Fault Location in Transmission Line Using Travelling Wave Detection Method in PSCAD

Transcription

1 Fault ocation in Transmission ine Using Travelling Wave Detection Method in PSCAD Karansinh M. Parmar 1, Dr. Rashesh P. Mehta 1 PG Student, Associate Professor, Department of Electrical Engineering Birla Vishwakarma Mahavidyalaya, Vallabh Vidhyanagar, ANAND, Gujarat k4karansinh@gmail.com 1, rpmehta@bvmengineering.ac.in Abstract: Exact and accurate finding fault location for a transmission line may decrease the search time for find out the fault to the recovery of the system. A modification of algorithm for transient detection base on Park s transmission is proposed. The transient detection technique is used more than one sample of voltage and current to make possible required transient analysis. This paper presents self-adapts to electrical noise, simple phase imbalances, and is appropriate to find the location of a fault on the transmission line. This proposed technique is evaluated through PSCAD program simulations. A 30-kV, 400km long transmission line system with various case studies carried out. From this power system voltage and current waveforms are picked up one sample at a time. The results reveal that the Park transformation (Tdq0) can detect accurately the transient occurrence. From this make support to different protective types of equipment based on the detection of the traveling wave. Keywords: Fault ocation, Transmission ine, Park Transformation, Traveling Wave. 08 I. INTRODUCTION The exact fault location of a transmission line is a problem that has been studied for decades. The increasing amount of power grids and the need to quickly recover the system after clearing faults boosted research regarding fault location method. The theory of traveling waves is used for finding the fault of a transmission line. Impedance-based algorithms have been used around since the early stages of power systems. [1] Such methods are based on the transient detection method as specified according to the number of monitored points. Several methods for transient detection have been proposed in order to enable an accurate and reliable fault location for a transmission line [3]. Methods based on traveling wave s theory utilize voltage and current data captured by Digital Fault Recorder (DFR) installed in the monitored transmission line terminals end [1]. Recently, the transient detection procedure was used in various protection applications as in methods for high impedance fault (HIF) detection in power quality distribution detection, usually in both assessing relay and CB performance, in the post analysis of fault records, [4] in fault location methods based on the theory of travelling wave and the triggering of some other protective functions. In [1] and in [5] is proposed a very simple method for fault location in a transmission line based on the transient fault location in transmission line based on the transient detection through the applying of the Park s transformation (Tdq0). such methods permit the monitoring of the three-phase signals simultaneously by the analysis of only one waveform- the direct axis signal which is computed using only one sample of each phase. The accuracy of traveling wave methods is a function of the sampling frequency, depends on the hardware used for data acquisition. In fact, it is referring to reliability than accuracy and errors as a function of the sampling frequency [1]. It is important to know that the transient detection is a critical step for fault location procedure. Thus the reliability of the fault point estimation by any algorithms depends on the reliability of the method used to estimate the initial transient instant. In this paper, a very simple method for transient detection is proposed. Conventional methods usually need more than one sample of voltage or current waveforms to perform the required analysis. Here, the window has only one sample, i.e., no need the previous voltage or current samples. All three phases are monitored by the direct axis voltage signals V d obtained through Park s transformation (Tdq0) and the method automatically determines the fault location point instant after the occurrence of the fault. II. TRANSIENT DETECTION USING PARK TRANSFORMATION The method for transient detection was presented by R. H. Park in the United States in 199. Tdq0 has been used in the whole area of electrical engineering, especially on research regarding salient pole synchronous machine where variable inductances in a static reference frame rotating at synchronous speed ω [figure 1 (a)]. For this end, a couple of rotating references are produced; the direct axis and quadrature axis. Here, the rotating reference system is used to remove power frequency signals from voltage waveforms allowing transient time detection. In the transient detection procedure, the Tdq0 is used to generate a reference frame which rotates in synchronism with the

2 three-phase voltage phasors at the power frequency. Consequently, the power frequency signal is removed from the analyzed waveforms and, thus, fault-induced transients are identified. In fact, for an observer on the rotating reference frame, variations near to zero are detected for the power system operating under normal conditions but, if a disturbance occurs, large variations are verified. Here, the rotating reference system is used to remove power frequency signals from voltage waveforms allowing transient time detection. An analogy between the Tdq0 s application in electrical machines research and in transient detection is shown in Fig (a) (b) Fig. 1. Park Transformation (Tdq0) application: (a) Electrical Machine research, (b) Transient detection technique For an observer on the synchronous reference frame, the steady-state portion of the signal has negligible values (i.e., TDQ blindfolds-rated frequency signals). Nevertheless, when a disturbance occurs, large waveform variations show up, enabling the faultinduced transient identification. It is important to point out the voltage or current imbalances due to faults detected by TDQ. This method is different from conventional high-speed fault detection algorithms. Where the beginning of the disturbance is taken only through the identification of the signal high-frequency components of fault. These features make TDQ very agreeable to be used in fault detection procedures. Also, the proposed technique can use voltage or current as individual input samples, or both quantities sampled simultaneously. In the latter case, the first transient time detection is assumed to be the moment in which the first voltage or current wave reaches the monitored bus. Such signals are called direct and quadrature axes components which will be represented from now on as V d and V. Both components may be used for transient detection but, here, only V q will be considered. For high impedance fault cases, V d coefficients present high attenuation. Thus, to increase the sensitivity of the proposed algorithm, difference coefficients (c) are calculated using Taylor s approximation: C dif = Vd( i) Vd( i 1) t Where V d is the direct axis component; i is the sample number; Δt is the time step (1) Expressions to compute the direct axis and quadrature axis components, which will be represented from now on as A d and A q, respectively, are shown in equation () and (3). Aa Ad Pdq * Ab Aq A c () Or A dq = P dq * A abc (3) Being P dq = cos( ) cos cos 3 sin( ) sin sin Where k is the kth signal sample, A abc is the monitored three-phase voltage V or current I, A dq is the directquadrature reference frame components (V d and V q for voltage or I d and I q for current), kwt, is the angular power frequency, Δt is the sampling interval used by ADCs, is the angle of A d, v, is the angle of A a (phase A monitored signal) Δ = - v and (angle between A d and A a ). III. PRINCIPES OF TRAVEING WAVE FAUT OCATION METHOD (4) Some drawbacks of impedance-based fault-location methods have encouraged research on traveling wave fault location (TWF) techniques [7]. As shown in [6], the fault distance given by TWF algorithms is a multiple of the sampling time interval Δt, in such a way

3 that the smaller Δt is, the greater the fault locator accuracy. The existing commercial digital relays normally work with sampling frequencies that range from 4 to 64 samples/cycle (40 to 3840 Hz) for 60-Hz power systems. These rates are too small to give good results using TWF methods and, so, DFRs are preferred to run the transient detection procedure. Although DFRs exist with a sampling frequency of 5 MHz [8], a typical DFR sampling frequency is 56 samples/cycle (15360 Hz), which is sufficient to provide a reasonable time resolution for TWF methods. Where d and d est are, respectively, the actual and the estimated fault location; is the line length; V propag and is the traveling wave propagation velocity. Here, V propag is taken as 98% of the speed of light, a value suggested by Zimath et al. [8], who successfully located faults in the field. IV. A. Simulation Model: SIMUATION STUDIES PSCAD simulation of the 30 kv system presented in the figure was performed to evaluate the Travelling wave fault location method. Fig.. Time-Space Diagram for a Transmission ine Monitored at Two Terminals To understand the basic principles of one- and twoterminal TWF methods, a time-space diagram, as depicted in the figure, is generally used. According to [7], despite the necessity of a global positioning system (GPS) to synchronize data from remote line ends, twoterminal methods are more reliable because they need to detect only the first incident traveling waves (i.e., instants t 11 and t 1 ) (figure). In fact, techniques that use measurements from only one terminal require the detection of wave s reflected at the fault point f. For instance, if Bus 1 is the monitored terminal, the fault point location d is determined by detecting, besides the instant t 11, the instant t 1, in which waves reflected at f reach the monitored line end. It is difficult to separate such waves from the refracted waves that reach Bus 1 at the instant t 1r (Fig. ). If reflected and refracted traveling waves are too close, the one-terminal method may present unacceptable errors. In this paper, a two-terminal method is implemented, allowing the evaluation of the proposed transient detection technique as if it was embedded in an actual traveling-wave-based fault-location system. The fault distance from Bus 1 is given by d d est ( t1 t11)* v 10 propag ( t t )* v, if 1 11 propag est, if d (5) d (6) Fig. 30Kv Power System Model The system parameters are as follows; transmission line length is 400 km, Base kv is 30 kv of both the buses 1 &, initial state of breaker 1 if OFF and the initial state of the breaker in ON, the time of the fault is 0. sec and duration of fault is 0.0 secs. It is important to point out that voltage and current instrument transformers have limited bandwidth so that they are used here in order to provide a more realistic evaluation of the proposed techniques. The three algorithm steps described were implemented using the MODES language to emulate the traveling wave detectors at buses 1 and (figure ). The algorithm was run for sampling frequencies of 0 khz and 50 khz. Third-order antialiasing filters with cut off frequencies that were slightly smaller than were implemented as well. Here, in accordance with the Nyquist sampling theorem, 8 khz for 0 khz, and khz for 50 khz were used. Case Studies Testing Results: Extensive fault simulation was carried out with following assumptions: Fault Distance: 5 km to 375 km from bus 1 with steps of 5 km. Fault Resistance: 1 ohm, 10 ohms, 50 ohms, 100ohm. Fault Type: AG, BC, ABC, BCG, and ABCG Sampling Rates: to address the effect of the sampling time interval Δt on the fault detection accuracy, sampling frequencies f s equal to 0kHz and 50kHz were used. Monitored Signals: For each simulated fault cases, the transient detection was performed picking up both

4 signals of voltage and current, individually and simultaneously. Fig. Waveforms of Voltage, Current, Fault, and Trip Signal for AG Fault Figure: ABC to Dq0 Transformation Waveform for AG Fault To calculate the performance of the traveling wave fault detection technique, the fault location errors 11 d dest r (%) *100 Where is the faulted line length, d is the actual fault location, and d est is the fault point computed using the TDQ- based traveling wave fault location method. Although the proposed algorithm may be built into fault-location devices which work with voltage and/or current samples, it is shown that the disturbance detection is more reliable when three-phase voltage and current waveform signals are analysed simultaneously, picking up the first transient time detection as the arrival time of the incident traveling waves at the monitored line ends. The higher the smaller the t and, consequently, the greater the accuracy of the traveling-wave detectors. V. CONCUSION In this paper, an approach for detection of fault location using traveling wave fault detection technique based on Park s transformation is presented. Only the currentvoltage samples are used to detect transients and all three phases are monitored by the direct axis voltage signal Vd obtained through Park s transformation (Tdq0). The simulation carried out for 30kV; 400 km long transmission line using this approach clearly shows that the fault location can be determined very accurately within a shorter period of time. The advantages of this techniques are effective, robust, simple, very reliable and quite appropriate for multiterminal TWF methods; including real-time fault induced transient detection procedure. VI. REFERENCES [1] F. V. opes, D. Fernandes, Jr., and W.. A. Neves, Fault location on transmission lines based on traveling waves, presented at the Int. Conf. Power Syst. Transients, Delft, the Netherlands, Jun [] F. V. opes, D. Fernandes, Jr., and W.. A. Neves, Transients detection in EHV transmission lines using Park s transformation, in Proc. IEEE Power Eng. Soc. Transm. Distrib. Conf. Expo., May 01, pp.1 6. [3] H. W. Dommel, J. Michels, High-speed Relaying Using Travelling Wave Transient Analysis. IEEE Conference, Paper No. A78, pp , January/February [4] M. Gilany, DK. Ibrahim, E. S. T. Eldin, "Traveling-Wave-Based Fault-ocation Scheme for Multiend-Aged Underground Cable System, "IEEE Transactions on Power Delivery, vol., pp. 8-89, January 007. [5] F. V. opes, D. Fernandes Jr, W.. A. Neves, A New Approach for Fault ocation in Transmission ines, (in Portuguese), 010 IEEE/PES Transmission and Distribution

5 Conference and Exposition: atin America, T&D-A 010, São Paulo, Brazil, November 8-10, 010. [6] A. O. Ibe and B. J. Cory, A traveling wavebased fault locator for two and three-terminal networks, IEEE Trans. Power Del., vol. PWRD-1, no., pp , Apr [7] P. Gale, P. Crossley, X. Bingyin, G. Yaozhong, B. Cory, and J. Barker, Fault location based on traveling waves, in Proc. 5th Int. Conf. Develop. Power Syst. Protect., 1993, pp [9] S.. Zimath, M. A. F. Ramos, and J. E. S. Filho, Comparison of impedance and traveling wave fault location using real faults, in Proc. IEEE Power Eng. Soc. Transm. Distrib. Conf. Expo., Apr. 010, pp [10] Sung-Duck Kim, An Improved Method for Fault ocation based on Traveling Wave and Wavelet Transform in Overhead Transmission ines, Journal of the Korean Institute of Illuminating and Electrical Installation Engineers (01) [11] Sergio uiz Zimath, Marco Antonio Ramos, Rio de Janeiro, Joaquim Moutinho Beck, Nei Mueller, Traveling Wave-Based Fault ocation Experiences, rd Annual Conference for Protective Relay Engineers [1] Y.-C. Kang1 T.-Y. Zheng1 Y.-H. Kim1 B.-E. ee1 S.-H. So P.A. Crossley3 1 The Department of Electrical Engineering, Chonbuk National University, Chonju , Korea Advanced Material Research Team, HANKOOK TIRE CO., TD., Daejeon , Korea 3 The School of Electrical and Electronic Engineering, University of Manchester, Manchester, UK, Development of a compensation algorithm for a measurement current transformer, Published in IET Generation, Transmission & Distribution Received on 10th March 010 Revised on 3rd July 010 doi: /iet-gtd [13] C. A. Silva, D. Fernandes Jr., Member, IEEE, and W.. A. Neves, Member, IEEE, Correction of the Secondary Voltage of Coupling Capacitor Voltage Transformers in Real Time, [14] Rucha V. Deshmukh, A.Prof. M. R. Salodkar, Combine Study of Transmission ine Fault Detection Techniques, International Journal of Advanced Research in Computer Engineering & Technology (IJARCET) Volume 3, Issue, February 014 [15] Iman Sadinezhad, and Vassilios G. Agelidis, An Adaptive Precise One-End Power Transmission ine Fault ocating Algorithm 1 Based on Multilayer Complex Adaptive Artificial Neural Networks, 009.ICIT 009.IEEE International Conference on Industrial Technology [16] S. A. Soliman, M. Belkhayat. Power Systems Fault Type Identification Based on Park's Transformation Algorithm, 006, arge Engineering Systems Conference on Power Engineering