- Email: [email protected]
Distance protection of block transformer units
Electrical Power and Energy Systems 102 (2018) 332–339
Contents lists available at ScienceDirect
Electrical Power and Energy Systems journal homepage: www.elsevier.com/locate/ijepes
Distance protection of block transformer units a
b
a,⁎
b
T a
D. Bejmert , M. Kereit , W. Rebizant , L. Schiel , K. Solak , A. Wiszniewski a b
a
Wroclaw University of Science and Technology, Poland Siemens AG, Berlin, Germany
A R T I C LE I N FO
A B S T R A C T
Keywords: Power transformers Distance protection Phase-to-ground faults Turn-to-turn internal short-circuits Measurement correction
The main protection of block transformer units is usually realized with use of differential relays. For back-up protection distance relays are mostly employed. Their second and third zones provide also coverage of faults along the adjacent line connecting the power plant with the nearest substation. This paper deals with a challenge related to proper impedance measurement across the transformer for the configurations where the currents from the HV star terminals are not available. Commonly used algorithms of the distance relays may be prone to serious under-reaching in such situations, which calls for a new approach with correction of the measurement procedure. The inclusion of zero-sequence current is proposed, and the improved protection algorithm is analyzed theoretically and thoroughly tested with the signals obtained from EMTP-ATP simulations. Rules for new protection settings are also provided.
1. Introduction In this paper problems related to protection of block units are discussed, where distance relay is usually a back-up for given transformer and adjacent lines [1]. Special attention is paid to the configurations when the distance relays are installed at the triangle side of a block transformer, as shown in Fig. 1. It is usually the case that the star side signals are not available or are not used, which may be a source of impedance measurement errors and relay maloperation [2,3]. It is even stated in [3] that the operation errors of the distance relays are unavoidable for behind located ground faults. Although the distance relay is meant as a back-up for faults along the adjacent line, its operation should be reliable and safe, which cannot be guaranteed with standard solutions and algorithms. There exist a number of decent publications related to distance protection principle. Nevertheless, not many of them are related to block transformer or in-zone transformer applications. If any, they mostly deal with different aspects of distance protection, e.g. discussing the protection under-reach for in-zone phase shifting transformers [4] or the influence of power swings on block unit distance protection [5]. Some proposals to overcome the problems with transformer distance protection operation may be found, introducing e.g.: - zero-sequence based compensation of the transformer tertiary winding influence (first zone is analyzed only) [6], - employing a three-relay arrangement based on resistance elements [7],
⁎
- distance protection coordination with use of the IEC61850 GOOSE message-based scheme [8], which do not solve the basic problem of wrong impedance calculation through the transformer for behind located faults. Since none of the abovementioned approaches is 100% efficient or the ideas described require a lot of effort and costly installations, an efficient and comprehensive approach to the problems with distance relay underreaching for single-phase faults behind the in-zone transformer is still to be developed. In this paper the following points have been addressed. First, present solution performance for d and Y transformer side faults is investigated. Then a proposal of improvements for correction of throughtransformer impedance measurement errors as well as development of settings recommendation for the new protection is discussed. It is shown that substantial improvement of the distance protection operation may be reached with introduction of the zero-sequence current and appropriate settings of the protection algorithms. 2. Distance measurement for ground and line faults at transformer star side The following analyses have been done for a selected configuration of Yd11 transformer, presented in Fig. 2. Similar investigations and results may easily be provided for other transformer connection groups. In relation to Fig. 2 one can introduce the winding turn ratio Nz and transformation ratio N defined as:
Corresponding author. E-mail addresses: [email protected] (D. Bejmert), [email protected] (M. Kereit), [email protected] (W. Rebizant), [email protected] (L. Schiel), [email protected] (K. Solak), [email protected] (A. Wiszniewski). https://doi.org/10.1016/j.ijepes.2018.05.012 Received 7 December 2017; Received in revised form 8 May 2018; Accepted 10 May 2018 0142-0615/ © 2018 Elsevier Ltd. All rights reserved.
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al. approx. 90% of Zgen (depends on x’d)
Z1 – impedance between the transformer delta terminals (measurement point) and the fault point (star side), Z1 = ZT1 + ZL1, Z0 – zero-sequence impedance to the fault point, Z0 = ZT0 + ZL0, ZT1 – transformer short-circuit impedance (referred to the star side voltage level), ZL1 – positive sequence impedance from the transformer star terminals to the fault point, ZT0 – transformer zero-sequence impedance (ZT0 = ZT1), ZL0 – zero-sequence impedance between the transformer star terminals and the fault point.
0,7s 0,4s
0,1s
Z2 appr. ZTr
Z1=0,7 ZTr
Z<
Z<
Fig. 1. Basic configuration of block unit with distance protection.
Then the measurement of phase A loop impedance is done with the following equations:
Za =
Uca 3 UA = I1−I3 N
1 N 3
(3IA−3I0)
=
UA N 2 (IA−I0)
(6a)
This may be rewritten in the form: I
(
)
Z
0 0 Z 1 + IA Z1 −1 Za = 12 I N 1− I 0
(6b)
A
which finally yields:
Z I Z Za = ⎛ 12 ⎞ ⎡1 + ⎛ 0 ⎞ ⎛ 0 ⎞ ⎤ − I I ⎝ N ⎠⎢ A 0 ⎠ ⎝ Z1 ⎠ ⎥ ⎝ ⎣ ⎦
Fig. 2. Yd11 transformer and line configuration studied.
⎟⎜
⎜
w Nz = Y wd N=
(1b)
With this in mind one can write the following relationships between triangle and star currents:
Ia = Nz IA
(2a)
Ib = Nz IB
(2b)
Ic = Nz IC
(2c)
I1 = Ia−Ib = Nz (IA−IB )
(3a)
I2 = Ib−Ic = Nz (IB−IC )
(3b)
I3 = Ic−Ia = Nz (IC−IA)
(3c)
2
Za =
1 3 UB = UB Nz N
Ubc =
1 3 UC = UC Nz N
1 3 Uca = UA = UA Nz N
UB = IB Z1 + 3I0
Z0−Z1 3
UC = IC Z1 + 3I0
Z0−Z1 3
⎜
⎟
(7a)
Z
Za =
ZL1 3 Z Z ⎛1 + ⎞ = 2.5 L1 = 2.5 1 N2 ⎝ 2⎠ N2 N2
(7b)
One can see that the measured impedance (7b) is 2.5 times higher than in reality and thus the relay is prone to serious under-reaching. In reality this coefficient may change between 1.5 and 2.5 and depends on the fault place. For faults at transformer terminals one should replace the total impedances with the ones of block transformer, which yields:
(4a)
Za = (4b)
ZT 1 ⎛ 1 ZT 0 ⎞ Z Z 1+ = 1.5 T21 = 1.5 12 N2 ⎝ 2 ZT 1 ⎠ N N ⎜
⎟
(7c)
One should of course understand that for the relay installed at the delta side the star side signals may not be available. Measurement of the phase current as well as more accurate impedance measurement during ground faults at Y side calls for “measurability” of the zero-sequence current 3I0. Then
(4c)
Measurement of impedance seen from the transformer delta side is further analyzed. For ground faults at the star side (at transformer terminals or along the adjacent line) the following relations for star side voltages are valid:
Z −Z UA = IA Z1 + 3I0 0 1 3
1
1 3 Z1 + 3 Z0 Z 1 Z0 ⎞ = 12 ⎛1 + 2 N2 N ⎝ 2 Z1 ⎠ 3
If the ratio of impedances ZL0 = 3 (it is so just for the line impedance L1 or for the case when the line is long and the transformer impedance can be neglected), then:
For the voltage signals it holds:
Uab =
(6c)
The formula (6c) shows the relative increase of the measured impedance caused by the current I0 and impedance Z0. I 1 For the A-phase faults it holds I 0 = 3 (the most probable value), A therefore:
(1a)
3 Nz
⎟
3IA = 3I0−
3 (I3−I1) N
(8)
With introduction of the corrected current Ir:
(5a)
Ir = (5b)
N N [3IA + 3I0 (n−1)] = I1−I3 + n3I0 3 3
(9)
one may derive the following relations for impedance measurement: (5c)
Za =
where 333
3 IA Z1 + I0 (Z0−Z1) N Ir
(10a)
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Za =
1+ 1 Z1 2 N 1+
I0 Z0 − Z1 IA Z1 I0 (n−1) IA
For the case when n =
Table 1 Input signals and impedance measurement results for various Yd transformers.
(10b) Z0 : Z1
1 Za = 2 Z1 N
Zab =
UA−UB IA−IB Uca−Uab = 3I1
3 N
(UA−UB )
3 N (IA−IB )
=
Z1 N2
3I0
Yd11 Yd5
AG BG CG
I1−I3 I2−I1 I3−I2
Uca Uab Ubc
− − −
AG BG CG
I1−I3 I2−I1 I3−I2
Uca Uab Ubc
+ + +
AB BC CA
I1 I2 I3
Uca−Uab Uab−Ubc Ubc−Uca
− − −
AG BG CG
I1−I2 I2−I3 I3−I1
Uab Ubc Uca
− − −
AG BG CG
I1−I2 I2−I3 I3−I1
Uab Ubc Uca
+ + +
AB BC CA
−I2 −I3 −I1
Uab−Ubc Ubc−Uca Uca−Uab
− − −
Ua Z = 12 I1 N
Uca−Uab U −U −U + Ua 3Ua Z = a c b = = 12 3I1 3I1 3I1 N
Uca−Uab 3Ua−3U0 U −U Z = = a 0 = 12 3I1 3I1 I1 N
Measured impedance Za or Zab I Z 1 + 0 ⎛ 0 − 1⎞ IA ⎝ Z1 ⎠ I 1− 0 IA I Z − Z1 1+ 0 0 IA Z1 Z1 2 N 1 + I0 (n − 1) IA Z1 N2 ⎜
⎟
Z1 N2
I Z 1 + 0 ( 0 − 1) IA Z1 I 1− 0 IA I0 Z0 − Z1 1+ IA Z1 Z1 N2 1 + I0 (n − 1) IA Z1 N2
Z1 N2
(11b) (a) with measurement of three phase currents at the transformer Y side (provided the currents are available), then: (12a)
3I0 = IA + IB + IC
(b) direct measurement of 3I0 with use of CT installed in the grounding wire at the Y side; (c) through measurement of current inside transformer delta connection, e.g. in phase a; in such a case for the transformer Yd11 the following equations hold:
(11c)
Ia + Ib + Ic = Nz (IA + IB + IC ) = Nz (3I0)
(12b)
Ia = Nz IA
Ic = Ia + I3
(12c)
1 3 (3Ia−I1 + I3) = (3Ia−I1 + I3) Nz N
(12d)
(11d)
3I0 =
However, the above is correct when Ua + Ub + Uc = 0 , which is usually true, but not always. E.g. in case when there is a ground fault at the delta side, the sum of triangle phase voltages amounts to Ua + Ub + Uc = 3U0 ≠ 0 . Therefore, correct measurement of impedance for the simultaneous fault at the star side would require subtracting the zero-sequence voltage, according to:
Zab =
Voltage
(11a)
which means that the phase quantities are used. Comparing (11b) and (11c) one can see that both versions are equivalent, since:
Zab =
Current
Yd1 Yd7
From Eqs. (11) it results that there is no influence of the zero-sequence current and thus no correction is required. The result of (11b) represents correctly measured impedance of the inter-phase A-B loop for faults at the star side. Traditional way of calculating the impedance of A-B fault loop with measurements taken on the triangle side is [9]:
Zab =
Fault type
(10c)
which means that the measurement is now errorless and the underreaching is eliminated. Since the exact fault place is unknown before taking the relay decision, it is proposed to use the ratio of impedances n referred to the end of given protection zone. The justification of this suggestion can be found in Section 3. For the 2-phase and 3-phase faults at the Y transformer side the following relations are valid:
Z1 =
Trafo
Ib = Ia−I1
The concepts (b) and (c) are illustrated in Fig. 3. Obviously, version with additional current transformer in the delta winding is only realizable for transformers consisting of separate units for particular phases. 3. Distance measurement – simulation studies
(13c) A number of simulation tests and investigations have been performed with use of EMTP-ATP software [10] in order to confirm the theoretical study results as well as to generate signals for the new relay prototype testing. According to the requirements, a system configuration with the synchronous generator, block transformer (500 MVA,
Because of that it can stated that one is always on the safe side and no error is introduced for the non-zero value of 3U0 when the algorithm with phase-to-phase voltages (11b) is applied. Otherwise, with 3U0 missing (not measured), serious error of impedance calculation may appear, which would lead to relay under-reaching (phase voltage Ua too large). Additionally, during the single-phase-to-ground short-circuits at the star side some zero-sequence voltage from the star side is transferred to the delta side via the capacitive link, which also may introduce an error. The delta side zero-sequence voltage may also be corrupted by higher harmonics, especially the third order ones. Therefore, although the equation (11c) is very elegant and simple, the Authors are of the opinion that the algorithm (11b) offers better accuracy. Similar relationships as above can also be derived for the other fault loops and other types of considered Yd transformers. The summary of the analysis for various transformer configurations is given in Table 1. When the transformer zero-sequence current is available and appropriate value of coefficient n is assumed one can reach correct impedance measurement for all fault types. The required 3I0 current may be reached in one of the following ways:
Fig. 3. Installation of CTs for zero-sequence current measurement. 334
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Fig. 4. Correct (stars) and measured (circles) impedances for line-line, line-lineto-ground and three-phase faults.
Fig. 6. Impedance measurement for phase-to-ground faults without (black stars) and with correction (red symbols) for various levels of n. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
short-circuit voltage 15%, Yd5, 400 kV/20 kV, Y-side solidly grounded) and transmission line (ZL0′ = (0.1604 + j0.7193) Ω/km, ZL1′ = (0.0288 + j0.3235) Ω/km, l = 150 km) has been modelled. It was assumed that third zone of distance protection should cover 1/3 of the line length (lIII = 50 km), therefore the fault spots were placed at locations lF every 10 km from transformer terminals up to the third zone reach (every 20% of lIII). The impedances seen from the generator side were calculated for the following fault types: phase-to-ground, phaseto-phase, phase-phase-to-ground, and three-phase. The way of simulation of transformer internal faults and appropriate ATP-EMTP model is described in our previous paper [11]. Below only selected results of simulations are shown in an aggregated form. In the following figures black star points represent exact locations of fault, while red circles mark the calculated fault impedances (steady-state of measurement). In Fig. 4 the impedance measurement results for multi-phase faults along the line are presented. It is seen that accuracy of measurement is good, especially as the reactance is concerned (certain error for fault resistance is seen); thus no correction for the zero-sequence current is required. In Fig. 5 the results for phase-to-ground faults are shown. Measured impedances are much higher than correct values, with factor 1.43 (at transformer star terminals) up to 1.55 (last measurement point at the
end of 3rd zone reach), which means that the relay sees the faults as located much further than in reality. Of importance is of course the reach in reactance axis, although also the resistance measured is much bigger than in reality (note that the axis range in R direction is different than in X direction, only for the sake of better picture readability). Introduction of the zero-sequence correction according to (10b) (see also Table 1) brings significant improvement of fault distance estimation accuracy, which is shown in Figs. 6–9. Unfortunately, it is impossible to reach perfect measurement for all fault places, since the ideal value of correction factor n depends on unknown fault location. In Fig. 6 the ratios of absolute values of measured impedance Za to the actual fault loop impedance Z1 are shown as a function of the fault distance from transformer Y terminals to the end of the 3rd zone reach. One can see that the errors of distance measurement without correction (black star points) are the highest. The results obtained with correction are much better; however, depending on the correction coefficient n they may be slightly overestimated (for n too small) or underestimated (n too large). The best results are obtained for the zero- to positive-sequence impedance ratio set for the end of the assumed third zone reach. In Figs. 7–9 the results of measurement correction expressed in R-X coordinates are shown for three levels of the zero- to positive-sequence
Fig. 5. Correct (stars) and measured (circles) impedances for phase-to-ground faults; calculation without correction (n = 0).
Fig. 7. Correct (stars) and measured (circles) impedances for phase-to-ground faults; calculation with correction (n = 1). 335
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Fig. 10. Measured impedances for transformer turn-to-turn faults for all six distance relay loops; no correction applied.
Fig. 8. Correct (stars) and measured (circles) impedances for phase-to-ground faults; calculation with correction (n = ZL0/ ZL1 = 3).
Fig. 9. Correct (stars) and measured (circles) impedances for phase-to-ground
(
faults; calculation with correction n =
ZT 0 + lIII / l ZL0 ZT 1 + lIII / l ZL1
)
Fig. 11. Measured impedances (magnified) for transformer turn-to-turn faults for A-G loop; no correction applied.
.
impedance ratio n. One can see again that the best results are obtained for n related to the end of the third zone reach, where all in zone faults are seen correctly, i.e. the impedance measured is equal or less than the zone settings. Further simulation cases have been generated for transformer internal faults. Since it is very difficult to derive theoretical relationships for inter-turn faults and ground faults (nonlinear dependencies with respect to fault spot and number of short-circuited turns) [11], the simulation results bring valuable information on the fault current levels and measured impedances. In Fig. 10 one can see the families of points for all six protection fault loops for simulated turn-to-turn faults with 10% up to 90% of short-circuited turns and their location with respect to transformer impedance ZT (black star). According to expectations, the smallest measured impedance values were noted for the phase A where the faults took place. Fig. 11 shows magnified part of Fig. 10 for the fault loop A-G only. One can see that all fault cases (10% up to 90% of shorted turns, seen from transformer terminals) are identified by the distance protection as located in the second zone. When the zero-sequence correction is applied (Fig. 12, n = 1), the measured impedances are only slightly changed (small influence of the zero-sequence current). The proposed settings of the relay with applied correction are inappropriate, the internal faults are still seen as located in the 3rd zone.
Fig. 12. Measured impedances (magnified) for transformer turn-to-turn faults for A-G loop; correction applied, n = 1.
336
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Fig. 13. Measured impedances for transformer internal ground faults for all six distance relay loops; no correction applied.
Fig. 15. Measured impedances (magnified) for transformer internal ground faults for A-G loop; correction applied, n = 1.
Similar analysis was performed for the cases of transformer internal ground faults spotted at 10% up to 90% of the star winding, counted from transformer terminals. Again, the smallest impedances are measured for the A-G loop where the faults took place, see Fig. 13. Closer analysis of the results for A-G loop indicate that such faults are seen mostly as external ones (out of the second zone reach), Fig. 14. The effects of zero-sequence correction (Fig. 15) is much stronger here due to higher level of the zero-sequence current. Yet, with the proposed relay settings the internal ground faults are still seen as external ones. One can state here that the operation of distance protection with/ without measurement correction for such internal faults is not satisfactory. Other protection principles (e.g. REF) ought to be applied to detect such events.
happen at the machine front terminals. Operation of the distance protection in such cases is very limited (CTs at the neutral point side, VTs at the terminals). Such faults should be cleared by the differential protection. If converted to ground faults (quite probable) it will be detected by ground fault relays. (c) The most likely to occur are double or three phase short-circuits between generator terminals and transformer triangle terminals. Such events should evoke firm operation of the distance protection. (d) Ground faults and phase-to-phase faults inside the block transformer are very unlikely, since they would also require breaking of the main insulation. Such faults may occur as a result of inter-turn faults (much more likely) that become developed into ground faults. Operation of distance protection for inter-turn transformer faults at any of transformer windings depends upon the number of short-circuited turns and their location. (e) Distance protection should pick up for phase-to-phase faults at the terminals of the triangle side as well as for phase-to-phase and ground faults at the transformer star side.
4. Protection logic and settings Before the protection logic is outlined one should be aware of the following. (a) Ground faults at the transformer triangle side (inside of generator, at generator-transformer terminals) cannot be detected by the distance protection. (b) Double-phase fault inside the generator is very unlikely, since it would require breaking of the main insulation. If any, then it can
The distance protection installed at transformer triangle side should be equipped with all six impedance loops operating independently, with measurements running in parallel. Three of the loops cover ground faults at the star side and phase-to-phase faults at the triangle side, whereas the remaining three are to respond to phase-to-phase faults at the star side. The input signals and impedance measurement results for various Yd transformers and particular loops are gathered in Table 1. The following distance protection logic is proposed: - the distance relay located at the delta side should have three adjustable zones of operation; the first zone with the shortest delay must not reach beyond the circuit breaker on the star side, the second zone with the delay 0.4 s must cover the transformer and the connection to the circuit breaker, the third zone with the delay 0.7 s must provide a back-up protection of the adjacent line, - in case of unavailable I0 current the protection reach is to be set as for phase-to-phase faults at the triangle side; should the protection operate, one has to diagnose fault at the triangle side, - in case of measurable zero-sequence current detection of significant I0 flow leads to confirmation of ground fault at the star side; then one has to increase the measurement current by the required part of I0 and proceed with the required zone reach for ground faults. With the above in mind and having observed the results of impedance measurement effects with and without zero-sequence current
Fig. 14. Measured impedances (magnified) for transformer internal ground faults for A-G loop; no correction applied. 337
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
In this case the ground fault loop impedance (7) is measured in excess and the relay is prone to under-reach (factor 1.5). Therefore the relay settings should be defined as:
ZI = 0.9Za = 1.35ZT1
(14a)
ZII = 1.2Za = 1.8ZT1
(14b)
ZII = 2.5Za = 2.5(ZT 1 + kZL1)
(14c)
4.2. Three loops covering the phase-to-phase short-circuits at the star side The relay settings for the inter-phase faults at the star side (no zerosequence correction needed) are defined in the same way as for option A2 (ground faults with correction), i.e.:
Fig. 16. Distance protection logic for ground faults at Y side and phase-to-phase faults at d side with I0 measurable.
(15a)
ZII = 1.2Za = 1.2ZT1
(15b)
ZIII = Za = ZT 1 + kZL1
(15c)
If all the three loops measure the same impedance, then one may diagnose a three-phase short-circuit at the star side. If only one loop measures the low impedance, then it is a phase- to-phase fault at the star side, and the faulted phases are the ones which correspond to the current of the loop.
available, the following zone settings are proposed. 4.1. Three loops covering the phase-to-ground faults at the star side and the phase-to-phase short-circuits at the delta side A1 – relay settings for the configuration with I0 measurable
5. Conclusions
In this case the fault loop impedance is correctly measured, both for inter-phase and ground faults. Therefore the relay settings can be defined as follows:
ZI = 0.9Za = 0.9ZT1
(13a)
ZII = 1.2Za = 1.2ZT1
(13b)
ZIII = Za = ZT 1 + kZL1
(13c)
In this paper the results of theoretical and simulative analyses of block transformer distance protection are presented. A proposal for improvement of protection operation for ground faults at the star side with use of zero-sequence current is developed. Relay logic and settings rules are given both for the configuration with I0 measurable and when the zero-sequence current is not available. The algorithms for impedance/distance calculation proposed assure correct relay operation, especially for the second and third zone, which provides useful backup for faults along the adjacent line. Thorough simulation tests performed with EMTP-ATP software have shown that the developed distance relay may face some problems under transformer internal inter-turn and ground faults, since such faults are most frequently not detected by the relay (they are mostly located in the second or even in the third zone). For such faults other principles and criteria should be used, employing e.g. the zero-sequence quantities [12] or the commonly applied restricted earth fault protection.
where k is the portion (0…1) of the line to be covered with the third zone. One should remember to select the right value of the correction coefficient n that should be adjusted to the reach of the third zone including protected transformer and the adjacent line or its part, i.e. Z +l /l Z n = ZT 0 + lIII / l ZL0 . If in (13c) one assumes full line length to be protected T 1 III L1 in the third zone (k = 1), then the correction factor should amount to ZT 0 + ZL0 n = Z +Z . T1 L1 For the phase-to-phase fault detected (I0 measured close to zero) there is only one zone to be set:
ZI = ZGT
ZI = 0.9Za = 0.9ZT1
References
(13d) [1] IEEE guide for protecting power transformers. IEEE Std C37.91-2008. [2] Ungrad H, Winkler W, Wiszniewski A. Protection techniques in electrical energy systems. New York: Marcel Dekker Inc.; 1995. [3] Ziegler G. Numerical distance protection. Principles and applications. 4th ed. Publicis Erlangen; 2011. [4] Ghorbani A. An adaptive distance protection scheme in the presence of phase shifting transformer. Electr Power Syst Res 2014;129:170–7. [5] Lizer M. Power unit impedance and distance protection functions during faults in the external power grid. Acta Energet 2012;4(13):22–33. [6] Bermann J, Prochazka M. Distance protection of big network transformers. In: Proceedings of the 17th international scientific conference on electric power engineering (EPE), Prague, Czech Rep.; 2016. [7] Frolova E, Osintsev A. Use of backup distance protection on a block transformer. Power Technol Eng 2016;50(2):220–3. [8] Kim HK, Kang SH, Nam SR, Oh SS. Improved Operating Scheme Using an IEC61850based Distance Relay for Transformer Backup Protection. In: Proceedings of the 2009 IEEE Bucharest PowerTech Conference, paper #131. [9] SIPROTEC 5, Transformer protection 7UT87 Manual, Siemens; 2016. [10] EMTP-ATP Manual. [11] Wiszniewski A, Solak K, Rebizant W, Schiel L. Calculation of the lowest currents caused by turn-to-turn short-circuits in power transformers. Int J Electr Power Energy Syst 2018;95:301–6. http://dx.doi.org/10.1016/j.ijepes.2017.08.028. [12] Wiszniewski A, Rebizant W, Schiel L. New algorithms for power transformer interturn fault protection. Electr Power Syst Res 2009;79:1454–61.
where ZGT is the impedance between relay location and transformer terminals. One should note that for the interphase short-circuits on the delta side the value measured represents the impedance between the fault and the location of the VTs. Therefore it is very small, unless the shortcircuits are inside the generator or transformer windings. Therefore, if the impedance measured becomes Za ⩽ ZT1 the fault may be diagnosed as an inter-phase short-circuit at the delta side. If all the loops measure low value of the impedance, then it is a three-phase short-circuit. If only one loop measures the low impedance, then the short-circuit may be diagnosed as a two-phase short-circuit between the phases which correspond to the voltage in the numerator of the impedance measurement algorithms. If the measured impedance is larger than ZT1, the shortcircuit may be diagnosed as a single-phase-to-ground fault at the star side. The faulted phase is the one, which corresponds to the voltages of the loop that measured the lowest impedance. The block scheme of the protection logic for this part is presented in Fig. 16. A2 – relay settings for the configuration with I0 not available 338
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al. Daniel Bejmert was born in 1979 in Walbrzych, Poland. He received his M.Sc., Ph.D. degrees from the Wroclaw University of Science and Technology, Wroclaw, Poland in 2004, and 2008, respectively. Since 2008 he has been a faculty member of Electrical Engineering Faculty at the WUST, at present at the position of the Assistant Professor. His research interests include: application of intelligent algorithms in digital protection and control systems, multicriterial and adaptive systems for power system protection and control and digital simulation of transient phenomena in power systems.
Ludwig Schiel was born in Weimar, Germany, in 1957. He studied Electrical Engineering at the Zittau Institute of Technology Zittau, Germany, finishing with the Dipl.-Ing. degree in 1984. In 1991 he received the Dr.-Ing. degree. In the same year he joined the Siemens AG, Germany. He is working in the Energy Management Division Digital Grid Automation Products concerning transformer protection.
Matthias Kereit was born in 1968 in Berlin, Germany. He studied electrical engineering in Berlin and graduated in 1992. He has been working with Siemens since 1992 and is now developing algorithms for digital protection devices, especially for distance protection relays. Since 2008 he has also been responsible for the university cooperations of the Berlin development department as a project manager.
Krzysztof Solak was born in 1982 in Sroda Slaska, Poland. He received his M.Sc. and Ph.D. degrees from the Wroclaw University of Science and Technology (WUST), Wroclaw, Poland, in 2006 and in 2010, respectively. At present he is an assistant professor in the Department of Electrical Power Engineering at WUST. His field of study is artificial intelligence (neural networks, fuzzy systems and genetic algorithms) for power system protection.
Waldemar Rebizant was born in Wroclaw, Poland, in 1966. He received his M.Sc., Ph.D. and D.Sc. degrees from Wroclaw University of Science and Technology (WUST), Wroclaw, Poland in 1991, 1995 and 2004, respectively. In 2012 he got the professorship degree. Since 1991 he has been a faculty member of Electrical Engineering Faculty at the WUST, at present at the position of the Dean of the Faculty. He is also Doctor Honoris Causa of the University of Magdeburg, Germany. In the scope of his research interests are: digital signal processing and artificial intelligence for power system protection. Prof. Rebizant has published more than 180 scientific papers, is an author of 11 patents and several patent applications. Details can be found at http://zas.pwr.edu.pl/WR.
Andrzej Wiszniewski graduated from the Electrical Engineering Faculty of the Wroclaw University of Science and Technology (WUST), Wroclaw, Poland, in 1957. In 1961 he received the Ph.D. degree, in 1967 the D.Sc. degree, and in 1972 the professorship in Electrical Engineering. All his life he has been working in the field of power apparatus and systems, being a specialist in protection and control of power systems. He is an author of 9 books and over 130 publications. All the time attached to the WUST he became the University Rector and had this position for 2 terms (1990 – 1996). In 1997 he took the position of the Minister for Science and had it for 4 years. Prof. Wiszniewski is a Distinguished Member of CIGRE and Honorary Member of the Polish Institution of Electrical Engineers.
339
Contents lists available at ScienceDirect
Electrical Power and Energy Systems journal homepage: www.elsevier.com/locate/ijepes
Distance protection of block transformer units a
b
a,⁎
b
T a
D. Bejmert , M. Kereit , W. Rebizant , L. Schiel , K. Solak , A. Wiszniewski a b
a
Wroclaw University of Science and Technology, Poland Siemens AG, Berlin, Germany
A R T I C LE I N FO
A B S T R A C T
Keywords: Power transformers Distance protection Phase-to-ground faults Turn-to-turn internal short-circuits Measurement correction
The main protection of block transformer units is usually realized with use of differential relays. For back-up protection distance relays are mostly employed. Their second and third zones provide also coverage of faults along the adjacent line connecting the power plant with the nearest substation. This paper deals with a challenge related to proper impedance measurement across the transformer for the configurations where the currents from the HV star terminals are not available. Commonly used algorithms of the distance relays may be prone to serious under-reaching in such situations, which calls for a new approach with correction of the measurement procedure. The inclusion of zero-sequence current is proposed, and the improved protection algorithm is analyzed theoretically and thoroughly tested with the signals obtained from EMTP-ATP simulations. Rules for new protection settings are also provided.
1. Introduction In this paper problems related to protection of block units are discussed, where distance relay is usually a back-up for given transformer and adjacent lines [1]. Special attention is paid to the configurations when the distance relays are installed at the triangle side of a block transformer, as shown in Fig. 1. It is usually the case that the star side signals are not available or are not used, which may be a source of impedance measurement errors and relay maloperation [2,3]. It is even stated in [3] that the operation errors of the distance relays are unavoidable for behind located ground faults. Although the distance relay is meant as a back-up for faults along the adjacent line, its operation should be reliable and safe, which cannot be guaranteed with standard solutions and algorithms. There exist a number of decent publications related to distance protection principle. Nevertheless, not many of them are related to block transformer or in-zone transformer applications. If any, they mostly deal with different aspects of distance protection, e.g. discussing the protection under-reach for in-zone phase shifting transformers [4] or the influence of power swings on block unit distance protection [5]. Some proposals to overcome the problems with transformer distance protection operation may be found, introducing e.g.: - zero-sequence based compensation of the transformer tertiary winding influence (first zone is analyzed only) [6], - employing a three-relay arrangement based on resistance elements [7],
⁎
- distance protection coordination with use of the IEC61850 GOOSE message-based scheme [8], which do not solve the basic problem of wrong impedance calculation through the transformer for behind located faults. Since none of the abovementioned approaches is 100% efficient or the ideas described require a lot of effort and costly installations, an efficient and comprehensive approach to the problems with distance relay underreaching for single-phase faults behind the in-zone transformer is still to be developed. In this paper the following points have been addressed. First, present solution performance for d and Y transformer side faults is investigated. Then a proposal of improvements for correction of throughtransformer impedance measurement errors as well as development of settings recommendation for the new protection is discussed. It is shown that substantial improvement of the distance protection operation may be reached with introduction of the zero-sequence current and appropriate settings of the protection algorithms. 2. Distance measurement for ground and line faults at transformer star side The following analyses have been done for a selected configuration of Yd11 transformer, presented in Fig. 2. Similar investigations and results may easily be provided for other transformer connection groups. In relation to Fig. 2 one can introduce the winding turn ratio Nz and transformation ratio N defined as:
Corresponding author. E-mail addresses: [email protected] (D. Bejmert), [email protected] (M. Kereit), [email protected] (W. Rebizant), [email protected] (L. Schiel), [email protected] (K. Solak), [email protected] (A. Wiszniewski). https://doi.org/10.1016/j.ijepes.2018.05.012 Received 7 December 2017; Received in revised form 8 May 2018; Accepted 10 May 2018 0142-0615/ © 2018 Elsevier Ltd. All rights reserved.
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al. approx. 90% of Zgen (depends on x’d)
Z1 – impedance between the transformer delta terminals (measurement point) and the fault point (star side), Z1 = ZT1 + ZL1, Z0 – zero-sequence impedance to the fault point, Z0 = ZT0 + ZL0, ZT1 – transformer short-circuit impedance (referred to the star side voltage level), ZL1 – positive sequence impedance from the transformer star terminals to the fault point, ZT0 – transformer zero-sequence impedance (ZT0 = ZT1), ZL0 – zero-sequence impedance between the transformer star terminals and the fault point.
0,7s 0,4s
0,1s
Z2 appr. ZTr
Z1=0,7 ZTr
Z<
Z<
Fig. 1. Basic configuration of block unit with distance protection.
Then the measurement of phase A loop impedance is done with the following equations:
Za =
Uca 3 UA = I1−I3 N
1 N 3
(3IA−3I0)
=
UA N 2 (IA−I0)
(6a)
This may be rewritten in the form: I
(
)
Z
0 0 Z 1 + IA Z1 −1 Za = 12 I N 1− I 0
(6b)
A
which finally yields:
Z I Z Za = ⎛ 12 ⎞ ⎡1 + ⎛ 0 ⎞ ⎛ 0 ⎞ ⎤ − I I ⎝ N ⎠⎢ A 0 ⎠ ⎝ Z1 ⎠ ⎥ ⎝ ⎣ ⎦
Fig. 2. Yd11 transformer and line configuration studied.
⎟⎜
⎜
w Nz = Y wd N=
(1b)
With this in mind one can write the following relationships between triangle and star currents:
Ia = Nz IA
(2a)
Ib = Nz IB
(2b)
Ic = Nz IC
(2c)
I1 = Ia−Ib = Nz (IA−IB )
(3a)
I2 = Ib−Ic = Nz (IB−IC )
(3b)
I3 = Ic−Ia = Nz (IC−IA)
(3c)
2
Za =
1 3 UB = UB Nz N
Ubc =
1 3 UC = UC Nz N
1 3 Uca = UA = UA Nz N
UB = IB Z1 + 3I0
Z0−Z1 3
UC = IC Z1 + 3I0
Z0−Z1 3
⎜
⎟
(7a)
Z
Za =
ZL1 3 Z Z ⎛1 + ⎞ = 2.5 L1 = 2.5 1 N2 ⎝ 2⎠ N2 N2
(7b)
One can see that the measured impedance (7b) is 2.5 times higher than in reality and thus the relay is prone to serious under-reaching. In reality this coefficient may change between 1.5 and 2.5 and depends on the fault place. For faults at transformer terminals one should replace the total impedances with the ones of block transformer, which yields:
(4a)
Za = (4b)
ZT 1 ⎛ 1 ZT 0 ⎞ Z Z 1+ = 1.5 T21 = 1.5 12 N2 ⎝ 2 ZT 1 ⎠ N N ⎜
⎟
(7c)
One should of course understand that for the relay installed at the delta side the star side signals may not be available. Measurement of the phase current as well as more accurate impedance measurement during ground faults at Y side calls for “measurability” of the zero-sequence current 3I0. Then
(4c)
Measurement of impedance seen from the transformer delta side is further analyzed. For ground faults at the star side (at transformer terminals or along the adjacent line) the following relations for star side voltages are valid:
Z −Z UA = IA Z1 + 3I0 0 1 3
1
1 3 Z1 + 3 Z0 Z 1 Z0 ⎞ = 12 ⎛1 + 2 N2 N ⎝ 2 Z1 ⎠ 3
If the ratio of impedances ZL0 = 3 (it is so just for the line impedance L1 or for the case when the line is long and the transformer impedance can be neglected), then:
For the voltage signals it holds:
Uab =
(6c)
The formula (6c) shows the relative increase of the measured impedance caused by the current I0 and impedance Z0. I 1 For the A-phase faults it holds I 0 = 3 (the most probable value), A therefore:
(1a)
3 Nz
⎟
3IA = 3I0−
3 (I3−I1) N
(8)
With introduction of the corrected current Ir:
(5a)
Ir = (5b)
N N [3IA + 3I0 (n−1)] = I1−I3 + n3I0 3 3
(9)
one may derive the following relations for impedance measurement: (5c)
Za =
where 333
3 IA Z1 + I0 (Z0−Z1) N Ir
(10a)
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Za =
1+ 1 Z1 2 N 1+
I0 Z0 − Z1 IA Z1 I0 (n−1) IA
For the case when n =
Table 1 Input signals and impedance measurement results for various Yd transformers.
(10b) Z0 : Z1
1 Za = 2 Z1 N
Zab =
UA−UB IA−IB Uca−Uab = 3I1
3 N
(UA−UB )
3 N (IA−IB )
=
Z1 N2
3I0
Yd11 Yd5
AG BG CG
I1−I3 I2−I1 I3−I2
Uca Uab Ubc
− − −
AG BG CG
I1−I3 I2−I1 I3−I2
Uca Uab Ubc
+ + +
AB BC CA
I1 I2 I3
Uca−Uab Uab−Ubc Ubc−Uca
− − −
AG BG CG
I1−I2 I2−I3 I3−I1
Uab Ubc Uca
− − −
AG BG CG
I1−I2 I2−I3 I3−I1
Uab Ubc Uca
+ + +
AB BC CA
−I2 −I3 −I1
Uab−Ubc Ubc−Uca Uca−Uab
− − −
Ua Z = 12 I1 N
Uca−Uab U −U −U + Ua 3Ua Z = a c b = = 12 3I1 3I1 3I1 N
Uca−Uab 3Ua−3U0 U −U Z = = a 0 = 12 3I1 3I1 I1 N
Measured impedance Za or Zab I Z 1 + 0 ⎛ 0 − 1⎞ IA ⎝ Z1 ⎠ I 1− 0 IA I Z − Z1 1+ 0 0 IA Z1 Z1 2 N 1 + I0 (n − 1) IA Z1 N2 ⎜
⎟
Z1 N2
I Z 1 + 0 ( 0 − 1) IA Z1 I 1− 0 IA I0 Z0 − Z1 1+ IA Z1 Z1 N2 1 + I0 (n − 1) IA Z1 N2
Z1 N2
(11b) (a) with measurement of three phase currents at the transformer Y side (provided the currents are available), then: (12a)
3I0 = IA + IB + IC
(b) direct measurement of 3I0 with use of CT installed in the grounding wire at the Y side; (c) through measurement of current inside transformer delta connection, e.g. in phase a; in such a case for the transformer Yd11 the following equations hold:
(11c)
Ia + Ib + Ic = Nz (IA + IB + IC ) = Nz (3I0)
(12b)
Ia = Nz IA
Ic = Ia + I3
(12c)
1 3 (3Ia−I1 + I3) = (3Ia−I1 + I3) Nz N
(12d)
(11d)
3I0 =
However, the above is correct when Ua + Ub + Uc = 0 , which is usually true, but not always. E.g. in case when there is a ground fault at the delta side, the sum of triangle phase voltages amounts to Ua + Ub + Uc = 3U0 ≠ 0 . Therefore, correct measurement of impedance for the simultaneous fault at the star side would require subtracting the zero-sequence voltage, according to:
Zab =
Voltage
(11a)
which means that the phase quantities are used. Comparing (11b) and (11c) one can see that both versions are equivalent, since:
Zab =
Current
Yd1 Yd7
From Eqs. (11) it results that there is no influence of the zero-sequence current and thus no correction is required. The result of (11b) represents correctly measured impedance of the inter-phase A-B loop for faults at the star side. Traditional way of calculating the impedance of A-B fault loop with measurements taken on the triangle side is [9]:
Zab =
Fault type
(10c)
which means that the measurement is now errorless and the underreaching is eliminated. Since the exact fault place is unknown before taking the relay decision, it is proposed to use the ratio of impedances n referred to the end of given protection zone. The justification of this suggestion can be found in Section 3. For the 2-phase and 3-phase faults at the Y transformer side the following relations are valid:
Z1 =
Trafo
Ib = Ia−I1
The concepts (b) and (c) are illustrated in Fig. 3. Obviously, version with additional current transformer in the delta winding is only realizable for transformers consisting of separate units for particular phases. 3. Distance measurement – simulation studies
(13c) A number of simulation tests and investigations have been performed with use of EMTP-ATP software [10] in order to confirm the theoretical study results as well as to generate signals for the new relay prototype testing. According to the requirements, a system configuration with the synchronous generator, block transformer (500 MVA,
Because of that it can stated that one is always on the safe side and no error is introduced for the non-zero value of 3U0 when the algorithm with phase-to-phase voltages (11b) is applied. Otherwise, with 3U0 missing (not measured), serious error of impedance calculation may appear, which would lead to relay under-reaching (phase voltage Ua too large). Additionally, during the single-phase-to-ground short-circuits at the star side some zero-sequence voltage from the star side is transferred to the delta side via the capacitive link, which also may introduce an error. The delta side zero-sequence voltage may also be corrupted by higher harmonics, especially the third order ones. Therefore, although the equation (11c) is very elegant and simple, the Authors are of the opinion that the algorithm (11b) offers better accuracy. Similar relationships as above can also be derived for the other fault loops and other types of considered Yd transformers. The summary of the analysis for various transformer configurations is given in Table 1. When the transformer zero-sequence current is available and appropriate value of coefficient n is assumed one can reach correct impedance measurement for all fault types. The required 3I0 current may be reached in one of the following ways:
Fig. 3. Installation of CTs for zero-sequence current measurement. 334
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Fig. 4. Correct (stars) and measured (circles) impedances for line-line, line-lineto-ground and three-phase faults.
Fig. 6. Impedance measurement for phase-to-ground faults without (black stars) and with correction (red symbols) for various levels of n. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
short-circuit voltage 15%, Yd5, 400 kV/20 kV, Y-side solidly grounded) and transmission line (ZL0′ = (0.1604 + j0.7193) Ω/km, ZL1′ = (0.0288 + j0.3235) Ω/km, l = 150 km) has been modelled. It was assumed that third zone of distance protection should cover 1/3 of the line length (lIII = 50 km), therefore the fault spots were placed at locations lF every 10 km from transformer terminals up to the third zone reach (every 20% of lIII). The impedances seen from the generator side were calculated for the following fault types: phase-to-ground, phaseto-phase, phase-phase-to-ground, and three-phase. The way of simulation of transformer internal faults and appropriate ATP-EMTP model is described in our previous paper [11]. Below only selected results of simulations are shown in an aggregated form. In the following figures black star points represent exact locations of fault, while red circles mark the calculated fault impedances (steady-state of measurement). In Fig. 4 the impedance measurement results for multi-phase faults along the line are presented. It is seen that accuracy of measurement is good, especially as the reactance is concerned (certain error for fault resistance is seen); thus no correction for the zero-sequence current is required. In Fig. 5 the results for phase-to-ground faults are shown. Measured impedances are much higher than correct values, with factor 1.43 (at transformer star terminals) up to 1.55 (last measurement point at the
end of 3rd zone reach), which means that the relay sees the faults as located much further than in reality. Of importance is of course the reach in reactance axis, although also the resistance measured is much bigger than in reality (note that the axis range in R direction is different than in X direction, only for the sake of better picture readability). Introduction of the zero-sequence correction according to (10b) (see also Table 1) brings significant improvement of fault distance estimation accuracy, which is shown in Figs. 6–9. Unfortunately, it is impossible to reach perfect measurement for all fault places, since the ideal value of correction factor n depends on unknown fault location. In Fig. 6 the ratios of absolute values of measured impedance Za to the actual fault loop impedance Z1 are shown as a function of the fault distance from transformer Y terminals to the end of the 3rd zone reach. One can see that the errors of distance measurement without correction (black star points) are the highest. The results obtained with correction are much better; however, depending on the correction coefficient n they may be slightly overestimated (for n too small) or underestimated (n too large). The best results are obtained for the zero- to positive-sequence impedance ratio set for the end of the assumed third zone reach. In Figs. 7–9 the results of measurement correction expressed in R-X coordinates are shown for three levels of the zero- to positive-sequence
Fig. 5. Correct (stars) and measured (circles) impedances for phase-to-ground faults; calculation without correction (n = 0).
Fig. 7. Correct (stars) and measured (circles) impedances for phase-to-ground faults; calculation with correction (n = 1). 335
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Fig. 10. Measured impedances for transformer turn-to-turn faults for all six distance relay loops; no correction applied.
Fig. 8. Correct (stars) and measured (circles) impedances for phase-to-ground faults; calculation with correction (n = ZL0/ ZL1 = 3).
Fig. 9. Correct (stars) and measured (circles) impedances for phase-to-ground
(
faults; calculation with correction n =
ZT 0 + lIII / l ZL0 ZT 1 + lIII / l ZL1
)
Fig. 11. Measured impedances (magnified) for transformer turn-to-turn faults for A-G loop; no correction applied.
.
impedance ratio n. One can see again that the best results are obtained for n related to the end of the third zone reach, where all in zone faults are seen correctly, i.e. the impedance measured is equal or less than the zone settings. Further simulation cases have been generated for transformer internal faults. Since it is very difficult to derive theoretical relationships for inter-turn faults and ground faults (nonlinear dependencies with respect to fault spot and number of short-circuited turns) [11], the simulation results bring valuable information on the fault current levels and measured impedances. In Fig. 10 one can see the families of points for all six protection fault loops for simulated turn-to-turn faults with 10% up to 90% of short-circuited turns and their location with respect to transformer impedance ZT (black star). According to expectations, the smallest measured impedance values were noted for the phase A where the faults took place. Fig. 11 shows magnified part of Fig. 10 for the fault loop A-G only. One can see that all fault cases (10% up to 90% of shorted turns, seen from transformer terminals) are identified by the distance protection as located in the second zone. When the zero-sequence correction is applied (Fig. 12, n = 1), the measured impedances are only slightly changed (small influence of the zero-sequence current). The proposed settings of the relay with applied correction are inappropriate, the internal faults are still seen as located in the 3rd zone.
Fig. 12. Measured impedances (magnified) for transformer turn-to-turn faults for A-G loop; correction applied, n = 1.
336
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
Fig. 13. Measured impedances for transformer internal ground faults for all six distance relay loops; no correction applied.
Fig. 15. Measured impedances (magnified) for transformer internal ground faults for A-G loop; correction applied, n = 1.
Similar analysis was performed for the cases of transformer internal ground faults spotted at 10% up to 90% of the star winding, counted from transformer terminals. Again, the smallest impedances are measured for the A-G loop where the faults took place, see Fig. 13. Closer analysis of the results for A-G loop indicate that such faults are seen mostly as external ones (out of the second zone reach), Fig. 14. The effects of zero-sequence correction (Fig. 15) is much stronger here due to higher level of the zero-sequence current. Yet, with the proposed relay settings the internal ground faults are still seen as external ones. One can state here that the operation of distance protection with/ without measurement correction for such internal faults is not satisfactory. Other protection principles (e.g. REF) ought to be applied to detect such events.
happen at the machine front terminals. Operation of the distance protection in such cases is very limited (CTs at the neutral point side, VTs at the terminals). Such faults should be cleared by the differential protection. If converted to ground faults (quite probable) it will be detected by ground fault relays. (c) The most likely to occur are double or three phase short-circuits between generator terminals and transformer triangle terminals. Such events should evoke firm operation of the distance protection. (d) Ground faults and phase-to-phase faults inside the block transformer are very unlikely, since they would also require breaking of the main insulation. Such faults may occur as a result of inter-turn faults (much more likely) that become developed into ground faults. Operation of distance protection for inter-turn transformer faults at any of transformer windings depends upon the number of short-circuited turns and their location. (e) Distance protection should pick up for phase-to-phase faults at the terminals of the triangle side as well as for phase-to-phase and ground faults at the transformer star side.
4. Protection logic and settings Before the protection logic is outlined one should be aware of the following. (a) Ground faults at the transformer triangle side (inside of generator, at generator-transformer terminals) cannot be detected by the distance protection. (b) Double-phase fault inside the generator is very unlikely, since it would require breaking of the main insulation. If any, then it can
The distance protection installed at transformer triangle side should be equipped with all six impedance loops operating independently, with measurements running in parallel. Three of the loops cover ground faults at the star side and phase-to-phase faults at the triangle side, whereas the remaining three are to respond to phase-to-phase faults at the star side. The input signals and impedance measurement results for various Yd transformers and particular loops are gathered in Table 1. The following distance protection logic is proposed: - the distance relay located at the delta side should have three adjustable zones of operation; the first zone with the shortest delay must not reach beyond the circuit breaker on the star side, the second zone with the delay 0.4 s must cover the transformer and the connection to the circuit breaker, the third zone with the delay 0.7 s must provide a back-up protection of the adjacent line, - in case of unavailable I0 current the protection reach is to be set as for phase-to-phase faults at the triangle side; should the protection operate, one has to diagnose fault at the triangle side, - in case of measurable zero-sequence current detection of significant I0 flow leads to confirmation of ground fault at the star side; then one has to increase the measurement current by the required part of I0 and proceed with the required zone reach for ground faults. With the above in mind and having observed the results of impedance measurement effects with and without zero-sequence current
Fig. 14. Measured impedances (magnified) for transformer internal ground faults for A-G loop; no correction applied. 337
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al.
In this case the ground fault loop impedance (7) is measured in excess and the relay is prone to under-reach (factor 1.5). Therefore the relay settings should be defined as:
ZI = 0.9Za = 1.35ZT1
(14a)
ZII = 1.2Za = 1.8ZT1
(14b)
ZII = 2.5Za = 2.5(ZT 1 + kZL1)
(14c)
4.2. Three loops covering the phase-to-phase short-circuits at the star side The relay settings for the inter-phase faults at the star side (no zerosequence correction needed) are defined in the same way as for option A2 (ground faults with correction), i.e.:
Fig. 16. Distance protection logic for ground faults at Y side and phase-to-phase faults at d side with I0 measurable.
(15a)
ZII = 1.2Za = 1.2ZT1
(15b)
ZIII = Za = ZT 1 + kZL1
(15c)
If all the three loops measure the same impedance, then one may diagnose a three-phase short-circuit at the star side. If only one loop measures the low impedance, then it is a phase- to-phase fault at the star side, and the faulted phases are the ones which correspond to the current of the loop.
available, the following zone settings are proposed. 4.1. Three loops covering the phase-to-ground faults at the star side and the phase-to-phase short-circuits at the delta side A1 – relay settings for the configuration with I0 measurable
5. Conclusions
In this case the fault loop impedance is correctly measured, both for inter-phase and ground faults. Therefore the relay settings can be defined as follows:
ZI = 0.9Za = 0.9ZT1
(13a)
ZII = 1.2Za = 1.2ZT1
(13b)
ZIII = Za = ZT 1 + kZL1
(13c)
In this paper the results of theoretical and simulative analyses of block transformer distance protection are presented. A proposal for improvement of protection operation for ground faults at the star side with use of zero-sequence current is developed. Relay logic and settings rules are given both for the configuration with I0 measurable and when the zero-sequence current is not available. The algorithms for impedance/distance calculation proposed assure correct relay operation, especially for the second and third zone, which provides useful backup for faults along the adjacent line. Thorough simulation tests performed with EMTP-ATP software have shown that the developed distance relay may face some problems under transformer internal inter-turn and ground faults, since such faults are most frequently not detected by the relay (they are mostly located in the second or even in the third zone). For such faults other principles and criteria should be used, employing e.g. the zero-sequence quantities [12] or the commonly applied restricted earth fault protection.
where k is the portion (0…1) of the line to be covered with the third zone. One should remember to select the right value of the correction coefficient n that should be adjusted to the reach of the third zone including protected transformer and the adjacent line or its part, i.e. Z +l /l Z n = ZT 0 + lIII / l ZL0 . If in (13c) one assumes full line length to be protected T 1 III L1 in the third zone (k = 1), then the correction factor should amount to ZT 0 + ZL0 n = Z +Z . T1 L1 For the phase-to-phase fault detected (I0 measured close to zero) there is only one zone to be set:
ZI = ZGT
ZI = 0.9Za = 0.9ZT1
References
(13d) [1] IEEE guide for protecting power transformers. IEEE Std C37.91-2008. [2] Ungrad H, Winkler W, Wiszniewski A. Protection techniques in electrical energy systems. New York: Marcel Dekker Inc.; 1995. [3] Ziegler G. Numerical distance protection. Principles and applications. 4th ed. Publicis Erlangen; 2011. [4] Ghorbani A. An adaptive distance protection scheme in the presence of phase shifting transformer. Electr Power Syst Res 2014;129:170–7. [5] Lizer M. Power unit impedance and distance protection functions during faults in the external power grid. Acta Energet 2012;4(13):22–33. [6] Bermann J, Prochazka M. Distance protection of big network transformers. In: Proceedings of the 17th international scientific conference on electric power engineering (EPE), Prague, Czech Rep.; 2016. [7] Frolova E, Osintsev A. Use of backup distance protection on a block transformer. Power Technol Eng 2016;50(2):220–3. [8] Kim HK, Kang SH, Nam SR, Oh SS. Improved Operating Scheme Using an IEC61850based Distance Relay for Transformer Backup Protection. In: Proceedings of the 2009 IEEE Bucharest PowerTech Conference, paper #131. [9] SIPROTEC 5, Transformer protection 7UT87 Manual, Siemens; 2016. [10] EMTP-ATP Manual. [11] Wiszniewski A, Solak K, Rebizant W, Schiel L. Calculation of the lowest currents caused by turn-to-turn short-circuits in power transformers. Int J Electr Power Energy Syst 2018;95:301–6. http://dx.doi.org/10.1016/j.ijepes.2017.08.028. [12] Wiszniewski A, Rebizant W, Schiel L. New algorithms for power transformer interturn fault protection. Electr Power Syst Res 2009;79:1454–61.
where ZGT is the impedance between relay location and transformer terminals. One should note that for the interphase short-circuits on the delta side the value measured represents the impedance between the fault and the location of the VTs. Therefore it is very small, unless the shortcircuits are inside the generator or transformer windings. Therefore, if the impedance measured becomes Za ⩽ ZT1 the fault may be diagnosed as an inter-phase short-circuit at the delta side. If all the loops measure low value of the impedance, then it is a three-phase short-circuit. If only one loop measures the low impedance, then the short-circuit may be diagnosed as a two-phase short-circuit between the phases which correspond to the voltage in the numerator of the impedance measurement algorithms. If the measured impedance is larger than ZT1, the shortcircuit may be diagnosed as a single-phase-to-ground fault at the star side. The faulted phase is the one, which corresponds to the voltages of the loop that measured the lowest impedance. The block scheme of the protection logic for this part is presented in Fig. 16. A2 – relay settings for the configuration with I0 not available 338
Electrical Power and Energy Systems 102 (2018) 332–339
D. Bejmert et al. Daniel Bejmert was born in 1979 in Walbrzych, Poland. He received his M.Sc., Ph.D. degrees from the Wroclaw University of Science and Technology, Wroclaw, Poland in 2004, and 2008, respectively. Since 2008 he has been a faculty member of Electrical Engineering Faculty at the WUST, at present at the position of the Assistant Professor. His research interests include: application of intelligent algorithms in digital protection and control systems, multicriterial and adaptive systems for power system protection and control and digital simulation of transient phenomena in power systems.
Ludwig Schiel was born in Weimar, Germany, in 1957. He studied Electrical Engineering at the Zittau Institute of Technology Zittau, Germany, finishing with the Dipl.-Ing. degree in 1984. In 1991 he received the Dr.-Ing. degree. In the same year he joined the Siemens AG, Germany. He is working in the Energy Management Division Digital Grid Automation Products concerning transformer protection.
Matthias Kereit was born in 1968 in Berlin, Germany. He studied electrical engineering in Berlin and graduated in 1992. He has been working with Siemens since 1992 and is now developing algorithms for digital protection devices, especially for distance protection relays. Since 2008 he has also been responsible for the university cooperations of the Berlin development department as a project manager.
Krzysztof Solak was born in 1982 in Sroda Slaska, Poland. He received his M.Sc. and Ph.D. degrees from the Wroclaw University of Science and Technology (WUST), Wroclaw, Poland, in 2006 and in 2010, respectively. At present he is an assistant professor in the Department of Electrical Power Engineering at WUST. His field of study is artificial intelligence (neural networks, fuzzy systems and genetic algorithms) for power system protection.
Waldemar Rebizant was born in Wroclaw, Poland, in 1966. He received his M.Sc., Ph.D. and D.Sc. degrees from Wroclaw University of Science and Technology (WUST), Wroclaw, Poland in 1991, 1995 and 2004, respectively. In 2012 he got the professorship degree. Since 1991 he has been a faculty member of Electrical Engineering Faculty at the WUST, at present at the position of the Dean of the Faculty. He is also Doctor Honoris Causa of the University of Magdeburg, Germany. In the scope of his research interests are: digital signal processing and artificial intelligence for power system protection. Prof. Rebizant has published more than 180 scientific papers, is an author of 11 patents and several patent applications. Details can be found at http://zas.pwr.edu.pl/WR.
Andrzej Wiszniewski graduated from the Electrical Engineering Faculty of the Wroclaw University of Science and Technology (WUST), Wroclaw, Poland, in 1957. In 1961 he received the Ph.D. degree, in 1967 the D.Sc. degree, and in 1972 the professorship in Electrical Engineering. All his life he has been working in the field of power apparatus and systems, being a specialist in protection and control of power systems. He is an author of 9 books and over 130 publications. All the time attached to the WUST he became the University Rector and had this position for 2 terms (1990 – 1996). In 1997 he took the position of the Minister for Science and had it for 4 years. Prof. Wiszniewski is a Distinguished Member of CIGRE and Honorary Member of the Polish Institution of Electrical Engineers.
339