Energy Education Science and Technology Part A: Energy Science and Research 2012 Volume (issues) 30(1): 75-88 ST and LSSVR-based the fault location algorithm for the series compensated power transmission lines Murat Uyar * Siirt University, Department of Electrical and Electronics Engineering, 56100 Siirt, Turkey Received: 06 September 2011; accepted: 18 November 2011 Abstract This paper presents an approach based on S-Transform (ST) and Least Square Support Vector Regression (LSSVR) techniques for predicting the fault location in a compensated power transmission line with the fixed series capacitor. The entropy features of ST matrix are extracted for reducing the dimension of three-phase current signal measured from the sending end of the transmission line. Then, the extracted features are applied as input to LSSVR for determining fault location on series compensated line (SCL). The presented method has been tested using model of a 400kV, 320km transmission line, which is compensated, by a three-phase capacitor bank in the middle. The results show that the proposed method is capable of determining fault location on SCL under wide variations in operating conditions (i.e. fault resistance, fault inception angle, fault distance, percentage compensation level, source impedance and load angle). Keywords: Fault location; Series compensated lines; S-transform; Least square support vector regression ©Sila Science. All Rights Reserved. 1. Introduction The series compensation increases the power transfer capacity of the transmission lines, voltage regulation and power system stability. However, it is also a possible source of error which causes undesirable operation of the transmission line protection system. As seen at the relaying point, traditional distance protection relays utilized in transmission lines are based on the measurement of impedance. However, in the case of series compensated transmission lines due to the presence of metal oxide varistors (MOV), this impedance measure is not a dependable parameter to predict the fault distance [1]. Because the series capacitor with the MOV introduces nonlinearity in the circuit and causes additional transients in the system, distance relays connected for protecting power system will tend to incorrect operation. Therefore, to improve reliability and sustainability of the total power system, the fault location must be identified accurately for simplifying fast repair and maintenance of the faulty line [2]. In literature, several algorithms have been proposed and applied for determining the location of faults taken place on series compensated transmission lines. They can be divided into three main categories, namely; phasor based, traveling-wave based, and intelligent methods based. In algorithms based on phasor, terminal voltage and/or current phasors are considered as input. The method covers one terminal algorithm, two or multi [3-5]. For phasor-based algorithms, the acquisition of high-accuracy phasor estimates needs to be obtained at least one cycle of data. Therefore, the algorithms in this category are incapable of coping with high-speed applications. Represented by references [6, 7], the traveling-wave based algorithms use the return time of the reflected waves traveling from the fault point to the line terminal as a measure of distance to the fault. Shortly, in these classical algorithms, the equivalent model of the SCs and MOVs are used for the calculation of the voltage drop across the compensation device [1]. This will inevitably lead to errors. _______________ * Corresponding author. Tel.: +90-484-223-1224; fax: +90-484-223-6631. E-mail address: muratuyar1@gmail.com (M. Uyar).