Comparison and performance analysis of closed loop controlled nonlinear system connected PWM inverter based on hybrid technique

Abstract This paper proposed closed loop control of nonlinear system connected inverter based on the optimal neural controller (ONC). The novelty of the proposed method rests on the hybrid technique which is the combined performance of both, particle swarm optimization (PSO) technique and Radial basis function neural network (RBFNN). It effectively optimizes the feasible solutions by updating the generations, by taking lesser time with greater reliability. In the proposed method, the PSO generates the dataset according to different loading conditions. The RBFNN is trained by using the target control signals along with the corresponding input load voltage error and change in error. Depending on the load variations, the RBFNN predicts the exact control signals of the inverter during the testing time. Since experimentation and comparison of such inverter models on hardware being relatively expensive, the proposed method is implemented in the MATLAB/Simulink platform and the performance has been validated through the comparison analysis with the conventional techniques. The comparison results have proved the superiority of the proposed method.