Intelligent switch control of an active power line conditioner

Article type: Research Article

Authors: O'Connell, Robert M.

Affiliations: Electrical and Computer Engineering Department, University of Missouri-Columbia, Columbia, MO, USA. Tel.: +1 573 882 8373; Fax: +1 573 882 0397; E-mail: oconnellr@missouri.edu

Abstract: his paper describes the application of various soft computing tools, including the genetic algorithm, artificial neural networks, and fuzzy logic to the control of the semiconductor switches in a novel active power line conditioner (APLC). The use of intelligent control in the APLC is warranted by the complexity and cost of conventional control schemes that could be used alternatively. MATLAB simulations of the APLC with either the neural network controller or the fuzzy logic controller predict that the system can improve the total harmonic distortion of a power line current from over 100% to well under 10% using the relatively low, and therefore relatively lossless, switching frequency of 3840 Hz. Measurements on an experimental prototype showed that in steady-state operation, the neural-network-controlled APLC improved the distortion of a specific line current from 114% to 12%, and the fuzzy-logic-controlled system improved it from 125% to only 30%. On the other hand, the fuzzy-logic-controlled system was able to smoothly track a sudden load change and convert the line current from one compensated steady-state and nearly sinusoidal waveform to another one in approximately one 60 Hz cycle.

DOI: 10.3233/KES-2007-11205

Journal: International Journal of Knowledge-based and Intelligent Engineering Systems, vol. 11, no. 2, pp. 129-137, 2007