A novel Counter based Neural Network for Switching Fault Identification in Power Inverter

This paper proposes a novel neural network for the detection of switching faults in the power converters and inverters. The highlight of this work is the enhanced counter based neural network utilized for the identification of switching faults in the 27-level Ladder Multi-level Inverter. Though there are several machine and deep learning algorithms, the neural network algorithm present high reliability, accuracy and consistency in the classification, identification and prediction. In this work, the neural network is combined with the enhanced counter circuit to detect the fault occurrence in the inverter circuit. The proposed algorithm is developed using the VHDL code and synthesized in the FPGA board. The performance analysis is evaluated for power, area and IC layout for the proposed method.