RVMs Based Method for Fault Analysis in Advanced Series Compensated Line

RVMs Based Method for Fault Analysis in Advanced Series Compensated Line

Abstract—This paper suggests an efficient method for fault classification and section identification (where the faults occur before or after of a compensation device) in advanced series compensated line with Thyristor Control Series Capacitor (TCSC). Different compensation level and non-linearity of protective equipment of TCSC appurtenance are the main reasons of fault analysis complexity. Relevance Vector Machines (RVMs) as a powerful classifiers have been applied for fault analysis purpose. Fault signals, which are related to the comprehensive combination of different parameters of a sample network, are analyzed by the RVMs. The results and performance of RVMs are compared with Support Vector Machines (SVMs). RVMs with lower Relevance Vectors (RVs) have a faster and more accurate performance in comparison to SVMs. The effectiveness of RVM classifiers is tested on middle compensated transmission line with TCSC, which is simulated by using PSCAD/EMTDC software.