부산대학교 도서관

Energy-saving and consumption reduction in HVAC systems is becoming increasingly critical for environmental and social sustainability. This paper proposes a novel approach that combines deep learning algorithms with model predictive control (MPC) for reduce energy consumption and improve efficiency in HVAC systems to achieve energy saving. A DNN algorithm is integrated into the MPC framework for HVAC control which trained by historical data to learn the complex relationship between environmental conditions, occupancy patterns, and HVAC system parameters. By predicting optimal control signals and accurate sensor measurements based on real-time data inputs, the DNN enables precise control and setpoints optimization of HVAC systems. Simulation and experimental studies were conducted to evaluate the effectiveness of the proposed algorithms, the results demonstrate that the integration of DNN-based MPC leads to significant energy savings and reduction compared to traditional control strategies. The findings highlight the potential of intelligent HVAC systems to achieve substantial energy savings while maintaining occupant comfort.