Power System Protection Using Machine Learning Technique

Abstract

This chapter presents a new approach for distance relaying of transmission line using machine intelligence technique such as Support Vector Machine (SVM). SVM is a relatively new computational learning method based on the statistical learning theory. The proposed technique is used for developing protection schemes for Thyristor Controlled Series Compensated (TCSC) Line using post fault current samples for half cycle (10 samples) from the inception of the fault and firing angle as inputs to the SVM. Three SVMs are trained to provide fault classification, ground detection and section identification respectively for the TCSC line. Also SVM is used for faulty phase selection and ground detection in large power transmission system without TCSC. The method uses post fault current and voltage samples for 1/4th cycle (5 samples) as inputs to SVM-1 to result faulty phase selection. SVM-2 is trained and tested with zero sequence components of fundamental, 3rd and 5th harmonic components of the post fault current signal to result the involvement of ground in the fault process. The polynomial and Gaussian kernel based SVMs are designed to provide most optimized boundary for classification. The classification test results from SVMs are accurate for simulation model as well as experimental set-up, and thus provides fast and robust protection scheme for distance relaying in transmission line.