Fault location methods

Fault location methods Abstract This paper presents a comparative study between two fault location methods in distribution network with Distributed Generation (DG). Both methods are based on computing the impedance using fundamental voltage and current signals. The first method uses one-end information and the second uses both ends. A 30 kV three-phase line was studied in the presence of a 3 MW fixed speed wind turbine. Index Terms-Fault location, Distribution network, Distributed Generation, Fixed Speed Wind Turbineà ¢Ã¢â€š ¬Ã‚ ¦ Introduction Fault location problem in transmission networks has been studied deeply because of its importance in the power system and because its difficult to physically check long transmission lines [1]. Nowadays, the problem of fault location was extended to distribution network in order to identify the fault location as quickly as possible to improve the power quality and the system reliability. The application of classical techniques, presented in Section 2, is not easy due to the complexity of the distribution systems which are characterized by the non-homogeneity of line, the load uncertainty, the phase unbalanceà ¢Ã¢â€š ¬Ã‚ ¦ [2] Fault location problem in the distribution network becomes more complicated with the presence of the Distributed Generation (DG). In fact, the DG resources, connected to the distribution system, which are in general, wind turbines and small hydro-electrical plants [3], contribute to the fault level of the network and their effect depends on their size, type and placement [4]. The infeed currents from the DGs cause errors in the estimation of the distance of the fault point since they can affect the amplitude, the direction and, indirectly, the duration of fault currents [3]. In this paper, we present two fault location methods which have been successfully applied to a real fault that occurred on a 225 kV transmission line in [5]. They will be, first, tested on a simple distribution line without DGs. Then, we consider a fixed speed wind turbine connected to the other side of the line. FAULT LOCATION TECHNIQUES Fault location methods can be classified into different categories: Methods based on travelling waves in faulted line: in [6], authors present a travelling wave based fault location method, which was successfully applied on a transmission line, and extended to distribution line with DGs. The main advantages of this approach are its insensitivity to the contribution of the DGs during fault and the requirement of fault signal only from the substation end of the faulted line. Methods based on harmonics analysis: those methods are not frequently used since grid operators have the aim to reduce harmonics in the power system. Method based on computing the short circuit power, using the voltage and the currents, to determine the fault location [5]. Method using instantaneous voltage and current available at both ends of the line [5]. Methods based on determining the apparent impedance using the fundamental components of voltage and current: this method is the most widely used because of its simplicity and efficiency, and it does not require a big investment in equipment [1], [2]. Those methods can be divided into two groups: Methods using one-end information and Methods using both ends of the affected line. Fig. 1 shows a simple three-phase distribution line with a load connected via a transformer (30 kV/ 575 V). The basic approach used for determining the fault distance d is to calculate the impedance seen from substation (NL) during the fault. This paper presents two fault location techniques. Method using one-end information Where: Vk is the voltage of the faulted phase and Vf is the fault voltage. Vk and Vf depends on the fault type as given in table I. Method using both ends information The voltage and the current of the two ends line are related with this expression: Where: VLi, VRi, ILi, IRi are resp. the voltage and the current of the left and the right side of the phase i. VLj, VRj, ILj, IRj are resp. the voltage and the current of the left and the right side of the phase j. Zik,Zjk: elements of the impedance matrix of the line L: line length d: fault distance Simulation and results In order to show the effectiveness of the presented algorithms, the system presented in Fig. 1 is tested with the source, the line and the load parameters given in table II. Where: dest and dreal are respectively the estimated and the real fault distance, and L is the line length. Fig. 2 shows the estimated error for the simulated system of Fig. 1 for different fault resistance value to compare the two techniques described above. Fig. 3 shows the fault location results for different load power. It can be seen that as the load power increases, the estimation error also increases. For the first technique, the estimation error is less than 1.5% and can still be acceptable, but for the second technique, the estimation error can reach 30% for a 5 MW load. In fact, the accuracy of the algorithm, for variable load, depends on the short-circuit power of the source. Fig. 4 shows that the estimation error decreases if the ratio between the load power and the source short-circuit power decreases. In order to study the influence of the integration of the DGs into the distribution networks on the fault location accuracy, a three-phase line integrating a 3MW fixed speed wind turbine at the right side of the line is considered in Fig. 5. A 1MW local load has been connected to the WT. Fig. 6, 7 and 8 present the WT characteristics: the nominal wind speed is 9 m/s; the wind speed is imposed equal to 8 m/s that makes the WT generating 0.66 pu of its active power. The reactive power is generated by an 800 kVAR capacitors. The WT speed is fixed to 1 pu. A one-phase fault, during 600 ms, that occurs on the line with different fault resistance value and different load power, is used to evaluate the presented methods. Fig. 9 presents the estimated error for the simulated line with wind turbine. It can be seen that the error is higher than in the first case because of the participation of the WT to the fault current which is not delivered only by the source. Then, the source voltage increases and the impedance seen from the source will be higher than the impedance of the same fault on the line without WT. Comparing with the results presented for a line without DGs, we can see that the contribution of the WT in the fault current increases widely the estimation error of both methods, especially the second one that uses the recorded information from the source bus and the WT connection point. The effect of the uncertainty of the load is investigated by varying its value from 0 to 5 MW, for a fault located at 20 km from the source. Fig. 10 presents the accuracy of the described methods while varying the load. Unlike the result presented for the line without GDs, the estimated error decreases while increasing the load impedance. This result shows that conventional methods cant be well used for network with DGs. It is known that an increase in generation capacity, increases the fault current, then the participation of the DGs to the fault level will increase too. For that, we consider two wind turbine of 3MW each one, connected to a distribution network at the same connection point. The wind reference of the first WT is fixed to 8m/s, and for the second, it starts with a wind speed of 8m/s then it increases to 9m/s to simulate the two wind sources with different rate of penetration. Fig. 12 shows the characteristics of both WTs. Conclusion This paper presents two impedance based fault location methods tested on a distribution line with and without distributed generation. The two techniques present an interesting precision for fault location in distribution system that does not integrate GDs. But, with the existence of the WT connected to the grid, those methods are not applicable especially for a high fault resistance value or variable load impedance. Thus, integration of the DGs into the distribution network requires further study on the existing fault location techniques to adapt them with the DGs state when a fault occurs. References J. Mora, J. Melendez, M. Vinyoles, J. Sà ¡nchez, M. Castro, An Overview to Fault Location Methods in Distribution System Based on Single End Measures of Voltage and Current, Journal Name, vol. 1, no. 3, pp. 1-10, Mar. 2000. Y.-J. Ahn, M.-S. Choi, S.-H. Kang and S.-J. Lee, An accurate fault location algorithm for double-circuit transmission systems, in Proc. IEEE Power Eng. Soc. Summer Meeting, vol. 3, 2000, pp. 1344-1349. TH. Boutsika, S. Papathanassiou, N. Drossos, Calculation of the Fault Level Contribution of Distributed Generation According To IEC Standard 60909, NTUA-Electric Power Division, Athens. V.R. Kanduri, Distributed Generation Impact on Fault Response of A Distrubution Network, Thesis of the Faculty of Mississippi State University, 2004. A. Abadlia,La Localisation des Dà ©fauts dans les Lignes Electriques, Thesis of the National School of Engineers of Tunis (ENIT), 2007. C.Y. Evrenosoglu, A. Abur, Fault Location in Distribution Systems with Distributed Generation, 15th PSCC, Liege, 22-26 August 2005, Session 10, Paper 5, p. 5.