Power system harmonic estimation using neural networks

The increasing application of power electronic facilities in the industrial environment has led to serious concerns about source line pollution and the resulting impacts on system equipment and power distribution systems. Consequently, active power filters (APFs) have been used as an effective way to compensate harmonic components in nonlinear loads. Obviously, fast and precise harmonic detection is one of the key factors to design APFs. Various digital signal analysis techniques are being used for the measurement and estimation of power system harmonics. Presently, neural network has received special attention from the researchers because of its simplicity, learning and generalization ability. This paper presents a neural networkbased algorithm that can identify both in magnitude and phase of harmonics. Experimental results have testified its performance with a variety of generated harmonies and interharmonics. Comparison with the conventional DFT method is also presented to demonstrate its very fast response and high accuracy.

I. Introduction

The difficulty in measuring power system harmonics comes from the fact that harmonic generating loads are dynamic by nature. Fast methods for measuring and estimating harmonic signals are thus required. Various digital signal analysis techniques are being used for the measurement and estimation of power system harmonics. These include FFT, Last Square, Least Absolute Value, Kalman filter, valve transformation etc.