Power Generation Nowcasting of the Behind-the-Meter Photovoltaic Systems

Rapid increase in the number of photo voltaic systems (PVSs) across the low voltage distribution systems requires real-time monitoring of those systems. However, considering the necessary investment cost, real-time monitoring infrastructure installation does not seem to be achieved in the near future. Thus, alternative approaches to achieve this necessity is crucial. In this study, a method is proposed to nowcast the power generation of the behind-the-meter PVSs. The method relies on the strong correlation between the harmonic current injection of the PVS inverter and the power generation of the PVS. The Artificial Neural Network method (ANN) is utilized to model the relation between the predictors and the aggregated PVS output, i.e. generated power. Finally, the effectiveness of the approach is assessed with real field data.