We apply Model Predictive Control (MPC) to a building emulator; the considered building is modelled as a 1200 m2 single space using verified high-fidelity models from the IDEAS Modelica library. The building is equipped with a ground source heat pump (GSHP) and a gas boiler for heat production, and passive cooling from the geothermal borefield and an active chiller for cold production. The building can also choose between thermally activating the building structure through embedded pipes and using air conditioning as the slow and fast-reacting emission systems, respectively. Thus, the building case study is hybrid both at the production and emission sides. The system components are sized based on a previous dynamic simulation of this emulator, and the geothermal borefield is deliberately undersized due to the hybrid production system of the building.
The first set of inputs from the optimization is used to advance the simulation, which starts at the beginning of the heating season, hence closing the loop. The optimization problem aims at minimizing the energy use of the installation while keeping thermal comfort in the space. Since the performance ratio between the GSHP and the gas boiler is around 5:1, MPC prefers the use of the GSHP to cover the heating energy needs of the building as much as it can. Computationally, solving this optimization problem is expensive and in the same order of magnitude as the full MPC simulation.
The energy distribution results of the Optimal Control Problem (OCP) revealed that the energy use of the building could be further decreased. To achieve this, the OCP increases the amount of cooling produced by passive cooling, thus reducing the thermal imbalance of the borefield and unlocking further use of the GSHP for the next heating season.
In summary, the shadow cost comprises the energy use of the building over a long-term horizon and can be computed using a set of prescribed predictions of the building heating and cooling needs and adding a set of static energy balance equations to the optimization problem. The short- and long-term optimizations are coupled by the load history of the borefield model, which in turn determines the predictions of the borefield outlet fluid temperature. This simulation study clearly shows that there is further potential in the optimal control of hybrid geothermal systems by accounting for the ground’s long-term dynamics.
Iago Cupeiro Figueroa, Lieve Helsen KU Leuven & DeltaQ
The text has been shortened by the HPC team
Read the full article here >
Read the full HPT Magazine here >
