PRODUCED BY THE OPERATIONS DIRECTORATE OF ENERGY NETWORKS ASSOCIATION
Engineering Recommendation G89
Issue 1 2011
Specification of DC Time Constants for
Switchgear
© 2011 Energy Networks Association
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system or transmitted in any form or by any means,
electronic, mechanical, photocopying, recording or otherwise, without
the prior written consent of Energy Networks Association. Specific
enquiries concerning this document should be addressed to:
Operations Directorate
Energy Networks Association
6th Floor, Dean Bradley House
52 Horseferry Rd
London
SW1P 2AF
This document has been prepared for use by members of the Energy
Networks Association to take account of the conditions which apply to
them. Advice should be taken from an appropriately qualified engineer
on the suitability of this document for any other purpose.
ENA Engineering Recommendation G89
Issue 1 February 2011
Contents
Foreword .................................................................................................................................... 2
Scope ......................................................................................................................................... 2
Report Recommendations .......................................................................................................... 2
Introduction ................................................................................................................................ 3
Key elements of fault current for switchgear ratings ................................................................... 3
Fault level studies undertaken by ENA members ....................................................................... 7
Analysis of the survey results ..................................................................................................... 8
Derivation of recommended ratings ............................................................................................ 9
Conclusions.............................................................................................................................. 12
Appendix 1 - Test Network for Benchmark Study ..................................................................... 13
Appendix 2 - Comparison of Study Results from Benchmark Network ...................................... 20
Appendix 3 - Network survey results ........................................................................................ 27
Appendix 4 - Benchmark network results ................................................................................. 28
Appendix 5 - Distribution of DC time constants ......................................................................... 32
Page 1 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Specification of DC Time Constants for Switchgear
Foreword
International standards define circuit-breaker terminal fault making & breaking capability in
terms of a rated ac RMS current and a 100% dc component decaying with a single time
constant. The standard value of dc time constant defined in circuit-breaker standard IEC 62271-
100 is 45ms; a value which corresponds well to the dc time constant of overhead lines but
which is significantly less than that of transformers (typically well in excess 100ms). Special
case alternative values of 75ms and 120ms which can be selected by users are also included in
the standard. In recent years ENA members have become concerned that the specification of a
45ms dc time constant may be inadequate for the UK network at 132kV and below and an ENA
working group was set up to establish the requirements of the network (via fault level studies)
and develop appropriate switchgear specifications. This report presents recommendations for
revised switchgear specifications for incorporation into ENA Technical Specification documents
in conjunction with the key elements of the work which justify these recommendations.
Scope
This Engineering Recommendation applies to switchgear at UK primary substations and above;
it does not apply to UK distribution substations.
Report Recommendations
ENA should specify switchgear short-circuit ratings at two values of DC time constant; 45ms
and 120ms.
ENA should accept that ratings proven at a DC time constant of 45ms can be extrapolated up to
a DC time constant of 90ms but not beyond.
ENA should accept that ratings proven at a DC time constant of 120ms can be extrapolated up
to a DC time constant of 270ms but not beyond.
ENA should accept that, providing adequate evidence of performance at 45ms and 120ms
exists, interpolation can be used to derive ratings in the range 90-120ms.
Peak make/withstand ratings should be specified in accordance with IEC requirements based
on the highest specified AC RMS short-circuit current and the associated dc time constant.
ENA should adopt the switchgear short-circuit ratings presented in Table 8 of this
recommendation for use within ENA Technical Specifications namely:
Time 132kV 3 132kV
66kV 33kV 20-25kV 11kV 6.6kV
constant ph 1ph
45 ms 40kA 40kA 40kA 31.5kA 20 kA 25kA 25kA
120 ms 31.5kA 31.5kA 31.5kA 20kA 12.5kA 16kA 16 kA
Switchgear short circuit breaking current ratings for incorporation in ENATS
Page 2 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Introduction
The following report is presented in three main sections which describe:
The key elements of fault current for switchgear ratings
The fault level studies undertaken by ENA members
The analysis of the survey results
The derivation of recommended ratings for inclusion within ENATS
It is important to note that this report is presented on the basis of worst case, single phase fault
conditions which result in the most onerous potential fault current interruption conditions for the
circuit-breaker.
Key elements of fault current for switchgear ratings
Whilst real fault currents are complex in nature, particularly when considering the DC elements,
they can be adequately described as the summation of three distinct components as follows:
AC synchronous fault level. This is the “steady state” fault contribution of connected sources
which does not vary within the timescale of interest within this report.
AC transient fault in-feeds. This is the contribution from local rotating equipment, such as
induction motors, which decays within the timescale of interest within this report. Further details
of this contribution and the modelling of it can be found in ER G74.
DC component. This is a component which is generated by inductive elements of the system in
response to the virtually instantaneous change in AC current at fault initiation. This component
decays with a time constant dictated by local system components and is typically in the range
40ms to 200ms depending upon the balance between resistive and inductive elements.
The AC contributions to the fault current waveform are illustrated in Figure 1. This figure shows
the fault occurring at the time of peak current (voltage zero) and these contributions add
together as shown in Figure 2.
Page 3 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Fault current AC components
60 kA
40 kA
20 kA
Synchronous
AC component
G74 infeed
0 kA
-20 ms 0 ms 20 ms 40 ms 60 ms 80 ms 100 ms
-20 kA
-40 kA
Figure 1: Typical AC fault current contributions
Total AC component
60 kA
40 kA
20 kA
Total AC
AC envelope
0 kA
-20 ms 0 ms 20 ms 40 ms 60 ms 80 ms 100 ms
-20 kA
-40 kA
Figure 2: Total AC fault current contribution
It can be clearly seen that the envelope of the AC current decays from its initial magnitude to
the steady state value over a short time following fault initiation.
Page 4 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
The above figures show the current instantaneously rising to a high value. However, due to the
system inductance, the current cannot instantaneously change as shown and a DC component
is generated to counteract this sudden change. The DC component decays in a relatively
complex manner however, for the purposes of simplification when specifying switchgear, it is
considered to decay with a single time constant governed by the X/R ratios of the circuits
feeding the fault.
Fault current AC and DC components
60 kA
40 kA
20 kA
Total AC
DC component
0 kA
-20 ms 0 ms 20 ms 40 ms 60 ms 80 ms 100 ms
-20 kA
-40 kA
Figure 3: AC & DC fault current contributions
Figure 3 shows both the AC and DC components and Figure 4 shows the resultant total fault
current and highlights the key features which are relevant for switchgear performance and
hence specification. These can be described as follows:
Page 5 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Total fault current
60 kA
Initial peak (Ipk)
Major loop
40 kA
20 kA
Total current
0 kA
-20 ms 0 ms 20 ms 40 ms 60 ms 80 ms 100 ms
-20 kA Minor loop
-40 kA
Figure 4: Total fault current
Initial peak current. This is the peak current, occurring at approximately 10ms from fault
initiation (50Hz system), which all series connected equipments must be able to withstand and
which circuit-breakers must be able to safely close onto (termed making). It is largely
determined by the values of the synchronous fault current and the G74 in-feed current. The
value of the time constant has a limited effect, with longer time constants increasing the value
slightly.
Total breaking current during arcing period. During opening the circuit-breaker will be
subject to the asymmetric current as shown in Figure 4 and, in particular, to a sequence of
major & minor current loops. The current behaviour during the very early part of the fault, when
the circuit-breaker contacts remain closed, is of limited interest. However, following contact
separation the circuit-breaker is required to control the resultant arcing until a suitable current
zero for interruption occurs and also to withstand that resultant transient recovery voltage. Due
to the relatively high energy associated with the major loops of fault current and the irregular
occurrence of current zeroes this is a potentially onerous requirement for circuit-breakers.
It is widely accepted [1] that, in order to assess the ability of circuit-breakers to operate where
the effective system DC time constant exceeds the tested value, a comparison of the first full
major loop of fault current following contact separation is appropriate. In doing this, the following
aspects of the major loop must be considered.
Peak breaking current. This is determined by the value of the synchronous fault current and
the DC time constant. A longer time constant results in a higher degree of asymmetry and has a
significant effect on the value of the peak current.
Arc energy. This is a function of the current and voltage in the major loop. It is widely accepted
that the arc voltage can be considered constant and hence that a comparison of the major loop
area is sufficient to compare arc energies.
Duration of major loop. The length of the major (and minor) loops and hence the arcing time
characteristics of the circuit-breaker during interruption are affected by the effective DC time
constant. Higher DC components significantly increase the length of major loops.
Page 6 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
On the basis of the above approach, fault level studies on the UK networks in the range
between 132kV and 11kV have been used to derive recommendations for future ENA
switchgear short-circuit current specification. Further details of the method used are presented
in later sections of this document.
Fault level studies undertaken by ENA members
In order to obtain relevant data for assessment each DNO, plus National Grid where relevant,
was asked to provide the fault level and X/R ratio for every circuit-breaker on their network
within the scope of the project.
Table 1 outlines the total number of circuit breakers and the major circuit configurations used in
the survey.
Voltage (kV) Number of Number of Breakers on Not on
Sites Breakers Transformer Transformer
LV LV
132 348 1648 915 733
66 140 787 311 476
33 1078 7457 5912 1545
20 - 25 95 596 548 48
11 2829 27603 27540 63
6 - 6.6 164 1645 1613 32
Totals 4654 39736 36839 2897
Table 1: Number of circuit-breakers used in the survey of fault data
It was identified early in the process that, if a national survey of fault levels was undertaken, the
results provided by each member company might not be directly comparable with each other
due variations in tools and methodologies. In an attempt to mitigate this, a benchmarking
exercise was undertaken to ascertain the variation in results for a pre-determined set of data.
A simple test network was devised and all member companies used their existing modelling
tools and methodologies to calculate the prospective make and break duties. A benchmark set
of results was also calculated using an electromagnetic transient programme to accurately
calculate the fault current characteristics of the network. The network used for the
benchmarking exercise is shown in Appendix 1. The results, including the reference results from
EMTP, are given in Appendix 2. The results of this indicate that:
1. Values of initial peak and ac currents in the majority of cases are close to the true values;
2. Values of %dc and peak break currents (which are directly related) show greater variation,
with some giving more onerous, and others less onerous, duties. However the majority are
close to the true results.
Note: the results of the benchmarking exercise led some member companies to review their
methodologies for calculating fault levels.
The results of the survey showed significant variations in the results obtained by member
companies. While the results of the benchmarking exercise demonstrated differences in
calculated fault levels and DC time constants, comparison of the benchmarking results indicates
that the effect of this will be to broaden the range of time constants reported, but not to affect
Page 7 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
the ac fault current values significantly. The working group agreed that it was up to individual
member companies to satisfy themselves that they were happy with their calculation methods
and associated results.
It should also be noted that, being based upon an intact network condition, not all system
conditions are fully addressed. In particular, recent experience suggests that the survey fault
levels do not adequately cover for 132kV National Grid/DNO interface sites in non-standard
configurations (e.g. four transformer sites which go from 2+2 to 3+1 running during outage
works). A 31.5kA, 120ms rating at 132kV has been incorporated into the recommendations of
this report to cater for this. This is not expected to alter the equipment offered to DNO’s or
ENASAP for assessment.
Analysis of the survey results
The key data from the survey is presented in Appendices 3, 4 & 5. Appendix 3 summarises the
results of the survey. Appendix 4 shows the fault level at each circuit-breaker plotted against the
effective DC time constant for the various cases (the DC time constant is 3.18 times the X/R
ratio). Appendix 5 presents the distribution of DC time constants analyses by frequency of
occurrence.
Tables 2 and 3 summarise the fault levels and DC time constants respectively falling within
90%, 95% and 99% percentiles i.e. those values below which 90%, 95% and 99% of the survey
results lie.
Percentile
values 132kV 132kV
66kV 33kV 20-25kV 11kV 6.6kV
3 ph 1 ph
90% 17.6 kA 21.3 kA 16.8 kA 16.0 kA 10.6 kA 11.5 kA 17.5 kA
95% 18.7 kA 22.8 kA 17.2 kA 17.7 kA 11.0 kA 12.5 kA 19.2 kA
99% 19.4 kA 24.5 kA 17.2 kA 21.5 kA 13.0 kA 14.8 kA 23.5 kA
Table 2: Analysis of Fault levels from survey
Percentile
values 132kV 132kV
66kV 33kV 20-25kV 11kV 6.6kV
3 ph 1 ph
90% 148 ms 120 ms 139 ms 110 ms 97 ms 93 ms 89 ms
95% 158 ms 135 ms 148 ms 131 ms 133 ms 124 ms 112 ms
99% 248 ms 157 ms 231 ms 172 ms 160 ms 235 ms 293 ms
Table 3: DC time constants
Page 8 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Derivation of recommended ratings
A CIGRÉ paper [2] has proposed that the capability of circuit-breakers to interrupt a fault current
with a higher DC time constant than tested may be based on the following criteria:
The major loop peak current during arcing is no higher than that tested.
The arcing energy (equivalent to the integral of current) is no higher than that tested.
The value of dI/dt (which determines the transient recovery voltage) is no higher than
that tested.
In practice, the last condition does not represent a limitation since a higher degree of asymmetry
results in current zero crossings which are closer to the current peaks and hence the rate of
change of current will be lower than would be the case for lower degrees of asymmetry.
Against these criteria, the following procedure has been used to evaluate the capability of
circuit-breakers at DC time constants higher than 45 ms (the standard value for IEC 62271-
100).
1. IEC 62271-100 specifies that circuit-breakers should be tested for asymmetrical duties at
an opening time equivalent to the minimum opening time of the circuit-breaker plus 10
ms. On this basis it is considered that the first realistic major loop to be considered will
be the loop centred on 50ms. To cover a variety of credible cases calculations have
been performed for the major loops centred on 50ms, 70ms and 90ms.
2. For time constants from between 45ms and 300 ms the times for zero crossings at the
start and finish of the major loops were calculated.
3. The integral of the arcing current for this period was calculated for a 1kA fault current.
4. For each major loop (50ms, 70ms, 90ms) the peak currents were calculated for the
range of time constants.
From this procedure it was concluded that the limiting factor in all cases is the arc energy. For
each value of time constant an AC de-rating factor was calculated to ensure that the integral of
arcing current (arc energy) was no greater than that for the reference current with 45ms DC time
constant. By plotting these values graphically it was shown that a lower rating (higher de-rating
factor) was derived based on the 90 ms major loop than the 50ms and 70ms loops. The 90ms
loop was therefore chosen as the basis for further work.
Using Microsoft Excel it was found that a polynomial equation of the following form gave a good
fit for the relationship between the de-rating factor and the DC time constant over the range of
time constants considered:
I % a b c 2 d 3 e 4 f 5 (Formula for extrapolation from 45 ms ratings)
Where the coefficients are:
a 1.5803
b -0.018476
c 1.5101 x 10-4
d -6.7533 x 10-7
e 1.5574 x 10-9
f -1.4456 x 10-12
Page 9 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Using this approach and the data from the fault level survey, AC RMS ratings (based on a
tested DC time constant of 45ms) were calculated that would be sufficient to meet the 90th,
95th & 98th percentile of studied system requirements. It is important to note that “negative” de-
rating factors i.e. increases in AC RMS capability on the basis of DC time constants of less than
the reference value (45ms) are not permitted in the assessment.
The resultant values are shown in Table 4 and the derating curves for these ratings are plotted
on the graphs in Appendix 4
Percentile
values 132kV 132kV
66kV 33kV 20-25kV 11kV 6.6kV
3 ph 1 ph
90% 23.1 kA 25.7 kA 24.4 kA 22.2 kA 14.4 kA 13.4 kA 19.2 kA
95% 25.7 kA 26.2 kA 26.3 kA 25.1 kA 15.1 kA 16.2 kA 21.3 kA
99% 26.6 kA 29.1 kA 26.4 kA 28.7 kA 18.0 kA 21.2 kA 25.6 kA
Table 4: Ratings (tested DC time = 45ms) to address 90/95/99% of study requirements
A further constraint placed upon extrapolation of type test evidence to address increased DC
time constants pertains to the percentage increase in length of the major loop. Whilst this
aspect has not been widely recognised internationally, an unlimited increase in major loop
length implies that circuit-breakers are not affected in any way by the time at which current
zeroes appear. Whilst this may be true for some technologies it is not a generally applicable
assumption and hence some provision should be made for it within the extrapolation & rating
assessment process. To date a value of 15% has been used by some ENA member companies
as the maximum allowable increase in major loop duration from type test to service conditions.
This allows, for example, an extrapolation from a tested time constant of 45ms to a service
condition of 90ms but not beyond.
The adoption of this limit of 15% is recommended here and thus, assuming standard type
testing based on 45ms, system requirements in excess of 90ms create the need for type testing
at a higher DC time constant.
IEC 62271-100 allows alternative time constants of 60ms, 75ms and 120ms to be specified.
Whilst a time constant of 120 ms does not directly cover all of the requirements of the networks,
applying the 15% increase in loop length would allow ratings base on 120ms to be extrapolated
up to DC time constants of 270ms; well in excess of the vast majority of system requirements.
In line with the earlier discussed procedure, a formula has also been derived for extrapolating
ratings from 120 ms ratings, as follows
I % a b c 2 d 3 (formula for extrapolation from 120 ms ratings)
Where the coefficients are:
a 1.609
b -0.007669
c 2.5056 x 10-5
d -2.9932 x 10-8
Page 10 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Table 5 below shows the equivalent ratings for the circuit-breaker sites in the DNO survey,
using a DC time constant of 120ms.
Percentile
values 132kV 132kV
66kV 33kV 20-25kV 11kV 6.6kV
3 ph 1 ph
90% 15.2 kA 16.9 kA 16.7 kA 14.6 kA 9.5 kA 8.8 kA 12.6 kA
95% 16.9 kA 17.2 kA 17.3 kA 16.5 kA 9.9 kA 10.7 kA 14.0 kA
99% 17.5 kA 19.2 kA 17.3 kA 18.9 kA 11.9 kA 14.0 kA 16.9 kA
Table 5: Ratings (tested DC time = 120ms) to address 90/95/98% of study requirements
Within IEC, standard ratings are selected from the R10 series. Tables 6 & 7 re-present the data
of Tables 4 & 5 with current values enhanced to an appropriate standardised value.
Percentile
values 132kV 132kV
66kV 33kV 20-25kV 11kV 6.6kV
3 ph 1 ph
90% 25 kA 31.5 kA 31.5 kA 25 kA 16 kA 16 kA 20 kA
95% 31.5 kA 31.5 kA 31.5 kA 31.5 kA 16 kA 20 kA 25 kA
99% 31.5 kA 31.5 kA 31.5 kA 31.5 kA 20 kA 25 kA 31.5 kA
Table 6 Circuit-breaker equivalent R10 ratings for 45 ms DC time constant
Percentile
values 132kV 132kV
66kV 33kV 20-25kV 11kV 6.6kV
3 ph 1 ph
90% 16 kA 20 kA 20 kA 16 kA 10 kA 10 kA 16 kA
95% 20 kA 20 kA 20 kA 20 kA 10 kA 12.5 kA 16 kA
99% 20 kA 20 kA 20 kA 20 kA 12.5 kA 16 kA 20 kA
Table 7 Circuit-breaker equivalent R10 ratings for 120 ms DC time constant
On the basis of a required level of coverage (90/95/99%) these tables provide the minimum
recommended short-circuit capability that should be incorporated into ENATS specifications.
However for 132kV and 66kV circuit-breakers higher ratings are recommended in order to allow
for abnormal running arrangements due to SGT outages.
Page 11 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Conclusions
ENATS should specify short-circuit ratings at two values of DC time constant; 45ms and 120ms.
ENA should accept that ratings proven at a DC time constant of 45ms can be extrapolated up to
a DC time constant of 90ms but not beyond.
ENA should accept that ratings proven at a DC time constant of 120ms can be extrapolated up
to a DC time constant of 270ms but not beyond.
ENA should accept that, providing adequate evidence of performance at 45ms and 120ms
exists, interpolation can be used to derive ratings in the range 90-120ms.
Peak make/withstand ratings should be specified in accordance with IEC requirements based
on the highest specified AC RMS short-circuit current and the associated dc time constant. As
an example, at 132kV, the peak requirements would be based on 40kA & 45ms giving a peak
factor of 2.5 (100kA.pk).
ENA should adopt the switchgear short-circuit ratings presented in Table 8 for use within
ENATS.
Time 132kV 3 132kV
66kV 33kV 20-25kV 11kV 6.6kV
constant ph 1ph
45ms 40kA 40kA 40kA 31.5kA 20 kA 25kA 25kA
120ms 31.5kA 31.5kA 31.5kA 20kA 12.5kA 16kA 16 kA
Table 8: Switchgear short circuit breaking current ratings for incorporation in ENATS
For 132kV and 66kV circuit-breakers higher ratings are recommended in order to allow for
abnormal running arrangements due to SGT outages.
References
[1] CIGRE Technical Brochure No. 304
[2] Consideration of X/R ratio in the application of high-voltage circuit-breakers. CIGRÉ
internal WG paper A3-03(WG11)17
Page 12 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
Appendix 1 - Test Network for Benchmark Study
Benchmark Network R6
11 January 2005
13201 13202
GENA10 GENB10
1 2
132 kV 132 kV
13203
GSPA10
13204 13205
3 GSPB10 4 GSPC10 5
132 kV 132 kV 132 kV
6
33007 7 8
GSPB30
33 kV 33 kV 33 kV
33006 33008
GSPA30 GSPC30
3309
DSBA30
9
11011 33 kV
BSPA50
11 13 14
11 kV 11 kV 11 kV
11013
BSPC50 11014
BSPD50
M
11012
12 10
BSPB50 11 kV 33010 33 kV
GENW30
BENCHMARK NETWORK DATA FOR ASG/OSG X/R SUB-GROUP
Page 13 of 37
ENA Engineering Recommendation G89
Issue 1 February 2011
All values are p.u. on 100MVA unless indicated otherwise
NODE DATA
Node PSS/E Name Voltage Busbar Voltage Loadflow Solution
Node (kV) Type Setpoint V (p.u.) Ang (deg)
1 13201 GENA10 132 Slack 1.0000 1.0000 0.00
2 13202 GENB10 132 PV 1.0000 1.0000 0.13
3 13203 GSPA10 132 PQ 0.9976 -0.35
4 13204 GSPB10 132 PQ 0.9963 -0.53
5 13205 GSPC10 132 PQ 0.9984 -0.40
6 33006 GSPA30 33 PQ 1.0029 -1.01
7 33007 GSPB30 33 PQ 0.9994 -2.55
8 33008 GSPC30 33 PQ 1.0023 -0.74
9 33009 DSBA30 33 PQ 0.9818 -2.28
10 33010 GENW30 33 PV 1.0100 1.0100 -0.58
Generator at busbar,
but not a PV busbar.
Generator Q fixed at
11 11011 BSPA50 11 PQ 0.1 MVAr 1.0080 -6.04
12 11012 BSPB50 11 PQ 1.0171 -4.90
13 11013 BSPC50 11 PQ 1.0117 -7.27
14 11014 BSPD50 11 PQ 1.0117 -5.79
Page 14 of 37
