2014 IEEE International Conference on Intelligent Energy and Power Systems (IEPS) 110 Improved ANN Based Impedance Estimator for Phase to Ground Faults In UHV Transmission Line Syed Furqan Rafique School of Electrical Engineering North China Electric Power University Changping District, Beijing 102206, China Z.Y. Xu Beijing Sifang Automation Co. Ltd, Shangdi District, Beijing 100085, China Akif Zia Khan School of Electrical Engineering North China Electric Power University Changping District, Beijing 102206, China Abstract — An accurate digital distance relaying scheme integrated with artificial neural network impedance estimator is developed using the weighted sum of the negative-zero sequence components of fault current assuming the voltage and current are in phase at the fault location. The modified scheme is immune to the variation in fault location for under reach and overreach problem caused by sequence current components. A series of test conducted on a 800kV, 400km transmission line for single phase to ground faults as well as simultaneous open conductor fault in PSCAD/EMTP and Matlab. Finally comparison has been done with the conventional methods in order to check the accuracy and robustness of the proposed scheme which is found to be 95%. Keywords — Negative Zero sequence current; Artificial neural network; simultaneous fault I. INTRODUCTION Embedded Systems technology extensively improved the performance of conventional distance relaying , it is possible to realize useful functions like event recoding of pre-fault data ,adaptive logic , multiple sensing in power system to ensure the protection of EHV or UHV transmission lines[1],[2]. However the basic principles of the distance relay has not been changed, which is based on the assumption that the fault location voltage is zero to measure the fault loop impedance according to the ratio of voltage to current. In fact, it is not possible that fault location voltage is zero provided the artificial short circuit happen[3], and voltage of fault point will affect the measurement of fault impedance, especially for the phase-ground faults with high resistance and heavy load[4],[5]. The transmission lines are likely to experience the environmental conditions and the possibility of occurring faults on the transmission system is generally higher than that on other important components of power system. Single line to ground faults are the most common faults around 75-80 %, they may be set off by lightning strokes, cause of heavy wind may fall trees across the lines, foggy environment and salt spray on dirty insulators may be the reason for insulator flash over, and ice and snow loadings on the insulator may cause strings to fail mechanically[6]. When a fault occurs on transmission system, it is the most critical to detect it and to find its location in order to make necessary measures and to restore power as soon as possible. The time needed to determine the fault point along the line will affect the quality and the performance of system [7]. The fundamental relation of protective system is to make sure the power availability by clearing faults as rapidly as possible in order to maintain stability and control. However still relays have under reach and over reach problems depends on inaccurate fault impedance calculation due to high path resistance and simultaneous open conductor fault .the important factors that creates an inaccuracy in the system like remote in-fed, ground faults with different fault resistances, inaccuracy of the measurements in Voltage and current at relay point. The other factors in reducing the performance of digital distance relaying are simultaneous open conductor with ground fault. This kind of fault frequently occurs on an overhead transmission system because of the breaking of one or more phase conductor at a point near to the transmission tower. This broken conductor from one side is connected to the tower with suspension insulator while the other side fallen to the ground or due to a failure of a line surge arrestor causing an explosion which resulted in the broken conductor coming into contact with earthed metalwork. Critical power systems disturbances are often take place during this type of simultaneous fault condition, which are the cause of in correct operation of the traditional digital distance relays Xu et al. [3] proposed a new approach of fault impedance calculation algorithm for phase to ground distance relaying. This method is primarily on the selection of three different combinations of sequence current components, namely, negative current component, zero current components, and comprehensive negative-zero sequence current components. However, in this algorithm, the method for selecting particular sequence current component has not clearly defined which is needed for the impedance calculation algorithm. As comprehensive negative zero components would also show inaccuracy based on the fault distance location and other system changes like simultaneous open conductor effect, this leave the scope of research in the aforementioned scheme. Artificial neural network technique is integrated with the above comprehensive negative zero algorithm in order to produce accurate results while calculating the faulted line impedance. Various methods have been proposed [8] describing applications of ANNs to fault detection and location in transmission system. In this paper, a doubly fed UHV 800kV and 400 km transmission line system is taken as a test object for simulations. A single-end fault estimator is proposed for on-line applications using ANN. A feed-forward neural network 978-1-4799-2266-6/14/$31.00 ©2014 IEEE