Predictive Control for TABS: Optimizing Building-Integrated Thermal Storage for Energy Efficiency and Comfort

Thermally activated building systems (TABS) coupled with heat pumps convert variable renewable electricity into heat stored in building mass, enhancing grid flexibility. Their high thermal inertia, however, complicates control and can waste energy under rule-based logic. A model-predictive control (MPC) framework employing a simplified resistance–capacitance (RC) thermal model is introduced to forecast temperatures and optimize heat pump operation for simultaneous energy and comfort objectives. Performance was benchmarked against conventional rule-based control (RBC) in a TRNSYS simulation of a Shanghai office across a 21-day winter period. Compared with RBC, MPC lowered heat-pump electricity consumption by 15%, reduced cumulative thermal discomfort from 19.4 to 5.7 K·h, and cut start-stop cycling by 45.5%, indicating steadier equipment operation. These results confirm MPC's ability to unlock the dual value of TABS as both an efficient HVAC solution and a responsive thermal-storage asset. Adopting such advanced control is pivotal for buildings supporting carbon-constrained, renewable-dominated power systems.