An AI approach for fault distance estimation in series compensated transmission line

The main objective of this paper is to accurately estimate the fault location in a series compensated transmission line using integration of discrete wavelet transform, particle swarm optimization (PSO) and support vector machine based algorithm. Estimation of accurate fault location will lead to quicker restoration of the supply. Instantaneous values of faulty current, voltage and power signals are available at the relay location. Using 10-level Discrete Wavelet Transform (DWT), the available signals are decomposed and thereafter the statistical features are obtained from the decomposed signals. Forward feature selection algorithm is used to select the best feature set. Selected features are then applied to the support vector machine (SVM) for estimating the fault distance. PSO algorithm is used to select the best SVM parameter by global searching technique. Proposed fault locator has been trained and tested for different fault scenarios (fault resistance and phase difference). The test results demonstrate that the adopted technique is a reliable method for estimating fault locations accurately.