E TEP zy Accurate Fault Location Algorithm for Power Transmission Lines J. Sadeh, A. M. Ranjbar, N. Hadjsaid, R. Feuillet Abstract zyxwvutsrqp In this paper we propose a new fault location algorithm for power transmission lines based on one terminal voltage and current data. zyxwvuts A distributed time domain model of the line is used as a basis for algorithm develop- ment. The suggested technique only takes advantage ofpost-faultvoltage and current samples taken at one end zy of the line and does not require filtering of DC offsetand high-frequency components of the recorded signals, which are present during transient conditions. Another advantage of the proposed method is the application of a very narrow window of data i. e. less than 1/4 of a cycle. The paper also proposes zyxw two different algorithmsfor lossless and lossv line models. Computer simulations approved the accuracy of the proposed methods. 1 Introduction zyxwvutsr Rapid detection, location and clearance of faults are essential factors of satisfactory operation of power sup- ply networks. When a fault occurs on the transmission line of an electric power system, it is very important to find the exact location of the fault. This can result in re- ducing the time required for repairing the damage caused by the fault and consequently improving reliabil- ity and continuity of energy supply. Fault location of transmission lines has always been a very well-known subject, being studied fora long time. However, this problem has nowadays been of more im- portance, due to deregulation and required quality of supply. Overhead lines in power systems have always been subject to faults, which can zyxwvuts be either transient or permanent and can be caused by various reasons. zyxwvu In case of permanent faults, locating the fault point allows one to go to the exact location and to take appropriate actions to restore the power. In the case of transient faults - which are usually cleared by automatic reclosing systems - locating the fault permits identifying critical points in the system and taking the necessary preventa- tive maintenance steps. These critical points may be caused by several reasons as pollution, insufficient insu- lation level, bird interference, etc. Many techniqueshave been proposed and applied for locating the exact point of fault on transmission lines [ 1 - 141. The main differences between these various al- gorithms are due zyxwvutsr to different transmission line models (lumped [l -31 or distributed [4-61). and the data re- quired(ofoneend[1,4,7]ortwoendsoftheline [2,8,9]). Sant and Paithankar [ 101 proposed a fault location algorithm technique that uses fundamental frequency voltage and current measured at one of the line terminals. This technique assumed that the line is connected to a source at one end only. The estimate of fault locations is not accurate if the fault current is supplied from both line terminals and some fault resistance is present. Takagi et al. [7], Wiszniewski [3], Eriksson et al. [ 11 and Cook [2] proposed methods which use fault current distribution factors, pre-fault and post-fault current and post-fault voltage from one line terminal. impedances of equivalent sources connected to the line terminals are required. In practice, the system configuration changes from time to time modifying distribution factors. Richards and Tan [I I] present a dynamic parameter estimation algorithm for locating faults, based on locally available currents and voltages. The differential equations are based on a lumped parameter line model. For modelling of the rest of the system at both ends of the line, Thevenin equiva- lents are utilised. The fault location problem is treated as a parameter estimation problem of a dynamic system, in which the response of the physical system is compared with the one of a lumped parameter model. The model parameters (location and resistance of the fault) are var- ied until an adequate match is obtained with the physical system response. The travelling wave approach was also studied [4-6.12.131. This approach is based either on the travel time measurements using correlation technique [ 121or on the calculation of voltages and/or currents pro- files along the transmission line [4-6,131. The travelling wave techniques offer some advantages but the compu- tational complexity is increased. Time domain represen- tation of a transmission line model has also been consid- ered [ 141. The model is obtained using the Laplace and 2 transforms. Data samples are considered as being avail- able from one end only. The voltage at the other end is es- timated using pre-fault data. In calculating the distance to a fault point, using voltage and current signals, it is advantageous to use ter- minal data of both sides of the faulty transmission line. However, from the practical point of view, it is more ap- propriate that the algorithm uses only one terminal volt- age and current data, which leads to simplicity of the equipment used. This paper introduces a new fault location algo- rithm that needs one terminal post-fault data and uses distributed time-domain model of transmission line to achieve required accuracy. The idea of the introduced ETEP Vol. 10, No. 5. SeptemberiOctober 2000 313