TLBO-trained ANN-based Shunt Active Power Filter for Mitigation of Current Harmonics

Abstract

The increased utilization of nonlinear devices is resulting in damage to power distribution infrastructure by introducing harmonics into power system networks, which in turn causes distortion in voltage and current signals. A novel solution called Shunt Active Power Filter (SAPF) has been developed to address this issue using power electronics. This study aims to provide a method that is efficient and cost-effective for lowering harmonics and improving power quality in distribution infrastructure. The proposed method combines the Teaching learning-based optimization (TLBO) technique with an Artificial Neural Network Controller (TLBO-ANN) in conjunction with SAPF. The primary objective of the TLBO-ANN algorithms in SAPF is to minimise total harmonic distortion (THD) for maximum system efficiency. Initially, Gain values (K_i, K_p) for a regular Proportional-Integral controller are optimised with the Particle Swarm Optimisation (PSO) technique. Those optimized parameters obtained from the PSO-tuned PI controller serve as input and target datasets for training the ANN controller. Subsequently, the TLBO algorithm is utilized to further refine the ANN controller by finding the optimal weight and bias values. Using MATLAB/SIMULINK software, we compare the performance of the proposed algorithm to that of the PSO-tuned PI controller and traditional PI controller. The findings from the simulation suggest that a SAPF utilizing a TLBO-trained ANN controller could improve THD in the supplying current while maintaining harmonics within IEEE-519 accepting levels.