Comparison of software prediction and measured performance of a grid-connected photovoltaic power plant

In this work, the performance of a photovoltaic (PV) installation is assessed. The plant consists of a grid-connected centralized system, where the supplied power is not associated with a particular electricity customer. Operational data from this PV plant, ground-mounted and located in the north of Portugal, are now available for a period of 3 years. The plant is equipped with PV modules (amorphous Si cells), with 60 Wp per module, and a total generating capacity of 124.2 kWp. In this installation, 24 inverters are used. To obtain an accurate prediction of the efficiency and power output, the characteristics of all plant components were introduced in the PVsyst software and TRNSYS software, together with meteorological data: either those collected at a local meteorological station or those provided by Meteonorm. The results obtained through the simulations and the measured output power values were compared. The results showed that both PVsyst and TRNSYS seem to be good tools to predict the annual electrical production of a PV plant, with an average relative difference in the results between both around 2%. In the simulations, parameters like orientation and inclination of the PV modules were analyzed and recommendations for improving the PV system production are given. The results showed that in the annual electrical production, the effect of panel tilt angle is more significant than the effect of panel orientation.