Fault Identification and Monitoring in Solar Powered VSI with Induction Motor using CNN

As renewable energy sources increasingly power motor-driven applications, solar-based inverter systems have gained significant attention. However, faults within power electronic devices, especially Voltage Source Inverters (VSIs), pose challenges by compromising performance and potentially damaging components. This study introduces a convolutional neural network (CNN) approach to detect and classify inverter faults in a solar-powered three-phase VSI system that drives an induction motor. A comprehensive simulation, developed in MATLAB/Simulink, integrates solar photovoltaic generation, a boost converter, and a VSI. The CNN model, trained on current and voltage signals under both normal and faulty conditions, achieved a classification accuracy of 99%. The findings highlight the feasibility of implementing a fast and accurate fault detection mechanism suitable for real-time applications.