ANN-based fault diagnosis technique for photovoltaic stings

This paper proposes an intelligent fault diagnosis technique to detect and classify possible faults occurring in Photovoltaic strings, based on the analysis of the symptoms observed in the I-V characteristic. The technique consists of two algorithms: The first one allows the classification of faults that have not the same symptoms evolution; whi le for the second one, two Artificial Neural Networks (ANNs) have been developed in order to classify faults that are not distinguished by the first algorithm. Results indicate that the designed technique accurately detect and classify the investigated faults.