This study is aimed at performing and analyzing the inverter sizing optimization process for large-scale grid-connected solar photovoltaics (PV). The local solar resource was evaluated and compared to the available satellite data. Analyses of the solar irradiance distribution and its potential effects on inverter sizing were performed. The performance of five different, commercially-available PV module technologies was also evaluated by means of remote data analysis of irradiation and temperature, as well as electrical parameters. This was performed using the metrics of energy yield and performance ratio (PR) for PV systems. A methodology was developed for estimating the optimal inverter sizing in the region considering overload losses and economic aspects, aiming at the optimization and cost reduction of PV-generated electricity. Considering the five technologies, the mean yield was 150 kWh/kWp on a monthly basis. The annual PR ranged from 77% to 85% depending on technology. The local solar resource, measured during three years, was compared to five distinct satellite-derived solar databases. Cloud enhancement and cloud edge effects were observed through the monthly solar irradiation distribution over the three years of measurements. Around 28% of the incident annual irradiation was at irradiance levels at or above 1000 W/m2. The methodology developed for the optimal inverter loading ratio (ILR) was applied over one full year of solar generation data for the five technologies. It was observed that for inverter loading ratios commonly used on utility-scale PV power plants (around 120%), the overload losses varied from 0.3% to 2.4%, depending on technology. The optimal ILR for the more traditional crystalline Si PV technology was estimated to be 126%.
MARTINS DESCHAMPS, EDUARDO ; Rüther, Ricardo