Measurements and simulations of the performance of the PV systems at the University of Gävle

Abstract: In the following years, the countries will have to face an increase in the energy demand. So far, the fossil fuels have been the main source to meet the energy demand, but they involve serious problems: they contribute to the climate change with high emissions of greenhouse gases, there is an uneven distribution of these resources and their reserves are finite. The renewable energies are the most reliable alternative, with a very low environmental impact in comparison. Among them, the photovoltaics seems to be the most promising emerging technology for the electricity generation. Its rapid growth in the last years has been due to the reduction achieved in the cost of the PV panels. When planning a PV installation, it is essential to be able to estimate the production. The power of a PV-module is given by the manufacturer at standard conditions (STC), which means that the irradiance is G=1000W/m2 at normal incidence and the temperature of the module is 25˚C. However, these conditions will never be reached in a real installation. Therefore, the measured power of the system has to be adjusted for the real conditions so that the real production and performance can be estimated. Today there exists no standard method for this procedure in Sweden. The main aim of this thesis is to develop a theoretical model for the four PV-systems installed at the laboratories (building 45) of the University of Gävle to estimate the performance and production, and prove its validity by comparing with real data measured with a short time resolution (second). This will also allow to know if the power generated by the modules is the promised one by the providers. Three of the studied systems have monocrystalline silicon modules, with different schemes: one system with bypass diodes, another with TIGO optimizers, and the third one with microinverters. The fourth system has thin film modules. The theoretical model considers correction factors for the cell temperature, the angle of incidence and the real irradiation reaching the modules’ surface; as all these aspects reduce the power obtained. When studying this model for clear sunny days, it can be observed that the theoretical model adjusts perfectly for the four systems in these conditions and almost a completely linear dependence is achieved between the measured and estimated power. The worse adjustment is obtained for the thin film system, for which the theoretical model gives lower values than the real ones. However, a better approximation can be obtained for this system by adjusting the value of the correction factor for the cell temperature. Moreover, the high values obtained for the maximum power during the clear days, very close to the peak power, indicates that the maximum power value provided by the manufacturers is in concordance with the real performance of the modules. In case of cloudy days, a small-time delay has been appreciated between the data recorded by both logger. The results have been studied with the raw data, obtained worse adjusting, and correcting this time discordance, getting again accurate results from the theoretical model.