Systematic photovoltaic system power losses calculation and modeling using computational intelligence techniques

The performance of a photovoltaic (PV) system is highly affected by different types of power losses which are incurred by electrical equipment or altering weather conditions. In this context, an accurate analysis of power losses for a PV system is of significant importance. Hence, the systematic calculation of the PV system power losses based on recorded data of the main electrical and meteorological parameters is investigated in this paper. The proposed approach for power losses calculation investigates both array capture losses (e.g. losses resulted from cell temperature, soiling, low irradiance, snow cover, mismatching, and module quality degradation) and system losses (e.g. losses resulted from cabling, inverter, etc.). Moreover, according to the best of our knowledge, the paper proposes for the first time in the literature the modeling of all types of losses based on computational intelligence techniques. The proposed models can predict the future daily values for each type of loss solely based on the main meteorological parameters. The proposed losses calculation approach is applied to 8 years of recorded data for a 1.44 kWp rooftop PV system located in Denver, CO. Several prediction models are built based on the calculated values of the losses. The models are not only validated on the PV system itself, but also on another PV system with different electrical characteristics in Las Vegas, NV. As the results show, the loss prediction models developed in this paper perform accurately for the main PV system and are also applicable to other PV systems.