Open-Circuit Fault Diagnosis for Three-Phase Inverter in Photovoltaic Solar Pumping System Using Neural Network and Neuro-Fuzzy Techniques

Abstract

This paper deals with the detection of single- and double-switching faults (open-circuit type "O-C") which appear in the inverter of a photovoltaic solar pumping system. Our system as a whole contains a photovoltaic module, a DC/DC step-up converter controlled by perturbation and observation maximum power point tracking technique, a three-phase DC/AC inverter controlled by the sinusoidal pulse width modulation technique, a three-phase induction motor, and a water pump. The used techniques to detect this type of faults are based on artificial intelligence (AI) (neural networks and neuro-fuzzy networks); we use AI as an observer to the inverter in order to detect the faults using extracted features from the inverter output currents. Both of the proposed fault diagnosis techniques show a good performance and high accuracy with less than ±5% of error for neuron-fuzzy and ±7% for artificial neural network and a response time of less than 0.1 s, which is a satisfying speed to detect the faults before a total degradation or any undesirable effects. This paper fulfills an identified need for faults diagnosis of a three-phase inverter in photovoltaic solar pumping systems using AI. The effectiveness of the AI techniques was evaluated for O-C fault detection by simulation tests using the MATLAB/Simulink environment.