Modeling and Simulation of Capacitor Voltage ... - IEEE Xplore

32 downloads 0 Views 860KB Size Report
Abstract--This paper describes the modelling of capacitor voltage transformer and simulation of its behaviour during transients using PSCAD/EMTDC. To damp ...
Paper accepted for presentation at the 2011 IEEE Trondheim PowerTech

1

Modeling and Simulation of Capacitor Voltage Transformer Transients using PSCAD/EMTDC Jayachandra Sakamuri and D. John Yesuraj, Member, CIGRE WG A3.22

Abstract--This paper describes the modelling of capacitor voltage transformer and simulation of its behaviour during transients using PSCAD/EMTDC. To damp out ferroresonance in CVT few ferroresonance methods have been proposed and used. In this paper, some of the suppressing circuits, series resonance type, parallel resonance type, and electronic type are reviewed. Specifications of these circuits and effects of various parameters on their performances are discussed. During line faults, when the primary voltage collapses, the CVT generates transients due to its energy storage elements. The effect of ferroresonance suppression circuits on the transient response of CVT is also studied and the proposed electronic type ferroresonance circuit is best suited to suppress low voltage transients and high voltage ferroresonance oscillations. Index Terms--CVT, Ferroresonance, Transient Response, Simulation.

E

PSCAD,

Modeling

I. INTRODUCTION

LECTRIC power systems are subjected to many types of disturbances that result in electrical transients due to lightning, fault, line energization/deenergization, switching of inductive/capacitive loads. When a sudden change in the system state occurs, the energy storage elements such as inductors and capacitors cannot allow instantaneous change to follow the new state and hence transients occur before it reaches a final steady state. Capacitor voltage transformers (CVTs) are widely used to transform the line voltages at transmission and sub-transmission levels to designated low voltage levels for monitoring, protection and control applications. The dynamic performance of protective relays depends on the signals produced by the CVTs, and these signals depend on the overall transient response of the CVT and the type of transients generated by the power system. Proper design and tuning of CVT assure that its output waveform is an exact replica of the input waveform under steady state conditions. However, under transient conditions, such as, faults and switching incidents, the CVT output deviates from its input due to the inductive, capacitive, and nonlinear components of CVT. Therefore, the response of the CVT during transients must be well known and quantified. The performance of a CVT is also affected by a phenomenon called ferroresonance, a special case of resonance, which can occur when a non-linear inductance of the intermediate

voltage transformer resonates with capacitance of CVT which generates high voltage across the components [1], [2]. Hence, the CVT is equipped with over voltage protection device and ferroresonance suppression circuit (FSC). The CVT FSC may be passive or active based on whether this circuit stores energy or not and these can be further categorized as (1) series resonance type, 2) power frequency blocking type , 3) fast saturation type , and 4) electronic type, based on the working principle [7]-[9]. However, these papers concentrated more on ferroresonance compared to transient response of CVT In this paper, a typical 420 kV CVT is used for analysis. The performance of the CVT with different FSCs is simulated for ferroresonance and transient response conditions. The existing FSCs can damp the ferroresonance oscillations but can not mitigate transient response oscillations due to CVT primary short circuit [3]-[5]. The proposed electronic type FSC can suppress the ferroresonance oscillation within short time as well as it can sense the transient condition and suppress the oscillations to get the better response. The electronic type FSC is best suited to meet the above two requirements according to IEC 60044-5 [12]. Section II explains the basic model of CVT. Different ferroreonance suppression circuits are explained in section III. Feroresonance analysis and simulation results using PSCAD/EMTDC are presented in section IV, transient response results with different FSCs are presented in section V, and the paper is concluded in section VII. II. CVT MODELING

Fig. 1. CVT schematic diagram Jayachandra Sakamuri and D. John Yesuraj are with R&D Department, Instrument Transformer Division of Crompton Greaves Ltd, Ambad, Nashik, India-422010(e-mail:[email protected]; [email protected]).

978-1-4244-8417-1/11/$26.00 ©2011

Schematic diagram of the typical 400 kV CVT used in this study is shown in Fig. 1. The CVT model composed of capacitive voltage divider (C1 and C2), compensating reactor (L), intermediate transformer (VT), over voltage protection

2

Voltage [p.u.]

device (S), drain coil (D), metal oxide varistor (MOV), ferroresonance suppression device (Zd),earth switch (ES), and burdens B1, B2.. The parameters of the CVT are given in the APPENDIX. The CVT transients are simulated using PSCAD. The VT losses and magnetization characteristics shown in Fig. 2 are considered in the model. A metal oxide varistor is used to limit the high voltages in the CVT secondary whose simulated characteristics are shown in Fig. 3 [6].

Fig. 4. CVT secondary voltage without any FSC

Current [p.u.] Fig. 2. Transformer magnetization characteristics

A. Series Resonance Filter This is an active FSC and involved energy storage elements. A series RLC filter shown in Fig. 5 can be used as FSC. Parameters of the filter are chosen in such a way that it resonates at 1/3rd of fundamental frequency at which dominant farroresonance oscillations occur [1],[2]and hence filter’s impedance, shown in (1) is equal to R and a maximum load (in W) is connected. The filter impedance is maximum at fundamental frequency and hence acts as an open circuit.

1 · § Z ( jω ) = R + j ¨ ωL − ¸ ωC ¹ ©

(1)

Current [A] Fig. 3. Voltage Vs Current characteristics of MOV

III. FERRORESONANCE SUPPRESSION CIRCUITS Ferroresonance oscillations occur in a CVT with no / low load condition when the equivalent capacitance of the coupling capacitor, in series with the nonlinear inductance (due to saturation of the magnetic core of the VT), resonates in the presence of periodic driving source at sub harmonic frequencies (majorly 1/3rd of fundamental frequency) [1]. The ferroresonance simulations were based on the test recommended by IEC60044-5 [12]. This includes short circuiting any one of the secondary winding of intermediate voltage transformer (VT) for a maximum duration of 100ms and opening the short after that while the CVT is kept energized. In order to analyze the importance of FSC in transient high voltage damping, the CVT is simulated for ferroresonance without any FSC and the overvoltage on the secondary side of CVT is shown in Fig. 4.

Fig. 5. Series RLC Filter.

Impedance [Ω], Power [W]

Voltage [V]

The filter parameters are chosen as R = 75Ω , L=1.825H, and C= 50μF at 50 Hz fundamental frequency connected across 200 V auxiliary winding. Therefore the variation of impedance and active power load of the filter with frequency is shown in Fig. 6.

16.66 Hz

Frequency [Hz]

Fig. 6. Impedance, power Vs frequency characteristics of series RLC filter

B. Power Frequency Blocking Filter The circuit shown in Fig. 7 is a series-parallel RLC filter consisting of two inductors Lf1 and Lf2 with the mutual coupling Mf, a capacitor Cf, and damping resistor Rf, tuned to the fundamental frequency with a high Q factor. The damping resistor is used to attenuate ferroresonance oscillations [9]. The filter impedance, shown in (2) is high at the fundamental

3

frequency. Therefore, this filter has no loading effect on the CVT under normal operating conditions. ª ( L + M )(1 − ω 2C ( L + M )) f f f1 « f2 Z ( jω ) = R + jω « −M f 2 « 1 − ω C f ( L f 1 + L f 2 + 2M f ) ¬

º » » (2) » ¼

At off-nominal frequencies, the impedance of the filter gradually approaches to the resistance of the damping resistor. The variation of impedance, active power load with frequency is shown in Fig. 8. But this filter impacts the transient response [3]. The filter is designed at 50 Hz frequency connected at 63.5 V using the method given in [9], and the values are

M f = 0.1146H , C f = 19.7μF , R f = 7.5Ω L f 1 = 0.2292H , L f 2 = 0.0573H

also damps the dangerous high voltage oscillations due to ferroresonance within less time compared to other techniques [10], [11]. As it does not involve any energy storage elements, it won’t affect the transient response of the CVT. Moreover, by switching on the electronic device in case of transient condition, it can further damp the low voltage oscillations by bringing the damping resistance across CVT secondary. To design the FSC, the following points are considered. • The magnitudes of high voltage oscillations without any FSC are approximately 2.5 p.u. and the CVT should withstand (a voltage of 1.5 p.u). for about 30s according to IEC 60044-5 [12]. Hence the FSC can be turned on if the voltage exceeds 1.5 p.u so that it will not affect the normal CVT operation. • The switch can be made ON for a specified duration until fault is cleared. • In case of fault, the damping burden should be connected in both positive and negative half cycles hence switch should be selected accordingly • The switch can also be turned on if the RMS value of the CVT voltage falls below certain low value to suppress the low transient oscillations. The schematic of the FSC with two back-to-back thyristor is shown in Fig. 9 .

(b) Digital model of the filter

Impedance [Ω], Power [W]

Fig. 7. (a) Power frequency Filter

Fig. 9. Electronic switch (Thyristor) FSC

IV. FERRORESONANCE SIMULATION RESULTS

Fig. 8. Impedance, power Vs frequency characteristics of power frequency blocking filter

Simulations using PSCAD/EMTDC have been carried out to test the effectiveness of introduced FSCs on suppression of ferroresonance of CVT. To establish the ferroresonance condition, the breaker S2 of Fig.1 is closed, whose ON resistance is 40mΩ, for 0.1s and then opened. Ferroresonance is said to be suppressed if the RMS voltage value deviation is less than 10% according to 60044-5 [12]

C. Electronic Switch The damping burden is connected across the CVT secondary by switching on a power electronic device in case of fault (either ferroresonance or transient condition). Under normal operating conditions there is very low and inherent burden connected, hence accuracy of the CVT also good. This technique leads to size reduction of the electromagnetic unit of CVT as it does not involve bulky inductor and capacitor. It

A. Performance of Series RLC Filter Simulation results of CVT with series RLC filter with the specified parameters in section 3.1 are given in Fig. 10. The ferroresonace oscillation got damped with in 0.5s. The disadvantage of this circuit is under normal operating conditions, there exists a current (0.32 A) flowing through the filter and hence a burden of 7.5 W at a power factor of 0.12 lag, which can affect the accuracy and transient performance of CVT. The size of the CVT is also high due to the bulky

Frequency [Hz]

4

inductor and capacitor, and the VA rating of the VT has to be increased to meet the specified accuracy requirements.

above 1.6 p.u. The enlarged view of currents through RLC series filter and MOV is shown in Fig. 12.

Fig. 12. Zoomed view of ILC and Ivar

Fig. 10. CVT secondary instantaneous voltage, RMS voltage, and current through the filter (ILC) with series RLC filter as an FSC

B. Performance of Power Frequency Blocking Filter With power frequency blocking filter (PBF) values given in section III, along with the MOV, the ferroesonance is suppressed within 0.15s as shown in Fig. 13. Here, ferroresonance over voltages are eliminated in fewer cycles and damping time of this FSC is lower than that of series RLC filter. But PBF affects the transient response as it involved energy storage elements [3],[4]. The size of this filter is high as it involves a bulk inductor and capacitor.

Fig. 11. CVT secondary instantaneous voltage in presence of MOV, RMS voltage, and current through the RLC filter (ILC) and metal oxide varistor (Ivar).

Fig. 13. CVT secondary voltage, its RMS value, current through the filter (IPBF) and MOV (Ivar), with power frequency blocking filter as an FSC

Simulation results with the addition of MOV are shown in Fig. 11. In this case the damping time is reduced to 0.3s as the MOV conducts immediately when the fault voltage exceeds

C. Performance of Electronic Switch The electronic switch FSC is connected at 200 V with a damping resistance R. Once the fault is detected, the switch is

5

ON for a fixed duration (say 80ms). After this time, if the fault still exists then the switch is continued to be ON until the fault is completely cleared .The MOV is also included in the circuit to damp the oscillations more effectively. The simulation results with a damping resistance of R = 80Ω is given in Fig. 14.The zoomed view of currents through thyristor switch and MOV are given in Fig 15 and from which one can see that the MOV is ON if the voltage goes beyond 1.6 p.u and thus it will limit the fault level. The thyristor switch is ON for duration of 80 ms once it detects the fault voltage above the set value (1.6 p.u) thereby it suppresses the oscillations completely. The performance of this FSC for different damping resistance values is shown in Fig. 16. With lower resistance, ferroresonance is more effectively damped out. However, there is an optimum resistance, which depends on circuit configuration, by which the best suppression is obtained. With R=80Ω, ferroresonance is suppressed within 0.1s

of a primary short circuit to the peak value of the secondary voltage before application of short circuit [12]. The CVT

Fig. 16. CVT RMS output voltages with damping resistances of R = 160Ω, 100Ω, 40Ω, and 20Ω with electronic switch FSC

response to a temporary, close in, line to ground fault is simulated by short circuiting the high voltage source of CVT with close-open operation of breaker S1 of Fig. 1, while S2 is kept open. During the test the burden shall be 100 %, 25% and 0% of rated burden and it can be controlled by using breakers S3 and S4.

Fig. 14. CVT secondary voltage, RMS voltage, and current through the switch (Ithy) and MOV (Ivar) with electronic switch as an FSC

Fig. 15. Zoomed view of Ithy and Ivar

V. TRANSIENT RESPONSE Characteristics of the transient response of the CVT is the ratio of secondary voltage at a specified time after application

The dependency of CVT transient response on different parameters such as point on the primary, transmission line voltage wave where the fault occurs, magnitude of stack capacitance, VT turns ratio, magnitude and power factor of burden and its composition and connection, exciting current of VT, and type of ferroresonance circuit have been studied in [3]. Among the above mentioned transient response deciding parameters, VT turns ratio, exciting current of VT, and Ferroresoannce suppression circuit are controllable. Remaining all parameters are either customer dependent or system fault dependent. The transient response is also critical if the fault occurs at zero crossing of the primary voltage. In this work the effect of Ferroresoannce suppression device, and VT magnetising current on transient response is studied. An attempt is made to improve the transient response with the help of thyristor based electronic FSC. A. Effect of Point on the Primary The transient response of the CVT without any suppression device but with 50 VA burden at 0.8 pf with fault applied at primary voltage zero crossing is shown in Fig. 17. Here the

6

secondary voltage of CVT oscillates at sub harmonic frequency and takes more time (more 0.3s) to get suppressed. The corresponding CVT primary and secondary voltage when the fault applied at primary crest voltage point is shown in Fig. 18. As seen from this Fig., at crest fault initiation, the discharge is a ringing high frequency oscillations caused by the parasitic capacitance of the compensating reactor and intermediate voltage transformer and due to lower VT burden. But after 25 ms of fault initiation, these oscillations were suppressed.

FSC on transient response with fault applied at primary zero crossing is shown in Fig. 19. Only CVT secondary voltage is shown here for better visibility of the results. The peak voltage is 20 V and it is taking more than 200 ms to suppress the oscillations and hence it can meet only 3PT1 of IEC 60044-5 [12].

Fig. 19 Transient Response of CVT with Series RLC FSC and 300 W Permanent Burden with a fault applied at primary voltage zero crossing

B2. Power Frequency Blocking Filter

Fig. 17. CVT Primary and secondary voltages without any FSC and fault applied at primary voltage zero crossing.

The PBF parameters explained in section III have been used in the simulation and the results are shown in Fig. 20. With this FSC the peak magnitude is high and steady state settling time is bit low compared to series RLC filter FSC. The results shows that CVT is only meeting 3PT1 transient response class of IEC 60044-5.

Fig. 20 Transient Response of CVT with Power Frequency Blocking Filter RLC FSC with a fault applied at primary voltage zero crossing.

B2. Electronic switch FSC Fig. 18. CVT Primary and secondary voltages without any FSC and fault applied at crest point of primary voltage.

B. Effect of Ferroresonance Suppression Circuit The effect of different FSC's on transient response is studied in this section. The effectiveness thyristor FSC to suppress transient oscillations is also explained. B1. Series RLC Filter with 300 W Permanent Burden A 300 W permanent burden along with series RLC FSC can be used to suppress Ferroresoannce oscillations as series RLC filter alone cannot suppress Ferroresoannce. The effect of this

The thyristor based Electronic switch FSC do not affect the transient response of CVT because it is a pure passive FSC without any energy storage elements. With proper sensing of transient response condition, the higher damping burden can be connected across CVT secondary by triggering the thyristor to suppress the transient oscillations. The transient response with a thyristor FSC connected at 200V with a damping resistance of 50Ÿ is shown in Fig. 21a. The transient response with the same FSC configuration but with damping burden of 25Ÿ is shown in Fig. 21b. The response is good with thyristor FSC and better with 25Ÿ burden compared to 50Ÿ. In this way by properly choosing the damping burden transient oscillations can be suppressed with thyristor FSC.

7

Fig. 21 (a) Thyristor FSC with 50Ÿ burden

Fig. 23 Transient Response of CVT with Electronic switch FSC with reduced VT magnetizing current

VI. APPENDIX Table I Typical 420 kV CVT Technical Data

Parameter

Value

Description

Vp

231 kV

Rated primary voltage of CVT (Ph-n)

Vs

63.5 V

Rated secondary voltage of VT (Ph-n)

Vt

200

Rated tertiary voltage of VT (Ph-n)

4361 pF

High voltage capacitance

113160 pF

Intermediate voltage capacitance

Fig. 21 Transient Response of CVT with Electronic switch FSC with a fault applied at primary voltage zero crossing.

C1 C2 Lc

86.21 H

Inductance of the compensating reactor

C. Effect of VT magnetising current.

Rc

320 Ÿ

Resistance of compensating reactor

Ld

15 mH

Drain coil inductance

Lp

8.5 H

VT primary leakage inductance

Ls

1 mH

VT secondary leakage inductance

Lm

2596 H

Magnetization inductance of VT

Rm

5.12 MŸ

Core loss resistance of VT

Fig. 21 (b) Thyristor FSC with 25Ÿ burden

If the VT magnetizing current is more, the more the energy stored in the capacitor and compensating reactor and hence more the subsidence transient. The magnetizing current of VT can be reduced by reducing the air gaps in the core. The magnetizing current considered for CVT discussed so far is 12mA at rated voltage. If the magnetizing current is reduced to 4 mA, then the transient response is studied with both series RLC FSC and Thyristor FSC. Fig. 22 shows the transient response results of a typical 400 kV CVT with reduced magnetizing current with series RLC filter as FSC. The transient response is better compared to higher magnetizing current shown in Fig. 17. The test results using Thyristor FSC is shown in Fig. 23. This is the best result among all. The result shows that the CVT is meeting all the transient response classes such as 3PT1, 3PT2, and 3PT3 of IEC 60044-5 [12].

Fig. 22 Transient Response of CVT with Series RLC FSC with reduced VT magnetizing current

VII. CONCLUSIONS This paper has presented the detailed model for transient studies and investigation of ferroresonance behaviour of typical 420 kV CVT with PSCAD/EMTDC. Various ferroresonance suppression circuits have been reviewed and their performance has been investigated for ferroresonance and transient response test conditions. The ferroresonance damping time using series RLC filter is more (0.5s) and it also affects the transient response and accuracy of the CVT due to low power factor permanent burden imposed by this filter. The performance of power frequency blocking filter for the ferroresonance (time =0.15s) and transient response conditions is better than the series RLC filter. But it still affects the accuracy and transient response due to its energy storage elements. The time domain simulation results show that electronic switch FSC can appropriately damp out ferroresonance within less time (0.1s). The transient response with this FSC is much better than the other two FSCs. The transient response study shows that it is critical for a fault at primary zero crossing. With the reduced VT magnetizing current, the transient

8

performance is better with series RLC filter. With Electronic FSC and with reduced VT magnetizing current, the transient response is much better and meeting all the classes, 3PT1 , 3Pt2 and 3PT3 of IEC 60044-5. This technique also leads to improved accuracy compared to the other FSCs. With all the three FSCs, adding a metal oxide varistor can limit overvoltage in the first cycle and significantly reduce the damping time. Although the existing FSC’s (series RLC, and power frequency blocking filter) perform well for damping ferroresonance, they can’t improve the transient response due to their energy storage elements. Hence, with the electronic switch FSC, the ferroresonance can be damped within five cycles as well as transient response can be improved by properly sensing the fault situation and controlling it accordingly. VIII. REFERENCES [1]

S. K. Chakravarthy and C. V. Nair, "Ferroresonant oscillations in capacitor voltage transformers," in Proc. 1995 IEE Circuits Devices Systems., vol. 142, pp. 30-36. [2] B. S. Ashok Kumar and Suat Ertem, "Capacitor voltage transformer induced ferroresonance – causes, effects and design considerations," Electrical Power Systems Research, vol. 21, pp. 23-31, Apr. 1991. [3] A. Sweetana, "Transient response characteristics of capacitive potential devices," IEEE Trans. Power Apparatus and Systems, vol. 90, pp. 19892001, Sept. 1971. [4] D. Hou and J. oberts, "Capacitive voltage transformers: Transient overreach concerns and solutions for distance relaying," in Proc. 1996 Canadian Conf. on Elect. and Comp. Engineering., pp. 119-125. [5] I. Sule, U. O. Aliyu, and G. K. Venayagamoorthy, "Simulation model for assessing transient performance of capacitive voltage transformers," in Proc. 2006 IEEE Power Engineering Society General Meeting, pp. 4 [6] B. Zitnik, M. Babuder, M. Muhr, M. Zitnik, and R. Tottapillily, "Numerical modelling of metal oxide varistors," in Proc. 2005 XIVth international Symposium on High Voltage Engineering, China. [7] M. Graovac, R. Iravani, X. Wang, and R. D. McTaggart, "Fast ferroresonance suppression of coupling capacitor voltage transformers," IEEE Trans. Power Delivery, vol. 18, pp. 158-163, Jan. 2003. [8] M. Sanaye-Pasand, A. Rezaei, H. Mohseni, Sh. Farhangi, and R. Iravani, "Comparison of performance of various ferroresonance suppressing methods in inductive and capacitive voltage transformers" in Proc. 2006 IEEE Power India Conf., pp. 8. [9] S. Shahabi, A. Gholami, M. Mirzaei, and Sh. Farhangi, "Investigation of performance of ferroresonance suppressing circuits in coupling capacitor voltage transformers" in Proc. 2009 IEEE 4th ICIEA Conf., pp. 216-221. [10] A. Abbasi, and A. Seifi, "Fast and perfect damping circuit for ferroresonance phenomena in coupling capacitor voltage transformers," Journal of Electric Power Components and Systems, Apr. 2009. [11] M. Sanaye-Pasand and R. Aghazadeh, "Capacitive voltage substations ferroresonance prevention using power electronic devices," in Proc. 2003 IPST Conf., USA. [12] IEC International Standard on Instrument transformers Part 5: Capacitor Voltage Transformers, IEC Standard, 2004.

Jayachandra Sakamuri received his M.Tech in Electrical Engineering from Indian Institute of Technology, Kanpur, India. He also worked on his M.Tech thesis at Technical University of Berlin, Germany. He is presently working as Sr. Executive (R&D), Instrument Transformers Division, Crompton Greaves Ltd, Nashik, India. His research interests include power system transient studies, smart grid, renewable energy integration, high voltage insulation monitoring.

D. John Yesuraj received his M.Tech in High Voltage Engineering from Anna University Chennai, India. He is presently working as Dy. General Manager (R&D), Instrument Transformers Division, Crompton Greaves Ltd, Nashik, India. He has vast experinece in design and development of HV and UHV instrument tarnsformers and lightning arresters. His research interests include power system transient studies, dielctric evaluation , condition monitoring of power system components.