New Intelligent Fault Diagnosis (IFD) Approach for grid-connected photovoltaic systems

Abstract The present work brings a new intelligent algorithm for PV system’s diagnosis and fault detection (IFD). At this stage of the study, this algorithm can detect and identify three recurrent cases between healthy and short circuit faults, as well as string disconnection in PV array using artificial neural networks (ANN). Both, detection and isolation are simple and fast. The developed model requires small training period and is based on only four inputs: the maximum power current and voltage from the output current-voltage (I-V) characteristic, the solar irradiation and the cell temperature. Experimental validation of the proposed IFD has been carried on small grid connected PV generator (PVG). The obtained results demonstrate that this approach can precisely detect and classify the existing faults with high accuracy (98.6%).