Portal de Periódicos da CAPES

Preventing maloperation of distance protection due to CCVT transients

2019; Institution of Engineering and Technology; Volume: 13; Issue: 13 Linguagem: Inglês

10.1049/iet-gtd.2018.6559

1751-8695

Seyed‐Alireza Ahmadi, Majid Sanaye‐Pasand, Mahdi Davarpanah,

Islanding Detection in Power Systems

IET Generation, Transmission & DistributionVolume 13, Issue 13 p. 2828-2835 Research ArticleFree Access Preventing maloperation of distance protection due to CCVT transients Seyed-Alireza Ahmadi, Seyed-Alireza Ahmadi School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, IranSearch for more papers by this authorMajid Sanaye-Pasand, Corresponding Author Majid Sanaye-Pasand msanaye@ut.ac.ir School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, IranSearch for more papers by this authorMahdi Davarpanah, Mahdi Davarpanah School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, IranSearch for more papers by this author Seyed-Alireza Ahmadi, Seyed-Alireza Ahmadi School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, IranSearch for more papers by this authorMajid Sanaye-Pasand, Corresponding Author Majid Sanaye-Pasand msanaye@ut.ac.ir School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, IranSearch for more papers by this authorMahdi Davarpanah, Mahdi Davarpanah School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, IranSearch for more papers by this author First published: 05 June 2019 https://doi.org/10.1049/iet-gtd.2018.6559Citations: 3AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Unsecured operation of backup distance relay during an external fault results in removing a needed energy link and can jeopardise overall system stability. This problem that is well known as distance relay overreach is very likely during coupling capacitor voltage transformer (CCVT) transients mainly under high system impedance ratio (SIR) conditions. To tackle this challenging problem, a new methodology consisting of three stages is proposed in this study. It is capable of solving the overreaching problem which occurs under high SIR conditions, without sacrificing the operation speed of distance relay for inside the first zone faults. Comprehensive simulation studies are carried out to test the proposed methodology. The obtained results demonstrate its feasibility and effectiveness for practical applications in the power system. 1 Introduction Coupling capacitive voltage transformer (CCVT) is generally used to create the precise copy of voltage signal from high voltage and extra high voltage power systems [1, 2], which is used in measurement and protection instruments. The distance relay, as a widely used transmission line protection device, needs voltage signals to identify the faulty point of the line accurately and make a proper decision quickly. For such an application, the accuracy of voltage measurement during the line short-circuit fault is of the most important requirements. In case of a fault incident, the CCVT output voltage is usually accompanied by some transients due to the sudden primary voltage drop. This phenomenon is the so-called CCVT subsidence voltage or CCVT transient response [3]. The CCVT output voltage under such transient period depends on its components type and the corresponding parameters on one side, and the system impedance ratio (SIR) and fault instant on the other side [4–8]. Under high SIR conditions, for considerable duration of time, these transients cause error in the phasor estimation procedure and result in decrement of the voltage magnitude and shift in the voltage phase angle [9]. These errors adversely influence distance relays and may lead to the relay overreaching problem [10]. This becomes worse particularly in the presence of a fault initiated at the zero voltage crossing instant [11]. To tackle this problem, lots of efforts have been done in the literature. Some of the suggested approaches are as follows: Utilising a band-pass filter designed to pass the nominal frequency of the power system in case of detecting high SIR condition is implemented in some commercial relays [12–14]. However, the filter should be chosen very narrow to properly resolve the overreaching problem, which is not practically possible due to the extra delay imposed on the relay operation [10]. Reducing the first zone of distance relay is another approach [15], used by some relay manufactures [16], which operates efficiently only under low SIR conditions. Besides, reducing the first zone reach may jeopardise the relay dependability for inside the first zone faults, mainly for the near end of zone faults. Applying a time delay for the first zone [10], after estimating the transient error of CCVT reported in [17], which may endanger coordination of distance relays. Employing least error squares (LES)-based approaches to more accurately calculate the voltage phasor value from the CCVT transient response is utilised in [7, 11, 18]. In these approaches, a comprehensive CCVT model is used and the characteristic of the transient response is extracted. Using these methods, the signal model of LES method has been modified to fully contain the transient signal components. Consequently, the phasor estimation is improved and the overreaching problem is reduced. Using the estimated magnitude of the decaying DC component is employed in [19] to detect whether the transient period of the CCVT is passed or not. This method behaves the same for all CCVT transients under different system and fault conditions. It is very suitable for high SIR conditions and effectively prevents the overreaching problem in these conditions. Each of the abovementioned approaches has its own merits which can be useful in some system conditions. However definitely, more efforts are required to find a more complete solution that could effectively solve the overreaching problem of distance relays caused by the CCVT transients, especially under high SIR conditions. To achieve this goal, a new methodology consisting of three stages is suggested in this manuscript. In the first stage, impedance trajectories approaching to the first zone of distance relay are detected. After that, in the second stage, high-risk data windows causing estimation error in the voltage phasor estimation procedure are intelligently detected using a proposed qualification index. In case of concluding a high-risk data window, in the third stage, some modification strategies are performed to prevent the maloperation of distance relay. Using the proposed methodology, the problem of distance relay overreaching under high SIR conditions is solved, without sacrificing the operation speed for inside the first zone faults mainly under low SIR conditions. Contrary to the previous methods which affect the relay operation for the whole CCVT transient period to prevent the overreaching problem, this method by using the proposed qualification index only affects the relay operation for high-risk data windows. Following is a summary of the paper sections. In Section 2, CCVT transient characteristic and its impact on the distance protection are discussed. Then in Section 3, the proposed methodology is presented and for testing it in Section 4 comprehensive studies are performed and the results are discussed. Finally, in Section 5, the paper is concluded. 2 CCVT transient characteristic and its impact on distance relay In Fig. 1a, the schematic of understudy CCVT model for transient studies of this paper is shown. This model consists of voltage dividing capacitors, compensating inductor, over-voltage protection device, which is a spark gap in series with a resistor, an intermediate voltage transformer, active and passive ferroresonance filters, which their models are expanded in Fig. 1b, stray capacitors, and a connected burden to the low voltage side [6, 19–23]. Fig. 1Open in figure viewerPowerPoint Schematic of the(a) Understudy CCVT model, (b) Ferroresonance suppression filters The active ferroresonance filter is designed for the CCVT system according to [6]. The hysteresis characteristics of the intermediate voltage transformer's magnetic core and the reactor of ferroresonance filters are modelled using the Tellinen method [24]. This method is capable of simulating symmetric and asymmetric minor hysteresis loops and remnant flux. In the Tellinen method, by knowing the ascending and descending branches of hysteresis major loop, the mathematical relationship between the flux density (b) and the field strength (h) is determined. Since measuring flux (λ) and hysteresis current (ih) are easier, they are used instead of b and h in the Tellinen model of this paper. At each time step of the simulation, according to the measured voltage of the magnetic core's primary side, the corresponding λ is calculated. Then, the values of ascending and descending curves of the hysteresis major loop (λ+, λ−) for the obtained hysteresis current (ih) of the previous time step are calculated. After that, according to the polarity of the magnetic core voltage, di/dt can be calculated as follows: if : (1)if : (2)where M0 is the slope of the major loop trajectory in the full saturated region, which is equivalent to the air core inductance. Finally, the new value of ih (ih new) is calculated as follows: (3)where Δt is the time step of simulation. It should be noted that the magnetic core is implemented through a controlled current source, in which its current is determined based on the obtained ih new from (3). In case of a fault occurrence, the primary voltage of CCVT is sharply reduced and the stored energies in the dividing capacitors and the compensating inductor start to discharge. As the mentioned energy storage elements cannot be discharged instantly, the secondary voltage of CCVT does not show the primary voltage replica accurately during the transient period. This phenomenon causes a transient damping voltage in the secondary side that can be oscillatory or unidirectional according to the transformer design, type of ferroresonance suppression filter, and connected burden on one side, and the SIR condition and fault incident angle from the other side [16, 19]. From the viewpoint of distance relay operation, CCVT transients for the faults initiated at the zero crossing instant of voltage signal are more destructive than the ones initiated at the peak instant [25, 26]. Simulations show that oscillations of the CCVT components are much higher when the fault initiates at the peak instant of voltage, but since the frequencies of theses oscillations are far from the antialiasing filter's cutoff frequency, the CCVT transients are filtered out and cannot result in maloperation of distance relay. Therefore, the major concern with CCVT transients is when the fault initiates at the zero crossing instant of voltage signal. In addition, the type of ferroresonance suppression filter, i.e. active or passive type, impacts the characteristic of CCVT's transient response. The reported studies reveal that the transient response of a CCVT equipped with passive filter, cannot noticeably affect the distance relay's performance, while this transient response for a one equipped with active filter is of major concern [19]. Therefore, this paper concentrates on the CCVTs equipped with active ferroresonance filter. Fig. 2 shows a typical time response of CCVT transients after a fault incident. In this figure, based on the research conducted in [11], the time response is decomposed to its constitutive parts of decaying (i) high frequency, (ii) low frequency, and (iii) DC components. As can be observed, this transient voltage has a considerable magnitude and a long damping duration. Fig. 2Open in figure viewerPowerPoint Typical time response of a CCVT after a fault incident [11] The CCVT time response can be mathematically characterized using its transfer function. The model depicted in Fig. 1 by using an active ferroresonance filter, has a high order transfer function. By performing some valid simplifications, which are neglecting the poles associated to very high frequencies (related to stray capacitors), and very fast dc poles accompanied by their corresponding zeros, the model becomes a five order system with the following transfer function [11]: (4)In the above equation, there are a couple of decaying high frequency components, a couple of low frequency components, and one DC component. pH, pL, and pDC are high frequency, low frequency, and DC poles, respectively, which are as follows: (5) (6) (7)where τH, τL, and τDC are the time constants of high frequency, low frequency, and DC poles, respectively. Similarly, ωH and ωL are the angular frequencies of high and low frequencies poles, respectively. During normal operation of CCVT, and when the system oscillations are damped, the output voltage is in the form of a pure sinusoidal. Nevertheless, immediately after the fault incident, due to the system inherent nature, overall CCVT response contains some additional frequency components similar to the example shown in Fig. 2. Due to existing such components in the overall response, phasor estimation is accompanied by error. Many numerical relays extract the fundamental frequency component of the input signal through phasor estimation digital filters and use it to monitor the system condition [27]. For a case in which the voltage and current signals phasor estimations are accompanied by error, the final decision made by the distance relay is not accurate and can cause unnecessary tripping of the relay. The CCVT transient response after a fault incident causes maloperation of distance protection relays due to the false estimation of voltage signals in the phasor estimation procedure [19]. Previous experiences show that under high SIR conditions, these transients for some considerable duration of time cause decrement in the voltage magnitude and shift in the voltage phase angle [9, 19]. In these periods, the impedance value seen by the relay becomes less than the accurate impedance and causes the distance relay overreaching problem [3]. To avoid the incorrect operation of distance relay for the faults out of the protected line, i.e. the faults beyond 100% of it, the reaching limit of the first zone is set typically from 80 to 90% of the protected line. This consideration in the protection scheme proves the unacceptability of the overreaching problem. The CCVT transients result in increment of the voltage estimation error and increase the risk of unsecured tripping decision for the out of the first zone faults. The incorrect decision of distance relay for these faults loses the fundamental protection requirement of selectivity and has an extensive impact on the power system stability. 3 Proposed methodology Flow diagram of the proposed methodology is depicted in Fig. 3 and its three main stages are as follows: Detecting impedance trajectories approaching to the first zone of distance relay and consequently starting the checking procedure of the next stage. Checking of the input data windows and detecting high-risk ones using a proposed qualification index. Modifying the estimated voltage and current phasors in case of concluding a high-risk voltage data window. Fig. 3Open in figure viewerPowerPoint Flow diagram of the proposed methodology The abovementioned stages are completely explained below. Stage 1 : In this stage, by using an overreaching zone bigger than the first zone of distance relay (e.g. 1.1 times bigger), which is called auxiliary zone in the rest of the paper, impedance trajectories approaching to the first zone of distance relay are detected. When the impedances calculated from the input data windows of voltage and current signals lay inside the auxiliary zone, the probable overreaching problem is warned. Then to decide about the quality of the input voltage data window, and detect whether it contains high-risk transients or not, the next stage programme is called. Stage 2 : In this stage, voltage data windows containing unsafe CCVT transients are intelligently detected through recognising high-risk ones. In the voltage phasor estimation, a data window is called high-risk, when the probability of fundamental frequency component false estimation due to the CCVT transients is high. In this stage, high-risk data windows are identified using a qualification index proposed here. This index is obtained by fitting the input voltage data window with a complete signal model of the CCVT time response. Doing so, important information from the quality of input voltage data window becomes available. The complete signal model of the CCVT time response contains its transient characteristic formulated in (4) and is as follows: (8)where VF, VH, VL, and VDC are the fundamental frequency, high frequency, low frequency, and DC components of the voltage signal, respectively. These components form the time response of the CCVT transient period and are unknown. Full cycle LES method is used to estimate these valuable components, needed to calculate the qualification index, based on the measured voltage v(t). Other parameters, i.e. τH, τL, τDC, ωH and ωL are defined before and can be obtained by forming the transfer function in (4). After performing the LES estimation procedure, the components of input voltage signal for each data window are calculated as follows: (9) (10) (11) (12)where CF, CH, CL, and CDC are the amplitudes of fundamental frequency, high frequency, low frequency, and DC components of the voltage data window, respectively. When CCVT is operating under the normal condition in which there is no fault in the system, the fundamental frequency component, i.e. CF is much larger than the other components. Nevertheless, in case of a fault incident, and mainly during the transient period of CCVT, non-fundamental frequency components of CH, CL, and CDC become considerably large. In this case, the phasor estimation procedure would be accompanied by gross error. In the proposed methodology, a high-risk data window is detected when the ratio of fundamental frequency component to the non-fundamental frequency components of the voltage becomes less than a minimum threshold. To do so, the following qualification index is used: (13)After the entrance of calculated impedances into the auxiliary zone, Q will be calculated for the input voltage data window. Then, if it becomes less than the minimum acceptable value (Qmin), a high-risk data window condition is concluded and the next stage is called to perform data modifications. Otherwise, the raw data of signals fundamental frequency phasor estimations are kept unchanged to continue using them in the distance relay algorithm. Stage 3: In this stage, modification strategies are performed after detection of a high-risk data window to prevent the overreaching problem. These strategies are for eliminating the effect of corrupted phasor estimation of the voltage signal in the final impedance calculation. To do so, not only voltage amplitude and angle should be modified, but also the current angle should be changed to prevent wrong calculation of impedance angle. Modification strategies are as follows: Estimated magnitude and angle of the voltage should be kept to the previous safe condition value which has an acceptable Q. Estimated magnitude of the current should not be changed, but the estimated angle of the current should be kept to the previous safe condition value to keep the resultant impedance angle the same as the previous safe condition value. The above modifications are performed to prevent the overreaching problem resulted from the false phasor estimation of high-risk data windows. Since most of the data windows in the transient period are not high-risk, the methodology does not decrease the relay's speed for inside the first zone faults. For low SIR conditions, as the non-fundamental frequency components are not high, nearly all of the data windows are qualified in the second stage procedure. Besides, for the probable few high-risk data windows, when a severe fault inside the first zone occurs, the amplitude of current signal becomes large enough to drag the impedance into the first zone, even with the previous voltage magnitude of the safe data window. Therefore, the operation speed of distance relay for inside the first zone faults under low SIR conditions is not increased. For high SIR conditions, modification strategies are not applied during the whole transient period and only high-risk data windows causing the overreaching problem are affected. Therefore, these modifications do not have a considerable effect on the relay's operation time for inside the first zone faults and regular distance relay operation in these conditions becomes possible. However, the overreaching problem is effectively solved through the proposed methodology. 4 Simulation results and discussion Extensive simulation studies are performed in this section to evaluate and analyse the performance of the proposed methodology. Fig. 4 shows the understudy test system [19]. The CCVT model is shown in Fig. 1 and its specifications are available in [6]. In order to model the hysteresis phenomenon using the Tellinen method, it is necessary to determine ascending and descending branches' curves for the main hysteresis loop. These curves should be selected accurately to facilitate the computation and prevent the creation of volatility or fluctuations resulting from numerical calculations in time domain simulations. In the major loop of hysteresis curve, with a good approximation, the ascending and descending curves can be considered similar to each other. In other words, if the equation is determined for one of them, then the next curve can be obtained by displacement along the horizontal axis (ih) with a certain offset. In this paper, the saturation curves of [6] are used for the intermediate voltage transformer's magnetic core and the reactor of ferroresonance filters. Fig. 4Open in figure viewerPowerPoint Understudy test system [19] To conduct the method proposed in the previous section, transfer function poles should be determined. Fig. 5 shows the pole-zero diagram of the CCVT transfer function. As it can be observed, there are a couple of decaying high frequency poles, a couple of low frequency poles, and one DC pole in this diagram. It should be noted that there are some other very fast, high frequency, and far DC components which are not observable in the current view of the pole-zero diagram and are neglected in the transfer function model due to their negligible impact on the transient response of the CCVT. Fig. 5Open in figure viewerPowerPoint Pole-zero diagram of the CCVT model Simulations are carried out in the Power system computer aided design (PSCAD)/Electromagnetic transients including direct current (EMTDC) environment and the output currents and CCVT voltages are used in MATLAB to perform the distance relay algorithm. Full cycle discrete Fourier transform is used for phasor estimation in all scenarios. Sampling frequency in the simulations is set to 2400 Hz. For antialiasing filter, second-order Butterworth low-pass filter with the cutoff frequency of 300 Hz is used. All of the simulations are conducted on a PC with Intel Core i5 2.50-GHz processor and 8.0-GB RAM. The programs are implemented in MATLAB R2013a. Fig. 6Open in figure viewerPowerPoint Time response of the CCVT equipped with(a) Active ferroresonance filter, (b) Passive ferroresonance filter As mentioned in the previous sections, CCVT transients for the faults occurring at the peak of the voltage signal contains very high-frequency oscillations, which are diminished after passing the antialiasing low-pass filter. Figs. 6a and b show the time responses of CCVTs equipped with active and passive ferroresonance suppression filters, respectively. As it can be observed, frequencies of the oscillatory components are very high for the voltage peak fault incident and are filtered by the antialiasing filter. Besides, the active ferroresonance filter damps the high-frequency oscillations of CCVT transients more effectively than the passive one. Existence of such high-frequency oscillations are also reflected in the CCVT transfer function poles. As mentioned in Section 2, there are some poles with very high frequencies in the CCVT transfer function which have very minor effects on the studies of the proposed method. From the conducted study it can be concluded that, even for the worst condition, voltage peak CCVT transients are not destructive and cannot affect the operation of distance relay. Maloperation of distance relay due to CCVT transients occurs when the fault is initiated at the zero voltage angle. Therefore, for the evaluation studies of the proposed method, the fault incident angle is considered to be zero. Studies are performed under the worst fault condition, but for various SIR values and fault locations. As fault resistance helps to mitigate the CCVT impact on the distance relay operation, zero fault resistance (Rf = 0) is considered in the simulation studies. Reach limit for the distance relay's first and second zones are set to 80% of line 1 (L1) and 50% of Line 2 (L2), respectively. 4.1 Faults out of the first zone To evaluate the performance of the proposed methodology in detecting high-risk data windows and preventing the overreaching problem, a single phase to ground fault at the beginning of the second line (0% L2) depicted in Fig. 4, is applied at t = 150 ms. This location is in the second protection zone of the distance relay. The relay algorithm is expected to detect the fault and send the trip command ∼300 ms later. The study is carried out under both low- and high-SIR conditions (SIR = 0.5 and SIR = 20). 4.1.1 Scenario A.1 (low SIR condition) Time response of the CCVT output is depicted in Fig. 7a. In this figure, the output is smooth and is not full of high-risk transients. Fig. 7b illustrates the calculated Q for the data windows entering the phasor estimation block of distance relay during the simulation time with focusing on some parts of the CCVT transient period. The red line in this figure is the Qmin which is selected based on numerous simulations as 1.1 to work properly for different conditions. Fig. 7Open in figure viewerPowerPoint Scenario A.1(a) Time response of the CCVT output, (b) Values of Q for the data windows, (c) Trajectories of the calculated impedance As can be seen in Fig. 7b, the values of Q for this low SIR condition are acceptable for a very high percentage of the data windows during the CCVT transient period. This means that the proposed methodology for this low SIR condition accurately recognises most of the data windows during the CCVT transient period as the safe windows. Fig. 7c shows the trajectories of the calculated impedance for both using and not using the proposed methodology. As it was expected, in the low SIR condition, impedance trajectories and final results of both cases are very similar. 4.1.2 Scenario A.2 (high SIR condition) Time response of the CCVT output and Q values are depicted in Figs. 8a and b, respectively. In high SIR condition of this scenario, the output contains lots of high-risk transients. In Fig. 8b, it can be observed that the number of high-risk data windows is much more than those of the low SIR condition of the previous scenario. For a data window that its calculated impedance lays inside the auxiliary zone, if Q is not acceptable, modification strategies are performed to prevent the overreaching problem of the distance relay. Fig. 8Open in figure viewerPowerPoint Scenario A.2(a) Time response of the CCVT output, (b) Values of Q for the data windows, (c) Trajectories of the calculated impedance Generally, by comparing Figs. 5b and 6b of the conducted studies, the effectiveness of the proposed methodology in differentiating between high-risk data windows of high SIR conditions and safe data windows of low SIR conditions by means of detecting high-risk data windows is confirmed. Fig. 8c shows the trajectories of calculated impedance for scenario A.2. As can be seen, in case of not applying the proposed methodology, the calculated impedances for considerable numbers of data windows, enter inside the first zone and cause unwanted tripping of the distance relay. This is due to the underestimation of voltage for high-risk data windows. While, in the case of using the proposed methodology, high-risk data windows are detected and their estimated voltages and currents are modified to prevent the overreaching problem. 4.1.3 Scenario A.3 (comprehensive testing) To further assess the performance of the proposed methodology in detecting high-risk data windows causing overreaching problem, comprehensive test studies for various fault types and locations under the worst fault condition (zero fault incident angle and resistance) are conducted and the results are summarised and tabulated in Table 1. In this table, results of the proposed methodology in comparison with those of the method proposed in [19] and also with those of the case in which no preventive method is used are illustrated. Table 1. Comprehensive test studies for the out of first z