Occupancy-based model predictive control for residential building cluster with photovoltaic and battery systems

The rapid development of renewable energy sources brings new challenges for power systems operation, referring to stability and reliability due to its’ neutrally intermittent and stochasticity. And improving energy flexibility is crucial for optimizing power system operations, achieving financial benefits, and making effective policies. However, most existing flexibility studies focused on individual building levels rather than clusters. It cannot well consider the uncertainty of building envelope, HVAC systems, and occupant behavior, which is important for aggregate energy flexibility management. In this study, we first generated occupancy presence data from their hourly power meter data, then the Monte Carlo method was used to generate different building envelope characteristics, and HVAC parameters to capture the uncertainty of building clusters. An occupancy-based model predictive control strategy was used for HVAC load shifting and finally the shifted load was integrated with the PV-Battery optimization framework. OBMPC demonstrated great energy flexibility potential which can shift the load during the on-peak hour, save energy costs and reduce the system ramping ratio. But it could also increase the load during the off-peak hour and increase the system ramping ratio. PV-Battery can provide more energy flexibility and improve users’ financial benefits. But the same flexibility strategy could have opposite performance for individual buildings and cluster buildings. This study demonstrated the benefits of predictive control in building HVAC and PV-Battery systems for grid flexibility. Additionally, the proposed hierarchy control strategy can be effectively implemented in real-world BES operations.