Comparative Study of Active Power Filters with Least Mean Square algorithm and conventional strategy in discrete-time domain

Delgado I.; Ponce G.R.V.; Machaca S.; Mendoza E.; Calcina E. · 2024 IEEE International Conference on Aerospace and Signal Processing, INCAS 2024 (2024) · DOI: 10.1109/incas63820.2024.10798559

Resumen

This paper compares the performance of an active power filter implemented using the Least Mean Square (LMS) algorithm, with that of the traditional filter. An active power filter is designed to improve power quality in an electrical system. This entails reducing harmonics and reactive power, as well as eliminating unbalance. Since the LMS algorithm is among the fastest and simplest filters, which tends to make it easy to implement, the performance is evaluated using the Fast Fourier Transfer. Performance indicators for this application considered are harmonic removal and the stability period. The simulations are implemented in Matlab software, and the results allow a comparison of the LMS-based APF and reflect when the method is better and when it is necessary to have to stick to the traditional one. The LMS algorithm demonstrates results due to its ability to adapt to changes in real-time operation even under varying load operation frequency. This paper presents the mathematical model of two strategies of power compensation in continuous and discrete-time domain.

Instituciones: Universidad Nacional del Altiplano