A New Experimentally Validated Archetype for Model Predictive Control of Residential Buildings with Heat Pump

Model Predictive Control (MPC) provides efficient control of building heating systems, integrating comfort and energy goals. However, widespread implementation remains constrained by the computational complexity of predictive models and the challenges of controlling advanced thermal distribution systems, particularly those with heat pumps. To address these limitations, the authors previously proposed several MPC archetypes that simplify design and enable large-scale use. Experimentally validated on a real heat pump system, they cover three heating setups, namely low-temperature radiators without storage, with integrated storage, and underfloor heating. The present study extends this framework by introducing a novel archetype specifically developed for heat pump systems with underfloor heating and an external thermal energy storage. The MPC minimizes an objective function configurable via a user-selected penalty signal, such as electricity price, primary energy consumption, or other operational criteria. This work presents the archetype and demonstrates its experimental validation using a Hardware-in-the-Loop setup, in which the building is emulated while the heat pump, thermal storage, and hydraulic circuits are physically implemented. The results demonstrate the accuracy and effectiveness of the MPC, with indoor temperature deviations remaining low (maximum 2.33 °Ch below and 4.40 °Ch above the set-point) and feasible solutions achieved in 98.52% of the test duration.