AI-Based Energy Management of Grid Connected PV Systems in Positive Energy Buildings

In response to the growing importance of renewable energy in the global energy transition, this study addresses the critical need for efficient, real-time energy management in positive energy buildings by focusing on ultra-short-term prediction with a 5-minute prediction horizon. Utilizing a novel application of AI-driven dynamic control (AIDC) for intelligent real-time control, we employ the “House Winter Plus Energy” (HWPE) in Switzerland a Minergie-certified building with a 400% annual energy surplus due to extensive photovoltaic (PV) coverage to demonstrate how to maximize solar self consumption while enhancing grid stability. Two innovative AIDC strategies for grid-connected PV-battery systems were developed and evaluated through MATLAB/SIMULINK simulations. The first strategy enables dynamic switching between the grid and PV sources, achieving 60.57% reliance on PV energy. The second strategy integrates optimized battery storage to capture excess PV energy during peak production, reducing grid dependency to only 29.71% and covering 70.29% of energy needs through PV/battery systems, thereby reducing grid dependency and optimizing battery usage. These strategies leverage ultra-short-term prediction and real-time regulation using hybrid CNN-LSTM and LSTM models to accurately predict PV production and energy demand. Our findings highlight the potential of combining AIDC with real-time control and optimized battery usage to create resilient and economically viable energy solutions for future positive energy buildings (PEB).