Team Members: Ian Bolliger, Alana Siegner, Rinitha Reddy Kothapalle
School: UC Berkeley

The rise in home solar generation is reducing emissions from the residential building sector. However, net energy metering (NEM) policies are being phased out in many states, making residential solar projects less attractive from a financial standpoint. EMPOWER is addressing this problem through the development of a modular, affordable, residential energy management system that leverages the predictive power of modern machine learning to simultaneously optimize local energy generation, storage, and flexible loads within a house or microgrid. It integrates weather, photovoltaic, and user behavior forecasts to increase efficiency and is retrofittable to existing homes. In areas without NEM policies, EMPOWER estimates savings of over $500/yr for a product with a target price of $200-300.