Parameter Estimation for an Electric Power System Using Genetic Algorithms

Real waveforms in power systems are characterized as a fundamental sinusoid waveform combined with undesired transient harmonics. This work presents a method based on Genetic Algorithm ‑ GAs, which is an optimization technique inspired by genetics and natural evolution, to estimate the parameters in waveforms, through two different approaches. The first one consists in the estimation of the harmonic components present in distorted waveforms from electrical systems. The second one consists in the estimation of the amplitude, frequency and phase angles of voltage and current waveforms, related to an application of a frequency relay. A specially designed intelligent algorithm for optimization problems, GOOAL - Genetic Object Oriented Algorithm, was successfully implemented and tested. The proposed algorithms for both situations were tested with simulated data as well as real data from CHESF ‑ Companhia Hidroelétrica do São Francisco. The results observed have characterized the efficiency of this computational tool for the estimation of the desired parameters.