“Model predictive control brought to practice”
The integration of renewable energy sources into the electricity grid requires flexible consumers to maintain grid stability. Heat pump systems, combined with thermal energy storage and building thermal inertia, can provide such flexibility. Model predictive control harnesses this potential by enabling predictive and costoptimized operation while maintaining comfort constraints. In this study, we present results from an industrial field deployment of MPC in a pool of residential heat pump systems. The analysis combines three levels: qualitative behavioural signatures, quantitative year-on-year comparison of a representative system against similar buildings, and pool-level aggregation. Results indicate significant potential for energy cost savings and improved PV self-consumption, while maintaining user comfort. The work demonstrates the practical feasibility and robustness of MPC in real-world operation.