Neural network based adaptive phase and amplitude equalizer for optimal and robust control of the current regulated PWM inverters

In this paper a novel neural network (NN) based adaptive control system for optimal control of current regulated PWM inverters (CRPWMI) has been proposed. The optimal tracking ability that is tracking the reference signal with minimum phase and amplitude errors is one of the well known desirable performance of the inverter systems. Conventional controllers such as proportional-integral (PI) controllers can not provide acceptable tracking ability for many applications. The paper presents an adaptive phase and gain equalizer system by using band pass filters (BPF) with adjustable center frequencies. The filters are adjusted by a neural network based adaptive system for perfect tracking. The simulation results obtained from MATLAB/SIMULINK based studies show the successful operation of the proposed system and it can be suggested as a very suitable and useful tool for controlling a wide range of the closed loop inverter systems.