Model Predictive Control Based Energy Management System for Home Energy Resiliency

As extreme weather events increase, so do the outages they cause. During these unplanned outages, maintaining habitable conditions in a house requires a reliable energy supply for thermal comfort and critical loads. A rooftop photovoltaic (PV) system with battery storage is an excellent carbon-free solution to enhance energy resiliency against such outages. However, to maintain habitable conditions, the system must meet both non-air-conditioning (non-AC) load demands and the high startup power requirement of the AC system. This necessitates an intelligent controller to manage the trade-offs between competing demands. This paper proposes an intelligent controller based on Model Predictive Control (MPC) and compares its performance with a Baseline controller and a Rule-Based controller using a three-dimensional resiliency metric. Extensive simulations for various energy system sizes and AC startup power requirements were conducted for a week-long period in a single-family house in Florida, post-Hurricane Irma in 2017. The results demonstrate that the MPC controller outperforms the other controllers in energy-constrained scenarios (smaller PV-battery size, larger AC startup power requirement), providing better thermal comfort and servicing non-AC loads more effectively.