Lawrence Berkeley National Laboratory

The growing interest in onsite solar photovoltaic and energy storage systems (PVESS) is partially motivated by customer concerns regarding grid reliability. It is crucial to consider both the unpredictability of such events and the location-specific patterns in their timing and duration to accurately evaluate the capability of PVESS in alleviating these interruptions. We developed the publicly available Power Reliability Event Simulation Tool (PRESTO) that simulates power interruptions at the county level to examine PVESS backup performance - the percent of otherwise interrupted load served - during simulated short-duration power interruptions in a representative single-family home located in three representative U.S. counties. A PVESS with a relatively small amount (10 kWh) of battery storage reliably serves essential non-heating and cooling loads in most simulated events, serving an average of 93% of the energy demand across all events and fully serving loads in 84% of total events. This occurs despite the fact that the battery is being used on a near daily basis for time-of-use (TOU) bill management, with a minimal reserve setting, and therefore may not be fully charged at the time the interruption occurs. In contrast, requiring this PVESS to meet heating and cooling loads significantly diminishes its performance, with an average of 70% of load served and only fully serving 43% of total events. These findings highlight the challenge of addressing short-duration interruptions and stress the importance of understanding factors influencing PVESS backup performance. Our study provides valuable insights into potential mitigation strategies, with a focus on grid charging and battery size optimization.