Predicting Subhourly Clipping Losses for Utility-Scale PV Systems

Utility-scale photovoltaic (PV) systems are generally designed with a greater DC power capacity than their inverters are capable of converting to AC power, resulting in system clipping. Modeling tools used to forecast the generation of photovoltaic systems predict the degree to which a system is expected to clip throughout the year. However, standard PV modeling software utilizes an input of hourly meteorological data to predict these losses. For periods of intermittent irradiance in which the incident irradiance at times may cause the system to clip, an average of this irradiance over the course of an hour may result in a prediction which fails to encompass clipping losses for those intermittent intervals. Presented within this paper is a model for predicting these losses by averaging high-resolution datasets and comparing losses through the NREL developed modeling tool System Advisor Model. This prediction model utilizes freely available data through the National Solar Radiation Database to predict these losses, so the methodology presented allows users and developers to predict subhourly clipping losses at a site where minute-resolution ground data would otherwise be unavailable.