FAULT LOOP IMPEDANCE ANALYSIS FOR A TRANSMISSION LINE WITH SERIES CAPACITORS AND THEIR OVERVOLTAGE PROTECTION

Eugeniusz Rosolowski, Bogdan Kasztenny, Jan Izykowski Wroclaw University of Technology, Poland Murari Mohan Saha ABB Network Partner, Vsters, Sweden

ABSTRACT This paper analyses the issues of protective relaying for a series compensated transmission line. Two conditional impedances of a fault loop have been proposed for use in the analysis of the overreaching phenomenon and in the distance relaying. The impedances were determined on the base of the fault data obtained from ATP-EMTP simulations. The locus analysis of the steady state fault loop impedances for a variety of conditions has been performed. Keywords: power system protection, impedance measurement, series compensated line

fault loop impedances, the ATP/EMTP model of the 400kV, 300km series compensated transmission system shown in Fig.1 has been developed. The HV equipment was appropriately modeled as well as the current and voltage measuring chains including the Current Transformers (CTs), the Capacitive Voltage Transformers (CVTs) and the anti-aliasing Analog Filters (AFs) [3-4]. Large number of fault cases have been simulated with the use of this model and the analysis of the fault loop impedances location has been performed.
CTs&AFs (F1) SCs (F2)

B
MOVs AIR-GAPs
THERMAL PROTECT.

INTRODUCTION Series Capacitors (SCs) when installed on long transmission lines for improving the power transfer and enhancing both power and voltage control, create, however, significant problems for protective relaying of such lines. In order to prevent overvoltages during short-circuits, the SCs are usually equipped with the Metal-Oxide Varistors switched in parallel. This combination of the capacitive reactance and the non-linear resistance is a source of phenomena not experienced in traditional lines. Nonlinear or subsynchronous oscillations contaminate the relay voltage and current input signals. Current and/or voltage inversion creates additional problems for protective relays [1-2]. The paper analyses the issues of protection for a line compensated in the middle (Fig.1). In order to provide reliable relay input signals for determining the

CVTs&AFs

RELAY

Fig.1. Single line diagram of the studied system.

FAULT LOOP IMPEDANCE MEASUREMENT - DYNAMIC AND STEADY STATE ASPECTS Two distinctive fault locations with respect to the SCs&MOVs position shall be considered for the system under study (Fig.1). The fault (F1) occurs in front of SCs&MOVs, while the fault (F2) behind them as seen from the considered substation A. Generally, the faults in front of SCs&MOVs (F1) are considered as not differing much from the faults on uncompensated lines. The certain difference has to be taken into account, only if the remote line infeed effect cannot be neglected, what is the case of fault

locators rather than protective relays. Therefore, fault loop impedance measurement under faults in front of SCs&MOVs can be performed with the variety of algorithms developed for traditional uncompensated lines [5]. However, the substantial difference appears for faults behind SCs&MOVs (F2). Fig.2 presents the possible compositions of fault loops under such faults. Depending on the fault type, one or even two complexes of a SC and its MOV (together with an air-gap) are involved in the fault loop. Moreover, for any type of a fault, the composition of a particular fault loop changes with passing of time due to the possibility of shunting the SC&MOV scheme by its air-gap (on the command of the thermal protection). Involvement of SC&MOV or even two such complexes in the fault loop makes it a strongly nonlinear circuit. As a consequence of that, a fault loop impedance measurement becomes entirely different from the case of uncompensated lines in both dynamic and steady state aspects.
a)

steady state aspect of measuring the fault loop impedance. APPLICATION OF TWO CONDITIONAL FAULT LOOP IMPEDANCES FOR PROTECTION OF A SERIES-COMPENSATED LINE Application of classical distance relays for protection of a series-compensated line is troublesome. Regardless the dynamic performance features, the new problem, related to the reach of a protective relay under faults behind SCs&MOVs arrives. In this case significant overreaching may take place what forces to set the first zone very short and degrades the protection quality. This drawback results from the fact that the fault loop model applied for the classical distance relays (resistive-inductive circuit) does not match the case of a fault with SCs&MOVs in the fault loop. As a result of that, the determined apparent impedance of a fault loop is not a strict geometrical measure of the distance to a fault. Therefore, to resolve this problem, this paper proposes to base the distance relaying on the model of the fault loop which is adequate for those troublesome cases. Fig.3. presents the models of fault loops for different locations and types of faults. The models of fault loops are derived under neglect of the line shunt capacitances what is justified for protection purposes. For faults in front of SCs&MOVs (Fig.3a) the fault loop apparent impedance (R, X) is determined with the signals composed analogously as for uncompensated lines: for the phase-to-ground fault: the faulty phase voltage (vph) and the current compensated for zero sequence component (iph+k0i0) are used, for the phase-to-phase fault: the differences in voltages (vph1-vph2) and currents (iph1-iph2), calculated for the phases involved in a fault are used. For faults behind SCs&MOVs there are two cases, as shown in Fig.3b and c: Fig.3b presents the fault loop model for phase-toground faults occurring behind SCs&MOVs. The resistive-inductive impedance of a fault loop (R, X) is determined with use of the faulty phase voltage compensated for the voltage drop across the SC&MOV (vph-vv) and the phase current compensated for the zero sequence component (iph+k0i0). Fig.3c, in turn, presents the fault loop model for phase-to-phase faults behind SCs&MOVs. In this case the resistive-inductive impedance of the fault loop (R, X) is determined with the differences of phase voltages but compensated for the voltage drops across the SC&MOV schemes, and currents from the faulty phases.

b) R

Fig.2. Fault loops under different faults occurring behind SCs&MOVs: a) phase-to-ground fault, b) phase-to-phase fault. The nature of transients in the voltage and current relaying signals under involvement of SCs&MOVs in the fault loop is completely different in comparison with the phenomena observed for uncompensated lines. Thus, fault loop impedance measurement in such the environment requires modifications of the classical algorithms. Adaptation of the measuring algorithms to the specific nature of transients in the relaying signals, relevant to the faults behind SCs&MOVs, is presented in [6-7]. This paper continues those research efforts but focuses on the

i ph (i=i ph1 -i ph2 ) v ph (v=v ph1 -v ph2 )

a)

Equivalenting of SC&MOV scheme has been done by a series of ATP-EMTP simulations (by scanning through different points on the equivalent characteristic). It was assumed that the voltage-current characteristic for the MOV is approximated by a single exponential formula: v (2) i MOV = P v V REF The coefficients of the approximation (2) were taken as: P=1 kA, VREF=150 kV, q=23. The results of equivalenting for this data are shown in Fig.4c.
a) Iv IC b) Iv
MOV
q

b) i ph vv v ph k 0 i0 v phc R X

C Rv Vv Xv

v v1 c) i ph1 v phc1 v ph1 (i ph1 -i ph2 ) R (v ph1 -v ph2 ) X

IMOV

Vv

c)

40

i ph2 v phc2 v ph2 v v2

20

Rv
0

Fig.3. Models of fault loops for: a) phase-to-ground (or phase-to-phase) faults in front of SCs&MOVs, b) phase-to-ground fault behind SCs&MOVs, c) phaseto-phase fault behind SCs&MOVs. Thus, application of the models of fault loops for measuring the distance to a fault behind SCs&MOVs (Fig.3b and c) requires compensating the phase voltages (vph) for the voltage across SCs&MOVs (vv). In terms of the instantaneous quantities the compensation is performed as:

[ohm]

-20

-40

Xv
-60 -80

2000

4000

6000

8000

10000

Amplitude of Current Entering SC&MOV : Iv [A]

v phc = v ph v v

(1)

Fig.4. Equivalenting of SC&MOV: a) the original scheme, b) the fundamental frequency equivalent circuit, c) the equivalent characteristic. Location of a fault in terms whether it occurs in front of SCs&MOVs or behind them is a random factor. Discrimination between these two cases is a difficult task and has to be performed on the base of transient features of the available relaying signals [9]. To avoid this, it is proposed to use the two conditional impedances of the fault loop in protective relaying of a series-compensated line. The first impedance is calculated without correction (identically as in the classical technique), while the second one with compensation for the SC&MOV equivalent. Then, these two conditional impedances are to be taken for making the final decision by a relay - in terms whether a fault occurred on a line (more precisely: in the protective zone covering the predefined segment of a line) or outside it. Let us denote: Z - impedance measured under assumption that a fault loop does not contain SC&MOV (or SCs&MOVs - for phase-to-phase faults),

As the voltage drops across the SCs&MOVs are not measured at the relaying point, they have to be estimated. The estimation can be performed on the base of the phase currents and the parameters of the approximation for the voltage-current characteristic of the MOVs as in [7]. The locus analysis of the steady state fault loop impedances is the primary aim of this study. For the analysis the compensation (1) can be replaced by the fundamental frequency equivalenting of SCs&MOVs. For this purpose, the parallel connection of a fixed series capacitor and its non-linear protecting resistor MOV (Fig.4a) was represented for the steady state by the fundamental frequency equivalent (Fig.4b). The equivalent is of the form of a series branch with the resistance RV and the reactance XV, both dependent on the amplitude of a fault current. Fundamental frequency currents and voltage drops denoted in the original scheme (Fig.4a) and in the equivalent (Fig.4b) match.

 Zc - impedance measured under assumption that a fault loop contains SC&MOV (or SCs&MOVs). The impedance without compensation (Z) is measured directly from the voltage and current signals, composed analogously as for traditional uncompensated lines. This impedance is valid for faults occurring between the relaying point (here the substation A) and the SCs&MOVs position as well as for backward faults. The impedance with compensation (Zc) is obtained by subtracting the equivalent (Zv) of the SC&MOV (or the equivalents for phase-to-phase faults) from the impedance without compensation (Z): ZC = Z Zv (3)

pedances for protection of a line leads to a danger of overreaching. The difference between the two conditional impedances for a particular fault case is equal to the impedance of the equivalent of the SC&MOV. This difference (represented by the connection line) depends on the operating point of the MOV.
120 X [] -Z - Zc

The impedance (3) has been considered for the faulty semi-steady state, i.e. assuming that the thermal protection of the MOVs does not operate. This impedance is valid for faults behind SCs&MOVs.

R [] 120

LOCATIONS OF TWO CONDITIONAL IMPEDANCES OF A FAULT LOOP In order to study the behavior of the seriescompensated line as a plant to be protected, a number of fault conditions were simulated using the developed ATP-EMTP model. The parameters changed during the simulation include: - fault location, - fault type, - fault inception angle, - fault resistance, - direction of pre-fault power flow. Location of both the conditional impedances for the simulated fault cases was studied. In the following figures, the "uncompensated" impedance is denoted by a circle, while a square stands for the "compensated" impedance. The two conditional impedances for the same fault case are connected in the figures by the line segment. Fig.5. presents the impedance locations for 7 different fault locations of a phase-to-ground fault with zero fault resistance. On the base of these results one can conclude that: For the faults applied in front of SCs&MOVs (at 50, 100, 150- km), the "uncorrected" impedances are located on the line impedance vector and deliver a strict geometrical measure of the distance to a fault. At the same time, the "corrected" impedances (which are not relevant for these faults) are located left from the line impedance vector. For the faults located behind SCs&MOVs (150+, 200, 250, 300km), the "compensated" impedances place on the line impedance vector. The "uncompensated" impedances are placed right and down from the line impedance vector. Therefore, utilization of these "uncompensated" imFig.5. Locations of impedances: Z - "uncompensated" and Zc - "compensated" for phase-toground faults at different points on the line AB; fault resistance: 0. Fig.6 displays locations of the conditional fault loop impedances under the faults as in Fig.5, but with 5 fault resistance. Comparing the locations of the fault loop impedances from Fig. 6 (fault resistance: 5) and Fig.5 (fault resistance: 0) it is seen that certain differences appear. For faults with higher fault resistance (Fig.6) one observes: The "uncompensated" impedances (circles) for the faults in front of SCs&MOVs, are moved right towards high resistances. The fault resistance additionally twisted due to the infeed effect contributes to the shift. For the faults behind SCs&MOVs, the "uncompensated" impedances are shifted from the line impedance vector not only right, but in addition very much down (for example, the fault at 300km is seen as at some 160km). As a consequence of that, the classical relay (utilizing "uncompensated" impedance) may see a distant fault closer than a close fault (for example the fault at 150km gives around 50 reactance, while the fault at 200km - only some 35 reactance). The "compensated" impedances which are relevant for these faults are also moved right and down. Moreover, the shift towards high resistances is disproportionally large (for example for the fault

at 300km, instead of 5 shift equal to the fault resistance, some 20 displacement is observed).
120 X [] -Z - Zc

[1] Application guide on protection of complex transmission network configurations, CIGRE, SC34-WG04, August 1990. [2] Krebs R., Retzman D., Ziegler G., Simulation and test of protection in FACTS environment, CIGRE Colloquium, Stockholm, paper 34-105, June 1995. [3] Esztergalyos J., Ohlen C., Nimmersjo G., Saha M.M., EMTP used in testing of a protection scheme for series compensated network, CIGRE Colloquium, Stockholm, paper 34-110 June 1995. [4] Dommel H., Electromagnetic Transient Program, BPA, Portland, Oregon, 1986. [5] Ungrad H., Winkler W., Wiszniewski A., Protection techniques in Electrical Energy Systems, Marcel Dekker Inc., New York, 1995. [6] Izykowski J., Rosolowski E., Saha M.M., Eriksson L., Study of transient simulation of series compensated network for investigation of digital protective relays, in Proceedings of Stockholm Power Tech Conference, Stockholm, pp.479-484, 1995. [7] Rosolowski E., Kasztenny B., Izykowski J., Saha M.M., Comparative analysis of impedance algorithms for series compensated lines, in Proceedings of 10th International Conference on Power System Protection, Bled, Slovenia, pp.21-26, 1996. [8] Ghassemi F., Johns A.T., Goddarzi J., A method for eliminating the effect of MOV operation on digital distance relays when used in series compensated lines, in Proceedings of 32nd Universities Power Engineering Conference UPEC97, Manchester, UK, pp. 113-116, 1997. [9] Novosel D., Phadke A., Saha M.M., Lindahl S., Problems and solutions for microprocessor protection of series compensated lines, in Proceedings of Sixth International Conference on Developments in Power System Protection, IEE Conference Publication No.434, Nottingham, UK, pp. 18-23, 1997. APPENDIX Basic parameters of the series-compensated system: line - 400kV, 300km, compensation rate: 70%, Z1=0.315/km, 85o; Z0=1.0265/km, 75o equivalent systems - identical at both the line ends: Zs1=15, 85o, Zs0=26.6, 85o

R [] 120

Fig.6. Locations of impedances: Z - "uncompensated" and Zc - "compensated" for phase-toground faults at different points on the line AB; fault resistance: 5. The obtained picture of the locations of the conditional impedances is a solid base for development of protective relaying algorithm for series-compensated lines.

CONCLUSIONS ATP-EMTP model of the series-compensated system has been developed. In order to study the behavior of the series-compensated line as a plant to be protected, a number of fault conditions were simulated using the developed model. The signals from the simulations were applied for calculation of fault loop impedances. Compensating capacitors and MOVs were equivalented for the fundamental frequency. The equivalents were applied in determining the apparent fault loop impedance valid for faults behind SCs&MOVs. The two conditional impedances of a fault loop: without correction and - with compensation for the SC&MOV equivalent have been proposed for application in protective relaying for seriescompensated lines. The impedances were calculated for the faulty steady-state of the large number of fault cases. The analysis regarding location of the conditional impedances has been performed. The obtained picture of the location of the conditional impedances is a solid base for development of efficient protective relaying algorithm for series-compensated lines. BIBLIOGRAPHY