Analyzing the efficacy of LMS-based control algorithms in enhancing power quality in three-phase grid-connected systems

Abstract

This paper introduces a novel method for controlling shunt active power filters (SAPFs) to improve network efficiency and reduce carbon emissions in the utility sector. It addresses the problem of current harmonics degrading system performance by employing reference current generation based on the least mean square (LMS) algorithm, it decomposes distorted current into fundamental active, fundamental reactive and harmonic components. Traditionally, LMS implementations suffer from poor dynamic response due to uniform learning rates. To overcome this, the method adjusts the learning rates for fundamental active and reactive components separately, improving dynamic response and reducing computational complexity. Extensive analysis validates the effectiveness of this approach under various conditions, demonstrating its superiority in enhancing system performance and reducing carbon emissions in electrical networks.