this is the fastest and most effective access method.
With the large amounts of data available, it is essential to have a search system that can guide a user to the time series relevant for a particular purpose. This is the role of the Web server which is the second and principal gateway to the time series data. Here all the background information can be viewed and the statistics and indexing parameters are used to search for data with specific characteristics. After finding relevant data, time series can be downloaded simply by clicking on the file name.
If a user requires many time series or has a slow internet connection, a third, less exotic possibility exists. On request (and for a handling charge), copies of cd-roms can be made and sent by conventional post. The web-server may still be used to choose the relevant data since the search and browse systems perform well even on dial-in internet connections.
The first two access methods are described in detail in the following sections. We conclude the chapter by examining a tool for visualising and performing spectral analysis on the time series.
4.1 Downloading time series using FTP.
Access to the time series data with direct ftp is straightforward and efficient but the user must know precisely which data are required. All forms of background information, statistics and indexing parameters reside on the web-server and are not visible with direct ftp access.
At present access to the ftp server is free of charge and the quantity of data that may be downloaded is unrestricted. This policy may be changed in the future when the database system is capable of being financially self-supporting.
Time series can be obtained in different ways depending on the required number of time series and their physical size:
1) The web-server (http://www.winddata.com/ ) contains a link to the ftp-server, enabling the browser to function as an ftp-client, giving direct access to the
32 time series files. Files downloaded in this way do not preserve the directory structure of the database unless this dicipline is manually imposed by the user.
2) Any ftp client program (on any operating system) can be used to access the ftp-server directly at address 130.225.71.50 (nt50.afm.dtu.dk). This is in principle the same as method 1) with the exception that the user must manually log on to the ftp server. At present, anonymous access is permitted with no download quantity limit. Files should be transferred as "ascii" in order to implement correct line termination conversion.
3) A hybrid of both web and direct ftp access is available, giving the advantages of both systems and in addition, automatically preserving the directory structure on download. Data are first identified using either of the web query systems (described in the following section). The "Files* button is then pressed, displaying a list of all the time series files fulfilling the search criteria. Using the browser File menu, this list is saved on the users computer. This file is then used as input to the (Windows 95 and Windows NT) utility program Dbwind.exe which creates the necessary local directories and transfers the files in the list, operating as a background ftp task.
4.2 Searching for data using a web browser.
The principal gateway to the time series data is through a web browser from which the user can access the wind database web-server at address http://www.winddata.com/ . Here the user can view background information and search for data using the query systems. The opening (home) page is shown in Figure 4-1.
Figure 4-1: Opening page for web server.
33 The design philosophy of this home page is to enable quick access to both the background information and to the database itself. The left frame, which is visible from every page on the server, contains links to additional pages with:
Welcome with the coordinators introduction. Site overview. Project descriptions. Getting started (tutorial and basic definitions), input definitions for the data providers. FTP-server access to time series. Software ready for download. Status for the database contents. News with information on the latest updates.
The most central link on the main frame is to the search system. Clicking this link opens the login form which presents two possibilities; login as unregistered or as registered user. Unregistered entry requires only an e-mail address (no password) and gives access to the "Simple" query system described below.
For access to the more sophisticated "Advanced" query system the user must first register (by filling out and submitting the registration form) and subsequently recieves a password. At the present time registration is free but registration tariffs may be imposed later.
4.2.1 Simple query.
The simple query has been designed as a straightforward and fast tool for finding time series. It contains only the most fundamental parameters (speed, turbulence intensity, direction, terrain and orography type) and uses only the "nominal" run statistics (one speed, one direction and one turbulence intensity value per time series).
Figure 4-2: Simple data selection form.
34 Despite the simplicity, it is anticipated that the simple query will be able to satisfy the majority of user requests. The data selection form is shown in Figure 4-2.
Figure 4-3: Matching runs
In order to prevent unnecessarily large network traffic, the user is not permitted to view results from a query until the number of resulting time series is less than 500. To enforce this discipline, only the "Count" button is visible until the parameters selected result in a number of time series under this limit. In other words, no other operations will be available until the "Count" button has been pressed and the number of resulting time series is under 500.
Having entered selection parameters resulting in under 500 time series, the user is presented with a further 3 operations; "View", "Files* and "Plot":
Pressing "View" results in a list of matching runs together with site name and "nominal" statistics as shown in Figure 4-3. On this page, site names and run names appear as hyperlinks. Clicking a site name enters the background information viewing system for the appropriate site. Clicking a run name results in a report giving an overview of the run including the run statistics for each channel and the names of the data files, as shown in Figure 4-4. By clicking on the file names, the time series can be directly downloaded.
35 Figure 4-4: Run statistics
Pressing “Files" from the query form results in a list of data files fulfilling the search criteria (Figure 4-2). Here time series can be downloaded by clicking on the individual data file names. Alternatively the file list can be saved locally and all the data files downloaded by using the Dbwind.exe utility described above.
Pressing "Plot ” from the query form calls a Java based plotting routine which is supplied with the mean speeds, directions and turbulence intensities of the time series satisfying the search criteria. Data can be viewed in various parameter combinations by selecting from the menu in the plot. Right clicking on a point in the plot displays an overview of the run (Figure 4-4) from which this particular time series may be downloaded by clicking the file name.
Figure 4-5: Resulting list of time series - available for download. 36 4.3 Advanced queries.
For searching using 10 minute statistics and the indexing parameters described in chapters the advanced query is used. Here the volume of statistics is many times greater since we are concerned with 10 minute rather than "nominal" run statistics and each channel is represented individually. Combining several parameters may result in a time consuming query. The combined search conditions are defined on three separate forms. The input variables are grouped in three and listed in figure 4-6:
1) ten-minute statistics. 2) transient events. 3) wind shear.
Input/Output to advanced query
Figure 4-6: Parameters used in the advanced queries.
As for the simple query, the user must first find a combination of selection parameters giving less than 500 results. Here the results are given as ten minute periods of a run so one run may well appear several times with a different ten minute period number. Pressing "View" results in a list of ten minute periods (Figure 4-7), sites and nominal statistics. As for the simple query, site and run names are given as hyperlinks having the same function as before. A ten minute number also appears as a hyperlink. Clicking here opens a report containing the ten minute statistics for all the channels together with links to the data files. At the right hand side of the "View" form are links to reports giving detailed information concerning the indexing parameters (gusts, direction change, shear, shear fit and gust directional index). Figure 4-7: Matching 10 minute periods
4.4 Visualising time series.
The data files contain [scaled] data, for example wind speed signals in [m/s] and wind direction signals in degrees. Since the file format is ascii, the time series can be analysed with many different software tools (after removing the file header). It has not been a goal of this project to produce data analysis software since most researchers would rather perform this task themselves.
However, as part of this project an existing software package has been adapted for visualising and performing spectral analysis on data in the common file format, including presentation of header information. Both compressed and uncompressed data files can be analysed directly. This software package (DANAP.EXE) is available in versions for both for UNIX and MS-Win95/NT systems. It can be downloaded together with a user guide from the software library on the web-server.
A number of different analysis tools are available in DANAP:
Plot the time signals including tools for zooming and exporting the signals. Displaying file header information in a separate window. Calculation and visualisation of density distributions.
38 Calculation of power spectra and cross correlation. Calculation of basic statistics.
Figure 4-8 shows a time series plot created with DANAP. The signal represents 1200 seconds of measurements recorded at 140m height with 2 Hz at the Cabauw site in The Netherlands.
39
% m 5. Contents of the database.
This chapter contains a site overview, table 5-1, and a summary of the database contents at the reporting time with reference to each site. This information is limited to a listing of the basic site characteristics while more site specific information are available on the web server.
Huge amounts of data, representing various terrain types, have been screened, indexed and included in the database. Medio 1999,23 sites are represented and further 6 additional sites are under preparation. These 23 sites represents more than 50000 hours of measurements using 10 GBytes of compressed storage.
Section 5.1 contains a instrument and data summary together with the data providers site and measurement description. These informations are listed for each site represented in the database medio 1999. Detailed site and instrument information are available on the web server together with site layout drawing. Site maps showing the exact measuring location are also available from the web server.
Section 5.2 cointains summary plots of the database contents (nominal and 10 minutes turbulence values) and extremes e.g. maximum gust, wind directional change,., registrered in the database during the indexing process. All extreme values are accessible from the Web server together with quality information for each of the time series.
40 Table 5-1: sites represented in the database
sites Inst. Country Hours avail. Alsvlk MIUU S 17500 yes Andros C.R.E.S. GR 625 yes A.Spruzza ENEL I 543 yes Cabauw EON NL 49 yes La Clape CSTB F 242 yes ECN/Petten ECN NL 49 yes Emden DEWI D 780 yes K-W-Koog DWD D 4 yes Jade DEWI D 344 yes Lamme Rise DK 663 yes Lavrio C.R.E.S GR 730 yes Lyse2 MIUU S 3355 yes Oak Creek Rise US 745 yes S.Jorge INETI PT 5 yes Skipheia NTNU N 13175 yes Sletringen NTNU N 3756 yes Tarifa Ciemat E 280 yes Tjasreborg ET/DTU DK 64 yes Toplou GRES GR 172 yes Vallersund NTNU N 5957 yes Vlndeby Rise DK 2060 yes Windy* GH UK >600 no Zeebrugge* VUB B >100 no Total 23 sites appr. 52000 Outstanding timeseries sites Inst. Country Hours avail. Ablsko2 MIUU S na. Marglarp2 MIUU S na. NSsudden2 MIUU S na. Sprogo2 Rise DK na. Tjaereborg NEG/Rise DK na. Utlangan2 MIUU S na..
na. = not available yet 'Ongoing measurements. 2Not converted to standard format or validated yet 5.1 Site information
Site: Alsvik, Sweden
Site_code Alsvik
Dominating terrain / orography coastal/flat
Number of masts / wind turbines 2/4
Number of channels wind speed / wind direction 16 (10,18. 24. 31,36, 41.47 & 56m) /16
Number of sonics, height of meteo.mast 0/54m
Hours of measurements / frequency 17500 /appr. 1 Hz
Comments The Intensive measurements covers 1991,1992,1993 and 1994 with an availability of 50%.The additional statistics includes temperatures and electric power from 4 wind turbines. Only time series from 1991 are indexed.
0 10 20 30 ______Wind speed [m/s]______Figure 5-1: Nominal turbulence for Alsvik site.
42 Site: Andros, Greece
Site_code Andros
Dominating terrain /orography pastoral / mountain
Number of masts / wind turbines 2/1 (+5)
Number of channels wind speed / wind direction 13(18m,32m &49m) / 6
Number of sonics, height of meteo.mast 0/40 m
Hours of measurements / frequency 625/1 Hz
Protect motivation The main objectives of the project were to ajdevelop draft design guidelines for wind • turbines in simple terrain wind farms, bjlnvestigate the behavior of wts in undulating and complex terrain wind farms and c)to extend simple terrain design guidelines to complex terrain.
Measurement system Full scale measurements at a wind farm sited In mountainous terrain consisting of seven V27-225kW fixed speed, variable pitch wts. Measurements of wind speed and direction were canted out at three heights a.g.l. on two meteorological masts using six cup anemometers and six vanes, In orderto define the operational characteristics of one of the seven wts In relation to the wind regime of the site.
Comments One electric power signal Is available as additional statistics. The time series are recorded in 1994.
10 20 30 Wind speed [mis] Figure 5-2: Nominal turbulence for Andros site.
43 Site: Acqua Spruzza, Italy
Site_code Aspruzza
Dominating terrain / orography pastoral / mountain
Number of masts / wind turbines 2 / (2)
Number of channels wind speed / wind direction 4 (30,33 & 40m) /1
Number of sonics, height of meteo.mast 0 / 40 m
Hours of measurements / frequency 543 /1.0 Hz
Project motivation The Acqua Spruzza wind turbine test site has been built by ENEL S.p.A. within the framework of a programme aimed at evaluating the technology of commercial medium-sized machines operating In complex terrain and very hostile climate, with special regard to availability, energy output, lifetime (through the monitoring of loads), operating and maintenance costs. The objective Is to assess the viability and the economic attractiveness of wind farms in hostile terrain and to understand the risks associated with the exploitation of these kind of sites. To this end, a suitable research programme has been outlined, comprising regular performance and load monitoring of the wind turbines and wind monitoring as well, through acquisition of both statistical and campaign series of data. The campaign wind data are available for the present Project ‘Database on Wind Characteristics"
Measurement system A number of data acquisition systems are presently Installed at the Acqua Spruzza test site, namely the Measurement Control and Monitoring system (MCM), the Scientific Measurement Systems (SMS) and the SQUIRREL data loggers. The MCM system allows the general monitoring of wind turbines and the wind measurements from conven tional anemometers installed on two wind masts, namely M1 and M2, through the recording of 10-min main statistics. The SMSs are specifically dedicated to the performance and load monitoring of three wind turbines, through the acquisition of extensive campaign data (20 - 60 Hz). The SQUIRREL data loggers are dedicated to summary wind data acquisition, some of which from sensors specially designed for operation in cold climate, and to the collection of wind data time series at 1 Hz fre quency from mast M1. The wind campaigns recorded by a SQUIRREL data logger are available for the present Project.
Comments The time series from this complex site Includes vertical wind speed recorded In 1997 and 1998.
Figure 5-3: Nominal turbulence for Acqua Aspruzza.
44 1
Site: Cabauw, KNMI Meteo Tower, Utrecht, The Netherlands
Slte_code cabauw
Dominating terrain / orography pastoral / flat
Number of masts / wind turbines 1/0
Number of channels wind speed / wind direction ' 4(20,40,80 & 140 m)/4
Number of sonics, height of meteo.mast 0 /140 m
Hours of measurements / frequency 49 / 2 Hz
Project motivation In the frame work of the project “a manual of design wind data for wind turbines ", measurements from the Royal Netherlands Meteological Institute (KNMI) were used extensively. In this project T a set of wind measurements of about 800 hr, measured in 1985 and 1986, was used. An important subject in 1 this project was the frequency of occurrence of wind gusts and of wind direction changes. Wind data files in which average values (averaging time 600s) greater than 15 m/s are present at 20 m height, were classified as wind files with strong winds. The wind data in this subset (50 hrs) are made available for the "Database on Wind Characteristics ’. The KNMI allows the use of these data under the following conditions: Users of the data refer to the source of the data e.g. with the following sentence: The Cabauw wind data were made available by the Royal Netherlands Meteological Institute (KNMI). The KNMI appreciates to be informed how the data were treated, and likes to get relevant reports. In general commercial use of the data is not allowed. Information about the terms for commercial use can be obtained i at the KNMI, attention to W.A.A. Monna, Section atmosferic research, P.O. Box 201,3730 AH De Bill, NL. •! ! Measurement system The Cabauw meteo mast Is a tubular tower with a height of 213 m and a diameter of 2 m. Guy wires are attached at four levels. From 20m upwards horizontal trussed measurement booms are 4 Installed at intervals of 20 m. At each level there are three booms, extending 10.4 m from the centerline of 4 the tower. These booms point to the directions 10,130,250 degrees relative to North. The SW and N i booms are used for wind velocity and wind direction measurements. These booms carry at the end two lateral extensions with a length of 1.5 m and a diameter of about 4 cm.
Comments The time series are recorded on a very high metmast during 1985 and 1986.
50 Site: Cabauw, Hie Netherlands 294 x 10 minute periods 40 based on 4 speed signals
30 i I 20 3 10
12 16 20 24 28 32 Wind speed [m/s]
Figure 5-4: Nominal turbulence for Cabauw site.
I
i
! 45
m lW7- TT 3: § m Site: La Clape near Narbonne, France
Site_code clape
Dominating terrain / orography scrub / hill
Number of masts / wind turbines 1/0
Number of channels wind speed / wind direction 6 (24m)/1
Number of sonics, height of meteo.mast 1 (40m)/40 m
Hours of measurements / frequency 242 / 2 & 8 Hz
Project motivation In the frame of the European Joule Program concerning wind resource assessment over. complex terrain, the purpose of this study is to provide a set of wind speed data for a hill in the South of Fance, a validation test for the software WASP and to improve the knowledge of turbulence over complex terrain.
Measurement system The hill called "La Clape" is situated in the North face of La Clape mountains, facing tramontana winds. The upwind plain Is very flat over more then 5 km, its altitude is 4 m above sea level; the hill summit is at 151 m a.s.l., which gives a difference altitude of 147 m and an average upwind slope greater than 0.2. Along hill slopes the ground is covered with bushes, vineyards and small pines. The plain Is typically an open country In the neighbourhood of the reference mast with some buildings and hedges, however, far away (3 km from the reference site) there are the town of Narbonne In the North West and the sea In the South. The site of the reference mast is called "Le Cercle". In the plain, a reference mast of 24 m height was equipped at two level (10 m and 24 m) with respectively a vane propeller and a three directional Gill anemometers and also temperature sensors, n the summit, a 40 m high mast had been erected and equipped at three levels: at 10 a vane propeller, at 24 m a three directional gill propeller and at 40 m a sonic anemometer.
Comments The time series are recorded during 1992 using 3-D Gill UVW Propeller anemometer.
50 Site: La Clape, France 738 x 10 minute periods 2. 40 based on 1 speed signal •1 30 8
1
0 5 10 15 20 25 30 Wind speed [m/s]
Figure 5-5: Nominal turbulence for La Clape site.
46 Site: ECN, Petten, The Netherlands
Slte_code ecn
Dominating terrain /orography costal/flat
Number of masts /wind turbines 4/1
Number of channels wind speed /wind direction 9 (17,27,37 & 41m)/3
Number of sonlcs, height of meteo.mast 0/37m
Hours of measurements /frequency 49/4 Hz
Protect motivation The primary goal was to design and build and test a flexible rotor (two blades, diameter 21.6" m) with passive tip-pitch control and a teetered hub with an elastomeric bearing. The FLEXTEETER rotor was tested at the 25m HAWT test facility of ECN. During the project the wind Input was measured with a set of four meteo masts. The wind data obtained as part of the test of FLEXTEETER are available for the "Database on Wind Characteristics".
Measurement system The measurement system consists of 4 meteorological mast, mast no. 1,2,3 were placed West of the turbine and mast 4 at the East side. The measurements system was based on a PDP datalogger and the data were transferred to a DIGITAL VAX system (VAX VMS operating system with data handling routines and programs developed at ECN). The data to be measured were divided In two groups: group 1, wind data with a recording frequency of 4 Hz; group 2, turbine operational data and mechanical load data with a recording frequency of 32 Hz. The maximum duration of a consecutive measurement with 32 channels in both groups was about 18 hrs.
Comments This site Is used for wind turbine testing and the time series are recorded during 1992.
50 I------, . Site: ECN, "The Netherlands OQA y 1fl mfnitfe nerinrie
4 8 12 16 20 24 28 32 Wind speed [m/s]
Figure 5-6: Nominal turbulence for ECN test site, Petten, The Netherlands.
47 Site: Emden, Germany
Site_code emden
Dominating terrain /orography costal/flat
Number of masts / wind turbines 1/11
Number of channels wind speed / wind direction 2 (46 & 68m) /1
Number of sonics, height of meteo.mast 0/68 m
Hours of measurements / frequency 780 / 20 Hz
Project motivation Due to the increase In size of today's series production wind turbines Into the MW-range optimised tower designs and minimised use of material are necessary for augmentation of these turbines’ economics. Exact knowledge of the loads is of vital importance In the design and certification process. To arrive at a dependable prediction of fatigue loads they are either calculated using elaborate computer simulations of the turbine’s dynamic behaviour or estimated by means of simplified fatigue load spectra. Lately, the use of load measurements becomes more and more Important in that issue. In Germany both, calculation and estimation are used. Recent measurements have given rise to a discussion about the adequacy of the use of simplified load assumptions. As a contribution to settle this discussion ‘real life’-measurements are carried out by DEWI within a research project which is coordinated by VDMA (the German machinery and plant manufacturers ’ association) and partly funded by the German Research Ministry (BMBF) and AVIF (Research society of the working alliance between the iron, steel and metal working Industries). Typical and statistically approved tower load spectra for most of todays serial produced wind turbine concepts are not available up to date. Procedures how measured load spectra may be modified to comply with the requirements of applicable standards are still under debate. Using the experience of the Monitoring Fatigue Loads - project which has been co-funded by the European Commission under JOULE II DEWI’s research work aims at filling in this lack of Information.
Measurement system A 68 m high mast served as bases for the meteorological measurements. The meteorological data was collected by two independent data loggers. An Ammonlt data logger was used for the long term recording of 5 minute averages of all data relevant for the power curve evaluation of a 1.5 MW WEC. In Addition a modular processor-controlled data acquisition system (MOPS) served for Investigations of the structural loads and dynamic behaviour of the turbine. Only the data gained from this system are relevant for the Data base on Wind Characteristics. The sample rate was 20 Hz and time series with a duration of 600 seconds have been stored.
Comments The time series are recorded during 1997.
0 5 10 15 20 25 30 Wind speed [m/s]
Figure 5-7: Nominal turbulence for Emden windpark.
48 Site: Kaiser-Wilhelm-Koog, Germany .
Slte_code kwkoog
Dominating terrain /orography pastoral/flat
Number of masts / wind turbines 3/(1)
Number of channels wind speed / wind direction 21 (10,50,75,100,125 & 150m)/18
Number of sonlcs, height of meteo.mast 0/150m
Hours of measurements /frequency 4/2 Hz
Project motivation The main goal Is to archive the wind data which were measured during the test programme' of the 3 MW wind turbine GROWIAN at Kaiser-Wilhelm-Koog at the North Sea coast of Germany from 1983 to 1987.
Measurement system Two masts of 150 m height are placed 65 m east-south-east of the 3 MW wind turbi ne, their lateral distance Is 52 m. 20 propellers and vanes are Installed at pairs of booms of 12 m length to the right and to the left of the two masts so that an area of 75m x 100m is covered. The measuring frequ ency is 2.5 Hz and the duration of one measururing run Is approximately 25 min. 300 runs are sampled be tween April 1984 and February 1987 at different Inflow conditions. 10 data sets are available In this data base, 9 for undisturbed Inflow conditions at various wind speeds and stability conditions and one for flow from the wind turbine. The other data sets are available from the German Weather Service, Hamburg.
Comments The measurement setup Is characterized with large spatial array of cups and vanes.
Sle:Kttkoog, Germany 25 x 10 minute periods based on 20 speed signals
Wind speed (m/s]
Figure 5-8: Nominal turbulence for Kaiser Wilhelms Koog. Site: Jade windpark, Wilhelmshafen, Germany
Slte_code jwe
Dominating terrain / orography rural / flat
Number of masts / wind turbines 1/ (4)
Number of channels wind speed / wind direction 3 (62,92 &126m)/1
Number of sonics, height of meteo.mast 0 /130 m
Hours of measurements / frequency 344 / 20 Hz
Project motivation In the framework of the EU funded program *WEGA II Large Wind Turbine Scientific . Evaluation Project* (Jou2-CT93-0349) a subproject CAN - Comparison of Aeolus II (located in Germany) and Naesudden II (located In Sweden) - was carried out to investigate the behaviour of two sister wind turbines with 3 MW rated power, but with different control mechanism and tower design. High resolution wind data recorded within this project at the Aeolus II to evaluate the mechanical loads also measured at the turbine are available for the "Database on Wind Characteristics'.
Measurement system A130 m high mast served as bases for the meteorological measurements. The meteorological data was collected by two independent data loggers. An Ammonit data logger was used for the long term recording of 5 minute averages of all data relevant for the power curve evaluation at the Aeolus II. In Addition a modular processor-controlled data acquisition system (MOPS) served for Investi gations of the structural loads and dynamic behaviour of the Aeolus II. Only the data gained from this sy stem are relevant for the Database on Wind Characteristics. The sample rate was 20 Hz and time series with a duration of 600 seconds have been stored.
Comments The time series are recorded in 1996 and 1997 as part of measurement programme for Aeolus II wind turbine which Is located near the DEWI wind turbine test centre.
Figure 5-9: Nominal turbulence for Jade Windpark, Germany.
50 Site: Lammefjord, Denmark
Slte_code lamme
Dominating terrain / orography pastoral/flat
Number of masts / wind turbines 4/0
Number of channels wind speed / wind direction 12(10,20 & 30m)/9
Number of sonics, height of meteo.mast 1 (45m)/30-45 m
Hours of measurements / frequency 663/8 & 16 Hz
Project motivation An attempt to gather data continuously for one year with sufficient spatial and temporal. resolution for wind turbine design studies. Completed with around 90% availability with a longest uninterrupted series of 103 days.
Measurement system An array of cups, vanes and one sonic anemometer were sampled by a pc. Data were stored on a magneto-optisk WORM drive. Data were recorded at 16 Hz for the sonic and 8 Hz for the cups and vanes.
Comments The time series are recorded during 1987 and are characterized with a large spatial array of cups and vanes.
Sle: Lammefjord, Denmark 2990 x 10 minute periods based on 10 speed signals
10 15 20 Wind speed [m/s]
Figure 5-10: Nominal turbulence for Lammefjord site. Site: Lavrio, Greece
Slte_code lavrio
Dominating terrain /orography pastoral /mountain
Number of masts / wind turbines 2/1
Number of channels wind speed / wind direction 4 (13 & 32 m) / 4
Number of sonics, height of meteo.mast 2 (24 m)/ 40 m
Hours of measurements / frequency 1 / 8 Hz
Project motivation The main objective of the project was to identify the mountainous terrain effects in ' the operation, safety and reliability of wind turbines. The methodology designed for the project included site wind and wt on site characterisations and complex terrain parameter (of importance to wt operation) identification through experimental and analytical work.
Measurement system Wind structure measurements on mountainous terrain: the 3D wind inflow to a wind turbine rotor was measured by GRES with a number of standard cups and fast sonic anemometers installed on a system of masts at GRES’ complex terrain wind turbine Test Station. The wind structure measurements were coupled to the response of a 110kW stall regulated wt (W110XT) through load and power measurements.
Comments
Figure 5-11: Nominal turbulence for Lavrio, Greece.
52 Site: Lysekil, Sweden
Site_code lyse
Dominating terrain / orography coastal / mountain
Number of masts / wind turbines 1/(2)
Number of channels wind speed / wind direction 12 (10,24,32,40,50,58 & 65m) /12
Number of sonics, height of meteo.mast 0/66m
Hours of measurements / frequency 3355 / app. 1 Hz
Project motivation At Lyse Wind Power Station the NWP 400 wind turbine was erected in 1992. The wind turbine is a two-bladed, upwind machine with a hub height of 40 m, and a rotor diameter of 35 m.
Measurement system Lyse wind power station Is situated at an artificial island created around two islets. On each of the rocky islets wind turbines are erected. The above mentioned NWP400 to the north and a Bonus 400 kW Mkll to the south.Between the two turbines a 66 m high meteorological tower is situated, which is equipped with wind speed and direction sensors of the MIUU type (Lundin et at., 1990) at 7 levels. At the uppermost 5 levels two anemometers are placed at each level. Temperature profile is also recorded. All measurements are sampled with 1 Hz and stored on 1 GB streamer tape.
Comments The time series are recorded during 1993,1994 and 1995 as part of a wind turbine measurement programme.
Figure 5-12: Nominal turbulence for Lyse site.
53 Site: Oak Creek, Tehachapl, CA, USA Site_code oak Dominating terrain / orography scrub/hill
Number of masts / wind turbines 2/(2) Number of channels wind speed / wind direction 8(10,50,65 & 79m)/SO Number of sonics, height of meteo.mast 2/80 m Hours of measurements / frequency 745 /8& 16 Hz
Project motivation Verification of the structural Integrity of a wind turbine involves analysis of fatigue loading as well as ultimate loading. With the trend of persistently growing turbines, the ultimate loading seems to become relatively more important. For wind turbines designed according to the wind conditions prescribed in the IEC-61400 code, the ultimate load is often identified as the leading load parameter. The objective of the Oak Creek project is to conduct a combined experimental and theoretical investigation of blade-, rotor- and tower loads caused by extreme wind load conditions occuring during normal operation as well as in stand still situations (where mean wind speeds exceeds the cut-out wind speed), with the purpose of establishing an improved description of the ultimate loading of three bladed pitch- and stall controlled wind turbines.
Measurement system The instrumentation of the meterological towers Included sensors at multiple levels. Basically, similar instruments on each of the two masts have been Installed in roughly the same level relative to the terrain level. The monitoring system is running continously, and the data are reduced and stored as 10- minutes statistics suplemented with intensive time series recordings covering periods where the mean wind speeed exceeds a specified threshold (15 m/s). Consequently, there are time gaps in the the time series. The monitoring sample rate is 32.
Comments The time series are recorded during 1998 and the measurement programme runs during 1999.
50 Site: Oak Creek 3960 x 10 minute periods based on 8 speed signals
0 0 10 20 30 Wind speed [m/s] Figure 5-13: Measured turbulence at Oak Creek.
54 Site: Calhete, S. Jorge, Azores, Portugal
Slte_code sjorge
Dominating terrain /orography costal /mountain
Number of masts / wind turbines 1/0
Number of channels wind speed / wind direction 5(10 &24m) /1
Number of sonics, height of meteo.mast 1 (27 m)/ 27 m
Hours of measurements / frequency 5 / 40 Hz
Project motivation Development and validation of wind park and local grid detailed dynamic models (INPark).'
Measurement system The measurement system was based on one 9200 PLUS NRG datalogger with NRG#40 cup anemometers and NRG#200P wind vane as transducers, and the 3D wind components were obtained through a solent research symetrlc head sonic anemometer, being the data aqulsition system a GH-Garrad Hassan T-DAS operated by the GH-MON software.
Comments Only a limited number of time series with high resolution recorded in 1996 are available from this site located In the Atlantic Ocean.
S*e:S. Jorge, Azores, Portugal 24 x 10 minute periods w based on 1 sonic speed signal "ST i 8 I * 3 ______1- 4 * ?♦
t) 5 10 15 20 25 30 Wind speed [m/s]
Figure 5-14: Nominal turbulence for San Jorge, Azores. Site: Skipheia, Froya, Norway
Site_code ski
Dominating terrain / orography coastal / hill
Number of masts / wind turbines 3/0
Number of channels wind speed / wind direction 17 (11, 21,41,45, 72 & 101m)/1
Number of sonics, height of meteo.mast 0/45 -100 m
Hours of measurements / frequency 13175 /Appr. 1 Hz
Project motivation The station was build as a part of the Norwegian Wind Energy Programme in 1980. The purpose was to study the wind structure In details. Particularly, a database for high wind speed condition was desired. Data were originally intended for wind energy production; the dimension af the masts corresponds to a large WECS. The station should also serve as a reference station for other measurement stations in the region. Together with a fourth mast 4 km further west (Sletringen) the station has provided data for calculation of dynamic wind loads on off-shore constructions.
Measurement system The measurment system consists of 3 meteorological masts (100,100 and 45 m) placed In a triangle 80-180 metres apart. Ten-minute average of wind speed and direction have been recorded since 1982. From 1988, the logging frequency has been 0.85 Hz. 40-60 channels have been recorded continuously. In a 45 m mast at Sletringen, data has been recorded for some periods since 1988.
Comments The time series are recorded during 1988 -1995 and covers large spatial array.
50 llfe^ Site: Skipheia, Norway 13175 hours based on 17 speed signals
i------'------1------'------r~ 10 20 30 40 Nominal wind speed [m/s] Figure 5-15: Nominal turbulence for Skipheia, Norway.
56 Site: Sletringen, Froya, Norway
Site_code sle
Dominating terrain / orography coastal/flat
Number of masts / wind turbines 1/0
Number of channels wind speed / wind direction 5(5,10,20,42 & 46 m /1
Number of sonics, height of meteo.mast 0/46 m
Hours of measurements / frequency 3736 /appr. 1 Hz
Protect motivation The station was build as a part of the Norwegian Wind Energy Programme In 1980. The purpose was to study the wind structure in details. Particularly, a database for high wind speed condition was desired. Data were originally intended for wind energy production; the dimension at the masts corresponds to a large WECS. The station should also serve as a reference station for other measurement stations in the region. Together with a fourth mast 4 km further west (Sletringen) the station has provided data for calculation of dynamic wind loads on off-shore constructions.
Measurement system The measurment system consists of 3 meteorological masts (100,100 and 45m) placed in a triangle 80-180 metres apart. Ten-minute average of wind speed and direction have been recorded since 1982. From 1988, the logging frequency has been 0.85 Hz. 40-60 channels have been recorded continuously. In a 45 m mast at Sletringen, data has been recorded for some periods since 1988.
Comments The time series are recorded during 1994 and 1995.
Figure 5-16: Nominal turbulence for Sletringen, Norway. Site: Tarifa, Spain
Site_code tarifa
Dominating terrain / orography pastoral / mountain
Number of masts / wind turbines 2/(10)
Number of channels wind speed / wind direction 14 (10,11 & 19m) / 3
Number of sonics, height of meteo.mast 3 (10 & 40 m)/ 40 m
Hours of measurements / frequency 280 / 4 Hz
Project motivation The primary goal was to investigate of design aspects and design options for wind turbines . operating in complex terrain environments within COMTER.ID project. The wind data, which were recorded as part of this measurement programme are available for the "Database on Wind Characteristics".
Measurement system The measurements system consist of two meteorological masts, mast n#1 located towards NNW direction and 15 degrees from the north, mast n§2located towards South direction and 190 degrees from the north The measurements system was based on Garrad Hasan acquisition system. The system operation was primarely focusing on recording structural loads including a number of metereological channels. Data recording frecuency was 40 Hz and duration of 600 seconds.
Comments The time series are recorded during 1997 and contains 3 sonic instruments.
0 5 10 15 20 25 30 Wind speed [m/s]
Figure 5-17: Nominal turbulence for Tarifa, Spain.
58 Site: Tjaereborg near Esbjerg, Denmark
Site_code tjare
Dominating terrain / orography coastal/flat
Number of masts / wind turbines 2/1
Number of channels wind speed/wind direction 12(10,30,45,60,75 & 90m) / 6
Number of sonics, height of meteo.mast 0/90 m
Hours of measurements / frequency 64/25 Hz
Protect motivation The primary goal was to design and build a 2MW wind tubine at Tj'reborg, Esbjerg. This project was accomplished with an intensive measuring programme where both mechanical loadsand the wind climate on the site was measured - and analyzed. The wind data were recorded as part of this measurement programme.
Measurement system The measurement system consists of 2 meteorological mast, mast no.1 was pla ced In front and mast no.2 behind the wind turbine - referring to the dominant wind sector. The measure ments system was based on two HP-dataloggers and the data was transferred to a central computer (HP operating HP-Basic with home developed data handling routines and programs). The system operation was primarily focusing on recording structural loads included a number of meteorological channels. Data recor ding frequency 25 Hz and durations 184,600 or 3600 seconds. Only data with a duration of 600 seconds or more are used.
Comments The time series are recorded during 1987-1993 as part of measurement programme for the 60m/2MW wind turbine.
Site:T]asfeborg. Denmaik 267 x 10 minute periods based on max. 12 speed signals
20 -
Wind speed [m/s]
Figure 5-18: Nominal turbulence for Tjaereborg, Denmark.
59 Site: Toplou, Crete, Greece
Site_code toplou
Dominating terrain / orography pastoral / mountain
Number of masts / wind turbines 1 /1 (+2)
Number of channels wind speed / wind direction 4 (35m) /1
Number of sonics, height of meteo.mast 1 (35m)/40 m
Hours of measurements / frequency 172/1 & 8 Hz
Project motivation The main objective of the project is to develop a cost effective and easy to apply method for further assessment of fatigue loading effects through measurements. Its simplicity and robustness will enable easy application by idustries. Obtained fatigue load "footprints" will be able to back up findings as derived from theoretical research work and will contribute to improved specifications on wind farm and mountainous terrain operation in international guidelines.
Measurement system Wind and wind turbine power and loads measurements on mountainous terrain at Toplou site, in Crete. Wind measurements are carried out using cups and vanes and a sonic anemometer for determining the 3D characteristics of the wind and these measurements are coupled to the response of a SOOkW stall regulated wt (TW-500) through load and power measurements.
Comments The time series are recorded during 1997 as part of another Joule programme. The additional statistics includes electrical power from nearby wind turbine.
50 Site: Toplou, Greece 1027 x 10 minute periods t based on 1 speed signal 130
0 0 10 20 3C Wind speed [m/s] Figure 5-19: Nominal turbulence for Toplou, Greece.
60 Site: Vallersund, Norway
Site_code vis
Dominating terrain / orography coastal/hill
Number of masts / wind turbines 1/0
Number of channels wind speed / wind direction 4 (5,10,18 & 30m)/1
Number of sonics, height of meteo.mast 1 /30 m
Hours of measurements / frequency 3756/appr1 Hz
Project motivation The project started in 1990, and the purpose was to study the wind structure,and to compare wind measurements to energy production from a neighbouring WECS.
Measurement system The measurement station consists of 30 m mast with wind speed sensors at 30,18, 10, and 5 m, and direction sensor at 12 m. Temperature has been measured at 30 and 10 m heights. The logging frequency has been 0.85 Hz.
Comments The time series are recorded during 1991,1992,1993 and 1994.
50 Site: Vallersund, Norway 35705 x 10 minute periods 40 based on 4 speed signals t m 30 c 10 15 20 25 30 Wind speed [m/s] Figure 5-20: Nominal turbulence for Vallersund, Norway. 61 Site: Vindeby, Offshore, Denmark Site_code vindeby Dominating terrain / orography offshore / flatl Number of masts/wind turbines 3/(10) Number of channels wind speed / wind direction 20/4 Number of sonics, height of meteo.mast 6/45 m Hours of measurements / frequency 1700 / 5 & 20 Hz Project motivation The primary goals of the Risoe Air Sea Experiment (acronym: RASEX) experiment were: 1) To carry out a "Kansas experiment over the sea"; that is, examine the validity of Monin Obukhov simi-larity theory over the sea using profiles and eddy correlation data. 2) To provide detailed Information on off-shore • wind climate characteristics of relevance for wind turbine design. Measurement system The instrumentation of the meterological towers included sensors at multiple levels. Basically, similar instruments on each of the three masts have been installed in roughly the same level relative to the mean sea level. The monitoring sample rate was 20Hz, and the data was reduced and stored as half-hourly means suplemented with representative time series covering roughly 10% of the total monitoring period. Comments Further details can be found on htto://mist.ats.orst.edu and note, that this home page also contains IMPORTANT information on ‘recalibration* of the old part of the RASEX dataset compared to the calibration applied for the present data. The modification of the data material comprises temperature sensors yielding absolute temperature as well as temperature diffe-rences, cup anemometers, sonics wind speed signals and sonic virtual temperature signals. Figure 5-21: Nominal turbulence for Vindeby, Denmark. 62 Site: Windy, Scotland Slte_code windy Dominating terrain / orography rural / hill Number of masts / wind turbines 1/ (?) Number of channels wind speed / wind direction 4 (7,16 & 33)/1 Number of sonics, height of meteo.mast 1/33m Hours of measurements / frequency >600/10 Hz Project motivation Not available yet Measurement system Not available yet Comments The time series are recorded in 1998 as part of Joule project: Windfarms In Hostile terrain. 50 Site: Windy, Scotland 2431 x 10 minute periods based on 3 speed signals 10 20 30 Wind speed fm/sl Figure 5-22: Measured turbulence at Windy Standard, Scotland. Site: Windpark Zeebrugge, Belgium Site_code zeeb Dominating terrain / orography coastal / flat Number of masts / wind turbines 1 / (23) Number of channels wind speed / wind direction 2/1 Number of sonics, height of meteo.mast 0/64 m Hours of measurements / frequency >100/2 Hz Project motivation On the sea-port of Zeebrugge in Belgium, 22 turbines are installed. 20 turbines of 200 kW . (1986), 1x 175 kW, 1x 400 kW (1997) and 1x 600 kW (1998) turbine. The University of Brussels does measurements on the 600 kW turbine to examine the dynamic behavior of the blades, nacelle, and mast. Measurement system The University of Brussels installed a 64 m wind-measuring mast. On top and on the intermediate height of 40 m, wind-speed is measured. Data is transferred to a central computer. The wind- speeds on 40 m and 64 m together with the winddirection are interpreted with LabView. Data recording fre quency is 2 Hz and the duration is 3600 sec. Comments The time series are recorded during 1998 and 1999 as part of the Turbowind 600kW wind turbi ne measurement programme. 64 \ 'i ! ! i 1 ( I J 6. Extremes. Introduction This chapter contains a summary of the mean speed, turbulence intensity and extremes data contained in the database. In presenting this summary we wish to demonstrate the volume and variety of the collected time series. Whilst the shear volume of time series represented gives the data a certain statistical respectability, the database is nevertheless in many ways unrepresentative. This is especially true for high wind speeds where the relatively sparse population can give undue weight to data from one site or terrain type. 6.1 Mean wind speed and turbulence intensity. The most concise summary of the database contents is achieved by plotting the “nominal" turbulence intensity as a function of the "nominal" mean speed as shown in figure 6-1. As described in chapter 3, these "nominal" values are averages for all the pertinent channels for the complete run duration and therefore one point on the plot represents one time series. Note that in figure 6-1 we have neglected points below 2 m/s or with a turbulence intensity greater than 50%. An interesting feature is that the turbulence intensity apparently converges to around 10% for extremely high wind speeds. 0 10 20 30 40 ______nominal wind speed [m/s]______Figure 6-1: Nominal turbulence intensity versus nominal wind speed for all runs. Attempting to repeat the foregoing plot, using 600 second statistics instead of "nominal" statistics results in over one million points since we now have typically several 600 second periods per run and each speed sensor is individually represented. Such a large number of points are very difficult to visualise in a scatter plot and are therefore shown as bin-averaged mean values (p) and standard deviation (o) against wind speed in figure 6-2. 66 0 -) 1 I 1 | I | ! 0 10 20 30 40 ______Wind speed [m/s]______Figure 6-2: Turbulence intensity for all windspeeds above 4 m/s Points in the extreme (> p + 3 x a) of the distribution are plotted in, as are all points for wind speeds greater than 34.5 m/s. From this plot can be seen that the mean turbulence intensity for 600 second periods falls steadily from 10% at low wind speeds to around 8% for the highest speeds. It should be noted that some of the high turbulence intensities may arise from wake conditions in wind farms. 6.2 Wind speed probability density distribution. The probability density distribution for all ten minute mean wind speed signals is shown in figure 6-3. This figure represents a population of more than one million observations in the range 2 - 36.5 m/s, the average wind speed being 9.7 m/s. 10 ml nut* m rages wind speed fm/sj Figure 6-3: Wind speed probability density distribution Since the high speed tail of the distribution is of particular interest but the probabilities are so small relative to the bulk of the distribution, a table with the number of ten minute observations for 0.5 m/s wind speed bins above 30 m/s is inserted in the plot. 67 6.3 Maximum wind speed. Maximum wind speed is illustrated by calculating the signal range (maximum - minimum) as a proportion of the standard deviation. This is plotted in Figure 6-4 for all ten minute periods with a mean speed greater than 4 m/s. For a normal distribution we expect the range to be approximately 6 standard deviations. From Figure 6-4 we can see that the mean range is sensibly constant for all wind speeds with a value around 5.6. On the other hand, the higher percentiles fall significantly with increasing wind speed indicating that the distribution of maximum speeds is much narrower for high wind speeds than at low speeds. distributions Wind speed [m/s] 6.4 Maximum gusts. Gusts are defined as the wind speed change over a 2, 5,10 or 30 second period. For each 10 minutes of times series, the largest gust is recorded for each time period. A summary of these results are given in Figure 6-5 for a 5 second period (rise time). Points falling above (p+3 x o) are plotted in as are all points for mean wind speeds greater than 34.5 m/s. Figure 6-5 indicates that the mean maximum gust increases in proportion to the mean speed. The outer percentiles of the distribution appear also to observe this roughly linear behaviour. At high winds, the population is too small for any meaningful statistical generalisations. 68 Figure 6-5 Maximum positive gust (5 sec. period) vs mean wind speed. 6.5 Wind direction change. Analogous to wind speed gusts, wind directional changes are determined and the maximum directional change over 2, 5,10 and 30 seconds are recorded for each 10 minutes of each direction signal. Once more these data have been binned for every 0.5 m/s of mean speed and the [p-o,p,p+o] from the distribution in each bin are shown as connected lines in Figure 6-6, for a time period of 5 seconds. As can be seen, up to about 40 m/s the average maximum direction change is sensibly constant at 20 degrees. Above 20 m/s the distribution becomes less broad. However it should be remembered that especially at high wind speeds, the distributions may be markedly influenced by wind from one particular site and generalisations may be of limited value. distributions Wind speed [m/s] Figure 6-6: Maximum wind direction change (5 sec period) vs mean wind speed. 69 6.6 Gust directional index. The gust directional index (gdi) conveys to what extent the maximum gust and the maximum direction change are correlated. A maximum gust occuring simultaneously with the maximum direction change (a potentially severe load case for a wind turbine) gives a gdi of 2. Completely uncorrelated maximum gust and maximum direction change gives a gdi of 1. As for both gust and direction change, this index is determined for different time periods (2,5,10 and 30 seconds) for each 10 minutes of time series. Here we clearly require a speed and direction signal pair (separated by under 1m). -2.0 Wind speed [m/s] Figure 6-7: Maximum gust directional index vs mean wind speed. The maximum gust directional index (5 second period) is shown in figure 6-7. Once more, the mean value from the binned distributions are shown as connected lines. Observations greater than 34.5 m/s are plotted in. Figure 6-7 shows that the average correlation between gust and wind directional changes is higher than 50 % for wind speed below 20 m/s and that a few of the events (=4%) are full correlated 5 according to the dotted curve. Above 20 m/s there is reduced correlation between gust occurrence and the change in winddirection. No events are fully correlated and the mean gdi falls beneath 1.5. Full correlation is defined for GDI a 1.99 70 7. Conclusion At the conclusion of the project the database contains more than 50000 hours of data from over 20 sites, resulting in more than 1 million individual ten minute periods of wind speed. This vast database of time series would be useless without adequate experimental documentation, effective quality control and indexing together with a flexible search system. We believe that all these criteria are amply fulfilled by the system in its current state. By basing the database around a web interface, the potential audience is maximised both geographically and in terms of computer technology. The Database on Wind Characteristics now exists and is a valuable and is ready for the wind engineering community. The task lying ahead is to capitalise on this achivement, to get the database employed to the degree it deserves. As a step in this direction, the database continues to operate as Annex XVII of the International Energy Agency ’s Wind Research and Development programme. Annex XVII aims to bring the database to a level where it can be financially self-supporting. This will be achieved by increasing the geographical coverage of the database and by disseminating knowledge of the database, particularly to the wind turbine industry. 71 Contents of annexes 1. Annex A1: Definitions ...... A2 1.1 Introduction ...... A2 1.2 Basic definitions ...... A2 1.3 Simple data selection ...... A3 1.4 Advanced data selection ...... A7 1.5 Transient events...... A8 2. Annex B1: Project description ...... A16 2.1 Template for the Background Information File...... A16 2.2 Description of Background Information File...... A17 2.3 Example Background Information File...... A18 3. Annex B2: Site description ...... A20 3.1 Template for the Site Description File ...... A20 3.2 Common format of the site description file...... A22 3.3 Example of Site Description File ...... A25 4. Annex B3: Master sensor file...... A27 4.1 Template to the master sensor file...... A27 4.2 Master sensor file description ...... A28 4.3 Example of master sensor file...... A32 4.4 Example of sensor file for additional statistics...... A35 5 5. Annex B4: Template for common file format ...... A36 5.1 Syntax of common file format ...... A36 5.2 Description of common file format ...... A37 5.3 Example of a data file...... A39 A1 1. Annex A1: Definitions 1.1 Introduction. This annex defines some of the terms used throughout this report and in the online search facilities implemented as part of the web server. The definitions are grouped in three parts.: - Basic definitions. - Simple data selection with specific definitions. - Advanced data selection with specific definitions. - Glossary of terms. The following syntax are used.: a) Bold face (eg. Run) item which are explaned. b) Normal face (eg. 199803422242) are used as an example. c) Items in brackets [test] refers to input or physical unit. d) Underlined items refers to additional information included as hyperlinks e) Italics face refers to sections explained otherplace. 1.2 Basic definitions. Run A run consists of one or more time series covering the same period [600 - 3600 seconds] with a unique starting time in terms of year, month, day, hour and minute. A run is identified with its runname and consists of one or more files. Runname (199705061503) The runname reflects the recording time for the time series. • Year: 1997 • Month: 05 • Day: 06 • Hour: 15 • Minut: 03 eg. the recording time is 6 May 1997 1 503 Time series A time series contains measurements from one or more channels. Each time series belongs to a particular run and is unique according to the sampling frequency. Each time series is saved i a common ASCII file format and includes a header with general information, basic statistics and a data field with measurements scaled in physical units. The basic statistics consists of channel identification, mean, standard deviation, turbulence intensity, minimum and maximum values for each signal referring to the whole run period. A2 The common file format is defined in Annex B4. 10 mins Statistics (mean, stdv., min, max., gust size,) are calculated for each channel referring to a 10-minutes period. Channel A channel is a derived or a direct signal from an instrument eg. wind speed, wind direction, temperature, sonic-x, sonic-u. The valid signal types are listed in table B3-2, page A33. Instrument An instrument is used to measure a quantity [wind speed, wind direction, temperature,..] and the valid types are listed in table B3-1, page A30. Ftp server This server [nt50.afm.dtu.dk] gives access to the raw time series which are stored on CD- ROM’s in a jukebox with online access. Access to the ftp server is made through anonymous login using your email address as password. Web server This server contains all documents and scripts. The documents are formatted in html and can be view with a standard web browser.The web server is used to access the indexed data located on the database server. Database server This server contains the database which consists of tables (> 30) with basic statistics, indexed values, background informations and information on the measurement instrumentation. 1.3 Simple data selection. The simple data selection is based queries directly in the nominal values. These are sums of mean values of equal signal type, belonging to the same run. Definition of the three different nominals are.: a) nominal wind speed is averaged wind speeds n %1 b) nominal turbulence intensity is averaged turbulence intensities n%1 n c) nominal wind direction is averaged wind directions n z 1 n A3 Note: all input values in the data selection form are optional. 1.3.1 Simple data selection form, min: [3] values < min are excluded in the query eq. min = 3 m/s means all values below 3 m/s are excluded max: [23] values > max are exclude in the query eq. max = 23 m/s means all values above 23 m/s are excluded wind speed [m/s] Select nominal wind speed Turbulence intensity [%] Select nominal turbulence Wind direction [deg] Select nominal wind direction i Duration [sec] < Select duration of time series range [600 - 3600]. ; Terrain The valid terrain classification should be according to the terrain type definitions listed in table ! B2-1, page A24. I Orography i Orography means terrain height variations and the dominant orography is selected according to table B2-2, page A24. site A site represents a unique location with measurements and it is possible to select data representing all sites or from a specific site. Count - (button) Number of matching runs fulfilling the requirements listed in the simple data selection form View - (button) List of matching runs fulfilling the requirements listed in the simple data selection form Plot - (button) Plot (graphic view) of nominal wind speed, wind direction and turbulence fulfilling the requirements listed in the simple data selection form Files - (button) A4 List of file names for the matching runs fulfilling the requirements listed in the simple data selection form. These files are available for download directly from the FTP-server. 1.3.2 Matching runs. Run: (199705061503) The runname refers to the starting time in terms of year(1997), month(05), day(06), hour(15) and minute(03) and includes an link to Report page with run statistics. Eg. recording of this example is initiated at 15:03 the 6 May 1997. Site =aspruzza Unique site name with link to the actual site description. Speed= 27.94 Nominal wind speed Dir= 233.8 Nominal direction Turb%= 11.45 Nominal turbulence intensity Durations 3600.00 Duration of the resulting run- in seconds 1.3.3 Report. The report present statistical informations identical to the file header for one single run consisting of one or more files. Run= 199705061503 The runname refers to the starting time in terms of year, month, day, hour and minute. Sites asoruzza Unique site name with link to the site description. Starts 06-05-97 15:03:19 Unique recording time for this run, 15:30 the 6 May 1997. Durations 3600.00 Duration of the resulting run - in seconds. Files /asoruzza/1997/dav1 26/1503 010.zio Unique filename and directory with information about site, date and time of recording, sampling frequency and compression technic. This label contains a link to the exact file location in the FTP-server. Each compressed file only contains one data file with the extension "dat".. A5 Example: file=1503_010.zip contains one ASCII file=1503_010.dat, recorded with a frequency of 1 Hz. The filename syntax is MS-DOS compatible and restricted to 12 characters *hhmm„fff.xxx" where • hh=hour (2 digits) • mm= minute (2 digits) • _ = underscore • fff = 10 x scanfrequency [Hz] • xxx=dat means file extension for ascii file • xxx =zip means file extension for pkzip compressed file Frequency: 1.00 Sampling frequency; unit = Hz Packed size=27684 Size of compressed file in Bytes. Channel: s11 Channel name with link to signal and instrument informations. Type:s Signal type according to table 1. Freq.: 1.0 Sampling frequency; unit: Hz Height: 33 Instrument height above ground level, unit = m. Mean: 3.80 Mean value representing the duration time and unit according to signal type. Stdv: 1.18 Standard deviation representing the duration time and unit according to signal type. Min: 0.8 Minimum value in duration time and unit according to signal type. Max: 8.3 Maximum value in duration time and unit according to signal type. Turbulence: 11. Turbulence intensity, unit [%]. A6 1.4 Advanced data selection. The advanced data selection form is used for queries based indexed values referring to 10 minutes statistics and consists of 3 different data selection forms (ten minute stats, transient events and wind shear). Note: the reference period is 10 minutes and all input variables are optional in these forms. 1.4.1 Ten minute stats. The 10 minutes values are all averaged for 600 seconds, min: [3] values < min are excluded in the query eq. min = 3 m/s mean all values below 3 m/s are excluded max: [23] values > max are exclude in the query eq. max = 23 m/s mean all values above 23 m/s are excluded Wind speed [m/s] Select 10 minutes values of wind speed [interval] Turbulence intensity [%] Select 10 minutes values of turbulence intensity [interval] Wind direction [deg] Select 10 minutes values of wind direction [interval] sensor height [m] Select sensor height [interval] (above ground level). Site [all] Select unique site name or all sites. Runname Select a specific or a group of runnames (eq. 199801 % results in timeseries recorded during January 1998). Note: wildcard is °%" Transient events Select transient events to include items concerning gust size, acceleration, wind directional changes. Wind shear Select wind shear to include items concerning wind shear. Options No options are available yet. A7 1.5 T ransient events. period [sec] Select data indexing period T, where T refers to a moving window "size*. The data indexing has been performed with four different periods (2,5,10 & 30 seconds) as defined on figure A1-1. Note: data representing T=2 are omitted for low sampled data (n<2 Hz). Gust [m/s] Select the positive gust size +VG [interval], as defined on figure A1-1. Moving 7=2,5,10 or 30 sec. 600 sek. Figure A1-1: Definition of gust indexing Acceleration [m/s/s] Select the positive gust acceleration, +Va [interval]. Negative gust [m/s] Select the negative gust size -VG [interval], as defined on figure A1-1. Negative acceleration [m/s/s] Select the negative gust acceleration, -rVa [interval]. Dir change [°] Select the size of the directional change aDir [interval] A8 Dir accel.[°] Select the size of the directional acceleration aDir/T [interval] Gust directional index Select the size of the gust directional index (GDI) [interval]. abs{U(t+dt)-U(t)) | abs{Dir(t+df)-Dir(f)) \ GDI max(abs((U(f+ df) -U(tj))) max(abs((Dir(t+df)-Dir{f)))) ) Note: 1.0 s GDI s 2.0 where GDI=2 indicates full time correlation between |max]gust and | max | directional change. Wind shear Wind shear [m/s/m] Select wind shear, [interval], reference period: 2, 5,10 30 seconds. The wind shear is calculated as a linear shear based on a vertical signal separation; Ah=h 2-h1; ah> 15 m; h, % 15 m; h2 s 100m Wind shear = &V / ah [ m/s/m] aV = wind speed difference [m/s] ah = distance [m] Shear factor and shear exponent Select shear factor and shear exponent [interval], reference period = 10 minutes. The shear factor and shear exponent defines the wind profile and are determined as a polynomial fit based on 3 or more wind speed measurements from to the same meteoro logical mast (and boom direction). Wind speed Vh (T=10 minutes) at height h is determined as Vh = factor x hexponem Shear monotonic Shear non monotonic Select whether the shear should be monotonic or non-monotonic. Shear monotonic means the wind speed increases with height. 1.5.1 Matching ten minute periods. The resulting screen contains links to several pages, each underlined. Run= 199807311453 A9 Runname with link to run statistics presented in the report, section 1.3.3. Ten min= 1 10 minute numbers [1,2,3,4,5,6] which fulfills the query conditions. Site= test Sitename with reference to site description. Speed= 19.4 10 minute speed [m/s] Dir= 140 10 minute direction [deg] Turb %= 12.7 10 minute turbulence [%] Gusts= G Reference to Ten minute speed gusts and accelerations, presented in section 4.6 Dir change= D Reference to Ten minute direction change and acceleration report, presented in section 4.7 Shear= S Reference to Ten minute shears report, presented in section 4.8 Shear Fit= F Reference to Ten minute shear fits report, presented in section 4.9 Gust directional index= Gdi Reference to report on gust directional indicies in section 4.10 1.5.2 Ten minute statistics. Site= test Reference to site description for site = test Run= 199807311453 Runname Start= 31-07-98 14:53:00 Start time of time series recording Ten min= 1 of 1 Ten minute period number and total number of ten minute periods in this run. A10 File= /aspruzza/1997/dav1 26/1503 010.zip Unique filename and directory with information about site, date and time of recording, sampling frequency and compression technic. This label contains a link to the exact file location in the FTP-server. Each compressed file only contains one data file with the extension "dat".. Example: f ile=1503_010.zip contains one ASCII file=1503_010.dat, recorded with a frequency of 1 Hz. The filename syntax is MS-DOS compatible and restricted to 12 characters "hhmm_fff.xxx" where • hh=hour (2 digits) • mm= minute (2 digits) • _ = underscore • fff = 10 x scant requency [Hz] • xxx=dat means file extension for ascii file • xxx =zip means file extension for pkzip compressed file Frequency=1 .00 Sampling frequency; unit = Hz Packed size=27684 Size of compressed file in Bytes. Channel= s11 Channel name with link to signal and instrument informations. Type = s Signal type according to table B3-2. Freq.= 1.0 Sampling frequency; unit = Hz Heights 33 Instrument height above ground level, unit = m. Mean= 3.80 Mean value representing 10 minutes and unit according to signal type. Stdvs 1.18 Standard deviation representing 10 minutes and unit according to signal type. Min= 0.8 Minimums value in 10 minutes and unit according to signal type. Max= 8.3 Maximums value in 10 minutes and unit according to signal type. A11 Turbulence= 11. Turbulence intensity in 10 minutes [%]. 1.5.3 Ten minute speed gusts and accelerations. Site= test Reference to site description for site = test Run= 199807311453 Runname Start= 31-07-98 14:53:00 Start time of time series recording Ten min= 1 of 1 Ten minute period number and total number of ten minute periods in this run. Channels cup 90 All channels contains a link to the signal and instrument documentation. Heights 90.0 Instrument height above ground level. Periods 2 Data indexing period T, defined in section 4.3 Gust+s 2.6 Maximum positive gust size, defined in section 4.3 referring to a period T. Gust-s -4.5 “Maximum" negative gust size, defined in section 4.3 referring to a period T. Acce!+: 1.3 Maximum positive speed acceleration, defined in section 4.3 referring to a period T. Accel-: -2.25 “Maximum" negative speed acceleration, defined in section 4.3 referring to a period T. 1.5.4 Ten minute direction change and accelerations. Sites test Reference to site description for site s test Run= 199807311453 Runname Starts 31-07-98 14:53:00 A12 Start time of time series recording Ten min= 1 of 1 Ten minute period number reference number. Channels dir60 All channels contains a link to the signal and instrument documentation. Heights 90.0 Instrument height above ground level. Periods 2 Data indexing period T, defined in section 4.3 Dir changes 19.8 Maximum absolute change in the wind speed direction, referring to the period T. Dir accels 9.9 Maximum absolute acceleration of the wind speed direction, referring to the period T. 1.5.5 Ten minute shears. Sites test Reference to site description for site = test Runs 199807311029 Runname Starts 31-07-98 10:29:00 Start time of time series recording Ten min: 1 of 6 Ten minute period number and total number of ten minute periods in this run Channel 1: cuo45 First (lower) wind speed channel used in shear calculation. Height 1:90.0 Height above ground level of (lower) instrument 1. Channel 2: cup90 Second ( upper) wind speed channel used in shear calculation. Height 2: 90.0 Height above ground level of (upper) instrument. A13 Period: 2 Data indexing period T, defined in section 4.3 Shear+: 0.03 Maximum shear value referring to a period of T seconds. Unit = [m/s/m] Shear-: 0.02 Minimum shear value referring to a period of T seconds. Unit = [m/s/m] 1.5.6 Ten minute shear fits. Site=test Reference to site description for site = test Run:199807311029 Runname Start: 31-07-98 10:29:00 Start time for recording Ten min: 1 of 6 Ten minute period reference number and total number of periods in this run. Mast: 1 Mast number used in shear fit calculation Speeds: 3 Number of wind speed signals used in shear fit calculations. Monotonlc: T Quality of wind speed used in shear fit calculations. T=True means Vh1 Shear factor: 6.15 Shear exponent: 0.14 The shear factor and exponent defines the average wind profile Vh averaging period is 10 minutes. Calculation of averaged wind speed Vh vs height Vh = factor x hexponent eg. V40 = 6.15 x 40°-u = 10.31 m/s A14 1.5.7 Ten minute GDI’s. Site=test Reference to site description for site = test Run:199807311029 Runname Start: 31-07-98 10:29:00 Start time for recording Ten min: 1 of 6 Ten minute period reference number and total number of periods in this run. Cup: cup45 Channel name for wind speed signal used in determination of gust directional index. Vane: dir45 Channel name for wind direction signal used in determination of gust directional index. Period: 5 Reference period for determination of gust directional index. Gdi: 1.77 Determined gust directional index value. A15 2. Annex B1: Project description This annex contains a description of the necessary project information. The project information included in the database is based this project description file and prepared as an ASCII text file. 2.1 Template for the Background Information File. [Basicjnformation] project_code = Institution = Person = E_mail = URL = Address = Telephone = Telefax = Collaborators = Funding_agencies = project_start_date = project_end_date = [Project Motivation] ..free text with an unlimited number of lines [Measurement_System] .. free text with an unlimited number of lines [Attachments] Number_of_pub!ications = Number_of_maps = Number_of_graphs = [publication^] Description = Reference = Unlimited number of publications [Map_1] A16 Description = Filename = Unlimited number of maps [Graph_1] Description = Filename = Unlimited number of graphs 2.2 Description of Background Information File. Naming convention for the Project Information File: project.pro (example) Note thatindicates a comment which may appear everywhere. The Project Information File is transferred as an ASCII text file [Basicjnformation] project_code = project name - indicating the coordinating project. If a project has only one site, the site_code and the project_code may be identical. Institution = Name of organisation which had primary responsibility for the measurement programme. Person = contact person E_mail = E_mail address for the contact person URL = web page[s] with additional information about this project. Address = Postal address of the contact person Telephone = Telephone number to the contact person/ responsible institution. Telefax = Telefax number to the contact person/ responsible institution. Collaborators = Additional project partners Funding_agencies = List of funding agencies e.g. EU, Ministry of Energy project_start_date = Start of measurement project (eg. 1-1-88) project_end_date = End of measurement project (eg.31-12-89) [Project Motivation] .. The motivation behind carrying out the measurements, given in free text [Measurement-System] .. Description of the measurement system, given in free text format. A17 I [Attachments] Number_of_publications= Number of publications listed below Number_of_maps = Number of maps showing the area of the project. The description includes information about the location, scales,.. Number_of_graphs = Graphs relevant in the domain of the project, for example: - measured wind speed distribution for all sites - measured wind roses for all sites Each of the graphs is supplied with information about the item e.g. distribution type, period The preferred format is GIF or JPEG [publication,!] Description = short summary of publication number 1. Reference = reference to publication number 1 . [Map_1] Description ; description of the map number 1. Filename = name of file containing the map number 1; preferred format '.GIF or '.JPEG - or other commonly used format [Graph,!] Description: description of content viz. on graph number!. Filename = name of file containing !. graph; preferred format '.GIF or '.JPEG - or other commonly used format. 2.3 Example Background Information File. [Basicjnformation] project_code = PO-Mistral Institution = INETI Person = Ana Estanqueiro e_mail = [email protected] URL = http://www.ineti.pt/ite/ite.html Address = Azinhaga dos Lameiros, !699 Lisboa .Portugal Telephone = 351.1.7162712(ext.2725) Telefax = 351.1.7163797 Collaborators = LNEC,lST,EDP,EDA •r I Funding_agencies = NATO SfS program, INETI, Ministry of Energy project_start_date = 01-01-94 project_end_date = 30-09-98 A18 ;|iamwmw .■IE [Project Motivation] Development and validation of wind park and local grid detailed dynamic models (INPark). [Measurement_System] The measurement system was based on one 9200 PLUS NRG datalogger with NRG#40 cup anemometers and NRGS200P wind vane as transducers, and the 3D wind components were obtained through a solent research symetric head sonic anemometer, being the data aquisition system a GH- Garrad Hassan T-DAS operated by the GH-MON software. [Attachments] Number_of_publications = 3 [publication^] Description = Describes the initial phase of the S. Jorge experimental campaign. Reference = Castro,R.M.,A.I.Estanqueiro,J.G.Saraiva,L.Gomes e J.M.Ferreira de Jesus(1996a). "Nato SfS project PO-Mistral: Status of the Experimental Validation Phase". A.Zervos, H.Ehmann e P.Helm(Ed.s). Proceedings of 1996 EUWEC H.Stephens and associates. Bedford. [publication_2] Description = Adresses the wind tunnel calibration of the ultrasonic anemometer digital and analog outputs includind the effect of the tilted incoming flow. Reference = Estanqueiro, A. I. e F. Marques da Silva (1996). Calibration Report of the "Solent-Research model" ultra-sonic anemometer, INETI/DER-LNEC.(ln Portuguese). [publication_3] Description = Describes the INPark model and characterizes its validation campaign. Reference = Estanqueiro, A. I., J. M. Ferreira de Jesus e J. G. Saraiva (1996a)."A wind park grid integration model for power quality assessment". A. Zervos, H. Ehmann et P. Helm (Ed.s). Proceedings of 1996 EUWEC, H. Stephens and Associates, Bedford. A19 3. Annex B2: Site description. This annex contains a description of the necessary site description including mast and wind turbine information. The site description included in the database is based the available site information and is prepared as an ASCII text file. 3.1 Template for the Site Description File. [Site_global_data] site_name = version = site_code = parent_project = longitude = latitude = altitude = country = dominant_terrain_type = dominant_orography = no_of_masts = no_of_wind_turbines = [mast_1] x = y= z = roughness_class = turbine_wakes = Description = [turbine_1] x = y = z = description = diameter = hub_height = rated_power = rated_wind_speed= [Attachments] no_of_site_maps = no_of_site_drawings = no_of_site_photos = no_of_site_graphs = no_of_mast_photos = A20 no_of_mast_drawings [site_map_1] Description = Filename = [site_photo_1] Description = Filename = [site_graph_1] Description = Filename = [site_drawing_1] Description = Filename = [mast_photo_1] Mast_no = Description = Filename = [mast_drawing_1] Mast_no = Description = Filename = [Wasp] Orographyjile = Raw_data_file = Obstacle_file = Roughness_file = 3.2 Common format of the site description file. Naming convention for the Site Desription File: site.sit, eg. tjare.sit and note that significant changes in the site characteristics (e.g. new mat. Mast or wind turbines) require a new site description file including a new site_code. [Site_g!obal_data] site_code = Site name (succinct - used as an overall site reference, where same site_code should be used in all definition files) parent_project = Name of the coordinating project (defined in project description file) site_name = Site name including near-by town and/or state version = Date of creation (format: day-month-year eg. 31-12-95) country = Name of country longitude = Longitude specification of referencepoint (usually mast 1) [deg,min,sec.dd,E/WJ latitude = Lattitude specification of referencepoint (usually mast 1) [deg,min,sec.dd, N/S] altitude = Altitude specification of referencepoint (usually the ground level of mast 1) [m] dominant_terrain_type = The dominant terrain type as summarized in table B2-1 Table B2-1: terrain types type description bridge measurements from a bridge coastal water and land forest forest ice snow and ice cover offshore open sea pastoral open fields and meadows rural agriculture with some buildings sand sand cover scrub bushes and small trees urban town dominant_orography = the dominant orography according to table B2-2 Table B2-2: Orography types type description flat flat landscape hill rolling hills mountain sharp contours - separation expected A22 no_of_masts= Number of different met. mast locations specified in this file. The location, roughness classes and possible wind turbine wake sectors are included for each meteorological mast. no_of_wi nd_turbines = Number of near-by wind turbines specified in this file. Include wind turbines which might influence the wind speed measurements (within a distance of approx. 1 km) [Mast_1] x = Relative x-coordinate of Mast_1 w.r.t. reference location ; positive => East [m] according to figure B2-1 y = Relative y-coordinate of Mast_1 w.r.t. reference location ; positive => North [m] according to figure B2-1 z = Relative z-coordinate of Mast_1 w.r.t reference location ; positive => Upwards [m] according to figure B2-1 roughness_class = Roughness classes for each of 12 sectors as seen from Mast_1 - according to the Wasp definitions - the Wasp sector sizes are defined in table B2-3. Y-axis; direction North X-axis; d recti on East Figure B2-1: Mast and wind turbine reference coordinate system turbine_wakes = Wake status for each of 12 sectors defined as F[alse] or T[rue] as seen from Mast_1. Description = Additional information - free text e.g. description of mast type (shape, construction, foundation or guy wires etc.) A23 Table B2-3: Definition of roughness classes Sector No. Sector 1 0°±15° 2 30°±15° 3 60° ±15° 4 90° ±15° 5 120° ±15° 6 150° ±15° 7 180° ±15° 8 210° ±15° 9 240° ±15° 10 270° ±15° 11 300° ±15° 12 330° ±15° [Turblne_1] x = Relative x-coordinate of Turbine_1 w.r.t. reference location; positive => East [m] according to figure B2-1 y= Relative y-coordinate of Turbine_1 w.r.t. reference location; positive => North [m] according to figure B2-1 z = Relative z-coordinate of Turbine_1 w.r.t. reference location; positive => Upwards [m] according to figure B2-1 description = Name and type of Turbine_1 diameter = Diameter Turbine_1 [m] hub_height = Hub height of Turbine_1 [m] rated_power = Rated power of Turbine_1. [kW] rated_wind_speed = Rated wind speed for Turbine_1. [m/s] [Attachments] no_of_site_maps = Number of site maps. no_of_site_drawings = Number of site drawings showing mast and wind turbine positions. no_of_site_photos = Number of site photos viz. the surrounding landscape 0 - 360 degrees with sufficient resolution. no_of_site_graphs = Number of site related graphs e.g. wind speed distribution. no_of_mast_photos = Number of mast photos, e.g.: viz. Mounting details. no_of_mast_drawings = Number of mast drawings, e.g. engineering drawings. [site_map_1] Description = Description of the map number 1. Filename = Name of file containing the map number 1; preferred format \GIF or ‘.JPEG - or other commonly used formats. A24 [site_photo_1] Description = Description of photo number 1. Filename = Name of file containing photo number 1; preferred format ‘.GIF or '.JPEG - or other commonly used format. [site_graph_1] Description = Description of graph number 1. Filename = Name of file containing the graph; preferred format '.GIF or '.JPEG - or other commonly used format. [site_drawing_1] Description = Description of drawing number 1. Filename = Name of file containing the drawing; preferred format '.GIF or '.JPEG - or other commonly used format. [mast_photo_1] Description = Description of photo number 1. Filename = Name of file containing photo number 1; preferred format '.GIF or '.JPEG - or other commonly used format. [mast_drawing_1] Description = Description of drawing number 1. Filename = Name of file containing the drawing; preferred format '.GIF or '.JPEG - or other commonly used format. [Wasp] Orography_file = Name of file containing the WAsP inputs Raw_data_file = Name of file containing the WAsP raw data Obstacle_file = Name of file containing the WAsP obstacle definitions Roughness_file = Name of file containing the WAsP roughness informations 3.3 Example of Site Description File. [Site_global_data] site_name = sjorge.sit version = 27-04-98 site_code = S_Jorge parent_proj ect = PO_Mistral longitude = 28.57 W lattitude = 38.36 N altitude = 711 country = Portugal dominant_terrain_type = coastal dominant_orography = mountain no_o£_measurements_location= 1 A25 no_of_wind_turbines = 5 no_of_masts = 1 [mast_l] x = 0 y = 0 z = 0 roughness_class = 1,1,1,1,1,1,1,1,1,1,1,1 turbine_wakes = t,t,t,f,f,f,f,f,f,f,f,f Description = Mast number one, located upwind to primary wind direction. Lattice tower with foundation, height 27 m with instruments mounted on booms, cup anemometers at 10 and 24 m, wind vane at 10m. The sonic anemometer at 27 m is topmounted. [Attachments] no_of_site_photos = 2 no_of_mast_photos = 2 [site_photo_l] Description = S. Jorge Island physical model, LNEC wind tunnel (topview of the wind park location) . Filename = sitel.jpg [site_photo_2] Description = S. Jorge Island physical model, LNEC wind tunnel (view of the island southern coast) . Filename = site2.jpg [mast_photo_l] Description = view of the 24 m cup anemometer and the top mounted sonic. Filename = mastl.jpg [mast_photo_2] Description = view of the mast and the whole wind measurement system. Filename = mast2.jpg A26 4. Annex B3: Master sensor file. This annex contains a description of the primary sensors and the mounting details. The sensor informations included in the database are based on the master sensor file and is prepared as an ASCII text file. Sensors added as additional statistics as defined in a separate sensor file. 4.1 Template to the master sensor file. [Master Sensor File] site_code = version = no_of_sensors = [sensor_1] sensor_type = serial_no = manufacturer = model_spec.= last_calibrated = sensor_height = boom_direction = sensor_direction = Top_mounted = mast_number = boom Jength = boom_shape = boom_dimension = mast_dimension = meas_distance = No_of_signals = [SignaM] Signal_name = Signal_type = time/Length_constant = MinMeasVal= MaxMeasVal= Units = Accuracy = A27 4.2 Master sensor file description. Naming convention : site.mxx site = site_name m ##= ## is a sequence number (sensor configuration number) starting with 01 and ending with 99 where “m" indicates it is a master sensor file, e.g. tjare.mOl - the primary master sensor file for the tjare site. Minor changes in the instrumentation, e.g. moving an instrument is incorporated as a new sensor/signal. Significant changes to the instrumentation are implemented with a new master sensor file, characterized with a new sequence number (e.g. tjare.m02). In this case, sensor and signal names from the previous master sensor file may be re-used. Note: remember to include a reference to the new master sensor file in the [Common File Header]. Instrument and signal identification An instrument [=sensor] is defined as a measuring device resulting in one or more signals [=channels]. The instrument is referred to by its model type and serial number (e.g. Rise P-1081 , SN.1234). The signal from a sensor is assigned a channel name which is used in the data files as part of the section named [Sensor Statististics]. Comments Comment lines are marked withat the first position. Comments may be included anywhere in the file. [Master Sensor File] - required information (indicates file_type) site_code = Site name (succinct - used as an overall site reference, where a unique site_code are used in all files) Version = Version date day-month-year (eg. 8-7-96 = 8 july 1996) No_of_sensors = Number of sensors defined in this file [Sensor_1] the following 6 lines are necessary and required for each sensor in the master sensor file. Sensor_name = Data providers name of the sensor. The sensor_name must be unique within this master sensor file. The signal_name and the sensor_name of a given sensor can be identical. Sensor_type = Type of instrument, according to the valid instrument types listed in table B3-1. A28 U 'P. Table B3-1: Instrument types Type Description Cup Cup anemometer Prop Propeller anemometer Vane Wind vane Cuva Cup/vane - combined sensor Sonic Sonic anemometer Pitot Pitot tube Rain Precipation, rain measurements. Seac Sea current registration system Term Thermometer Wave Wave height recorder Inclin Speed inclination registration system New types have to be agreed on ksh/msc Sensor_height = Instrument height above ground; required [m] Boom_direction = Boom direction with reference to North; required [deg] Sensor_direction = Sensor direction [Free text] Top_mounted = T [=True] or F [=False]; indication of whether the sensor is mounted on top of the mast. Mast_number = Mast number, referring to mast number defined in the site description file; required [-] boomjength = Distance between instrument and mast centerline, defined on figure B3-1. Note: boomjength = 0 fora top mounted instrument boom_shape = Shape of boom (eq. circular, square, ..) boom_dimension = Equivalent boom "diameter", equal to side-length facing the wind. mast_dimension = Side length (or major diameter) of the mast at measurement height. A29 Mounting Details Boomjcflmenthn Equivalent boom diameter equal to the side-length feting the wind H*st_dimenshn tide length or major diameter at sensor height. 77777777777777 Figure B3-1: definition of mounting details. meas_distance = Distance between measuring point (=p!ane) and the upper surface of the mounting boom, according to figure B3-1. Serial_no = Serial number of instrument. Manufacturer = Manufacturer of instrument. ModeI_spec. = Specification of instrument. Last_calibration_date = Calibration date for instrument. No_of_signals = Number of signals originating from this instrument. [Signal_1] Signal_name = Unique signal name. Signal_type = Type of signal according to table B3-2 Table B3-2: signal types Slgnaljypes Description Units ahum absolute humidity kg/m**3 bare barometric pressure hPa evut sonic covariance u-T °Km/s evuv sonic covariance u-v m*'2/s**2 evuw sonic covariance u-w m**2/s**2 cwt sonic covariance v-T °Km/s cww sonic covariance v-w m**2/s**2 cvwt sonic covariance w-T °Km/s cx Sea current speed, x-dir cm/s cy Sea current speed, y-dir cm/s d wind direction deg Signal types Description Units grad global radiation W/m”2 nrad net radiation W/m**2 pitd wind speed based on dynamic pitot tube m/s measurement pow wind turbine power (active) kW powa wind turbine power (active) kW rain precipitation mm/hr rhum relative humidity % rich richardson number - s wind speed from cup or propeller m/s sd derived, zeroed, sonic direction (mean=0) deg six Horizontal speed inclination angle, x-dir deg siy Horizontal speed inclination angle, y-dir deg spx speed logitudinal direction m/s (propeller anemometer) spy speed lateral direction (propellor anemometer) m/s spz speed vertical direction (propellor anemometer) m/s ss derived sonic speed (=SQRT(uA2+vA2)) m/s St sonic (virtual) temperature degC su sonic component aligned in m/s mean wind direction sv sonic component aligned in horizontal cross-wind m/s direction sw sonic component aligned in vertical direction m/s sx unaligned sonic horizontal component #1 m/s sy unaligned sonic horizontal component #2 m/s sz unaligned sonic vertical component m/s tabs absolute temperature degC tdif temperature difference degC teta sonic anemometer coord, rotation about z axis deg tilt sonic anemometer tilt deg ustr friction velocity u* m/s wave wave height m Note: new signal types have to be agreed on ksh/msc MinMeasVal = Minimum measurement value. MaxMeasval = Maximum measurement value. Time/Length_constant = Time or length constant for the instrument i.e. description of the temporal or spatial resolution. Units = Units according to signal_type in table B3-2. Accuracy = An estimated signal accuracy. A31 4.3 Example of master sensor file, MASTER SENSOR FILE Site = sjorge = S_Jorge, Azores Prepared by INETI, Lissabon Modified : 3/6-98 ksh [Master Sensor File] Site_code = sjorge Version = 27-04-98 No_o f_s ens ors = 4 [sensor_l] Sensor_name = wsOlO Sensor_type = cup Sensor_height = 10 Boom_direction = 225 Sensor_direction 0 Top_mounted = F Mast_number = 1 Boom_length = 1.13 Boom_shape = circular Boom_dimension = 0.013 Mast_dimension = 0.30 Meas_distance = 0.07 Serial_no = n.a. Manufacturer = NRG Model_spec .= NRG#40 Last_calib. = may 1995 No_of_signals = 1 [Signal_l] Signal_name = wsOlO Signal_type = s Time/Lenght_constant=3m MinMeasVal 0 MaxMeasVal 40 units = [m/s] accuracy = Flow distortion due to terrain [sensor_2] Sensor_name = wdOlO Sensor_type = vane Sensor_height = 10 Boom_direction = 225 Sensor_direction 0 Top_mounted = F Mast_number = 1 Boom_length = 0.43 Boom_shape = circular Boom_dimension = 0.012 Mast_dimension = 0.30 Meas_distance = 0.14 Serial_no = n.a. Manufacturer = NRG Model_spec .= NRG#200P Last_calib. = May 1995 No_of_signals = 1 [Signal_l] Signal_name = wdOlO Signal_type = d Time/Lenght_constant=n.a MinMeasVal = 0 MaxMeasVal = 360 units = [deg] accuracy = Flow distortion due to terrain [sensor_3] Sensor_name = ws024 Sensor_type = cup Sensor_height = 24 Boom_direction = 255 Sensor_direction = 0 Top_mounted = F Mast_number = 1 Boom_length = 1.13 Boom_shape = circular Boom_dimension = 0.013 Mast_dimension = 0.30 Meas_distance = 0.07 Serial_no = n.a. Manufacturer = NRG Model_spec . = NRG#40 Last_calib. = May 1995 No_of_signals = 1 [Signal_l] Signal_name = ws024 Signal_type = s Time/Lenght_constant=3m units = [m/s] MinMeasVal = 0 MaxMeasVal = 40 accuracy = Flow distortion due to terrain [sensor_4] Sensor_name = sonic Sensor_type = sonic Sensor_height = 27 Boom_direction = n.a. Sensor_direction = 0 Top_mounted = T Mast_number = 1 Boom_length = 0 Boom_shape = circular Boom_dimension = n.a. Mast_dimension = 0.30 Meas_distance = 0.49 Serial_no = 0173R2 Manufacturer = Solent Model_spec . = Research, symetric head Last_calib. = May 1996 No_of_signals = 6 [Signal_l] Signal_name = s27x Signal_type = sx Time/Lenght_constant= 6 0_msec MinMeasVal = -40 MaxMeasVal = 40 A33 units = [m/s] accuracy = Flow distortion due to terrain [Signal_2] Signal_name = s27y Signal_type = sy Time/Lenght_constant =60_msec MinMeasVal = -40 MaxMeasVal = 40 units = [m/s] accuracy = Flow distortion due to terrain [Signal_3] Signal_name = s27z Signal_type = sz Time/Lenght_cons tant=6 0_msec MinMeasVal = -40 MaxMeasVal = 40 units = [m/s] accuracy = Flow distortion due to terrain Aligned sonic channels su,sv,sw added 3-6-98/ksh [Signal_4] Signal_name = s27u Signal_type = su Time/Lenght_cons tant=6 0_msec MinMeasVal = -40 MaxMeasVal = 40 units = [m/s] accuracy = Flow distortion due to terrain [Signal_5] Signal_name = s27v Signal_type = sv Time/Lenght_constant=60_msec MinMeasVal = -40 MaxMeasVal . = 40 units = [m/s] accuracy = Flow distortion due to terrain [Signal_6] Signal_name = s27w Signal_type = sw Time/Lenght_cons tant=60_msec MinMeasVal -40 MaxMeasVal 40 units [m/s] accuracy Flow distortion due to terrain A34 4.4 Example of sensor file for additional statistics MASTER SENSOR FILE - ADDITIONAL SENSORS Site = sjorge = S_Jorge, Azores Created : 3/6-98 ksh [Master Sensor File] Site_code = sjorge Version = 27-04-98 No_of_sensors = 1 [sensor_4] Sensor_name = sonic Sensor_type = sonic Sensor_height = 27 Boom_direction = n.a. Sensor_direction 0 Top_mounted = T Mast_number = 1 Boom_length = 0 Boom_shape = circular Boom_dimension = n.a. Mast_dimension = 0.30 Meas_distance = 0.49 Serial_no = 0173R2 Manufacturer = Solent Model_spec. = Research, symetric head Last_calib. = May 1996 No_of_signals = 2 [Signal_l] Signal_name = s27d Signal_type = sd Time/Lenght_constant= 6 0 msec units = [m/s] accuracy = Flow distortion due to terrain [Signal_2] Signal_name = s27s Signal_type = ss Time/Lenght_constant=60 msec units = [m/s] accuracy = Flow distortion due to terrain A35 5. Annex B4: Template for common file format. This annex contains a description fo the common file format used for all data files. The description includes all the main items inciudend in the data files. 5.1 Syntax of common file format. [Common File Header] site_code = date = time = project_file = sitejile = sensor_file = sequences frequencies: file_names= durations sensor_cfgs run_names site„names versions [File Header] datajile - frequency s no_of_signals s no_of_scans = [Sensor Statistics] type,qa,height,wake,name,mean,std.,min,max,unit [Additional Statistics] type,qa,height,wake,SignalName,mean,std.,min,max,unit [Data Field] x(1,1) x(1,2) x(1,3) x(1,4) x(1,5) x(2,1) x(2,2) x(2,3) x(2,4) x(2,5) A36 5.2 Description of common file format. Comments Comment lines are marked with at the first position. Comments may be included anywhere in the file. [Common File Header] - required information site_code = Short name (succinct - used as an overall site reference). date = dd-mo-yy; date when the time series was recorded eg. 31-12-93 time = hh:mm:ss; time when the time series was recorded eg. 17:38:12 project_file = Name of project information file [=site_code.pro] with extension "pro" (e.g. tjare.pro = project description for the Tjaereborg project). site_file = Name of site information file [=site_code.sit] with file extension = .sit (e.g. tjare.sit refers to the Tjaereborg site description file). sensor_file = Name of master sensor file [=site_code.m01] with file extension = ,m## (e.g. tjare.mOl refers to the Tjaereborg master sensor list). sequence = Number in measuring sequence (e.g. = 2/3 means this is the second time series af 3 consecutive time series). frequencies = All available frequencies covering this particular period [Hz]. file_names = All available files covering this particular period duration = Duration of time series [seconds]. sensor_cfg = Sensor configuration number (e.g. 1) run_name = Runname based in recording time, defined on page AS, Annex A1. site_name = Site name and country [File Header] - required informations data_file = Name of current data file, according to the definition in page A6, Annex A1. frequency = Frequency [Hz]. no_of_sensors = Number of sensors in this data file. no_of_scans = Number of scans in this data file. Version = Text string with information about post processing performed (e.g. sonic signal alignment). [Sensor statistics] - required information. This section contains one row corresponding to each of the available sensors [no_of_sensors] which are present in the [Data Field] The sensor statistics are given in terms of: A37 Type = Signal type, referring to the definition in table B3-2. qa = Quality index for the sensor (-1,0=bad,1=good) height = Heigth above ground level [m] Wake = Wake status (-1,0=nowake,1= sensor inside wake). The wake sectors referring to a mast are defined in Annex B2. Name = Signal_name, according to the master sensor list. Mean = Average value of the recorded time series. Std. = Standard deviation of the recorded time series. Min = Minimum value of the recorded time series. Max = Maximum value of the recorded time series. Unit = Unit of the recorded signal, according to table B3-2. [Additional statistics] - optional information. The additional statistics covers signals not present in the [Data Field] which are of significant interest. This output format is identical to the [Sensor statistics]. The signal types are also in agreement with table B3-2. Type = Signal type, referring to the definition in table B3-2. qa = Quality index for the sensor (-1,0=bad,1=good) height = Heigth above ground level [m] Wake = Wake status (-1,0=nowake,1= sensor inside wake). The wake sectors referring to a mast are defined in Annex B2. Name = Signal_name, according to the master sensor list. Mean = Average value of the recorded time series. Std. = Standard deviation of the recorded time series. Min = Minimum value of the recorded time series. Max = Maximum value of the recorded time series. Unit = Unit of the recorded signal, according to table B3-2. [Data field] - required information. All data are scaled and stored in physical units [m/s], [deg], [degC] and the numbers are seperated with empty ‘spaces'. Each line contains only one scan with a set of numbers equal to the number_of_sensors given in the [File Header] section. The number of lines are equal to the number_of_scans also given in the [File Header] section. 5.3 Example of a data file, [Common File Header] site_code = sjorge date = 1- 8-96 time = 18: 8: 0 project_file = s_jorge.pro site_file = s_jorge.sit sensor_file = s_jorge.m01 sequence = 1/1 frequencies = 40.0000 file_names = 1808_400.dat duration = 600.00 sensor_cfg = 1 run_name = 199608011808 site_name = Calheta, S.Jorge Island, Azores, Pt [File Header] data_file = \sjorge\1996\day214\1808_400. dat frequency = 40.0 no_of_scans = 24000 no_of_sensors 8 version = aligment; 1.1a d. 19/3-1997 Hans E. Joergensen [sensor statistics] sx 1 27 0 s27x -1.71 0.97 -4.7 1.1 [m/s] sy 1 27 0 s27y -7.24 0.89 -10.3 -4.3 [m/s] sz 1 27 0 s27z 1.71 0.79 -0.6 4.6 [m/s] sd 1 27 0 s27d 226.64 7.40 204.0 247.6 [deg] ss 1 27 0 s27s 7.51 0.89 4.6 10.4 [m/s] su 1 27 0 s27u 7.64 0.84 4.9 10.1 [m/s] sv 1 27 0 s27v 0.00 0.99 -2.9 3.0 [m/s] sw 1 27 0 s27w -0.00 0.82 -2.6 2.9 [m/s] [Additional Statistics] cvuv 1 27.0 0 cvuvl 0.0583 .0 0.06 0.06 [M**2/S**2] cvuw 1 27.0 0 cvuwl -0.1119 .0 -0.11 -0.11 [M**2/S**2] cvut 1 27.0 0 cvutl 0.0000 .0 0.00 0.00 [Km/s] cww 1 27.0 0 cwwl -0.0258 .0 -0.03 -0.03 [M**2/S**2] cvwt 1 27.0 0 cvwtl 0.0000 .0 0.00 0.00 [Km/s] cwt 1 27.0 0 cwtl 0.0000 .0 0.00 0.00 [Km/s] tilt 1 27.0 0 tiltl 12.97 .0 12.97 12.97 [deg] teta 1 27.0 0 tetal -103.27 .0 -103.27 -103.27 [deg] ustr 1 27.0 0 ustrl 0.3388 .0 0.34 0.34 [m/s] [data field] -1.70 -7.27 1.04 226.83 7.47 7.51 0.01 -0.66 -1.80 -7.50 0.95 226.47 7.71 7.73 -0.03 -0.80 -1.79 -7.53 1.06 226.65 7.74 7.78 -0.01 -0.70 A39