Presentation Details
| Improving Utility Visibility of Behind-the-Meter PV and Battery Storage Using Net Load Upama Nakarmi, Marc Perez, Philip Gruenhagen. Clean Power Research L.L.C., Napa, CA, USA |
Abstract
The rapid growth of residential solar‑plus‑storage systems have created a large, distributed fleet of behind‑the‑meter (BTM) PV and battery systems with the potential to provide peak load reduction, local grid congestion relief, and enhanced distribution‑level reliability. However, most residential installations lack real‑time telemetry for these systems, leaving utilities with only net load measurements that combine customer consumption, PV production, and battery charge/discharge measurements. This lack of operational visibility limits the ability of utilities and aggregators to quantify true PV output, characterize battery behavior, and incorporate distributed energy resources into planning and grid operations. This paper presents a novel methodology for detecting operational battery storage systems and inferring their charge/discharge time series using only net load data and system location. Building on our prior work that infers PV system specifications from net load, the method first simulates PV production using inferred PV specifications and irradiance data. The simulated PV signal is then removed from measured net load to help isolate the residual battery signal. A random forest–based regression model is trained to estimate hourly battery charge/discharge timeseries profiles. Then, the inferred battery timeseries is used to estimate key battery specifications—including effective capacity, peak power rating, and round‑trip efficiency. The methodology is demonstrated on a residential solar‑plus‑storage site in California with five years of measured PV, battery, and net load data. Results show that the approach successfully detects operational battery behavior and reconstructs battery charge/discharge profiles with reasonable accuracy, enabling estimation of battery specifications from net load alone. The full paper will extend these results to a large, synthetic dataset across 18 U.S. locations, highlighting the potential for utilities to gain meaningful visibility into BTM PV and storage systems without additional metering infrastructure.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.