Deep Learning Based Fault Diagnostic Technique for Grid Connected Inverter

The global paradigm shift from conventional energy resources to the renewable resources has propelled the adoption of Photovoltaic (PV) as an alternative energy source. With the increased focus on PV systems, the reliability and stability of grid-connected solar inverters is a major area of interest. The occurrence of fault in any part of the inverter may severely affect the operation of the whole system leading to adverse effects at the grid end. To improve the system reliability, it is imperative to have a Fault diagnostic mechanism that is capable of identifying and classifying such failure conditions. This paper proposes an intelligent data driven Deep Learning based fault detection and classification technique for grid connected single phase PV Inverters. The fault diagnostic scheme is implemented in MATLAB. The training accuracy obtained is 99% and testing accuracy is 98.3%. Hence, the proposed technique is effective and satisfies the requirements of a fault diagnostic classifier.