Automatic Failure Diagnosis of PV Array by Neural Network Using String I-V Curve and Differential Data from Reference

Abstract

This study uses string I-V curve to automatically diagnose failures in photovoltaic (PV) array. Detecting PV array failures is necessary for the safe operation of PV systems. In this study, a neural network is used to automatically determine the cause of independent and combined failures in PV array with high accuracy. Furthermore, after determining the defects, the rate of decrease of the failed cell current is detected and the series resistance value is estimated to quantitatively diagnose the failure. For this purpose, this study uses the difference data between the string I-V curve and the reference I-V curve, which is the I-V curve during normal power generation, to efficiently extract the characteristic values of the failure. As a result of verification using the measured I-V curves, the factors of single and combined failures were automatically determined with high accuracy (97.8%), and the rate of decrease of the failure cell current was successfully detected and the series resistance value was estimated.