Improving the performance of PV systems by faults detection using GISTEL approach

In this paper, we present a new approach for detecting the faults in the photovoltaic systems based on the satellite image approach for estimating solar radiation data and DC output power calculations for detecting the failures. At first stage, the estimation of the hourly global horizontal solar radiation data has been evaluated by using the GISTEL (Gisement solaire par teledetection: Solar Radiation by Teledectection) model improved by the fuzzy logic technique. Thus, the results were compared with the ground solar radiation measurements. On the other hand, the comparison between the simulated and measured output DC powers was reached to find the nature of the faults in the PV array. The results showed a good accuracy and the simple implementation of the proposed approach. The estimation of the hourly solar radiation presents an NRMSE <0.22 using GISTEL model improved by fuzzy logic comparing with the estimation without fuzzy logic with an NRMSE = 0.2885 for clear sky and NRMSE = 0.2852 comparing with NRMSE = 0.3121 for cloudy sky. (C) 2014 Elsevier Ltd. All rights reserved.