Realtime Implementation of Multiple Fault Detection Methodology for Solar PV Farms using Hybrid AMH and Deep Learning Techniques

In this article, an intelligent fault detection mechanism is proposed to classify multiple electrical faults. Various electrical faults like line-line fault, open circuit fault, short circuit faults, and line-to-ground faults are studied in detail. Voltage and current sensors placed at each solar photovoltaic (SPV) array string collect the data during these fault conditions. Machine learning and Deep learning techniques are used to classify the type of fault and can even detect the exact location of the fault. A mismatch detection circuit consisting of a boost converter and optocoupler is designed to differentiate between mismatch faults and electrical faults; this circuit design also indicates the exact location of a mismatch fault. The proposed method is implemented in the Python platform and the results are analysed.