You are on page 1of 9

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/274246580

RELIABILITY OF HV/MV SUBSTATIONS WITH AIR-ISOLATED AND GAS-


INSULATED SWITCHGEAR

Conference Paper · August 2014

CITATIONS READS

2 1,700

3 authors:

Dragoslav Perić Miladin Tanaskovic


Higher Education Technical School of Professional Studies in Požarevac, Serbia Electric Power Distribution Company Belgrade
12 PUBLICATIONS   166 CITATIONS    13 PUBLICATIONS   112 CITATIONS   

SEE PROFILE SEE PROFILE

Nebojša Petrović
JSC "Elektromreža Srbije" Belgrade - Serbian Transmission and Market Operator
5 PUBLICATIONS   2 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Substation Reliability View project

All content following this page was uploaded by Dragoslav Perić on 23 December 2016.

The user has requested enhancement of the downloaded file.


2014 Paris Session
SC B3 SUBSTATIONS
http : //www.cigre.org
PS2 LIFE-CYCLE MANAGEMENT OF SUBSTATIONS
B3-216

RELIABILITY OF HV/MV SUBSTATIONS


WITH AIR-ISOLATED AND GAS-INSULATED SWITCHGEAR

DRAGOSLAV PERIĆ1, MILADIN TANASKOVIĆ2, NEBOJŠA PETROVIĆ1


1
Serbian Transmission System and Market Operator,
2
Utility Company for Distribution in Belgrade

SERBIA
dragoslav.peric@ems.rs

ABSTRACT

High-voltage and medium-voltage (HV/MV) substations (typically 110kV/10kV and 110


kV/35 kV substations) connect transmission and distribution systems with consumers of electric
energy. The selective search method was used for calculation of substation reliability, where all
configuration elements were grouped into blocks. Failure rates and durations were used as main HV/MV
substation equipment reliability indices. Subject to the analysis were single busbar H configuration and
configurations with double busbars on the high voltage side of HV/MV substations having air-isolated or
gas-insulated switchgears. These arrangements are typical for distribution and transmission grid.

Key words — Reliability, HV/MV Substations, Gas-Insulated Switchgear (GIS), Air-Isolated


Switchgear (AIS)

1 INTRODUCTION
The reliability of electricity supply to customers (e.g. the rate and duration of power outages)
depends on the reliability of HV/MV substations, namely on a single-pole diagram and on the
reliability of substation equipment. The reliability of HV/MV substations is determined by means of
the selective search method [1], in which coincidences (overlapping) of up-to-second-order events
leading to interruption of certain functions are taken into account. Reliability indices for equipment of
the HV/MV substations were taken over from [3], [4].
Reliability calculations for the single busbar "H" configurations (arrangements) were conducted to
analyse the cases: with one disconnector in a bus coupler; with two disconnections in a bus coupler; and
with two disconnectors and a circuit breaker in a bus coupler. Cases with or without longitudinal
sectioning of both busbar systems were analyzed for configurations with a double busbar system. The
calculations performed were aimed to determine the reliability indices for the transit of electric energy,
outage of an entire capacity or of a half of it. In dealing with double busbar configurations we also
considered the reliability of different switching states (normal state, busbar refurbishment). In addition,
sensitivity analysis of reliability results of studied functions was made for different equipment reliability
input data.
The performed analyses have shown that the active faults and durations of substation component
restoration have significant impact on their reliability indices so that it would be necessary to
adequately expand future questionnaires for equipment reliability surveys.

2 SUBSTATION REALIBILITY CALCULATION


For a substation reliability calculations in this paper the selective search method [1] was used,
which is a simplified version of minimal cut- sets, method previously used for analyses of this type
[2]. As to apply this method the following steps are to be made. The first step is to define the functions
of a facility whose reliability needs to be determined. In the second step, the functional blocks for the
facility are formed. A functional block consists of those facility elements which are functionally

*Kneza Miloša 11, Beograd, dragoslav.peric@ems.rs


connected in such way that any failure or planned maintenance on them causes other block elements to
trip out. Block elements are put back into service only after the restoration of the affected block
element. Failure rate and mean time to repair of the block are calculated from the reliability indices of
all the elements constituting the block. In the third step, we tried to determine, by means of reviewing
the facility arrangement, which of the block outages, alone or in coincidence with outages of other
blocks, causes the interruption of a function under consideration. In the fourth step: by summing up
the unavailability indices of blocks that caused interruption of the given functions, the corresponding
resultant reliability indices are determined for each function. The calculations paid respect to the
active failure rates, minding that circuit breaker faults are also a part of the active failures
For all substation arrangements and switching states corresponding tables are set up, which show
block outages and function interruptions caused by these failures. Procedure for reliability calculation
by means of the selective search method is demonstrated for the facility having H2 arrangement
(Fig.1). On the high voltage side of the scheme in Figure 1 the following elements are not shown:
instrument voltage and current transformers and surge arresters in outgoing feeders; voltage
transformers in bus sections; and surge arresters in power transformer bays. Indices of substation
equipment reliability are given in Table I for all elements in configuration, in conformity with [3], [4],
including instrument transformers and surge arresters, which are not shown in the schematic diagrams.
For the equipment failure rate and duration symbols fk and rk, symbols, are used, respectively.
Table I – Iindices of HV/MV Substation Equipment Reliability
Equipment component (source for fk, rk) fk (1/year) rk (h)
Transformer HV/MV [4], [1] 0,0045 200
Voltage transformers MV [3]*, [1] 0,000874 10
Voltage transformers HV [3], [1] 0,000874 10
Circuit breaker MV [3]*, [1] 0,0014 24
Circuit breaker HV [3], [1] 0,0014 24
Disconnector MV [3]*, [1] 0,0029 5
Disconnector HV [3], [1] 0,0029 5
MV busbar with 4 ports [1] 0,00005 10
HV busbar with 2 ports [1] 0,000025 10
HV busbar with 4 ports [1] 0,00005 10
Current transformers MV [3]*, [1] 0,000874 10
Current transformers HV [3], [1] 0,000874 10
Surge arresters MV [1] 0,0003 10
Surge arresters HV [1] 0,0003 10 * data [3] refer to the
GIS circuit breaker 0,001038 30 voltage levels in the range
GIS disconnector except busbar disconnector 0,001149 35 60kV to 100kV (MV
GIS busbar with terminals and disconnector 0,000645 35 equipment)
GIS current and voltage transformers 0,000168 45
Failure rate f, unavailability u and failure duration r of blocks are calculated with the formula (the
sum (∑) over k comprises all elements (components) of the block equipment):
f k rk
f = fk , (1) u= f k rk , k u (2)
k k
r= .
f f
By adding failure rate values of circuit breaker, disconnector, instrument transformers and surge
arrester the sum for block B1 is equal to:
1
f = f k 0,0014 0,0029 0,000874 0,000874 0,0003 0,006348 ,
k year
For the duration of the fault we obtain:
f k rk
k 0,0014 2,2 0,0029 4 0,000874 4 0,000874 4 0,0003 4
r= 3,6030 h .
f 0,006348

2
Fig.1 –HV/MV substation arrangements which are subject of analysis
Reliability indices for other blocks of the H2 arrangement (Table II) are calculated in the same
way. The letter "a" in the block failure designations active failures, or block faults which, due to
activation of the relay protection, cause, by rule, tripping of other blocks, which remain out of service
for the time s required to make necessary changes in topology and to restore facility functions .
Table II – Failure rate and duration for all blocks of H2 configurationt
Block fBk rBk Block fBk rBk Block fBk rBk
B1 0,0063 3,6030 B62 0,0154 4,1005 B32 0,0067 4,0503
B2 0,0063 3,6030 B63 0,0072 3,6500 B33 0,0029 4,0000
B4 0,0096 91,1642 B71 0,0207 3,5129 B34 0,0029 4,0000
B5 0,0096 91,1642 B72 0,0207 3,5129 B4aHV 0,0019 0,1
B61 0,0154 4,1005 B31 0,0067 4,0503 B5aHV 0,0019 0,1
Time for restoration is presented by value s = 0.1 h, and ratio of active failures in the total number
of failures by a value fa/f=0.25. Block B63 operates with a normally opened circuit breaker, and its
failures, therefore, do not affect the analyzed functions of the facility.
Events that should be analyzed are the following: interruption of entire load, interruption of a half
load and suspension of the transit on the HV side. Table III shows the lists of functional blocks the
outage of which leads to the interruption of analyzed functions.
Table III – Block failures that lead to interruption of certain functions for the case of H2 arrangement
Function Block failures that interrupt a function
Interruption of the entire load B1a, B2a, B31a, B32a, B33a, B34a, B4aHV, B5aHV
Interruption of half load B31, B32, B33, B34, B4, B4aMV, B5, B5aMV, B61, B62, B71a, B72a
Suspension of transit B1, B2, B31, B32, B33, B34, B4aHV, B5aHV
Reliability indices of analyzed facility functions are calculated using the relationships of the same
type as equations (1) and (2), by taking into account that the sum (∑) over k covers all failures that
interrupt the considered function. Hence, the failure rate of the transit suspensions is calculated by
adding the corresponding values for failure rates of blocks B1, B2, B31, B32, B33, B34, B4aHV,
B5aHV (Table II):

3
1
f = f Bk 0,0063 0,0063 0,0067 0,0067 0,0029 0,0029 0,0019 0,0019 0,0356 ,
Bk year

The unavailability and average annual duration of transit suspension are:


u= f Bk rBk 0,00630 3.6030 00630 3.6030 00670 4.0503 00670 4.0503
Bk
,
h min
00290 4 00290 4 00190 0,1 0019 0,1 0,124 7 ,42
year year
u 0,123
r= 3,46 h .
f 0,036

3 RESULTS OF CONFIGURATION ANALYSES

3.1 Input Data and Assumptions


The analyzed substation arrangements share a common MV single pole diagram with single
busbars, as the aim was to analyze the impact of selected single pole diagram of the HV side upon the
substation reliability. Bus section bay on MV side is normally open, and it can not be allowed to be
out of service. H arrangements differ in the bus-section configuration on HV side: one or two
disconnectors, or complete bay equipped with circuit breaker and two disconnectors.
Configurations with double busbars differ from each other in the way how the bus coupler or bus-
section bay is designed. Three switching states are analyzed for double busbar configurations:
1. In arrangements D1 and D2 all four HV lines and both transformers are connected to the same
busbars which are decoupled from the other busbar. These states for the considered arrangements
are designated as states D11 and D21, respectively.
2. Per two HV lines and a transformer are connected to both busbar systems with closed bus coupler
in both considered arrangements and designated as states D12 and D22.
3. Two lines are connected to one busbar system and other two lines and transformers to the second
busbar system. For D1 configuration this is state D13 and for D2 configuration this is state D23.
In both arrangements the corresponding bus couplers are closed.
The first of those switching states is, for example, used in regular maintenance of one busbar
system, and the second switching state is symmetrical and commonly used in normal operating
conditions. In the third switching state, transit and transformation functions are maximally separated,
e.g. in a situation where two lines were built mainly to supply important power consumers connected
to the HV network. For the switching states D12 and D22 a function of partial transit is introduced,
when electricity transit is suspended on one or two HV lines.

3.2 Comparative Analysis of the Configuration Reliability


Unavailability index, namely annual function interruption time expressed in minutes was selected for
comparative analysis of the configuration reliability, of switching states and AIS, GIS technologies
(Table IV). Markings from Figure 1 were used in the first column of Table IV. For the case where the
primary (HV) section of the configuration is equipped with GIS technology, the markings have
additional prefix „G". Typical values marked in the table by a bold text are shown in Figure 2’s diagrams
The extremely longest duration of interruption of the entire load is characteristic of H1 and GH1
configurations, because the entire load is being interrupted until its restoration in the case of failure of
B33 disconnector, since being the only disconnector in the bus section. When the H1 and GH1
configurations were exempted from the analysis, it could be seen that the switching states D12 and
D22 are by far the best ones, because an interruption of the entire load occurs only in the case of active
failures of B303a, or B310a and B63a.
Outage of a half capacity is approximately equal for all of the analyzed arrangements and switching
states, because it largely depends on MV arrangement being the same for all the cases in this paper.

4
Table IV Unavailability – Annual interruption time of functions [min/year]
Suspension of Partial Transit Interruption of Entire Load Interruption of a half load Suspension of transit
H1 0,94 133,34 13,69
H2 0,08 135,08 14,56
H3 0,05 133,37 17,10
GH1 2,46 143,74 27,89
GH2 0,05 148,57 30,30
GH3 0,02 143,77 27,79
D11 0,10 128,78 0,10
D12 0,13 0,01 128,78 0,07
D13 0,08 128,78 0,05
D21 0,11 128,78 0,11
D22 0,10 0,01 128,89 0,05
D23 0,09 128,78 0,04
GD11 0,05 122,80 0,05
GD12 0,06 0,01 127,76 0,03
GD13 0,0 255,53 0,03
GD21 0,06 127,76 0,06
GD22 0,06 0,01 127,83 0,03
GD23 0,06 127,76 0,04

Fig. 2 – Suspension of transit and interruption of entire load


The likelihood of transit interruption is about 100 times higher for H arrangement than for double
busbar arrangement. On the basis of an assumption that there are no simultaneous failures of multiple
blocks, full transit interruption cannot happen in cases of switching states D12, D22, GD12 and GD22.
In order to achieve consistency in diagram presentation, duration times of partial transit interruption in
these switching states are divided by 2. When considering only the double busbar configurations it is
evident that, because of a higher number of elements over which transit occurs, the longest time of
transit interruption can be verified for switching states of D11, D21, GD11 and GD21.
Unavailability does not depend so much on the technology applied (AIS, GIS) because there are
higher rate of power outages in the first case, and duration of the fault in the second case. Application
of GIS technology gives slightly better results, except in the case of outage of a half load and in the
case of transit suspension for the H configurations.

3.3 The Sensitivity Analysis


The objective of the sensitivity analysis is to determine the effect of input data for components,
(Table I) upon the calculated substation unavailability. Variations of input data of components are
marked in the following way:
3fPR – 3 times higher circuit breaker failure rate;
3fRS – 3 times higher disconnector failure rate;
3fPR&RS – 3 times higher circuit breaker and disconnector failure rates;
1/8rTR – 8 times less time needed for transformer renovation (shortened to 24 hours);
AK50% – active failures increased to 50% of the total number of failures.

5
As can be seen, ratings of unavailability i among studied configurations and switching states are
maintained for all considered variations of the input data, though differences in the substation
unavailability in absolute terms become more obvious. From this we can draw an important
conclusion about the validity of the comparative reliability analysis of configurations from the
previous subsection, because the obtained ratings are also valid for significant variations of input data.

Fig. 3 - Sensitivity of the entire load outage function to the change of the equipment reliability
indices

3.4 Overall Assessment of Configurations


It can be stated that, as regards the analyzed H configurations, H1 configuration is significantly less
reliable and therefore it should not be used, although it can be, at a relatively fair price, modified to the
H2 configuration by adding a disconnector in a bus section bay, which is far more reliable. The H3
configuration is somewhat better than the H2 configuration. Nevertheless, each concrete case should
be subject to an analysis whether an extra circuit breaker would justify the reduced outage time of
entire load. The results of reliability calculations of "H" configuration confirmed that experts [7] who
worked in 1980’s with the Utility Company for Distribution in Belgrade were right when decided to
use two sectioning disconnectors in HV bus couplers, during the construction of the first 110/10 kV
substation in Belgrade with 110 kV GIS.
Double busbar configurations enable more reliable transit than the H2 and H3 configurations, and
significantly more reliable supply of entire load when operating in switching states - D12 and D22.
Also, the D2 configuration does not have significantly better indicators than the D1 configuration,
which could, in general, justify the introduction of extra 4 disconnectors and consequently more
complex manipulations with the switchgear equipment.
Reliability of the supply of a half of load outage is approximately the same for all configurations,
since it, as mentioned above, largely depends on the MV configuration which was taken to be the
same for all studied cases.
A technology (AIS or GIS) that is applied on HV side does not dominantly affect the arrangement
reliability, and the decision on the application of particular technology will therefore depend more on
other circumstances, such as a project feasibility and the price of a build able land. However, for the
transit function of double busbar configurations GIS technology provides approximately two times
lower unavailability than the AIS solution that can be important in some cases.

4 QUESTIONNAIRES
In order to carry out a calculation of substation reliability it is necessary to have data on reliability
of substation components. Those data are collected from surveys, and the most relevant among them
are the CIGRE surveys [3], [4]. However, majority of the surveys deals mainly with component failure
rate, while the component restoration time and a share of active faults in the total number of failures
are analyzed in a much lesser extent.

6
4.1 Restoration Time
When a failure of substation component occurs, relay protection system triggers the opening of a
corresponding circuit breaker, and procedure for renovation of a faulty element starts thereafter. Steps
of element restoration are schematically shown in Fig. 4.
failure element (disconnection) t 1 t10 switching element (end of recovery)
restoration of power supplay of consumer (without faulty element) t2 t20 returning to the basic operation state
„open“ work permit t3 t30 „close“ work permit
start of work t4 t40 end of work
start effective work t5 t50 end of effective work

Fig. 4 – Timing diagram for procedure of renovation of a faulty element


After disconnecting the faulty element a first thing to do is to restore the power supply to
consumers (e.g., transfer the load from the incapacitated power transformer to another transformer,
MV network reconfiguration, topology change after active fault, etc). In some cases, restoration of
power supply to consumer does not happen when there is no alternative power supply route.
In cases when it is possible to restore immediately the supply to consumers, a postponement of the
corrective work is possible. This is usually done so as to create better conditions for work (e.g. to
begin with work on the next working day, wait for better weather conditions, or to get more time for
preparations). In such cases, the time t20 required to reinstate the normal switching state can also be
postponed to be conducted after restoring the supply to consumers t10 (end of renovation actions on the
faulty element).
Time for renovation (repair or replacement) of the faulty element is calculated as:
r (t10 t1 ) (t3 t2 ) (3)
,
where t3-t2 is the time of delay of start of work on renovation of element.
To obtain a reliable data for statistical analysis and calculation of restoration time, survey
respondents need to provide all relevant times as shown in Fig. 4. The data collected in that way would
enable the analysis of technology and organization of element restoration, which could lead to shorten
the restoration time.
4.2 Active Failures in the Total Number of Component Failures
Equipment element anomalies observed during equipment inspection will be analyzed as a failure,
if they need to be shut down urgently, i.e. within a half an hour [3]. However, most of the equipment
failures are identified by a relay protection system, which activates adjacent circuit breakers to
suspend any kind of supply to the faulty element. Bearing in mind that the number of circuit breakers
in a substation is limited, and if they are in position OPEN that would disconnect not only the faulty
component but, as a rule, the functional block. So their exclusion does not exclude the only element
of order but, as a rule, a functional block. Disconnection of elements of a functional block is usually
consequence of an element failure. Furthermore, additional consequences may occur when tripping out
other elements that do not experience any failure. For example, this can happen when there is a circuit
breaker fault or failure (or of a corresponding relay protection) which automatically disconnects one
other (healthy) circuit breaker. Transformer failures not cleared by operation of the protection system
are also classified in the active faults. The effects of active faults are eliminated within the time period
needed to isolate the block with an faulted element by changing the topology, while other correct
elements are put back into operation.
More complex faults in the power network may cause multiple operation of protection system and
outages of various elements, either because of unselective protection operation or due to occurrence of
multiple successive failures. When complex faults have occurred, SCADA systems generates huge
number of records of events, which are, however, not sufficient for a proper analysis of the problem.
That is why electricity transmission and distribution companies generally set up a failure analysis
team, which strives to shed light on events in the power system from different angles, where
sometimes active failures are not clearly identified.
The following recommendations should be used for failure analyses and survey development:
1. Observing the N elements of the same type and category (for example, 110kV circuit breakers,
SF6) within the period of T years, during which n failures were recorded, out of which na can be

7
classified to active failures. The failure rate and schare of active failures is calculated as:
n fa na
f (4) (5)
NT f n
2. The values n and na are determined by means of forming a relevant list (list n and list na) and any
element failure is added to the list n, and if the failure is classified as active it adds to the list na.
The number of list items at the end of the period represents the value n and na.
3. If it is a complex fault with a higher number of outages, it is necessary to identify the element
affected by a failure which is to be added to its list n and na.
4. Failure duration time should be recorded for each active fault of element till putting the unaffected
elements back into operation. These times are by a rule, approximately the same for all elements,
because they primarily depend on the time required to make necessary changes in equipment
topology and can be categorized according to voltage level, of facility design, network
management organization, etc on.

5 CONCLUSION
The selective search method was used to carry out calculation of the HV/MV substation reliability
for different variants of a single busbar (H) configurations and configurations of double busbar on HV
side. Reliability indices were taken into consideration for electric power transit, outage of entire load
or of a half load. Also, reliability indices for double busbar configurations were taken into
consideration for different switching states (normal state, busbar refurbishment) which is a novelty in
the application of these methods. We conducted an analysis of sensitivity of results to changes of input
data on the reliability of equipment. The analysis confirmed the relative relations of unavailability of
different configurations and switching states. Basic practical conclusions of this analysis are:
1. H configurations should have two disconnectors in the bus section bay in HV busbars, and no
significant improvements to the substation reliability were achieved with a circuit breaker
added to the bus section.
2. Configurations with a double busbar installed on the HV (line) side of the substation are more
reliable than H configurations.
3. Configuration with a double busbar having complex bus section and bus coupler (circuit
breaker and six (6) disconnectors) is not significantly more reliable than the configuration
with a simple bus coupler (circuit breaker and two disconnectors).
4. The most reliable switching state out of all switching states for double busbar configurations
is the one where power lines and transformers are symmetrically (per half) arranged in two
busbar systems, and bus coupler is in position CLOSED.
5. Reliability of configurations performed in GIS technology proves to be significantly better for
the function of power transit for double busbar system configurations.
6. Time for renovation of elements and the share of active failures in the total number of failures
are not sufficiently covered by the surveys on substation component failures. This paper
highlights the need for collection and processing of these data in future surveys.

6 REFERENCES
[1] Jovan M. Nahman, Dependability of Engineering Systems, Springer-Verlag, Berlin, Heidelberg,
New York, 2002.
[2] J. Nahman, N. Mijušković, Reliability analysis of EHV substations, CIGRE Conference, Paris,
paper 23-05, 1980.
[3] Cigre TB 509, Final Report of the 2004 - 2007 International Enquiry on Reliability of High
Voltage Equipment Part 1 - Summary and General Matters Working Group A3.06,October 2012
[4] Cigre WG A2.37, Transformer Reliability Survey: Interim Report, Electra No 26, pp. 46-49,
April 2012
[5] C. R. Heising at al., Final report on high-voltage circuit breaker reliability data for use in
substation and system studies, CIGRE Conference, Paris, paper 13-201, 1994
[6] E. D. Tweed, C. E. Withers, D. J. Hughes, A distribution substation rating strategy based upon
reliability analysis, IEEE Trans. PAS, Vol. 102, No.9, pp. 2893-2897, Sept.1983.
[7] N. Mijušković, Calculation of reliability SF6 gas insulated plants and a proposal to amend the
"H" schemes in the SS 110/10 kV, 18. Conference JUKO CIGRE, R.23.02, 1987.

View publication stats

You might also like