The US Energy Information Administration (EIA) estimates that residential PV systems accounted for 61 percent of total annual generation from small scale distributed PV systems in 2021. The amount of power coming from these Behind-The-Meter (BTM) PV systems emphasizes the need for utilities and system operators to accurately predict PV system performance. This report describes a low-cost, accurate, and scalable software solution that infers PV system specifications from net load data and PV system location. The inferred system specifications include azimuth, tilt, and PV system size. Net load includes the combined effect of a customer’s production and consumption (i.e., gross load). Many utilities use only a single meter for their customers. These utilities collect net load data for billing purposes. We apply our methodology to 8,324 PV sites in the territory served by the Sacramento Municipal Utility District (SMUD). SMUD customers who own BTM PV systems have two meters. The primary meter measures net load and the additional meter measures PV production. As a result, SMUD represents a valuable test case to validate the method’s accuracy and compare it to other methods.
Inferring PV system specifications from net loads and reconstructing gross loads
