An Extensive Analysis of Power Quality Improvement in a Three-Stage 15-Level Inverter Using Various ANN Algorithms

The performance of multilevel inverters (MLIs) in microgrids plays a vital role in managing daily fluctuations in power demand and supporting the reliable operation of renewable energy sources (RES). To enhance reliability, a 3-stage, 15-level inverter is designed, offering improved performance with a reduced number of switches and lower Total Harmonic Distortion (THD). The number of switches is lowered by using a better Pulse Width Modulation (PWM) method, and Artificial Neural Network (ANN) algorithms help with finding faults. The research integrates both PWM and ANN techniques to enhance the power quality of multilevel inverters (MLIs). A controller using an ANN and a Field-Programmable Gate Array (FPGA) is suggested to correctly identify problems and measure Total Harmonic Distortion (THD) in different fault situations. Datasets for training and testing the ANN models are collected from MATLAB simulations and experimental setups. The performance of different ANN algorithms is analyzed to identify the most effective approach. Fault classification is conducted based on THD measurements for various fault types, including first module failure, second module failure, third module failure, single switch failure, two switches failing in the same leg, and two switches failing in different legs.