See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/267324262

Power swing protection of series compensated transmission line with novel

fault detection technique

Conference Paper · March 2014

DOI: 10.1109/ICGCCEE.2014.6922359

CITATIONS

12READS

791

2 authors:

Some of the authors of this publication are also working on these related projects:

TRANSMISSION LINE PARAMETER INDEPENDENT FAULT LOCATOR DESIGN IN CRITICAL CONDITIONS OF NETWORK View project

Condition Monitoring of Transformer View project

Saptarshi Roy

National Institute of Technology, Warangal

26 PUBLICATIONS   42 CITATIONS   

SEE PROFILE

P. Suresh Babu

GATE Educational Institutions

48 PUBLICATIONS   69 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Saptarshi Roy on 27 December 2014.

The user has requested enhancement of the downloaded file.

POWER SWING PROTECTION OF SERIES COMPENSATED TRANSMISSION LINE WITH NOVEL FAULT DETECTION TECHNIQUE

Saptarshi Roy1, Dr. P. Suresh Babu2, 1 Student, M.Tech 2nd year,PSE, NITWarangal, Warangal-506004, A.P.India e-mail: saptarshi.roy.ju@gmail.com 2 Assistant Professor, EED, NITWarangal, Warangal-506004, A.P.India e-mail: drsureshperli@nitw.ac.in ABSTRACT-Fault detection during power swing condition is a challenge for stable operation of power system due to several reasons. This paper proposes a negative-sequence current–based technique for detecting presence of fault, classification of fault occurred, estimated zone and location of the fault occurred and the fault inception time with respect to system reference clock during the power swing condition in a series compensated line. The technique is tested for different series -compensated power systems including a SMIB system and a WSCC-9bus-3machine power system with their different configurations and contingency combinations. Power swing caused by various faults are simulated with PSCAD / EMTDC and MATLAB/SIMULINK and compared with available techniques to prove the effectiveness of the proposed technique algorithm. INDEX TERMS: fault detection, power swing, series compensation, negative sequence current, fault classification, fault location.

I. INTRODUCTION

Fast and accurate determination of a fault in electrical power system is a vital part in power restoration [20].Power Swing is caused by large disturbances in power system, which if not blocked caused mal-operation of distance relays and undesired tripping of circuit breaker, change in load impedance, thus induce unwanted relay operations at different locations of the network, major power-outages or even power blackout..If a fault occurs during the power swing, the relay must detect the fault and operate quickly. This paper proposes a technique for detecting faults in a series-compensated line during the power swing. During unbalanced faults, the negative-sequence components become significant and due o transients in current signals in the initial period, a negative sequence component is observed even for three-phase faults. To discriminate the faults during swing in a series-compensated line, a cumulative sum (CUSUM) of change in the magnitude of the negative-sequence current-based approach is proposed in this paper [16]. The performance of the algorithm is tested for numerous cases for an SMIB system and a 9-bus system simulated with SIMULINK/PSCAD. Further we detected the classification of the fault using fine tuning method and detected the location of the fault.

II. FAULT DETECTION CHALLENGES DURING POWER SWING IN SERIES-COMPENSATED LINES: Series compensation imposes protection problems .The use of series capacitors in transmission lines results voltage/current inversion, sub harmonic oscillations, transients, sub-synchronous oscillations [11].

Fig1: 400KV SMIB Test System.

A test system shown in Fig. 1 is considered [1]. Both Line-1 and Line-2 are 40% compensated and the capacitors are placed [3] .The system details are provided in Appendix A. The system with the distributed line model is simulated using SIMULINK/PSCAD. The traditional fault detection techniques, like the sample-to-sample or cycle-to-cycle comparison of current (or voltage) signals cannot be reliable during the power swing [2].

Fig2:Current Waveforms during Power Swing

Fig3: Voltage Waveforms during Power Swing III.PROPOSED FAULT DETECTION TECHNIQUE : The computation steps for the method are provided

-----(1) I2=negative sequence current; alpha=exp(j2*pi/3) and Ia IbIc are line currents

Sk=∆(I2k) = I2k -I2k-1 ---(2) For Sk > €, gk =max (gk-1+Sk-€,0) ---(3) Where, gk = fault index and € = the drift parameter in gk . A fault is registered if gk>h, ---(4) Where h is a constant and ideally equals zero. Epsilon provides the low-pass filtering effect and influences the performance of the detector. When Sk > epsilon, the gk value increases by a factor of difference between Sk and epsilon. After each fault detection index g is reset to zero. The selection of epsilon and h is important for determining the performance of the algorithm. Epsilon is a very small quantity which having a value greater than zero. The value of h is set such that the algorithm can maintain the balance between dependability versus security and speed versus accuracy requirements of the relaying scheme. Considering all extreme fault situations during the power swing we worked with hmax=0.48 and €max=0.05 value. The proposed method is based on the CUSUM approach and, therefore, a distinctly much higher index value is obtained during the fault.

IV. RESULTS The algorithm for is tested for different conditions including balanced and unbalanced faults etc. Using SIMULINK/PSCAD with distributed parameter line model data was generated. The data-sampling rate was maintained at 4 kHz for the 50-Hz power system. A .Line-to-Ground Fault in the Series-Compensated Line

Fig4 : Implementation of SMIB Test System Using PSCAD

Algorithm is tested for a line to ground fault of ag –type with a fault resistance of 0.1 ohm initiated at 0.34 sec for 0.04 sec duration at a distance 240KM from bus1.

a)g - t plot b) output-t plot

Fig5 :(a &b):LG fault Analysis

The fault is detected after 4.75 ms of fault initiation. At the time of fault inception the fault index g glows enormously high. Output shows 1 after inception of the fault and 0 before inception.

a)healthy Signal g -t plot b)LG Fault :I2 With Low R c) LG Fault :I2 with fault resistance100 ohm Fig :6(a,b,c)Performance Analysis of a healthy signal and a signal containing LG fault

A. Fault Detection During Single-Pole Tripping Initially phase-a of Line-1 is out of service following an ag- fault occurring during normal operation. It introduced swing into the system. A line-to-ground fault of bg-type with a fault resistance 100 Ohm is created at 0.2s at a distance of 160 km from the relay location toward bus N during swing. For this case, and the fault detection is possible after 0.6-ms fault inception.

a) g - t plot during single pole tripping

b) I2 - t during Single Pole tripping

Fig: 7: Performance Analysis during Single Pole tripping condition

B. Fault Detection During Double-Pole Tripping

a) g2 vs t plot b)Fig: I2 - t plot Fig: 8(a, b) Here fault detection is possible after 0.55 ms. V. CLASSIFICATION OF THE DETECTED FAULT: Faults are classified into two categories.

1) Ground Fault 2) Phase Fault. Classification or type of detected fault can be detected by tracking the absolute variation of fault index g and the variation of variation of fault index g (fine tuning of g).

a) ∆g-t for LG fault b) ∆g-t plot for LLLG

c) ∆g vs t plot for LL

d) ∆g-t plot: LLG fault e) ∆g-t plot : LLL

Fig: 9(a,b,c,d,e)Variation of g w.r.t time

From the variation of g itself, we can segregate an LG fault from a non-LG fault. The variation of g in case of a LG fault is very less with compared to other case. The

maximum variation of fault index g in case of LG fault 0.16 unit here, for other faults the value of gmax variation goes beyond 0.3 and above unit.

The trick of the detection is the nature of the ‘g’ variation - t curve. The variation of g gradually diminishes to zero towards the end of fault duration and this tendency of variation declining towards zero rapidly increases with respect to if any additional line or ground involved in the fault. See the similarities between the plot of g-variation in case of LL and LLL & LLG and LLLG .In case of LL and LLL there is a set of local maxima other than the global maxima variation (local maxima approx. 60% of global max.).This is missing in case of LLG or LLLG. So, by seeing the local maxima in the plot it is possible to segregate LL, LLL from LLG, LLLG fault.

IDENTIFIATION OF PHASE FAULTs:

The trick of the detection is the nature of the (variation of(Variation of g)) -t plot.

The variation of (Variatiion of) g gradually diminishes to zero towards the end of fault duration and this tendency of variation declining towards zero rapidly increases w.r.t if any additional Line or ground involved in the fault.

a) ∆(∆g) vs t plot for LL fault

b) ∆(∆g)-t plot in case of LLL fault Fig : 10(a,b): LLL fault the local maxima is missing.

IDENTIFIATION OF GROUND FAULTs : The trick of the detection is the nature of the (variation of(Variation of g)) -t plot. The variation of (Variation of) g gradually diminishes to zero towards the end of fault duration and this tendency of variation declining towards zero rapidly increases, if any additional Line or ground involved in the fault.

a) LLG ∆(∆g)-t

b) LLLG ∆ (∆ g) - t plot Fig: 11(a,b) :∆(∆g)-t plot for ground faults VI. PREDICTION OF ZONE AND LOCATION OF THE DETECTED FAULT: Trick to find zone is to analysis the time delay sensed by the system to detect the fault. It is a function of fault distance. By this way we can predict the Zone of the fault.

Fig: 12: LLG fault detection -time delay vs percentage length Plot for the SMIB System

Line Length= 320 KM Fault impedance Zf =( Vf/If) Ohm Distance of Fault Location= (Zf/ Z1) KM Where Z1 is Positive sequence impedance Since it is a 3-Zone Protection Scheme, Impedances and Time Constants of the relays varies Zone to Zone, so, sometimes it becomes problematic to detect the exact location of the fault but of course Zone of the fault we can predict. More the length increases the algorithm for fault location identification gives more accurate value. In case of fault impedance, during healthy state positive sequence impedance (Z1) present in the system and the value is given in the paper. Vf and If can be obtained from the sampling of the waveforms we got. VII. APPLICATION TO THE MULTIMACHINE 9-BUS SYSTEM: A multi machine power system with series compensation WSCC 3-machine, a 9-bus configuration is considered for the study where a modification has been incorporated by providing 40% compensation at the beginning of line 7–8.

The system details are given in Appendix B.

Fig: 13:WSCC-9 bus System Implementation Using PSCAD A three-phase fault is created in line 5–7 at 0.1 s. The fault is cleared at 0.3 s by opening breaker B3 and B4. The removal of the line causes a swing condition. Different faults are simulated on line 7–8 to test the algorithm.

Fig14:Current and Voltage Waveforms of relay bus during the power swing

a) I2 vs t Plot During Single Pole Tripping Condition

b) g - t plot during Single Pole Tripping Fig:15. Single pole tripping

Fig16: Single-Line Diagram of the Modified WSCC 9-bus System

1) Case-1: Line-to-Ground Fault: An ag- fault with a fault resistance of 0.1 is created during the power swing on line 7–8 at a distance of 160 km from the relay location at 0.3 s. The index increases to a higher value at the inception of the fault. The technique is able to detect the fault within half-a-cycle of its inception.

a) g - t plot b) I2 - t plot Fig17:LG fault analysis

2) Case-2: Double Line-to-Ground Fault:

The performance of the algorithm for a double-line-to-ground fault of with a ground fault resistance of 0.1 is created on a and b-phases at 0.3 s on line 7–8 during the power swing at a distance of 160 km from the relay location.

a) g2 vs t b) I2 vs t Fig:18. Fault detection.

VIII. CONCLUSIONS In this paper a novel fault detection technique for detecting the classification and location of the fault is proposed. LLG fault is created at 240 KM distance during power swing condition. After analysing it we got below waveform.

a) Voltage and Current Waveforms

b) ∆ g - t plot

c) ∆ (∆ g) - t plot Fig: 19:LLG fault Analysis By analysing the ∆g and ∆(∆g) - t plot we got this is a non-LG ground fault and less inclined to zero . i.e from ∆g vs t plot it is clear that it is either LLG or LLLG fault and from ∆(∆g) vs t, it is coming as LLG fault. By sampling the fault voltages, and performing calculations, it was obtained the fault location as 241.07 km for actual fault location 240Km.

Accuracy: Error in measurement of fault distance from the proposed algorithm is computed as follows. Accuracy= (100 %-(( 241.07-240)x100%)/240)= 99.554% So, the proposed fault detection technique for the series compensated line during the power swing is presented in this paper to detect the location and classification of the fault. It uses fine tuning or the nature of variation of fault index to detect the classification of the fault and the fault impedance calculation for the location of the fault. The performance of the proposed algorithm is tested for balanced and unbalanced faults for different series compensated systems.

APPENDIX A System data for SMIB Geneator: 600MVA,22KV,50HZ,H=4.4MW/MVA Xd=1.81p.u, Xd’=0.3p.u, Xd’’=0.23p.u, Td0’=8s,Td0’’=0.03s,X0=1.76p.u. ,Xq’’=0.25p.u,Tq0’’=0.03s,Ra=0.003p.u, Xp(Potier reactance)=0.15p.u. Transformer: 600MVA,22/400KV,50HZ,D/Y,X=0.163p.u, Xcore=0.33p.u,Rcore=0.0p.u,Pcopper=0.00177p.u Transmission lines: Length=320Km Positive-sequence impedance=0.12+j0.88 Ohm/Km

Zero-Sequence Impedance=0.309+j1.297 Ohm/Km Positive-sequence capacitive reactance=487.723x1000 Ohm-Km Zero-sequence capacitive reactance=419.34x1000 Ohm-Km.

APPENDIX B

System data for 3-machine 9-bus configuration: Gererators Gen-1: 600 MVA,22KV,50HZ Gen-2: 465 MVA,22KV,50HZ Gen-3: 310 MVA,22KV,50HZ Transformers T1: 600 MVA,22/400KV,50HZ,D/Y; T2: 465 MVA,22/400KV,50HZ,D/Y; T3: 310 MVA,22/400KV,50HZ,D/Y; Transmission line: Length of line 7-8=320Km.,line 8-9=400Km,line 7-5=310Km.,line 5-4=350Km,line 6-4=350Km,line 6-9=300km. Loads Load A=300MW+j100MVAr. Load B=200MW+j75MVAr. Load C=150MW+j75MVAr. Other parameter used are same as APPENDIX A

REFERENCES [1] S. M. Brahma, “Distance relay with out-of-step blocking function using wavelet transform,” IEEE Trans. Power Del., vol. 22, no. 3, pp.1360–1366, Jul. 2007. [2] S. R. Mohanty, A. K. Pradhan, and A. Routray, “A cumulative sum-based fault detector for power system relaying application,”IEEE Trans. Power Del., vol. 23, no. 1, pp. 79–86, Jan. 2008. [3] Abdolamir Nekoubin,”Simulation of Series Compensated Transmission Lines Protected with Mov”, World Academy of Science, Engineering and Technology 58,pp.1002-1006 2011. [4] Paresh Kumar Nayak, Ashok Kumar Pradhan, and Prabodh Bajpai, “A Fault Detection Technique for the Series-Compensated Line During Power Swing”, IEEE Trans. Power Del.,, VOL. 28, NO. 2, APRIL 2013,pp.714-722. [5] P. K. Nayak, A. K. Pradhan, and P. Bajpai, “Detecting fault during power swing for a series compensated line,” presented at the Int. Conf. Energy, Autom., Signal, Bhubaneswar, Odisha, India, Dec. 28–30,2011. [6] A. Esmaeilian, A. Ghaderi, M. Tasdighi, and A. Rouhani, “Evaluation and performance comparison of power swing detection algorithms in presence of series compensation on transmission lines,” in Proc. 10th Int. Conf. Environment Elect. Eng., May 8–11, 2011, pp. 1842–1848. [7] S. Lotfifard, J. Faiz, and M. Kezunovic, “Detection of symmetrical faults by distance relays during power swings,” IEEE Trans. Power Del., vol. 25, no. 1, pp. 81–87, Jan. 2010. [8] X. Lin, Y. Gao, and P. Liu, “A novel scheme to identify symmetrical faults occurring during power swings,” IEEE Trans. Power Del., vol.23, no. 1, pp. 73–78, Jan. 2008. [9] R. J. Marttila, “Performance of distance relay mho elements on MOVprotected series-compensated lines,” IEEE Trans. Power Del., vol. 7,no. 3, pp. 1167–1178, Jul. 1992. [10] H. J. Alture, J. B. Mooney, and G. E. Alexander, “Advances in series compensated line protection,” 2008. [Online]. Available: www.selinc.com/20081022, TP6340-01 [11] A. Newbould and I. A. Taylor, “Series compensated line protection:System modeling and relay testing,” in Proc. 4th Int. Conf.Develop.Power Protect., 1989, pp. 182–186.

[12] G. Benmouyal,D. Hou, andD. Tziouvaras, “Zero-setting power-swing blocking protection,” presented at the the 31st AnnualWestern Protective Relay Conf., Spokane, WA, Oct. 2004. [13] A. Mechraoui and D. W. P. Thomas, “A new principle for high resistance earth fault detection during fast power swings for distance protection,”IEEE Trans. Power. Del., vol. 12, no. 4, pp. 1452–1457, Oct.1997. [14] A. P. Apostolov, D. Tholomier, and S. H. Richards, “Superimposed components based sub-cycle protection of transmission lines,” in Proc.IEEE Power Eng. Soc. Power Syst. Conf. Expo,, Oct. 2004, vol. 1, pp. 592–597. [15] T. S. Sidhu and M. Khederzadeh, “Series compensated line protection enhancement by modified pilot relaying schemes,” IEEE Trans. PowerDel., vol. 21, no. 3, pp. 1191–1198, Jul. 2006. [16] F. Gustafsson, Adaptive Filtering and Change Detection. NewYork: Wiley, 2000. [17]John P.Nelson,”System Grounding and Ground-Fault Protection in Petrochemical Industry:A need for a better Understanding”,IEEE Transactions on Industrial Applications,Vol.38,No.6,Nov.-Dec 2002,pp.1633-1640. [18]Normann Fischer and Daqing Hou,”Methods For Detecting Ground Faults in Medium Voltage Distribution Power Systems”,[2006].[Online].Available:https://www.selinc.com/WorkArea/DownloadAsset.aspx?id=2385 [19]A.M.Sharaf,L.A.Snider and K.Debnath,”A Neural Network Based Relaying Scheme for Distribution System High Impedance FaultDetection”,[1993].[Online].Available: www.researchgate.net/...fault_detection/.../79e41510b7791025bb.pd [20]Choowong-Wattanasakpubal and Teratum-Bunyagal,”Algorithm for Detecting,Identifying,Locating and Experience to Develop the Automate Faults Location in Radial Distribution System”.Journal of Electrical Engineering and Technology ,Vol.5,No.1,pp.36-44,2010.

BIBLIOGRAPHY

Saptarshi Roy: Received the B.E degree in Electrical Engineering from Jadavpur University, West Bengal, India in2009. He was an Engineer in Crompton Greaves Limited from 2009 to 2010 and associated with Tata Consultancy Services Limited from

2011 to 2012. Currently he is pursuing second year of M.Tech in Power System Engineering in the department of Electrical Engineering, National Institute of Technology, Warangal,India. His area of interest is power system protection.

Dr. Suresh Babu Perli: Currently he is working as an Assistant Professor in Department Of Electrical Engineering, National Institute of Technology, Warangal. His area of interest is Power System Protection with digital

multifunction relays, Development of Adaptive protection schemes and Digital filtering algorithms

View publication statsView publication stats