Wavelet Transform and Artificial Intelligence for Unbalanced Current Protection of 230kV Capacitor Switching Transient Inrush Current

Abstract

High transient inrush currents from back-to-back capacitor switching can cause the maloperation of protective relays, especially the instantaneous overcurrent (50) and unbalanced current protection relays (60C). This paper presents a combination of discrete wavelet transforms and artificial intelligence as an efficient technique to analyze the inrush current switching. The discrete wavelet transform is used to detect and classify either isolated or back-to-back capacitor switching transient signals. After that, the output from wavelet coefficients acts as the artificial intelligence input for discriminating the 6-difference cases of transient inrush current mitigation methods by using the combination of discrete wavelet transform and fuzzy inference system and discrete wavelet transform and probabilistic neural network. The proposed technique of discrete wavelet transform for detection and classification shows enhanced performance accuracy of 100 %. The fuzzy inference system and probabilistic neural network that can discriminate the inrush current mitigation methods have a high accuracy of 90.57 % and 96.72 %, respectively.