This study proposes an MPC-based control method that considers uncertain disturbances as an optimal control method for Thermally Activated Building Systems (TABS) with a slow thermal response. Specifically, the load prediction results are corrected in real-time by data assimilation. The effectiveness of the proposed method was demonstrated by validating the accuracy of three load prediction models: one for the planning phase, one for the early operation phase, and one for the late operation phase. The co-simulation results using CFD and MATLAB/Simulink showed that the integrated water flow rate of TABS can be reduced by correcting the load prediction.
