Method for Setting the Resistive Reach of Quadrilateral Characteristics of Distance Relays Elmer Sorrentino Universidad Simon Bolivar, Venezuela
[email protected]
Eliana Rojas ABB, Venezuela
[email protected]
Abstract-A new method for setting the resistive reach on quadrilateral characteristics of distance relays is proposed in this article. The method is based on: a) analysis of the impedances seen by the relay (apparent impedances); and b) explicit definition of the protection desirable attributes for each analyzed zone (selectivity, sensitivity). In the proposed method, the resistive reach setting is calculated assuming that the reactive reach setting has been previously defined. The proposed method was applied in an example with 18 distance relays and its solution was compared with a conventional simplified solution. The conventional simplified solution consists in setting the resistive reach by multiplying the reactive reach by a constant factor. The result obtained with the proposed method is different since there is not a constant factor among the settings of the reactive and resistive reaches. Index Terms-Distance characteristic.
relay
setting,
Quadrilateral
I. INTRODUCTION
Traditionally, the distance relay zones have been set according to simple rules [1-4]. The non-traditional options can be grouped according to their conceptual basics: based on expert systems, mathematical optimisation, adaptive protection or probabilistic methods [5]. The well-known traditional setting rules have been developed to have a specific reactive reach for solid faults [1-5]. In the case of relays with quadrilateral characteristics, the reactive and resistive reaches can be set independently. For these relays, it is desirable to defme the resistive reach by an analysis of faults through impedance. Some traditional setting methods consider faults through impedance, but usually just a typical resistance value is considered [1,2]. Such methods usually do not consider that the apparent impedances are affected by diverse factors [6]. A still more simplistic option is to set the resistive reach by multiplying the reactive reach by a constant factor [7]. The main purpose of this article is to present a new method for setting the resistive reach on quadrilateral characteristics of distance relays. The proposed method is based on: a) analysis of the impedances seen by the relay (apparent impedances); and b) explicit definition of the protection desirable attributes for each analyzed zone (selectivity and sensitivity). II.
PROPOSED SETIING CRITERIA
A. Preliminary basic considerations
-It is considered that the reactive setting (XR) has been evaluated previously by traditional rules. Such rules are not universal, especially for the delayed zones; due to this fact, the rule used for each zone in this article will be specifically described here.
Jesus Hernandez Seneca, Venezuela
[email protected]
-The time delays for zones 2, 3 and 4 are assumed to be predefmed and fixed. Additionally, it is considered that there are not teleprotection schemes (communication-assisted trip), breaker-failure protection (50BF), line differential protection (87L), or other line protection functions. These considerations affect the criteria for setting the relay reaches. -The proposed method for setting the resistive reach (R0 can be adapted to the different ways of setting XR. The criteria used in this article for XR setting are relatively complex. This helps to better explain the proposed method. B. Analyzed quadrilateral characteristics
The quadrilateral characteristics may have different shapes; Fig. I shows the first quadrant for 3 cases. It will be assumed that the settings are defmed by the first quadrant in the R-X plane. For the sake of simplicity, shape of Fig. la will be used in this article. However, the techniques developed here can be applied to other characteristic shapes (such as those shown in Figures Ib and lc) by using an adaptation to the geometry of each particular characteristic.
Fig. I. Examples of different shapes for quadrilateral characteristics.
C. Setting the zone 1 resistive reach C.l. Criterion used/or the reactive reach The first criterion states that zone I only has to operate for faults on the line since this zone is instantaneous. Zone I should not operate for faults at the remote bus, by selectivity. Zone 1 reactive reach (XR1) will be set to 80% of the reactance of the protected line (X L+): XR1= 0.8 XL+. C.2. Criterion used/or the resistive reach According to the previous paragraph, zone 1 resistive reach (RRI) must be set in a way that assures that the relay first zone will not trip for faults at the remote bus. Considering the effect of the fault resistances (~) on the apparent impedance (ZAP), there are 3 cases: a) Faults at the remote bus whose ZAP tends to fall within zone I (Fig. 2a). To have a safety margin, the resistive setting will be limited to the value of the real part of ZAP (RRI-A) where the imaginary part of ZAP is 90% of XL+. b) Faults at the remote bus whose ZAP tends to be parallel to XR1 (Fig. 2b). It will be assumed that the possible error of measurement of the relay is proportional to the ZAP module. For this reason, when the ZAP imaginary part minus 5% of
978-0-947649-44-9/09/$26.00 ©2009 IEEE
ZAP module is 85% of X L+, the corresponding real part of ZAP limits the resistive setting (~I-B)' c) Faults at the remote bus whose ZAP tends to separate of X RI (Fig. 2c) : RRI is not limited by ZAP. RRI setting will be the smaller value of ~I-A and RRI-B, if both situations can happen. If the RRI setting is not limited by ZAP, RRI could be set to a very high value.
setting for the zone 2 reactive reach (X Z-MIN-Z) will be computed as 110% erx., (Xz_MIN_z=I ,IXL+)' -If X Z-AVG is greater than X Z-MIN-Z, there is not conflict between those values and the setting is: XR2=XZ-AVG. -IfXz-AVG is less than XZ-MIN-Z, it will be assumed that it is not possible to guaranty selectivity with that short adjacent line and the setting will be: XR2=XZ-MIN-Z. Actually, the solution for these cases is to implement a unit protection scheme for the short adjacent line (line differential and/or a scheme with teleprotection) and/or a change in the zone 2 delay for the line in study . The analysis of such solutions is beyond the scope of the present work .
a
If .Im{ZAP}=0,9XL+, => RR I .A = ~e{ ZA P }
Faults within the protected line
~:: : : ".':') .------------~
XR1=0,8XL+
R
If [.Im{ZAP }-0,05IZAPI1=0,85XL+, => RRI .B =~e{ ZA P }
- - - - - - - - - - - - -- ,
: I
RZ.MIN.Z
RZ.MIN.l
I
I I
ZL+,ADJ
R j)
I
Zona I of the adja cent line, as it is viewed by the analyzed relay
c
RRI is not limited by ZAP
RZ· MAX=0' 9RZ·RI·ADY
IFaults out of. the I protected lme
R
RZ-RI-ADY R Fig. 2. Criteriaused for settingthe zone 1 resistive reach.
D. Setting the zone 2 resistive reach D.l. Criterion usedfor the reactive reach It will be considered that the main objective of zone 2 is to cover the sector of the line that is not covered by zone I . This implies that the reactive reach should be set to cover more than 100% of the protected line impedance, in order to guaranty sensitivity for internal faults . This criterion is frequently used; however, it is usually necessary to take precautions that guaranty selectivity when there are adjacent short lines at the remote bus . This is because the beginning of zone 2 of the relay of the adjacent short line could overlap with the zone 2 of the relay in study . The setting of the zone 2 reactive reach (Xd will be done thus: -The desirable minimum setting for the zone 2 reactive reach (X Z-MIN-I) will be computed as 120% of the reactance of the protected line: XZ-MIN-I = 1.2 XL+. -The desirable maximum setting for the zone 2 reactive reach (X Z-MAX) will be computed as 80% of the total reactance seen by the relay for a fault at the beginning of zone 2 of the adjacent line protection at the remote bus (XL+,ADJ,SHORT). The case with the smaller additional reactance will be used: X Z-MAX = 0.8 (X L++0 ,8XL+, ADJ,SHORT). -If X Z-MAX is greater than X Z-MIN-" then there is not conflict between those values and the setting will be : X R2 = XZ-MIN-I . -If X Z-MAX is less than XZ-MIN-" the desired sensitivity is not possible without a lack of selectivity, and : -The average of the previous values will be computed (XZ-AVG=[XZ-MIN-I+XZ-MAX]/2). The allowable minimum
Fig. 3. Limits for the settingof the zone2 resistive reach.
D.2. Criterion usedfor the resistive reach The setting criterion of zone 2 resistive reach (RR2) is similar to the criterion described for X R2 . A desirable sensitivity will be defined to cover faults in 100% of the protected line, with the typical fault resistance value (R F-TyP) multiplied by a safety factor (F sl) . An allowable minimum sensitivity will be defmed using a smaller safety factor (Fsz): Fsz
-If RZ- MAX is greater than RZ-MIN-Z, then there is not conflict between those values and the setting will be: RR2=Rz-MAX • -If RZ-MAX is less than RZ-MIN-Z, then it is not possible to guaranty selectivity for some values of fault resistance and the setting will be: RR2=Rz-MIN-Z. D.3. Comments about both criteria In both setting criteria, if the first condition is satisfied (X Z-MAX>XZ-MIN-" or RZ-MAX>Rz-MIN-I), then a different action could be taken, in order to increase still more the zone 2 sensitivity. For example, in such cases the setting could be the maximum value instead of the minimum, or an average of both values. An analysis of those options is outside the scope of the present work; however, Fig. 4 helps to illustrate this concept. In the example of Fig. 4, it is assumed that the resistive setting (Rd has been limited by the allowable minimum sensitivity (RZ-MIN-Z) ' In such case, an increase of the reactive reach sensitivity (to use XR2-CASEZ, instead of XR2-CASEI) would imply a greater lack of selectivity for resistive faults in the adjacent line if the fault is out of the zone 1 of the adjacent line protection.
~
I Zone I of the adjacent line I jx XR2-CASE2 ~~ ~~~~~~~~~~ -~ - ~ ~ ~ ~ ~ ~ ~ B Increase er x, implies , XR2-CASEI : a lack of selectivity if ZL+ : ZAP .ISIn . th'ISregIon . ZL+,ADJ
~
eL+
i
of the apparent impedance (ZAP) seen by the relay in study is 110% of XR3, or if the imaginary part of the apparent impedance minus 5% of the ZAP module is 105% of XRJ , the corresponding real part of ZAP will be the value of resistive reach (Rd. As zone 1, if the RR3 setting is not limited by ZAP, then RRJ could be set to a very high value.
I ZAP(varyIng . R) F
jX (FINFEED)ZL+,AD.J
~
I
XR3=O,8(X L++(f' INFEED)XR2-ADJ,SHORT) If: (Im( ZAP}=I,IXR3) R
OR ([Im( ZAP} -O,051ZAPI]=I,05XR3), ~
RR3=>\'f(ZAP}
Fig. 5. Criteria used for setting the zone 3 resistive reach.
III. SYSTEM USED AS EXAMPLE A. Power system description
Figure 6 shows the power system used as example and its data are in Tables I, II and III. C I:26km
R LCA
C I:3.37km PLM Fig. 6. Power system used as example (II5kV).
RR2-MIN-2
Fig. 4. Example of a lack of selectivity by increasing XR2 sensitivity .
E. Setting the zone 3 resistive reach £.1 . Criterion usedfor the reactive reach It will be assumed that the main objective of zone 3 is to operate as backup protection for faults in adjacent lines [8]. However, selectivity between zones 3 of different lines will have priority because zone 3 is the faster backup function. This criterion presupposes that the faults non-covered by a zone 3 as backup will be covered by its zone 4, that is more sensitive (zone 4 has a greater reach or it is simply a directional function). Zone 3 reactive reach (XR3) should be set at 80% of the lowest total apparent reactance seen by the relay in study for faults at the end of zone 2 of the relays that protect adjacent lines. Worst case combines the smaller zone 2 reactive reach of the relays of the adjacent lines (XR2-ADJ,SHORT) with the smaller infeed (FINFEED, due to the current contributions at the remote bus): XR3=0.8 (XL++ (FINFEED) XR2-ADJ,SHORT). To fmd the previous value may not be simple. For the sake of simplicity, XRJ was set to 75% of the smallest total apparent reactance for faults at the end of the adjacent lines to the remote bus: XR3 = 0.75 (XL++XAP-ADJ-LOWEST). Base Case of load flow was used; therefore, the FINFEED values correspond to the Base Case. E.2. Criterion usedfor the resistive reach Zone 3 resistive reach (RRJ, Fig. 5) is set similarly to RRI. The resistive faults were computed at the end of the same adjacent line used for the XR3 setting. If the imaginary part
TABLE I: LINE PARAMETERS (r,x in Q/km; b in umho/km), r, bo b+ x+ ro Xo 0.1211 0.4959 3.347 0.3160 1.102 1,938 Cl C2 0.1714 0.4928 3.421 0.3630 1.151 1.860
LCA GUA LM
P(MW) cos(
TABLE II ' EQUIVALENT GENERATOR DATA P(MW) X+=K(Q) Xo(Q) 7.3 3.3 Slack 15.9 15.9 120 120,0 53.0 20 LCA 73 0.900
TABLE Ill' Loxn DATA LM LA PMT LR 48 31 38 56 0.900 0.900 0.900 0.936
Q(MVAR) Slack
74.37 12.39 PLM 38 0.900
GUA 30 0.850
B. Ground distance function description The apparent impedance seen by ground distance function (ZPh-G) depends on its polarisation method [9]. It is assumed that the relay uses the following form of polarisation: ZPh-G = VPh-G / (IPh+KoI~ (I) VPh-G: Phase-to-earth voltage of the faulted phase. Iph: Current of the faulted phase. IR: Residual current (IA + IB + Ie). Ko: Residual compensation factor. It is assumed that Ko is set exactly to see the positive sequence line impedance for solid faults: Ko = (ZLO - ZL+) / (3 ZL+) (2) ZL+: Positive sequence line impedance. ZLO: Zero sequence line impedance.
c.
Pre-fault loadflow
With fault resistance (R F), apparent impedance depends on the pre-fault load flow, measured in the relay locality [6],[9]. The determination of the worst possible condition for each zone of each relay is outside the scope of the present article. By simplicity, a simple preliminary analysis of the system in study suggested the use of the following conditions of prefault load flow: -Base Case: It is the system described in section III-A. -Case 1: It is the Base Case without one transmission line. For the system in study, the approximated load flow values at the relay localities are in Table IV (QMAX). -Case 2: This case is as Case 1 and, additionally, this case presupposes that the system operators can control the reactive power flow. For the system in study, a half of the previous reactive power value was assumed; the values are in Table IV (QMIN). These cases were used thus: a) For setting zone 1, Case 1 was used when the pre-fault load flow is positive and Case 2 when it is negative; for the system in study, by coincidence, P and Q have the same sign in the simulated cases. b) For setting zones 2 and 3, Base Case was used.
D. Typical groundfault resistance value The ground fault resistance value depends on multiple factors. Each R F value has a probability of occurrence [10]; however, a typical value is required in the present work in order to compute the desirable zone 2 reach. Such value was supposed arbitrarily (R F_TYP=50). Using safety factors, the results are: RFl=(Fsl)RF-TYP=200; RF2=(Fs2)RF-TYP= 100.
The settings obtained for zone 2 are in Table VI. RR2 is, in some cases, less than RRI (it happens at bus 2 of the lines LM-LA, PLM-LR, LCA-PLM, LCA-LR). This result is illustrated in Fig. 7, it is not conventional and it happens since RR2 must be limited to reduce the risk of lack of selectivity (since X R2 is greater than X R1). The settings obtained for zone 3 are in Table VII. RR3 is, in some cases, less than RR2 (it happens at bus 1 of the lines GUA-LM, LA-PMT, and at bus 2 of LA-PMT, LR-PMT). This result is similar to the result described for zone 2. If RR3 is less than RR2' then zone 2 is more sensitive for faults with fault resistance. In such cases, as zone 3 is not so sensitive, the backup function for faults with a high value of fault resistance is the zone 4. By this reason, zone 4 would have to be sufficiently sensitive in these cases. In the cases where RR2 is less than RRI (or RR3 is less than RR2), this nonconventional result could be avoided if the desired condition were imposed. Such condition is: the greater the value of the reactive setting, the greater must be the value of the zone resistive reach. There are different ways for imposing such condition. Special care should be taken to update the resistance setting values since there are dependences among the reach settings; for example, Fig. 3b shows how the RR2 setting of a relay depends on the value of RRI of the relay of an adjacent line. IV: PRE-FAULTLOAD FLOW, AT THE LOCALITY OF THE RELAYS IN STUDY, FOR THE CASES 1 AND 2 (QMAX AND QMIN, RESPECTIVELY)
TABLE
Prefault load flow (line in study)
Line
Line
in
out of
PandQ
P
OMAX
OMIN
study
service
direction
MW
MVAR
MVAR
IV. RESULTS FOR THE RELAY SETTINGS
GUA-LM
PLM-LCA
GUA->LM
The settings obtained for zone 1 are in Table V. The X R1 values are identical to those obtained in another study [11] since the used criterion is exactly the same. However, in that study a unique factor of RIJX R was used (RIJXR=2). Table 5 shows clearly that RIJX R is not constant with the developed method: RIJX R varies between 0.77 and 33.41. On the other hand, the maximum value of fault resistance, at the remote bus, for which the adjustment of the RRI was defmed, varies between 0.970 and 17.390. However, there are 5 cases where the resistive reach was limited by the line thermal capability and not by the apparent impedance locus.
LM-LA
PMT-LR
LM->LA
LA-PMT
PMT-LR
LA->PMT
LA-PMT
LM-LA
LA->PMT
PMT-LR
PMT-LR
LR->PMT
LR-PLM
PLM-LCA
PLM->LR
90 70 38 -31 70 87 -38 125 90 68 125
56 33 18 -15 33 36 -18 54 56 33 54
28.0 16.5 9.0 -7.5 16.5 18.0 -9.0 27.0 28.0 16.5 27.0
TABLE V : ZONE
GUA-LM LM-LA LA-PMT LR-PMT PLM-LR LCA-PLM LCA-GUA LCA-LM LCA-LR
LCA-LR
PLM->LR
LCA-LR
LCA->PLM
LCA-GUA
GUA-LM
LCA->GUA
LCA-LM
GUA-LM
LCA->LM
LCA-LR
PLM-LCA
LCA->LR
SETTINGS. THE FAULT RESISTANCE THAT DEFINED THE RESISTIVE REACH SETTING (FIG. 2) IS SHOWN. VALUES IN PRIMARY OHMS.
Line (Bus 1Bus2J
LR-PLM PLM-LCA
Bus 1
XR1 10.31 4.81 2.41 2.71 1.34 3.95 10.31 8.73 3.98
Bus 2
RRl
R F MAX
RR1/ XRl
12.11 7.75 4.39 4.95 2.98 8.30 16.60 14.17 8.39
3.86 8.71 4.98 5.67 2.70 8.62 6.16 8.77 8.54
1.17 1.61 1.82 1.83 2.23 2.10 1.61 1.62 2.11
. .
XR1 10.31 4.81 2.41 2.71 1.34 3.95 10.31 8.73 3.98
RRl
R F MAX
RR1/ XRl
18.06 88* 5.06 88* 104.79 132* 7.89 132* 132*
5.56
1.75 18.30 2.10 32.49 78.38 33.41 0.77 15.12 33.17
(*): These values were only limited by the line thermal capability (Fig. 2c).
N/A 5.92
N/A 17.39
N/A 0.97
N/A N/A
TABLE VI: ZONE 2 SETIINGS (SEE FIG. 3). VALUES INPRlMARY OHMS.
Line
Bus 2
Bus 1.
(Bus 1.Bus2J GUA-LM LM-LA LA-PMT LR-PMT PLM-LR LCA-PLM LCA-GUA LCA-LM LCA-LR
X R2 14.82 6.98 3.61 4.06 2.01 5.47 14.47 12.83 5.51
X R2 15.47 7.21 3.61 3.92 2.01 5.93 14.47 12.49 5.97
RR2
25.28 15.75 42.03 11.14 19.34 13.77 31.42 22.19 14.62
RR2
46.42 40.39 27.97 88* 31.33 42.20 60.14 132* 76.59
(") : These values were only limited by the line thermal capability. .I·x
X R2
i~: - - - - - - - - - - - - - - - - - - - - - - - - - - :
~
- - - -x;~ - - - - - - - - - - - - - - - - - - - - t - - - - - - - -:
8 L+
I
I
I
I
R
: : R R2
R RI
Fig. 7. Example of a nonconventional result: XR2 can be smaller than XRl. TABLE VII: ZONE 3 SETIINGS(SEE FIG. 5), VALVES INPRlMARY OHMS.
Linea (Ext. 1.Ext.2J GUA-LM LM-LA LA-PMT LR-PMT PLM-LR LCA-PLM LCA-GUA LCA-LM LCA-LR
Ext. 2
Ext. 1. XR3 18.01 7.55 5.18 5.33 8.26 5.57 21.92(1) 22.34 6.05
RR3
19.58 19.88 11.32 14.28 43.46 132* 50.24 67.41 132*
X R3 21.92(1) 44.89 7.36 4.39 4.19 8.43(1) 33.16 13.47(1) 8.43(1)
the variations in the results. On the other hand, the effect of the inclusion of more cases for the pre-fault load flows should be studied .
RR3
132* 88* 15.32 39.25 132* 132* 132* 132* 132*
(*): These values were only limited by the line thermal capability. (1): These values were not found by the criterion indicated in section Il-E, since it was not being possible (the reactances would be negative). To find them, the smaller reactance ofadjacent lines was added to XL+ and the total value was multiplied by O.75.
V. CONCLUSION -A novel method for setting the resistive reach of quadrilateral characteristics in distance relays was presented. The method is based on the analysis of the apparent impedance seen by the relay, and the explicit definition of the protection desirable attributes for each analyzed zone. -The proposed method was applied to an example with 18 distance relays and the resistive reaches for 3 relay zones were calculated. The results obtained with the proposed method were compared with a conventional simplified solution. The conventional simplified solution is to set the resistive reach by multiplying the reactive reach by a constant factor . The results obtained with the proposed method are substantially different since there are particular solutions for each relay location . -In the future, this work could be complemented of diverse ways. For example, an analysis of other criteria to set the reactive and resistive reaches could be done in order to study
REFERENCES [I] [2]
Areva T&D, "Network protection & automation guide," 2002. G. Ziegler , "Numerical distance protection . Principles and application s," Siemens AG, 1999. [3] ABB, "Protective relaying. Theory and applicat ions," Marcel Dekker Inc, 1994. [41 R. Mason, 'T he art and science of protective relaying," John Wiley & Sons Inc, 1956. [5] V. De Andrade, E. Sorrentino, "Revision bibliografica sobre los metodos para ajustar el alcance de los reles de distancia" (in Spanish), Proceedings of the I CVREE , Lecheria, Venezuela, 2007. [6] T. Rodolakis , D. Crevier , "Effect of loads, shunts and system uncerta inties on short circuit relay settings," IEEE Trans. on PAS, Dec. 1981, pags, 4701-4709 . [7] ABB Relay, "Distance Relay Type Razoa," 1985. [8] S. Horowitz , A. Phadke, "Third zone revisited," IEEE Trans. on PWRD,Jan. 2006, pags. 23-29 . [9] E. Sorrentino , "Polarizacion de la funcion de distancia ante fallas a tierra y su efecto sobre el alcance resistivo en zonas cuadrilateras" (in Spanish), Proceedings of the XII ERIAC, Foz do Iguazu, Brasil, 2007. [10] J. Barnard, A. Pahwa, "Determination of the impacts of high impedance faults on protection of power distribution systems using a probabilistic model ," EPSR, 1993, pags. 11-18. [11] E. Rojas, "Coordinacion de las protecciones de distancia del sistema a 115 kV de Seneca incluyendo El Guamache " (in Spanish), Final Project for BEE, Universidad Simon Bolivar, Venezuela, 2007.
