Title: Intelligent active power filter based on ADALINEs and PI-neural under unbalanced and distorted voltages

Authors: Ouarda Lahmadi; Louiza Benfarhi; Aziz Chaghi

Addresses: Department of Electrical Engineering, Batna University, Algeria ' Department of Electrical Engineering, Batna University, Algeria ' Department of Electrical Engineering, Batna University, Algeria

Abstract: This paper presents an intelligent active power filter with the different part of the filter based on a simple type of artificial neural network called ADALINEs (APF-neural) to improve its performance. The goal of our approach is to propose a new PLL-neural using the ADALINEs for the online extraction of the voltage component to recover an equilibrated voltage system and a PI-neural regulator for estimating the phase of a time-varying, then, the application of this PLL for assuring good filtering characteristics in power systems. The harmonic currents identification is based on the neural diphase currents method by using two ADALINEs. In the control scheme of the inverter and in the regulation of the dc voltage, the injected harmonic currents use a PI-neural. Our proposed APF-neural has been evaluated in simulations under unbalanced and polluted voltage. The results show excellent behaviours and performances, as well as robustness and usefulness.

Keywords: active power filters; APF; artificial neural networks; ANNs; ADALINE; power quality; learning rules; harmonic distortion; diphase currents; unbalanced voltage; distorted voltage; phase-lock loop; PLL control; PI control; intelligent power filters; dc voltage regulation; simulation.

DOI: 10.1504/IJPEC.2016.079889

International Journal of Power and Energy Conversion, 2016 Vol.7 No.4, pp.336 - 360

Received: 26 May 2015
Accepted: 26 Oct 2015
Published online: 20 Oct 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article