Classificação de distúrbios de energia elétrica baseada em sistemas imunológicos artificiais

Abstract

Nowadays, electricity assumes an essential role in the sustainability of modern society. The requirement and consumer demand for power quality are growing along with the advancement of technology and the increasing use of non-linear loads. This paper presents an application of artificial immune systems, focusing on the clonal selection algorithm, for power quality disturbances classification. The algorithm uses an initial population of antibodies to generate high affinity memory cells capable of recognizing antigenic electrical disturbances during each half cycle of the fundamental frequency voltage signal. The results demonstrate the algorithm's ability to classify disturbances such as sag, swell, outage and harmonics, with 100% efficiency rating. Another important feature of this approach is that it can be embedded, since the online stage classification has a low computational complexity with processing time around 103 μs. Based on comparative study with other studies, the results showed up best.