Beruflich Dokumente
Kultur Dokumente
Level Reduction
January 2003
(Revision 1)
September 2003
Executive Summary
The UK government has set targets that by the year 2010, 10% of the electrical energy
consumed in the UK will be provided from renewable sources and 10GWe capacity of
CHP plants will be installed by the same date. Achievement of these very ambitious
targets will most likely entail the connection of a very large number of distributed
generation to existing distribution networks. Connection of distributed generation on
the scale that is envisaged will require practical solutions to a number of identified
technical, commercial and regulatory challenges. One of the technical challenges is
fault level management as connection of distributed generation often results in
increased fault levels beyond the capacity of existing switchgear, especially in urban
areas.
In this report, results of investigations carried out to assess the effectiveness of the
options available for fault level management as well as, where necessary, their impact
on voltage profiles and system stability are presented and discussed.
Five main methods for fault level reduction namely current limiting reactor, Is-limiter,
superconducting fault current limiter, solid-state fault current limiter and network
splitting have been reviewed and discussed.
Network splitting was found to have the greatest potential for fault level reduction in
the short term as it is relatively inexpensive and furthermore it has high reliability and
flexibility. Modelling and simulations for various network-splitting scenarios
performed using PASCAD/EMTDC confirm that fault level can be reduced
significantly by network splitting. Furthermore, the quality of power supply to the
customer and the transient stability of the DG can be improved by using fast-closing
switchgear or a static transfer switch in the bus-section. The analysis shows that the
fault current is sensitive to closure of the normally open point (NOP) on the ring
effectively connecting the cable network in parallel with the bus-section breaker or
the reactor.
Contents
1. Introduction..............................................................................................................2
2. Methods of fault level reduction .............................................................................3
2.1 Current limiting reactor ....................................................................................3
2.2 Is-limiter ..............................................................................................................5
2.3 Superconducting fault current limiter..............................................................7
2.4 Solid-state fault current limiter.......................................................................11
2.5 Network splitting ..............................................................................................11
3. Modelling and simulations ....................................................................................13
3.1 Modelling...........................................................................................................13
3.2 Simulations ........................................................................................................14
3.2.1 Fault current reduction .................................................................................16
3.2.2 Voltage drop.................................................................................................18
3.2.3 Transient stability.........................................................................................22
4. Conclusions.............................................................................................................23
5. References...............................................................................................................24
1
1. Introduction
In 1997 the Kyoto Protocol to the UNFCCC was adopted. It binds industrialised
countries that sign it to achieve quantified GHG reduction targets. Overall, these are
designed to reduce industrialised country emissions by at least 5% from their 1990
levels by the commitment period 2008 to 2012 [2]. The targets for individual
countries range from an 8% cut for the European Union to a 10% increase for Iceland.
At present, there is some way to go before the Kyoto Protocol becomes legally
binding. It is required that at least 55 countries ratify the Protocol, including
industrialised countries (known as Annex I countries) accounting for at least 55% of
industrialised GHG emissions. The latest information shows that 97 countries have
ratified, but these only cover 37% of Annex I country emissions. Since the USA has
decided not to ratify the Protocol, it will only enter into force if Russia signs up. The
European Union has ratified the Protocol, and its Member States have agreed national
targets. The UK share of the EU ‘burden’ has been translated into a 12.5% GHG
emissions reduction target by 2008-2012.
As part of its strategy to meet its 12.5% target, the UK Government has set itself two
further targets to encourage the deployment of low carbon distributed energy sources.
The first one is that by the year 2010, 10% of the electricity consumed in the UK
should come from renewable sources and the second is that there should be 10GWe of
Combined Heat and Power (CHP) installed by the same date [3]. The Government’s
Performance and Innovation Unit (PIU) has suggested a more ambitious target of 20%
electricity from renewable sources by 2020 [4]. The forthcoming Energy White Paper,
due in early 2003, will confirm whether or not this further target will be adopted.
Achieving these targets will require connection of a large amount of distributed
generation (DG) into existing distribution networks. The 2010 targets alone will
require the connection of up to 14GW of new distributed generation to the UK’s
electricity distribution networks [5].
This report is concerned with cost effective management of fault levels as one of the
many technical challenges that must be resolved in order to connect a significant
amount of DG to existing distribution networks. The fault level management problem
is particularly acute in large conurbations where fault levels are already close to the
design limits of switchgear. The connection of DG in these areas requires some action
to be taken to ensure that fault levels remain within the design limits of existing plant.
This is an important safety issue which if unresolved would severely limit the amount
of DG that can be connected.
2
In this report, results of investigations carried out to assess the effectiveness of the
options available for fault level management as well as, where necessary, their impact
on voltage profiles and system stability are presented and discussed. The report begins
by reviewing the methods that can be used to reduce fault levels. It then focuses on
network splitting, modelling the impact of this technique on fault currents, voltage
profiles and stability as well as investigating the issues that must be considered and
required safeguards in applying network splitting.
Fault level is a measure of network robustness. A high fault level is a good indicator
of the strength of the system suggesting close proximity to generating stations or a
highly interconnected system. A high fault level implies low impedance between
source and load and hence is associated with good system voltage profiles and low
magnitudes of voltage dips when they occur. It also has a beneficial influence on the
speed of operation of protective devices under fault conditions. Therefore on the
whole a high fault level is not a bad thing. However, these benefits come at a price as
high fault levels require switchgear and other equipment with high rupturing
capacities, which is expensive. A balance must therefore be struck between the
benefits of high fault level and the cost of necessary switchgear and other plant. In
general short circuit breaking capacities of switchgear are standardised. In the event
that the connection of DG causes the fault level to rise above the existing switchgear
rating, it becomes necessary to find ways to reduce the fault level as a cheaper
alternative to replacement of the switchgear, which in most cases is a costly solution.
There are several methods that can be used to reduce fault levels in power systems.
Some of the notable ones are listed below:
• Current limiting reactor
• Is-limiter
• Network splitting
The basic principles underlying each of the above methods as well as their typical
application are discussed briefly in the following sections.
3
DG
CLR
CLR
(a) Bus section (b) Generator connection
CLR
CLR
CLR
DG
CLR
Applications of the CLR for limiting fault level are varied. One of the more attractive
applications is often the bus section application; see Figure 2(a). For this arrangement,
the CLR is placed between two busbars to connect them together. This arrangement
will be analysed in more detail later in this report.
The main disadvantage of the CLR is a high reactor impedance existing in normal
operation. As indicated earlier, this could degrade the voltage supplied to the
customers, and also influence the effective operation of tap-changing transformers.
4
2.2 Is-limiter
The Is-limiter shown in Figure 2 is a fault current limiter developed and supplied by
ABB Calor Emag [8, 9].
Parallel fuse
Current
transformer
Main conductor
and interrupter
(b) Load current path uninfluenced (c) Fault current switched to fuse
Figure 2 Diagram of the Is-limiter
(Source(s): www.abb.de/calor)
The Is-limiter consists of two parallel conductors: a main conductor and a parallel
fuse. Under normal operation, the load current flows through the main conductor.
During a fault, a tripping device disconnects the main conductor, transferring the fault
current to the parallel fuse with a high breaking capacity, which limits the fault
current during the first rise (in less than 1 ms) [9]. The main applications of the Is-
limiter are shown in Figure 3.
5
(a) Is-limiter coupling two parallel systems (b) Is-limiter in the generator circuit
(c) Is-limiter coupling a generator to the public (d) Is-limiter and reactor connected in parallel
network with current-direction comparison
The Is-limiter is often used to couple two systems or parts of systems (as shown in
Figure 3(a)), whose short-circuit withstand capability would not be sufficient if
connected in parallel via a circuit breaker. The Is-limiter separates the system
electrically into two parts before the fault current can endanger the system
components.
Figure 3(b) shows the Is-limiter protecting a public supply system from the short
circuit current supplied by the DG.
Figure 3(c) shows the Is-limiter installed in the connection between the public supply
and the DG. The Is-limiter can interrupt the fault current supplied by the DG with a
current-direction comparison.
Figure 3(d) shows the Is-limiter connected with a reactor in parallel. This arrangement
can avoid the copper losses, large voltage fluctuations and the electromagnetic fields
caused by the reactor.
6
2.3 Superconducting fault current limiter
Superconductor-based fault current limiters offer an alternative solution to controlling
fault levels on the network [10, 11, 12]. A superconducting fault current limiter
(SFCL), unlike reactors or high-impedance transformers, will limit fault current
without adding impedance to the circuit during normal operation.
Most SFCLs are based on the “superconducting and normal” (SN) transition property.
Superconductors are the only materials that change their resistance automatically from
zero to a high value when a certain ‘critical current’ is surpassed. Early
superconducting fault current limiters were too expensive for wide application in
electrical utilities, since they were based on superconducting materials, which can
only operate under extremely low temperatures (-269°C). With the discovery of high
temperature superconductors (HTSs) fifteen years ago, the cooling problem has been
greatly reduced. These new materials can be operated at much higher temperatures (-
196°C) and can be cooled simply by using liquid nitrogen.
Figure 4 shows the typical resistivity (Ωm) of (a) Bi-2212 (Bi2Sr2CaCu2Ox) bulk
superconductor and (b) YBCO (YBa2Cu3Ox) coated superconductor as a function of
current density (A/m2) and temperature (K) [13].
In the event of a fault, the SFCL develops a high resistance limiting the fault current.
With the SFCL, the utility can be provided with a low-impedance, stiff system with a
low fault level, as shown in Figure 5.
7
Normal Faulted
40MVA 40MVA
I_FL=2.1kA Zt=5% I_SC=8.4kA Zt=5%
11 kV 11 kV
There are two major types of the SFCL namely resistive and magnetic-shielded
inductive [14]. In the resistive type of SFCL, the superconductor, using the YBCO
thin film, makes use of the property of changing from a superconducting state to a
resistive state and produces a resistance when an overcurrent flows. A diagram of the
resistive SFCL is shown in Figure 6. The fault current pushes the superconductor into
a resistive state directly and a resistance appears in the circuit. The advantage of the
resistive SFCL is that the superconductor absorbs the energy of the fault current
directly.
SFCL
8
Figure 7 Schematic diagram of the CRIEPI inductive SFCL
(Ichikawa and Okazaki 1995)
Ideally, this type of SFCL has the characteristics of a transformer with a primary
winding and a ‘shorted’ secondary winding [15], as shown in Figure 8. The secondary
winding is the HTS cylinder whose function under normal conditions is to shield the
flux produced by the primary winding from entering the iron core. The primary
winding is connected directly to the circuit. If the secondary winding is driven beyond
the critical current of the superconductor, it reverts to a resistive state. Hence, the flux
coming from the primary winding enters the iron core. The inductance and hence
impedance of the primary winding rapidly increase. Thus the fault current in the
circuit is limited by the inductance of the SFCL.
9
SFCL
SFCL
(a) SFCL in the transformer circuit (b) SFCL in the feeder circuit
SFCL
Figure 9(b) shows the SFCL in a feeder circuit. Individual feeder equipment, that is
difficult to replace, such as underground cables or distribution switchgears, can be
protected by the SFCL.
Figure 9(c) shows the SFCL connecting two busbars. The busbars are only separated
by the SFCL during a fault.
10
costly and on present projections, an economical SFCL will not be available for
approximately another 5 to 10 years.
Z1 SW1
C1
L1
ZnO
BPS
Z2
The low impedance Z1 limits the inrush current through SW1. The over voltage
protection device ZnO (zinc-oxide arrester), and the bypass switch BPS, which backs
up the switch SW1 are also connected in parallel to the capacitors C1. The low
impedance Z2 restrains the inrush current when BPS closes.
However, the SSFCL is still not widely used in practice due to its high cost, low
reliability, and complicated auxiliary system.
11
Infinite system 33kV
SB00 SB01
F3
SB11
SB12
Bus1 11kV Bus2
11kV
LB1 LB2
Rated interrupt
EB0 Rated Interrupt
capacity capacity
250MVA EB1 250MVA
F1
Load1 Load2
DG
5MVA x d′′ = 0 . 2
The network splitting scheme shown in Figure 11 uses the bus-section circuit breaker
EB1 between Bus1 and Bus2. By splitting the network in this way, the impedance
between the 33kV and 11kV systems increases from 5% to 10%, reducing the fault
current coming from the public supply (33kV) significantly.
However, this scheme may decrease the flexibility of the DG. When busbar Bus2
needs maintenance, the DG has to be disconnected from the network, and therefore an
alternative network splitting arrangement that avoids this problem is shown in Figure
12.
12
Infinite system 33kV
SB00 SB01
F3
SB11 SB12
Bus1 11kV Bus2
11kV
LB1 LB2
Rated interrupt EB1 EB2 Rated Interrupt
capacity capacity
250MVA 250MVA
F1
EB0
Load1 Load2
DG
5MVA x d′′ = 0 . 2
In this section, modelling and simulations for various network-splitting scenarios are
presented and discussed. The simulation results were performed using
PASCAD/EMTDC. The primary aim of this work is to study the efficacy of network
splitting in fault current reduction by use of bus-section circuit breaker and reactor. A
secondary but equally important objective is to study the impact of these network
splitting options on voltage profiles and generator stability.
3.1 Modelling
Network:
The following system parameters were used in the simulations: Network primary
voltage is 33kV, 50Hz, with an infinite short circuit power.
- Main transformers are each 12.5 MVA, 33/11 kV, reactance 10%.
13
- Load1 and Load2 are each 5MVA and power factor 0.85.
- The feeder impedance between LB2 and Load2 is 0.1+j0.5 Ω.
Generator:
The following typical generator (synchronous machine) parameters were used to
represent the distributed generator [22]:
- Rated capacity: 4.51 [MVA],
- Rated voltage: 11 [kV],
- Frequency: 50 [Hz],
- Inertia constant: 1.05 [sec.],
- Armature resistance (Ra): 0.01 [p.u.],
- Potier reactance (Xp): 0.03 [p.u.],
- D-axis synchronous reactance (Xd): 2.95 [p.u.],
- D-axis transient reactance (Xd’): 0.25 [p.u.],
- D-axis transient time (Tdo’): 5.50 [sec.],
- D-axis sub-transient reactance (Xd”): 0.17 [p.u.],
- D-axis sub-transient time (Tdo”): 0.05 [sec.],
- Q-axis synchronous reactance (Xq): 1.35 [p.u.],
- Q-axis sub-transient reactance (Xq”): 0.31 [p.u.],
- Q-axis sub-transient time (Tqo”): 0.27 [sec.].
3.2 Simulations
Three network models were used to investigate network splitting as an effective
means of fault current reduction. The first network is the base case (case a), as it
represents the normal network configuration. This case is shown in Figure 13 where
the 11kV busbars Bus1 and Bus2 are shorted. In the second case (case b), the
network is split by installing a reactor between Bus1 and Bus2 (see Figure 14). The
last case (case c) is shown in Figure 15 depicting a split 11kV bar with a bus-section
circuit breaker that is operated normally open.
Infinite system 33kV
SB00 SB01
F3
SB11
SB12
Bus1 11kV Bus2
11kV
LB1 LB2
Rated interrupt
EB0 Rated Interrupt
capacity capacity
250MVA 250MVA
F1
Load1
DG
5MVA x d′′ = 0 . 2
F2
Load2
Figure 13 Case a -11kV busbar connected solidly
14
Infinite system 33kV
SB00 SB01
F3
SB11
SB12
Bus1 11kV Bus2
11kV
LB1 LB2
Rated interrupt
EB0 Rated Interrupt
capacity capacity
250MVA Reactor 250MVA
20% F1
Load1
DG
5MVA x d′′ = 0 . 2
F2
Load2
SB00 SB01
F3
SB11 SB12
Bus1 11kV Bus2
11kV
LB1 LB2
Rated interrupt EB1 EB2 Rated Interrupt
capacity capacity
250MVA 250MVA
F1
EB0
Load1
DG
5MVA x d′′ = 0 . 2
F2
Load2
Figure 15 Case c: Two 11kV busbars are split by
bus-section circuit breaker EB1
15
For each of the above cases the following studies were performed for various fault
locations as appropriate:
- Fault current reduction
- Voltage drop
- Transient stability
The results of these studies are presented and discussed below.
The fault currents flowing through LB2 are shown in Figure 16.
Case a
Case b
Case c
Figure 16 shows that the peak values of fault current have been reduced from 25kA to
17kA by using network splitting via a bus-section circuit breaker and 20kA using a
reactor.
If the busbar is separated (e.g. by a reactor), the fault current will be increased by
connection of cable network through closing the normal opening point (NOP) (see
Figure 17).
16
33kV, infinite system
Is_rms
F1
NOP
RC XC
DG
Cable network 5MVA, 20%
close
Similarly, if the busbar is split by a bus-section circuit breaker, the fault current will
also be increased by connection of the cable network through closing of the NOP (see
Figure 18).
17
33kV, infinite system
Is_rms
NOP F1
RC XC DG
With reference to Table 2, it is noted that the first peaks of fault currents increase
from 13.3kA to 24.5kA (almost double), when the cable lengths decrease from
infinite to zero. The SCLs at the 11kV busbar increase from 148.8MVA to
274.5MVA.
From the above analysis, it clear that, the effectiveness of reactors and network
splitting in fault current reduction will be limited by connection of a cable network in
parallel, particularly a short cable length.
18
When a three-phase short circuit fault occurs at 2 seconds at F1 in Figures 13, 14 and
15, the 11kV busbar will experience short-term voltage drops, as shown in Figure 19.
The fault F1 is cleared by opening LB2 after 150ms.
Case a
Case c
Case b
Figure 19 shows little effective difference of the voltage responses of 11kV busbar
after clearance of the fault in three cases a, b and c.
Similarly, for a fault occurring at 2 seconds at F2 in Figures 13, 14 and 15, the 11kV
busbar will also experience voltage drops, as shown in Figure 20. The fault F2 is
again removed by opening LB2 after 150ms.
Case a
Case b
Case c
19
Figure 20 shows the difference of the voltage drops during the fault due to feeder
impedance between switch LB2 and the faulty point. The voltage drops are about
50%, 41% and 36% for case a, b and c respectively. The voltages at the 11kV busbar
are almost the same after fault clearance.
Additionally, for a fault occurring at 2 seconds at F3 in Figures 13, 14 and 15, the
11kV busbar will also experience voltage drops. Fault F3 is cleared by opening SB01
and SB12 after 150ms. Then, the DG and Load2 are switched from Bus2 to Bus1 by
closing EB1 in 50ms after clearing the fault.
Case a
Case b
Case c
Figure 21 shows the differences of the recoveries of the 11kV busbar voltage in the
three cases. In general, the voltage recovery is better when the network is split by a
bus section circuit breaker than when a reactor is connected across the two sections of
the 11kV busbar.
Different types of electrical equipment will have different tolerances to voltage drop.
The CBEMA (Computer Business Equipment Manufactures Association) curve has
been widely quoted as providing some voltage quality guidance for other types of
equipment [24]. This curve has recently been revised and is now known as the ITI
(Information Technology Industry Council) CBEMA curve [25], shown in Figure 22.
20
Figure 22 ITI (CBEMA) curve
Published by:
Information Technology Industry Council (ITI)
1250 Eye Street NW Suite 200 Washington DC 20005
202-737-8888
http://www.itic.org
The ITI (CBEMA) curve was developed to be used as a guideline for manufacturers
in designing power suppliers for use with sensitive electronic equipment. The vertical
axis of the graph is the percent of rated voltage applied to a circuit. The horizontal
axis is the time which the voltage is applied. The graph is divided into three regions:
no interruption in function region, prohibited region and no damage region. Normally,
equipment is expected to operate within the “no interruption in function” region.
Equipment can be damaged when voltage spikes are severe enough to enter the
prohibited region. In contrast, if the voltage drops are below the lower limit and the
times which the voltages are continuously applied are exceeded, then the equipment
can enter the no damage region. In this area, the normal function of the equipment is
not to be expected, but no damage to the equipment should result.
21
Given the simplifying assumption of the ITI (CBEMA) curve there is little effective
difference between the voltage drops in this study. The fault duration times were
assumed to be 150ms in the case a, case b and case c. According to the ITI (CBEMA)
curve, the voltage drops and duration times in Figures 19, 20 and 21 will all be
located in the “no damage” region.
For transient stability investigation of the DG, a three-phase short circuit fault was
applied at 2 seconds at F3 and lasted 150ms. The closing times of the EB1 were
assumed to be 60ms and 50ms, respectively. The rotational speed of the DG is shown
in Figure 23 (PSCAD/EMTDC simulation).
III II
Curve I in Figure 23 shows transient instability of the DG after EB1 closing in 60ms.
The DG can be stable if the closing speed of EB1 is fast enough, as shown the curve
II in Figure 23.
A static transfer switch connected in parallel with EB1 is shown in Figure 24. It can
allow a fast transfer of the DG from a one busbar to the other within a quarter of a
cycle [26, 27].
22
Static transfer switch
EB1
4. Conclusions
This report has reviewed the basic principles underlying the methods that can be used
to reduce fault levels as well as their typical application. Five main methods namely
current limiting reactor, Is-limiter, superconducting fault current limiter, solid-state
fault current limiter and network splitting have been discussed.
The current limiting reactor is an effective means of fault current reduction in the
network. However, it suffers from the disadvantage that it could degrade the voltage
supply to the customer due to the voltage drop across it in normal operation. The Is-
limiter is not presently used widely on the UK public distribution systems. The
superconducting fault current limiter (SFCL) is an effective method to limit fault
current. It is however still is too expensive to be applied in practice, at least not in the
next 5 to 10 years. The solid-state fault current limiter (SSFCL) is also an efficient
fault current limiting device. However, the SSFCL is not widely used due to its high
capital cost and complex auxiliary system.
Network splitting was found to have greatest potential for fault level reduction in the
short term as it is relatively inexpensive and furthermore it has high reliability and
flexibility. This method was therefore analysed in greater detail to determine its
effectiveness as well as the impact on power quality and transient stability. Modelling
and simulations for various network-splitting scenarios were performed using
PASCAD/EMTDC. These scenarios confirm the expected result that the fault level
can be reduced significantly by network splitting via a bus-section circuit breaker.
Furthermore, the quality of power supply to the customer and the transient stability of
the DG can be improved by using fast-closing switchgear or a static transfer switch in
the bus-section. However, analysis shows that the fault current is sensitive to closure
of the normally open point (NOP) on the ring effectively connecting the cable
network in parallel with the bus-section breaker or the reactor. Care should therefore
be taken in cases where the cable impedance is low (short cable length) to ensure that
the network is operated in a radial configuration rather than as a ring.
Given the simplifying assumptions of the ITI (CBEMA) curve there is little effective
difference between the voltage drops in this study.
23
5. References
24
26. Sannino A., Bollen M., Mitigation of Voltage Sags and Short Interruptions
through Distribution System Design, www.elkraft.chalmers.se/publikationer
27. Inverpower Controls Limited, Medium Voltage Static Transfer Switch,
http://www.inverpower.com/products/sts/sts.html
25
The inter-disciplinary Tyndall Centre for Climate Change Research undertakes integrated
research into the long-term consequences of climate change for society and into the
development of sustainable responses that governments, business-leaders and decision-
makers can evaluate and implement. Achieving these objectives brings together UK
climate scientists, social scientists, engineers and economists in a unique collaborative
research effort.
Research at the Tyndall Centre is organised into four research themes that collectively
contribute to all aspects of the climate change issue: Integrating Frameworks;
Decarbonising Modern Societies; Adapting to Climate Change; and Sustaining the
Coastal Zone. All thematic fields address a clear problem posed to society by climate
change, and will generate results to guide the strategic development of climate change
mitigation and adaptation policies at local, national and global scales.
The Tyndall Centre is named after the 19th century UK scientist John Tyndall, who was
the first to prove the Earth’s natural greenhouse effect and suggested that slight
changes in atmospheric composition could bring about climate variations. In addition, he
was committed to improving the quality of science education and knowledge.
The Tyndall Centre is a partnership of the following institutions:
University of East Anglia
UMIST
Southampton Oceanography Centre
University of Southampton
University of Cambridge
Centre for Ecology and Hydrology
SPRU – Science and Technology Policy Research (University of Sussex)
Institute for Transport Studies (University of Leeds)
Complex Systems Management Centre (Cranfield University)
Energy Research Unit (CLRC Rutherford Appleton Laboratory)
The Centre is core funded by the following organisations:
Natural Environmental Research Council (NERC)
Economic and Social Research Council (ESRC)
Engineering and Physical Sciences Research Council (EPSRC)
UK Government Department of Trade and Industry (DTI)
For more information, visit the Tyndall Centre Web site (www.tyndall.ac.uk) or contact:
External Communications Manager
Tyndall Centre for Climate Change Research
University of East Anglia, Norwich NR4 7TJ, UK
Phone: +44 (0) 1603 59 3906; Fax: +44 (0) 1603 59 3901
Email: tyndall@uea.ac.uk
Recent Working Papers