Master's thesis Martin Stiller
Model predictive control of a geothermal field for the cost-optimal use of its heating and cooling supply potentialCopyright: EBC
The trend of the recent past in the energy sector shows a steadily growing importance of renewable energies to prevent greenhouse gas emissions from fossil fuels. The building sector accounts for a high amount of energy consumption in Germany and offers a high potential for optimization. Ground coupled heat pumps are a promising technology for the simultaneous heat and cold supply of buildings. Compared to other regenerative energy sources such as solar thermal systems, the ground source is characterized for its permanent availability.
The challenge of using the geothermal potential is to keep the soil temperature within small fluctuations, so the potential of the ground source is not overused regarding the long-term behavior. The slow-reacting geothermal field system with its large time scales and the fastreacting building system have to be coordinated. A promising approach for this coordination is the Model Predictive Control (MPC). The MPC relies on dynamic models of the process to be repeatedly optimized in defined time slots while it is possible to include future predictions such as weather forecasts. Thus it anticipates the buildings’ demand and optimizes its thermal behaviour based on specified control goals by minimizing a cost function. A special type of MPC is the distributed model predictive control (DMPC). It computes every subsystem model separately, which can simultaneously reduce the computational effort.
The goal is to develop a way to make long-term use of the cost-free geothermal energy possible without reaching the limits of its capacity and to achieve cost-optimal operation by low CO2 emissions in combination with secondary suppliers. An existing DMPC algorithm developed in previous works is adapted to the usage of a geothermal field for building’s heat and cold supply. The flow and return temperatures of the geothermal field vary considerably depending on the soil conditions and the type and way of use of the building, which is why a tailor-made, individual strategy is necessary. A Building’s individual energy demand curves can be used to derive an annual temperature curve for the flow temperature into the field. The operation of the geothermal field should follow this temperature curve, whereby deviations from the determined temperature are penalized or rewarded by the MPC algorithm.
The approach developed in this thesis to achieve sustainable operation of the field by penalties and rewards turns out to be promising. Cost factors for past, present and future deviations are introduced. By weighting the factors, a tracking of a ground setpoint trajectory is possible. The setpoint trajectory is provided by long-term optimization from averaged building loads in order to prevent overheating or cooling down the field over several years. The short-term optimization ensures that the soil temperature is maintained despite dynamic building loads and the use of fossil fuels is reduced.