Frequency and harmonics evaluation in power networks using fuzzy regression technique

Abstract

A new technique for frequency and harmonic evaluation in power networks is proposed in this paper. This technique is based on fuzzy linear regression and uses the digitized voltage samples, which are fuzzy numbers, to estimate the frequency and harmonic contents of the voltage signal. In this technique we formulate a linear optimization problem, where the objective is to minimize the spread of the voltage samples at the relay location subject to satisfying two inequality constraints on each voltage sample. Effects of sampling frequency, data window size, and the degree of fuzziness on the estimated parameters are investigated. The performance of the proposed technique is illustrated using simulated data.