Detection and Classification of Voltage Disturbances in Electrical Power Systems Based on Multiresolution Analysis and Negative Selection Algorithm

Abstract

Early detection of threats to electrical energy distribution systems helps professionals make decisions and mitigate interruptions in supply and improper activation of the protection system. Biologically inspired methods, e.g., artificial neural networks, genetic algorithms, and ant colonies, solve optimization problems and facilitate pattern recognition and decision-making. The present work presents a tool for detecting and classifying voltage disturbances based on the negative selection algorithm, which identifies and eliminates self-reactive cells, associated with multiresolution analysis, which analyzes the signal at different scales of detail, allowing a more complete understanding and detailed description of the phenomenon in question. The negative wavelet selection algorithm demonstrates robustness to detect and classify disturbances.