Model Predictive Control of a Heat Pump System Integrated with PVT Collectors and Ice Storage

This paper focuses on optimizing a building heating system consisting of a heat pump, photovoltaic collectors (PVT), an ice storage tank, and a buffer tank. The aim is to assess the potential of using Mixed-Integer Nonlinear Model Predictive Control (MI-NMPC) in a novel heat pump system with dual heat sources - PVT collectors and ice storage, along with buffer storage. The paper investigates dynamic interactions, algorithm development, and performance evaluation, emphasizing energy efficiency, cost-effectiveness, and demand response. We find that control-oriented dynamic simulation accurately captures the system’s behavior. The application of NMPC in this context highlights its potential for advanced control strategies in hybrid (switched) energy systems. The simulation results demonstrate the improved performance of the system using the MI-NMPC strategy compared to a reference Rule-Based Control (RBC) strategy. This improvement is reflected in a 17.4% reduction in electricity costs over the heating season and a 15.4% decrease in net costs, including revenue from selling excess energy to the grid. In addition, the MI-NMPC strategy increases the self-consumption rate of the generated PV power to 38.2%, further improving the economic and energy efficiency of the system. The application of MI-NMPC in this context underscores its potential for advanced control strategies in hybrid energy systems, particularly in multi-source heat pump systems.