________________________________________________________________________________________________ INDIGO Project: A Simulation Based Approach To Support District Cooling Design And Operation Andrea Costa1, Andrea Bassani1, Jesus Febres2, Susana López2, Saioa Herrero2, Miika Rama3, Krzysztof Klobut3, Yves Stauffer4, Rafael E. Carrillo4, Gorka Naveran5, Marcus Keane6, Raymond Sterling6 1 2 R2M Solution, Pavia, Italy / IK4-Tekniker, Eibar, Spain 3VTT, Espoo, Finland / 4CSEM, Neuchâtel, Switzerland 5Giroa Veolia, Zamudio, Spain / 6NUI Galway, Galway, Ireland Abstract All the developed models will be validated, both independently and considering the integration, using data In a district cooling system different types of cooling acquired at the test-site. generation can be combined (e.g., vapour compression The validated models Will be considered as reference for chillers, absorption chillers, and free cooling). Controlling the development of the innovative controllers, of the such complex systems in an efficient Way is challenging: management strategy, and of the planning tool. The neW the cooling demand is much more difficult to predict than models developed in Modelica Will be part of a District the heat demand and, as for absorption chillers, heat cooling open-source library (DCOL). sources such as the solar energy and the Waste heat are not predetermined by the designers. Introduction The EU project INDIGO deals With the improvement of This paper presents the role of the energy models in the District Cooling (DC) systems. Its main goal is the development of INDIGO, a research project that is funded development of a more efficient, intelligent, and cheaper by the European Union in the context of the Horizon 2020 generation of DC systems (Loureiro, 2018; Costa, 2017). research and innovation program. The project started in The results of INDIGO Will include: March 2016 and it is going to finish in August 2019. Its - predictive controllers; main goal is the improvement of the district cooling (DC) systems. - system management algorithms; In comparison With the heating systems, the cooling - an open-source planning tool. systems present problematic aspects related to the greater To validate the results, the consortium is analysing some difficulty in the prediction of the energy demand. The case studies. The proposed solutions for DC systems Will reason is that the energy demand can change quickly be installed in the Basurto Hospital campus in Bilbao. because it is influenced by factors such as the solar Different models regarding the buildings and all the radiation and the internal heat loads. relevant components of the DC system are being Another problematic issue is that the difference between developed: the supply Water temperature and the return Water 1. generation systems; temperature for DC loops is generally around 8°C, While 2. distribution and storage systems; in district heating loops the difference is usually greater 3. HVAC systems; than 40°C. That causes an increase in the cost of the 4. thermal behaviour of the buildings, considering also piping system and of the energy requested for pumping. internal loads and building use. The improvement of the DC systems shall be achieved The first three parts are being simulated by means of through the development of new tools for the design and Modelica (Modelica, 2016), an open-source object- for the management of district cooling systems: oriented modeling language that provides dynamic • predictive controllers that Will be responsible for simulation models for building energy and control systems. The fourth part is being modelled With determining the set-points of the mechanical systems EnergyPlus (EnergyPlus, 2016). They are going to be (some of them Will include self-learning algorithms), integrated through the Functional Mock-up Interface at all levels (generation, distribution, storage, AHUs, (FMI) for co-simulation (Nouidui, 2014). and fan-coil units). They Will be developed on the basis of “reduced” and validated versions of the The models of certain building envelope elements are energy models (Hoyo Arce, 2017); validated using experimental measurements (heat flow rates, temperatures, entering solar radiation). • an innovative management strategy, on Which the controllers Will be based, Will consider energetic Component models for the air handlers and for the fan- aspects, environmental aspects, and economical coils found in the studied buildings are developed in aspects; Modelica. Different kinds of chillers are modelled too. • an open-source planning tool for the evaluation of Particular attention is paid to the distribution system, existing systems and for the design of new district where thermal and hydraulic effects are coupled and thus must be considered jointly. cooling systems. It Will be based also on the results of simulations; ________________________________________________________________________________________________ Proceedings of the 16th IBPSA Conference 1913 Rome, Italy, Sept. 2-4, 2019 https://doi.org/10.26868/25222708.2019.210705 ________________________________________________________________________________________________ • an open library including physical/mathematical 4. building geometry, thermal behaviour of the building models developed in Modelica of the components that structures, internal loads, and building use. compose a DC system that will be called District 5. Learning algorithms, control and optimisation. cooling open-source library (DCOL). 6. Planning tool. The models regard mainly the real situation of the Basurto The fourth part Will be developed in EnergyPlus and Will Hospital, in Bilbao, Which is divided in 20 buildings, be converted into functional mock-up units (Blochwitz, among them 9 are mechanically cooled. The results of 2012) that Will be then imported and used in a Modelica INDIGO are going to be applied and measured in the model that includes the AHUs. The Weather data Will be plants relative to that facility. The DC system Was integrated in the same functional mock-up unit (FMU) of installed in 2003 and extended in 2011 by Veolia-Giroa. the building. The FMU Will have as input variables latent The chilled Water is produced by tWo absorption chillers and sensible heat gains provided by the AHUs to each and by four vapour compression chillers. The system conditioned thermal zone and as output variable room supplies chilled Water alWays at 7°C (as setpoint conditions (zone air temperature and relative humidity) temperature) and it modulates the Water floW rate. The and outdoor air conditions. absorption chillers are cooled by Water loops, one vapour The combined Modelica-AHUs with EnergyPlus-FMUs compression chiller is Water-cooled While the other three will be then converted into another FMU and used for electric chillers are air-cooled (Fig. 1). The hot Water, training an advanced control algorithm to be developed in which is used by the absorption chillers and to heat the INDIGO and for the integration With the distribution- buildings, is heated by the Waste heat of a cogeneration storage-generation models, Which are being created in system. That heat can be stored in a Water tank. The Modelica. cogeneration system is activated based on the economic The integration of different tools alloWs the exploitation convenience of producing electricity. of their different capabilities. EnergyPlus is a Whole In this case the cooling demand is particularly variable building energy simulation softWare, whose development because it depends on the number of scheduled surgeries is funded by the U.S. Department of Energy – Building and on the occupancy level as Well (Passerini, 2017). Technologies Office. It is free, open-source, and cross- The internal zones are conditioned mainly though Air platform. In EnergyPlus many physical-mathematical Handling Units (AHUs) or through fan-coil units. Since models relative to the building physics (as Well as to the the local regulations for hospitals do not permit air HVAC systems) are already available and validated. recirculation, an important amount of energy is requested Modelica is a non-proprietary, object-oriented, equation- for the treatment of the outdoor air. based language. The use of a Modelica library (Wetter, In the other two pilot cases, located in Barcelona, the 2014) and the development of new models require the use models and the INDIGO solutions (controllers and of a Modelica simulation environment (some management system) Will be tested Without a physical environments are commercial, a few environments are installation. They are not considered in this paper. free). The use of Modelica alloWs a greater flexibility than EnergyPlus. A similar approach has been applied to Water netWorks control optimisation (Perfido, 2017). Validation The models Will be validated, at different levels, using measured data. The quality of the measured data Will be checked in order to eliminate the unreliable ones. The validation process Will consider statistical test methods, with the definition of an acceptable range of accuracy with respect to the model goal. It will be an iterative process. The validated models Will be then used as a tool for the improvement of the plant and for the development of other tools, such as the controller, the management Figure 1: Scheme reporting the situation of the cooling strategy, the planning tool. production/distribution plant in the hospital A monitoring campaign has started in summer 2016 and The integrated detailed modelling it has been completed