An ANN-based harmonic mitigation and power injection technique for solar-fed distributed generation system

Abstract

In this article, a fifteen-level cascaded H-bridge multilevel inverter with active power filtering capability is suggested to increase the power quality in a single-phase distributed generation (DG) system. Integrating renewable power sources in the distribution line becomes easier with advancements in power electronic converters. At the same time, increase in usage of nonlinear loads leads to electronic pollutions like harmonics and power factor issues. Since most of the DG systems have interfacing inverters, it can be utilized for power quality improvement. This proposed work presents a control scheme to gain maximum utilization of these grid interfacing inverters with artificial neural network (ANN). An instantaneous p-q theory in a-ß-0 reference frame-based control algorithm is derived to control the inverter in the single-phase distribution system. In the ANN-based approach, the control algorithm is solved and the ANN is trained based on the switching angles obtained. The trained system is integrated with DG system to operate as a multitasking circuit by adding active power filter (APF) operation. It is proposed to utilize the interfacing inverter as: (1) interfacing inverter to add power produced from renewable sources and (2) APF to mitigate the harmonics. Both functions are accomplished simultaneously. The proposed work is verified with extensive MATLAB/Simulink, and the obtained results demonstrate that the proposed approach delivers a notable improvement in power quality concerning reduction in total harmonic distortion and injection of real power generated through RES into the distribution line. A 3-kWp photovoltaic panel with multifunctioning inverter designed utilizing ANN is executed in an experimental setup to prove the efficiency of the proposed approach.