Networks
March 2016
Mott MacDonald, Victory House, Trafalgar Place, Brighton BN1 4FY, United Kingdom
T +44 (0)1273 365 000 F +44(0) 1273 365 100 W www.mottmac.com
Low Energy Automated Networks
Work Package Two: Technical Analysis
This document is issued for the party which commissioned it We accept no responsibility for the consequences of this
and for specific purposes connected with the above-captioned document being relied upon by any other party, or being used
project only. It should not be relied upon by any other party or for any other purpose, or containing any error or omission
used for any other purpose. which is due to an error or omission in data supplied to us by
other parties.
Contents
Executive Summary i
1 Introduction 1
1.1 Project Introduction__________________________________________________________________ 1
1.2 Summary Conclusions of Work Package One (WP1)________________________________________ 1
1.3 Work Package Two: Technical Analysis __________________________________________________ 2
4.1 Introduction_______________________________________________________________________ 32
4.2 Discussion of condition monitoring technologies __________________________________________ 32
4.2.1 Oil sampling and Dissolved Gas Analysis (DGA) __________________________________________ 32
4.2.2 Partial Discharge (PD) Analysis _______________________________________________________ 33
4.2.3 Current and Voltage Monitoring _______________________________________________________ 34
4.2.4 Acoustic Analysis __________________________________________________________________ 34
4.2.5 Sweep frequency response analysis (SFRA) _____________________________________________ 35
4.3 Option Comparison_________________________________________________________________ 35
4.4 Conclusion _______________________________________________________________________ 37
4.4.1 Physical damage has resulted from magnetising inrush current ______________________________ 37
4.4.2 Migration of moisture between paper and the oil __________________________________________ 37
6 Conclusion 41
Appendices 43
Appendix A. PSCAD-EMTDC Simulation Methodology _______________________________________________ 44
A.1 Study Methodology and Approach _____________________________________________________ 44
A.1.1 Analysis Software __________________________________________________________________ 44
A.1.2 System Description_________________________________________________________________ 44
A.1.3 PSCAD-EMTDC Modelling ___________________________________________________________ 45
Appendix B. Detailed system parameters __________________________________________________________ 49
Appendix C. Detailed Studies of Option 1 __________________________________________________________ 50
C.1 33/11 kV 15 MVA Dyn11 Transformer energisation at Hedge End 33 kV substation assuming worst case
condition (system minimum fault level, 80% residual flux in the transformer iron core) ____________ 50
C.1.1 Scenario 1 – The 15 MVA 33/11 kV TASS-Transformer is energised against the 33 kV systems at
minimum fault level without the availability of the Primary-Transformer________________________ 51
C.1.2 Scenario 2 – The 15 MVA 33/11 kV TASS-Transformer is energised against the 33 kV systems at
minimum fault level with the Primary-Transformer is online. ________________________________ 53
C.2 33/11 kV 20 MVA Yy0 Transformer energisation at Rose Green 33 kV substation from the Hunston 33
kV Substation _____________________________________________________________________ 55
C.2.1 Scenario 1 – The 20 MVA 33/11 kV TASS-Transformer is energised against the 33 kV systems at
minimum fault level with the Primary-Transformer is online (assuming no residual flux). ____________ 57
C.2.2 Scenario 2 –The 20 MVA 33/11 kV TASS-Transformer is de-energised to calculate the prospective
residual flux and subsequently energised against the 33 kV systems at minimum fault level when the
Primary-Transformer is online ________________________________________________________ 59
Appendix D. Detailed Results for study Option 2 at Rose Green 33 kV ___________________________________ 61
D.1.1 Option 2 – Rose green results ________________________________________________________ 61
Appendix E. Reference Documents ______________________________________________________________ 65
Executive Summary
This report presents the analysis and conclusions of Work Package Two (WP2) of
Mott MacDonald’s involvement with Southern Energy Power Distribution’s (SEPD)
project: Low Energy Automated Networks (LEAN).
The main focus of the LEAN project is Transformer Auto Stop Start (TASS).
TASS is the controlled switching out of a 33/11kV transformer. TASS seeks to
reduce the total losses. Where transformers are installed in pairs or groups
(primarily for security of supply), under low loading the no-load (iron) losses saved
by switching off a transformer can be greater than the increase in on-load (copper)
losses in the remaining transformer(s). In these cases electrical losses may be
reduced by having one transformer switched out during low load periods.
The aim of this WP2 report is to provide a technical analysis of a number of issues
related to the application of the LEAN project including:
Consideration of required transformer energisation procedures to ensure
compliance with the relevant codes and practices.
Identifying potential failure modes in the transformers with TASS applied, and
the selection of condition monitoring procedures to allow management of the
risk of failure.
Identifying the risks to the security of supply from applying TASS and ANT, and
finding potential ways of ensuring that the security of supply is maintained, and
identifying any risks in applying these methods
Transformer Energisation
From the studies documented in this report, we can make the following
conclusion: for the majority of installations, uncontrolled switching of 33/11 kV
transformers in the SEPD network will not comply with the requirements for
normal routine switching operations given in the Distribution Code.
A controlled switching relay design applicable for use with a ganged operation
circuit breaker such as Vizimax SynchroTeq MV [1] may provide a distribution
code compliant solution. This would not require the installation of any additional
primary equipment (other than measuring devices such as VTs if these are not
presently installed). However, as SynchroTeq MV does not yet have an extensive
operational experience within the power industry in the UK, it is recommended
that this conclusion be verified through discussions with the SynchroTeq
manufacturer - Vizimax Inc, and through on-site testing.
For detecting the migration of moisture between paper and oil it is recommended
to carry out oil testing for moisture content and dielectric strength. This may be by
regular sampling or on-line analysis.
We have carried out an initial review and risk assessment of the issues related to
the application of TASS and ANT. This is provided in Figure 5.4 and addressing
these issues will form a key part of the work package three scope of the LEAN
project.
1 Introduction
This report presents the analysis and conclusions of work package two (WP2) of Mott MacDonald’s
involvement with Southern Energy Power Distribution’s (SEPD) project Low Energy Automated Networks
(LEAN).
The LEAN project is based on a solution called Transformer Auto Stop Start (TASS) which seeks to reduce
the total losses incurred by SEPD.
TASS is targeted on the substations where transformers are installed in pairs or groups connected in
parallel jointly supplying loads. Under light loading the no-load (iron) losses saved by switching off one
transformer can be greater than the increase in on-load (copper) losses in the remaining transformer(s). In
these cases, electrical losses may be reduced by having one of the transformers switched out.
The LEAN project has a second component. This component is alternate network topologies (ANT). ANT
is the application of new network topologies (primarily interconnection and auto-switching) to ensure that
the application of TASS does not result in a reduction in the security of supply of loads.
WP1 involved the analysis of information on 33/11 kV sites in the SEPD network to identify the most
suitable substations for the application of TASS. There were 21 sites selected based on energy savings,
number of required switches and the condition of the transformer. The final selection is shown in Table 1.1
Energy Age of
Substation Savings Number of ON / Location of 33 kV Number of Transformers
Name (MWh) OFF Switches Circuit Breakers Transformers (Years)
MAIDEN 68.28 39 Local 2 55 / 54
NEWTON
PEWSEY 45.18 217 Local 2 55
REDLANDS 43.65 183 WINTERBOURNE 2 50
ABBAS & CHICKERELL
ROSE GREEN 82.66 34 HUNSTON S/S 2 49
SOUTHCOTE 41.32 101 Local 2 52
TRADING 58.75 14 SWINDON S/S 2 51
ESTATE
UPPER 42.19 52 BICESTER S/S 2 46 / 44
HEYFORD
WESTON 76.52 76 Local 2 49
WINDRUSH 87.93 0 WITNEY S/S & 2 17
PARK BURFORD S/S
This document is the summary report for WP2. WP2 considers the technical feasibility of applying LEAN.
Including:
The impact to the distribution network in terms of grid code compliance.
The technical feasibility of applying TASS to a substation with existing plant.
What, if any, additional plant is required to provide a grid code compliant application of TASS to the
substation.
How to monitor any potential damage to a transformer from applying TASS.
Identifying the risks to the security of supply from applying TASS and ANT.
Identify potential ways of ensuring the security of supply is maintained, and identify risks in applying
these methods (this task will feed into WP3).
This WP2 report is structured in four sections where these issues are addressed:
TASS control strategy review
Transformer energisation study
Transformer condition monitoring
Risk Register for the application of TASS and ANT
TASS requires the switching of power transformers. A power transformer being switched on (energised)
has the potential to cause large voltage and current transients as the network energises the transformer.
The energisation transient current inrush may cause voltage harmonics, voltage sags and, ultimately,
electrical and mechanical stresses on equipment (both customer and DNO owned). To limit these
negative effects the distribution code (section DPC4.2.3.3) gives the limit for voltage step changes caused
by transformer energisation.
The Distribution Code section DPC4.2.3.3 states that for normal routine operations of energising
transformers; the voltage step change limits are as defined in Engineering Recommendation P28 (ER
P28).
ER P28 defines a general limit of 3% on the allowable magnitude of rms voltage drop for switching events
occurring with a period exceeding 30 ms as shown in Figure 2.1.
DPC4.2.3.3 does make allowance for “rare” events, defined as not more frequent than once a year. For
these rare events, a larger limit of 10% is applied. This has typically been the assumed value for
conventional distribution transformer energisation. This would only be applicable for a TASS scheme
where the switching was rare, such as those substations where we do not anticipate more than one or two
switches per year.
Note that the requirements in the Distribution Code apply to ‘Users’ of the system. They do not explicitly
apply to the DNO. However, it can be considered good industry practice and varying from these
requirements will require detailed consideration of the statutory requirements for DNOs.
To limit the voltage step changes caused by transformer energisation, a number of strategies can be
considered.
In this report we have considered the three following options for energising a transformer:
Option 1: Uncontrolled switching using the existing equipment
Option 2: Controlled switching using the existing primary equipment
Option 3: Controlled switching using new primary equipment
Option 4: Switching using a pre-insertion resistor
This is the simplest TASS strategy and it involves switching the transformer at a random point in time. This
option is suitable for sites which require infrequent switching (less than 10 per year) or where the voltage
drop is not expected to exceed the 3% limit recommended by ER P28.
To understand the design of the 33kV circuit breakers within SEPD’s network we reviewed all the
information provided by SEPD for details of the circuit breaker operating mechanism. In particular we
attempted to understand if the circuit breakers could be configured to provide individual or staggered pole
operation. Where this information was not discoverable within SEPD’s data we contacted manufacturers.
From this review, we concluded that none of the 33kV circuit breakers energising a transformer with TASS
employed would currently have individual or staggered pole operation.
As such for Option 2, we considered only options that could be applied for circuit breakers with ganged
operation (all phases operated together). This initially proved difficult and we could not find any control
relays with operational experience in the UK which offered the capability to reduce inrush transients for
ganged operation circuit breakers.
In order, to give a wider market view we contacted a number of companies operating outside of the UK.
One of these, Vizimax [1], claimed that one of their products, namely the SynchroTeq device, can be used
for controlled switching and inrush current limitation without using a staggered or independent pole
operated breaker.
Generally, a controlled switching device measures the power system voltage to synchronise the circuit
breaker’s mechanical operations. SynchroTeq MV also does this, but in addition it calculates the
transformer residual flux remaining after the last isolation.
The idea of “controlled switching taking into account the residual flux” was presented by Brunke and
Fröhlich [2]. This is the promising approach because if the residual flux is known exactly and the closing of
the breaker poles is timed precisely, the transient inrush current can be eliminated completely.
The optimal time of energisation is when the transformer core residual flux prior to the energisation is equal
to the transformer core prospective flux. The residual fluxes are obtained by integrating the corresponding
phase voltages. The prospective flux is the flux which would exist if the supply source was connected to
the transformer at the steady state condition. At the point on wave which the residual flux and the
prospective flux would coincide (i.e. the magnitude of the residual flux is equal to the prospective flux), the
effective core flux corresponds to that which would exist in the transformer during normal operation.
Hence, when energised at this moment on the point-on-wave, there would practically be no inrush
transient. Figure 2.2 shows the illustration of the transformer energisation optimal switching based on the
flux requisition method (knowledge of instantaneous residual flux and prospective flux).
Figure 2.2: Transformer energisation optimal switching based on the flux requisition method (knowledge of
instantaneous residual flux and prospective flux) (point A- random switching point, point B- optimal
switching point)
Therefore regardless of the type of winding connection of the transformer, as long as the voltage at the
transformer terminals can be measured before and after the de-energisation an optimal switching angle
can be estimated. Figure 2.3 shows the application schematic of the Vizimax SynchroTeq© for power
transformer application.
For a gang-operated circuit breaker, the knowledge of residual flux and prospective flux at each phase is
required. From this the optimal switching angle where the residual flux at each phase is approximately
equal to its core prospective flux can be obtained. Figure 2.4 shows how an optimal angle is obtained for a
simultaneous switching operation.
Figure 2.3: Vizimax SynchroTeq Application Schematic For No-Load Transformer Inrush Current Mitigation and
Switching
Figure 2.4: Optimal switching angle (point-on-wave) that satisfies the requirement of residual flux ≈ prospective
flux. Energising a transformer at this angle will ensure minimal transient inrush
Conventionally applied controlled switching usually requires independent poles or staggered pole circuit
breakers.
The most conventionally applied controlled switching for transformer energisation is point-on-wave
controlled circuit breakers. In this case, each pole is controlled independently to minimise inrush
transients. This is a technology commonly applied at voltages in excess of 66kV.
We contacted a number of leading manufacturers to determine the market availability of single pole
controlled circuit breakers rated for a 33kV. We could find no manufacture that provided a conventional
distribution 33kV circuit breaker with single pole controlled operation.
In order to implement point on wave control for 33kV switching an unconventional approach would be
required such as one of the following
66kV or 132kV circuit breakers with single pole controlled operation. This would be effective but would
be a high cost solution
Single pole circuit breakers - these are available for the traction industry. We do not anticipated that
these could be easily adapted for three phase operation.
The Fuse Saver device which is a line mounted circuit breaker which has independent phase operation
or soft three phase ganged switching. However, this is only applicable up to 24kV at the moment.
Generator circuit breakers such as the Siemens 3AH376 breaker. These circuit breakers have high
load and fault ratings and feature phase segregated tripping. This may be effective but would be a
high cost solution.
In order to mitigate the risks arising from transformer inrush currents the magnitude of these currents has
to be reduced. This can be achieved by using pre-insertion resistors in the circuit breakers.
This method includes the insertion of a three phase resistance in series with the energising breaker. The
voltage drop across the pre-insertion resistor produced by the inrush current will decrease the voltage on
the transformer windings, which in turns reduces the magnetic over flux in the core resulting in the
decrease of the magnitude of the transient magnetising currents. Once the inrush transient current has
decreased, the resistance is then bypassed. The time of the resistance insertion along with the time at
which it is bypassed are the parameters that influence the effectiveness of this method. The circuit breaker
could end up being severely damaged if the pre-insertion resistor is not successfully bypassed after the
switching operation. The pre-insertion resistor is insulated for the full system voltage. It needs to have a
high thermal mass to absorb the energy involved in the short time that the inrush currents are present.
However, this method has some undesired features. The resistors have to be included in the circuit
breaker design, thus leading to additional maintenance requirements and while it effectively reduces the
inrush currents during energisation, it does not address the root cause. This is why this solution has been
implemented in the past but is not a preferable method in modern power systems where new alternative
solutions have been introduced. This view is reflected in the 2014 report of CIGRE Working Group
C4.307, “Transformer energization in power systems” which states:
“Pre-insertion resistors require relatively large resistors to be installed in series to the main circuit breaker
interrupters and an effective reduction of inrush current is achieved only by an optimal choice of the
resistance value and the pre-insertion time. However, the addition of switching devices equipped with pre-
insertion resistors in series with the energizing breaker is probably not practical because such devices
would have to be specially ordered from the manufacturer at a high cost. Circuit breakers equipped with
pre-insertion resistors are also no longer available off-the-shelf for voltages less than 500 kV since modern
EHV breakers are designed for use with POW closing.”
3.1 Introduction
This section of the report will detail the transformer energisation studies for simulating the four switching
strategies described in Section 2. We will present the results of these studies and provide conclusions
from these results.
These studies required detailed Electromagnetic Transient (EMT) models. These models are developed in
the EMT program PSCAD-EMTDC. Details of the models and the simulation methodology are presented in
Appendix A.
In the following sections, the results of the studies are assessed against the operational limit for RMS-
voltage step change specified by the Distribution Code (discussed in Section 2.2) and the referenced
requirements of ENA ER P28.
3.2.1 Introduction
Transformer core residual flux amplitude is determined by the position on the voltage wave (switching
angle) during switching out. When this residual flux matches the prospective flux when the transformer is
energised, the inrush current is minimised. The larger the difference between the residual flux and the
prospective flux when the transformer is energised the greater the inrush current and voltage step change.
Transformer switching, especially in DNO networks, is usually performed randomly, i.e. the opening and
closing of the circuit breaker is performed based on the conditions it is subject to either by CB protection
trip or a periodic maintenance schedule.
If a transformer is energised at a random instant, it is possible that no transient inrush current will occur;
but in most cases a transient inrush current will arise [3]. This is because the transformer transient inrush
currents depend not only on the instant of energisation, but also on the residual flux of the previous de-
energisation. As mentioned in [4], the worst case transformer core residual flux that yields the transformer
peak inrush current is about 80%.
Based on the control scenario and transformer residual flux described above; an initial transformer
energisation study has been carried out to assess the worst case voltage drop at the 33 kV busbar due to
normal transformer energisation.
The major factors in determining the voltage drop (and inrush current) are the transformer rating, the
system fault level and the system X/R ratio.
When considering which Substations and TASS transformers to model we have endeavoured to make
sure any of the results can be extrapolated and general conclusions can be drawn. To do this, we have
considered four sites with four different transformer sizes (5 MVA, 10 MVA, 15 MVA and 20 MVA). To
allow these sites to represent all the transformers in the SEPD 33/11 kV primary distribution substations,
we have considered the effect of varying the winding connections for the 5 MVA transformer to
demonstrate the effect this will have.
Table 3.1 shows the location of the transformers (i.e. the SEPD substations), nameplate rating and the
estimated peak inrush current (assuming that the transformer core residual flux is at ±80%) of the
substations modelled.
A PSCAD-EMTDC test circuit, as shown in Figure 3.1, has been developed to carry out the transformer
energisation study. The transformer residual flux is set at 0.8,0 and -0.8 pu respectively for each of the
transformer cores (numbered 1, 2 and 3 in the model). The transformer is assumed to be unloaded. The
33 kV network is represented by a Thevenin equivalent source model where the network impedance
represents the system fault level with an assumed system X/R ratio of 15. This is a simplistic model that
does not consider parallel circuits.
Figure 3.1: PSCAD–EMTDC test model for the transformer energisation study
33.0
0.0
15 [MVA]
V Ph
33 [kV] / 11 [kV]
I_HV1
#1 #2
Ea BRK 1.0e9 [ohm]
33, 50 [Hz]
100 [MVA]
Z1 = 0.1751 [ohm] + j2.4083 [ohm]
To allow the results for these sites to be extrapolated for sites with different system fault levels we have
varied the considered system fault level from 200 MVA to 1000 MVA with an incremented step of
200 MVA.
To find the worst case voltage drop, the switching was performed by varying the switching angle over one
power frequency AC cycle (0 to 360 degrees) with an incremented step of 5 degrees to calculate the worst
case switching angle that yields the worst case voltage drop. It should be noted that the worst case
switching angle may differ from one type of transformer to another. The worst case voltage drop at the 33
kV busbar was analysed and recorded.
3.2.3 Results
Figure 3.2 shows the worst case system voltage drop against the system fault level when energising an
unloaded three phase two winding transformer with 80 % residual flux in its iron core prior to the
energisation. For operating limit compliance analysis, a vertical line corresponding to the ER P28 3%
voltage drop limit is plotted against the curves.
Figure 3.2 allows the estimation of the worst case system voltage drop for a transformer energisation given
a transformer rating, winding configuration and system minimum fault level at the energising busbar.
This can be done by selecting the most suitable curve based on the transformer rating and winding
configuration. The worst case system voltage drop can then be read by selecting the point on the curve
with the correct system fault level. Where a specific curve is not suitable, a figure can be estimated by
averaging between the closest curves.
Figure 3.2: Transformer energisation assuming 80% residual flux - Prospective worst case voltage drop at the energising bus against the system fault level
3.2.4 Results applied to WP1 nominated substation for the LEAN TASS method
Table 1.1 presents the sites selected in WP1 for detailed analysis. We have estimated the worst case
voltage drop for each of these substations using Figure 3.2 and the substation information. The results for
each substation are presented in Table 3.2. Table 3.2 shows that, at each of the SEPD 33 kV distribution
substations nominated for the TASS method, the prospective worst case system voltage drop during
transformer energisation exceeded the 3% voltage drop limit specified by the ER P28. For TASS to be
employed in the SEPD distribution network, more sophisticated switching procedures will be required to
keep the magnitude of the system voltage drop within the specified limits.
Table 3.2: Lists of SEPD 33 kV substation and its associated 33 kV transformer rating, minimum fault level and
prospective worst case system voltage drop
Prospective worst case
system voltage drop
33/11 kV Transformer System Minimum fault based from Figure 3.2
SSEPD 33 kV Substation Rating (MVA) level (MVA) (ΔV %)
BEMERTON 15 551 7.8
BLACK BOURTON 10 178 14.6
BOURTON 10 173* ~14
FARNBOROUGH 20 401* 18.8
AIRFIELD
GILLINGHAM 15 220 16.5
GREEN PARK 20 510 15.9
HEDGE END 15 408 9.9
HOLES BAY* 15 330 12.6
HONDA 15 230* 16.2
LANGLEY 10 458* 7.2
LEAFIELD 5 264 5.9 - 8.6**
MAIDEN NEWTON 10 167 >14
PEWSEY 10 103* >17
REDLANDS 15 293* 13.7
ROSE GREEN 20 283 >20
SOUTHCOTE 15 140* >19
TRADING ESTATE 15 469 9
UPPER HEYFORD 7.5 242 10.5
WESTON 20 378 19.9
WINDRUSH PARK 15 330 12.6
(*) Minimum fault level data were not available for all the selected sites. Fields marked with an asterisk
represent estimated minimum fault levels calculated as 65% of the fault level data taken from SEPD’s Long
Term Development Statement.
(**) The vector group for the transformers at Leafield was not considered to have an impact at the final
result. The prospective worst case system voltage drop for this substation is presented as a range of
values. From this we have concluded that the 3% limit in ER P28 will not be complied with regardless of
the vector group.
To verify the results obtained from this initial assessment, detailed PSCAD-EMTDC switching studies have
been carried out for the Hedge End 33/11 kV and Rose Green 33/11 kV substations. These are presented
in Appendix C.
These substations were chosen because Hedge End has local 33 kV circuit breakers and Rose Green has
remote 33 kV circuit breakers for the switching of the transformers. The presence of cable in-between the
circuit breakers and the transformers is not considered in the general cases studied above.
In addition, the detailed study of Hedge End allowed for consideration of the sympathetic inrush and its
effects on the voltage step at energisation.
A detailed PSCAD-EMTDC model of Hedge End 33 kV substation has been developed. A transformer
energisation study was carried out to confirm the conclusions given in Section 3.2.4. The worst case
calculated voltage drop at the Hedge End 33 kV substation when energising the 15 MVA 33/11 kV
transformer is 9.3% which is consistent with the 9.8% voltage drop estimated using the curve in Figure 3.2.
When considering the switching angle for Hedge End it can be seen that the ER P28 limit of 3% would be
exceeded on approximately two out of every three random switching events. The results for Rose Green
were consistent with the Hedge End results.
Another aspect of the TASS method is that it requires a transformer being energised or de-energised whilst
another transformer is online. This condition may be unfavourable to transformer energisation as it may
induce sympathetic inrush in the other transformer and may cause nuisance to the system.
Transformer energisation studies were carried out to investigate the effect of sympathetic inrush. The
Hedge End 15 MVA 33/11 kV transformers were energised with and without the other 15 MVA 33/11 kV
transformer online. The sympathetic inrush phenomenon can be observed on the resultant transformer
inrush current as shown in Figure 3.3 below. The sympathetic inrush reduced the damping of the transient
inrush current, producing a longer current decay time. However, the inrush current decays to a reasonable
level in 1.5s.
Were the TASS method to be implemented at the Hedge End 33 kV substation the impact of sympathetic
inrush would be minimal and negligible. Sympathetic inrush is a localized problem and, depending on the
winding connection of the transformer and grounding of the system, the impact may differ slightly from one
substation to another. The sympathetic inrush does not influence either the peak inrush current or the
magnitude of the voltage dip due to transformer inrush current. As far as the requirements of ER P28 are
concerned, the sympathetic inrush phenomenon will not cause significant issues when implementing the
TASS method.
Figure 3.3: Transient inrush current when energising the Hedge End 15 MVA 33-11 kV TASS-Transformer with
and without the Primary-Transformer. Illustration of longer transient current decay due to sympathetic
inrush.
Transformer transient inrush current
I_HV1_A - single I_HV1_A - Parallel
2.00
1.75
1.50
1.25
Current (kA)
1.00
0.75
0.50
0.25
0.00
-0.25
Time (s) 0.50 0.60 0.70 0.80 0.90 1.00 ...
...
...
Some of the SEPD transformers are remotely switched from other substations. Transformer energisation
studies were carried out to investigate the performance of the SEPD 33 kV network when implementing
TASS to a substation with remote breakers. For this, a detailed EMT model of SEPD Rose Green 33 kV
substation has been developed. The 20 MVA 33/11 kV transformers at this substation were energised
remotely from Hunston 33 kV substation via a 5 km feeder cable. Two energisation scenarios have been
simulated and reported:
1. energisation of the 20 MVA 33/11 kV transformer assuming that the transformer residual flux is fully
demagnetised due to ring-down transients (resonance phenomenon); and
2. energisation of the 20 MVA 33/11 kV transformer with detailed simulation of the ring-down transient to
calculate the prospective residual flux in the transformer iron (a more accurate calculation of the effect
of the ring down/resonance phenomenon).
We consider the second of these studies to be more valid as it provides a more accurate calculation. The
result of the detailed study shows that there will be residual flux in the 20 MVA 33/11 kV transformer iron
core when de-energised. The study shows that the voltage drop at the Hunston 33 kV substation when
energising the Rose Green 20 MVA 33/11 kV transformer is 10%.
This is significantly less than the >20% calculated in Table 3.2 but still exceeds the 3% voltage drop limit
specified by the ER P28. As such, for TASS to be employed at a substation with remote breakers,
sophisticated switching procedures are likely to be required to keep the magnitude of the system voltage
drop within the specified limits of ER P28.
3.2.6 Conclusions
We have produced generic curves for the worst case voltage drop against minimum system fault level (see
Figure 3.2) in order to assess the ability of any TASS transformers to comply with the ER P28
requirements for voltage step changes. We have verified the accuracy of these curves by carrying out
detailed studies.
When considering this curve against the WP1 selected sites, we can see that no sites will be compliant
with the ER P28 when using existing switchgear with random switching.
During our detailed studies we considered the effect of a long section of cable between the circuit breakers
and the transformers. This section of cable introduces the possibility of ring-down reducing the magnitude
of transformer residual flux. This will have a beneficial effect on the voltage step change during
energisation of a transformer. However, in the detailed case we studied, the effect was not sufficient to
ensure compliance with the Distribution Code requirements for normal routine switching operations.
We conclude that for the all substations considered, the worst case voltage step change for transformer
switching will not be compliant with ER P28 infrequent switching.
To implement TASS for sites with switching occurring more than once per year, it will be necessary to
consider other switching strategies.
3.3.1 Introduction
The use and theory of applying controlled switching to the existing ganged 33 kV circuit breakers is
discussed in Section 2.3.2.
To consider the effectiveness of the selected controlled switching device (SynchroTeq MV), this section will
describe the procedures necessary to simulate its operation using models in the EMT program PSCAD-
EMTDC; describe the results; and comment on the effectiveness of operation.
To ensure that the SynchroTeq MV switching relays are effective, their operation has been simulated in the
two detailed PSCAD-EMTDC models developed for Hedge End and Rose Green.
This section will detail the simulation for Hedge End. The PSCAD-EMTDC model used for this simulation
is described in detail in Appendix C.
In the simulation, the TASS-Transformer (as defined in Appendix C) circuit breaker is programmed to close
when the phase A system source voltage reaches a zero crossing. The simulation assumes that the
primary-transformer (as defined in Appendix C as the transformer without TASS switching) is online and
that the LV load at the 11 kV bus is at the TASS cross-over point (consistent with switching on the TASS
transformer).
In the simulation the 15 MVA TASS-Transformer was energised and monitored until it reached a steady
state from which the transformer prospective flux could be calculated based on a voltage feedback loop.
Next, the transformer was de-energised to calculate the transformer residual flux. Figure 3.4 shows the
simulated prospective flux plotted on top of the residual flux. The optimal switching angle can now be
obtained.
Figure 3.4: Optimal switching angle (point-on-wave) for energising the 15 MVA 33/11 kV transformer
The TASS-Transformer is energised again at the optimal switching angle as identified in Figure 3.4. At this
angle, the residual flux at each phase matches its prospective flux. Figure 3.4 shows that at the instant of
the energisation, the effective core residual flux corresponds to that which would exist in the transformer
during normal operation (no residual flux offset that would initiate the transient inrush was observed).
Figure 3.5 shows the transformer flux of all three phases after the energisation.
Figure 3.5: Transformer Core residual flux associated with the voltage drop
Figure 3.6: Peak Instantaneous Inrush Current associated with the worst case voltage drop
Figure 3.6 shows the instantaneous peak transformer inrush current. The maximum instantaneous peak
inrush current is almost negligible. Figure 3.7 shows the retained RMS voltage at the Hedge End 33 kV
bus. There is virtually no voltage drop observed. The instantaneous phase to ground voltage at the 33 kV
bus shown in Figure 3.8 further confirms that optimal switching angle has been achieved. Virtually no
voltage distortion can be observed when the 15 MVA 33/11 kV transformer is energised.
Conceptually, controlled switching with a flux requisition method can be simulated using PSCAD-EMTDC.
The simulation shows that the method (used by the control switching device SynchroTeq) can greatly
reduce transient inrush due to transformer energisation. This study was repeated for Rose Green
substation with similar results observed. The results for Rose Green are presented in Appendix D.
We would anticipate that the 33kV circuit breakers that SynchroTeq may be used to control would be
spring operated. The time delay in the operation for spring-type circuit breakers can be considered
predictable and dependant on the mechanical properties of the circuit breaker. We note that the
SynchroTeq device includes the capability to monitor accuracy and as such can tune the delay timing for
the circuit breaker based on operational experience.
It is noted that the circuit breaker poles seldom close simultaneously. Typical spread of closing time
between three phases is in the order of 2.5-3 ms between the first and the last pole.
Worst case transformer switching generally produces the same switching pattern as shown in the figure
below. For a span of time (a set of switching degree - not necessarily 180 degrees as shown in the figure)
the residual flux is what would normally exist in the transformer iron core. During this time, the inrush would
be very minimal.
The figure shows that, for the Hedge End substation, even if the CB scatter time is 3 ms between the
closing the Circuit breaker first pole to the last pole, the voltage drop would be still be less than 3%. We
would expect similar results from the other substations under consideration.
Figure 3.9: Impact of Circuit Breaker time delay and inconsistency on effectiveness of operation at Hedge End
3.3.4 Conclusion
In order to assess the effectiveness of a SynchroTeq MV controlled switching device in meeting the
voltage step requirements of ER P28, we have modelled the operation of the switching device as
described by the manufacturer. Our simulations show that during energisation using such a switching
device the voltage step seen reduces substantially to a negligible level. We have carried out this study for
two sites with the same results found.
We can conclude that a SynchroTeq MV controlled switching device is theoretically capable of reducing
the voltage step seen at energisation to below the levels required to comply with ER P28.
However, as the operational experience of SynchroTeq MV within the power industry in Great Britain is
minimal, it is recommended that this conclusion be verified through discussions with the SynchroTeq
manufacturer - Vizimax Inc and through on-site testing.
3.4.1 Introduction
The use and theory of applying controlled switching using independent pole operated circuit breakers is
discussed in Section 2.3.3.
To consider the effectiveness of this point on wave control, this section will describe the procedures
required to simulate its operation using models in the EMT programme PSCAD-EMTDC. We will describe
the results and comment on the effectiveness of operation.
We have modelled the operation of the ABB T183TM [5] point on wave control device in the simulation.
Other manufactures offer similar devices. Figure 3.10 shows the application schematic of a controlled
switching device for no-load transformer inrush current.
Figure 3.10: ABB’s Switchsync T183TM Application Schematic For No-Load Transformer Inrush Current Mitigation
and Switching [2]
To ensure that the ABB T183TM [5] point on wave control device is effective; its operation has been
simulated in the two detailed PSCAD-EMTDC models developed for Hedge End and Rose Green.
This section will detail the simulation for Hedge End. The PSCAD-EMTDC model used for this simulation
is described in detail in Appendix C.
In the simulation, the TASS-Transformer (as defined in Appendix C) is energised. The energising circuit
breaker poles are programmed to close independently. The circuit breaker closing time was programmed
according to the simplest method in controlled switching strategy specified in reference documents 4 and
5. The strategy is to close one phase at peak voltage and subsequently, 4.3 ms later, close the other 2
phases almost simultaneously as shown in Figure 3.11.
The results of the transformer energisation are shown below including the transformer peak inrush current,
the system voltage drop and the transformer terminal voltage waveforms. Figure 3.13 shows the
instantaneous peak transformer inrush current. The maximum instantaneous peak inrush current is
approximately 4 A. Figure 3.14 shows the retained RMS voltage at the Hedge End 33 kV bus. There is
virtually no voltage drop observed. The instantaneous phase to ground voltage at the 33 kV bus shown in
Figure 3.15 further confirms that the optimal switching angle has been achieved. Virtually no voltage
distortion can be observed when the 15 MVA 33/11 kV transformer is energised.
4.3 ms
Figure 3.12: Transformer Core residual flux associated with the voltage drop
Figure 3.13: Peak Instantaneous Inrush Current associated with the worst case voltage drop
3.4.3 Conclusions
For analysis purposes, the concept of single-pole controlled switching was simulated. The transients
simulations performed on the Hedge End 33 kV substation were repeated using a simple control closing
strategy described in [6]. The 15 MVA transformer was energised by closing the first pole of the circuit
breaker at maximum peak voltage and subsequently closing the other two poles simultaneously 4.3 ms
later. The results show that the voltage drop at the Hedge End 33 kV substation was reduced from 9.3%
to almost negligible. These simulations were repeated for Rose Green substation with similar results.
As such, we can conclude that a point on wave controlled switching device (such as ABB T183TM) would
be capable of reducing the voltage step seen at energisation to below the levels required to comply with
ER P28.
Point on wave controlled switching devices have been on the market for a long time and are widely used.
However, they can only be used with single-pole operated circuit breakers. Currently, there are no
conventionally used single pole operated circuit breakers for 33 kV nominal voltage equipment.
3.5.1 Introduction
The use and theory of applying pre-insertion resistors to reduce the inrush current and voltage step change
is discussed in Section 2.3.4. In general, a pre-insertion resistor limits inrush currents by the momentary
insertion of a resistive device into the circuit before full energisation of the transformer. The insertion of the
resistor is a two-step process. The initial circuit is made through the pre-insertion resistor in series with the
circuit breaker environment (i.e. SF6 environment). The resistor is then shunted as the main contacts
close.
To consider the effectiveness of pre-insertion resistors, this section will describe the procedures to
simulate their operation using models in the EMT programme PSCAD-EMTDC, describe the results and
comment of the effectiveness of their operation.
To model the use of a pre-insertion resistor, we have simulated the energisation of the TASS transformer
at Rose Green. The PSCAD-EMTDC model used for this simulation is described in detail in Appendix C.
In the simulation, the TASS-Transformer (as defined in Appendix C) is energised via a 5 Ω pre-insertion
resistor programmed to the PSCAD-EMTDC circuit breaker model and taken out 5 ms after.
The results of the transformer energisation are shown below including the transformer peak inrush current,
the system voltage drop and the transformer terminal voltage waveforms. Figure 3.16 shows the
instantaneous peak transformer inrush current. The maximum instantaneous peak inrush current is
approximately 210 A. Figure 3.17 shows the retained RMS voltage at the Rose Green 33 kV bus. The
instantaneous phase to ground voltage at the 33 kV bus is shown in Figure 3.18. Virtually no voltage
distortion can be observed when the 20 MVA 33/11 kV transformer is energised. An instantaneous
transient overvoltage of about 35 kV-peak can be observed at the instant of the energisation. This
overvoltage is due to the presence of the cable circuit, we would not expect this in substations with local
circuit breakers. The overvoltage quickly decays and no sustained overvoltages were observed.
The simulation was repeated for the Hedge End substation with similar results.
Figure 3.16: Transformer Core residual flux associated with the worst case voltage drop
Figure 3.17: Peak Instantaneous Inrush Current associated with the worst case voltage drop
3.5.3 Conclusions
From our studies and industry experience we can conclude that pre-insertion resistors can be used to
reduce transient inrush due to energisation of an unloaded transformer. However, it is no longer
considered to be a preferred solution by many network operators. This view is reflected in the 2014 report
of CIGRE Working Group C4.307, “Transformer energization in power systems” [4] which states:
“..Pre-insertion resistors require relatively large resistors to be installed in series to the main circuit breaker
interrupters and an effective reduction of inrush current is achieved only by an optimal choice of the
resistance value and the pre-insertion time. However, the addition of switching devices equipped with pre-
insertion resistors in series with the energizing breaker is probably not practical because such devices
would have to be specially ordered from the manufacturer at a high cost. Circuit breakers equipped with
pre-insertion resistors are also no longer available off-the-shelf for voltages less than 500 kV since modern
EHV breakers are designed for use with POW closing”.
From the studies described in this section we can conclude the following:
Option 1: Uncontrolled switching using the existing equipment cannot comply with the requirements for
normal routine switching operations given in the Distribution Code.
Option 2: Controlled switching using the existing primary equipment can comply with the requirements
for normal routine switching operations given in the Distribution Code.
Option 3: Controlled switching using new primary equipment can comply with the requirements for
normal routine switching operations given in the Distribution Code.
Option 4: Switching using a pre-insertion resistor can comply with the requirements for normal routine
switching operations given in the Distribution Code.
Option 3 does not appear to be an economically viable solution as the primary equipment required is not
conventionally available at 33 kV substations.
Option 4 may be feasible but will require the installation of site specific primary equipment for each
application of the TASS scheme.
Option 2 does not require the installation of any additional primary equipment (other than measuring
devices such as VTs if these are not presently installed). As such it can be considered a favourable
option. However, as the modelled device (SynchroTeq MV) does not yet have an extensive operational
experience within the power industry in the UK, it is recommended that this conclusion be verified through
discussions with the SynchroTeq manufacturer - Vizimax Inc and through on-site testing.
4.1 Introduction
As part of the Project Background, Southern Electric Power Distribution issued Mott MacDonald with a
report by High Frequency Diagnostics on the potential effects on transformer health of TASS. This
included consideration of the effect on transformers of repeated energisations/de-energisations and
prolonged de-energisation.
The report by High Frequency Diagnostics identified two major risks to primary 33/11kV distribution
transformers introduced by TASS [7] [8]. These are:
Physical damage resulting from the magnetising inrush currents that occur during energisation. The
cause of this physical damage would typically be inrush currents causing a mechanical shock to the
core and windings as the wrapped windings attempt to straighten in response to the current driven
alternating magnetic field. This has the potential to damage the internal structure or windings due to
the considerable forces involved. In addition, there is the potential that the solid winding insulation
could become damaged.
Migration of moisture from the oil to the paper insulation due to the more frequent thermal cycling of the
transformer causing the cellulose insulation eventually to breakdown. In particular, when the
transformer has prolonged periods of de-energisation this has the potential to allow the transformer to
cool substantially followed by sudden flows of energy during re-energisations having an effect on
moisture dynamics within the transformer. Of particular concern is that during the cool-down period
after switch off, the relative saturation of oil will increase. At its extreme at 100% relative saturation,
water will precipitate out and greatly reduce the dielectric strength of the oil.
To consider how best to identify if these or other failures are occurring in a transformer, this report will
consider the available condition monitoring options and comment on their suitability.
One of the simplest methods for transformer asset health monitoring is oil sampling.
The oil sample can be directly tested for a number of indicators of the transformer health, these include:
moisture content
dielectric strength
acidity
surface/interfacial tension
the presence of contaminants (such as PCBs or metals)
Furfuraldehyde analysis (to detect the breakdown of solid insulation)
These samples can be tested remotely from site, or if required, on-line sensors can be fitted to the
transformer.
A more sophisticated method of oil sampling is dissolved gas analysis (DGA). DGA relies on the fact that
different gases are known to be produced in transformer oil when it is exposed to excessive insulation
degradation or heating (i.e. due to partial discharge events, flashovers, internal arcing or other causes). By
understanding the composition, ratios and volumes of the involved gases, an idea of the health of the
transformer can be inferred.
DGA can be considered to be effective in detecting thermal and moisture issues (these affect the
composition of the oil). However DGA may only detect mechanical issue where insulation or paper is
damaged. DGA may not detect damage to the core or windings.
DGA involves taking a series of oil samples from a transformer and logging the occurrence and volumes of
different gases as a function of time. It is important to understand that you cannot assign a particular state
of health for a transformer given a single DGA result. Instead, a baseline of gas composition ratios is
defined and any change or rate of change of these values is of great interest.
DGA analysis is typically carried out on an annual basis for primary transformers. This typically involves a
sample of oil taken from transformer and analysed remotely at laboratory.
To understand the condition of a transformer better, the frequency of testing could be increased.
Alternatively, on-line DGA systems with local gas analysis are available in two types:
Single output systems, that generate a single output signal that is proportional to one or several of the
present gasses. These systems are useful for identifying issues and correlating them with specific
operating conditions, giving clues as to the origins of a fault.
Multiple gas analysers, that is able to identify the level of several diagnostic gases. These systems are
capable of providing more detailed information on the type of fault and its development.
For the systems with local gas analysis the sensor designs require direct contact with the transformer oil in
order to complete the analysis process. The sensors consist of two types:
Direct, mounted on a valve integrated with the side tank wall or in a pipe containing oil. Single output
systems are primarily this type.
Oil sample loop, where the unit takes an oil sample via small bore pipework in either a continuous or
periodic manner. These systems return the oil at the same or a different point. Multiple gas analysers
are typically of this design.
On-line DGA sensors with local gas analysis are expensive units, particularly multiple gas designs. They
also require maintenance and may require replacement within the transformer’s life.
Partial discharge analysis involves capturing and analysing the high frequency signals (electro-
magnetic/audible) that are generated by an incomplete breakdown of a dielectric material (oil or paper) in
the transformer.
This partial discharge may occur in a number of different locations in a transformer (including the winding,
tank or oil). Partial discharge can be caused by either mechanical or thermal/moisture issues as such this
can be considered to be an effective test for detecting all of the major issue due to TASS.
PD analysis has been applied more commonly for condition monitoring in transmission applications but still
retains the following operational challenges:
noise rejection
capital cost of data acquisition hardware
memory constraints
locating of the PD source within the transformer
the lack of a generalised rule based system that related activity to asset health
Analysis is typically completed on a unit by unit basis by an experienced Engineer, with a profile of the
activity assessed over time and variations from this, “normal” behaviour triggering an alarm.
Monitoring the current and voltage experienced by the transformer has the potential to provide the
following benefits:
The data can be used in a risk mitigation process. In particular, monitoring the magnitude and
frequency of inrush currents and over voltages.
Once a baseline of the response of the transformer to the switching events has been completed,
analysis of any trends in data could be used to identify if an issue has occurred. This idea relies on the
assumption that voltage and current waveforms are responses of the transformer to known switching
conditions. If the responses start to change and the external conditions are the same, there may be an
issue with the transformer.
Another advantage of this technique is that the existing SCADA protection system could potentially be
re-configured to generate the required data. This will require further consideration.
This test does not directly monitor any damage but will help build understanding of the mechanical
loads placed on the transformer during inrush.
Acoustic analysis is widely used for the monitoring of rotating machines where the kinetic energy involved
is distinctly different from that which a transformer experiences. However, it has been proposed that in the
event of a shift in the iron core within the transformer, the vibration signature (associated with
magnetostriction) is a potential identifier for a core shift event. Having an array of mechanical sensors
attached to the transformer may also allow us to gain an idea about the mechanical shock introduced to
the unit during switching.
Sweep frequency response analysis (SFRA) is a widely accepted technique for probing the mechanical
integrity of power transformers. It relies on the understanding that as the distributed RLC properties of a
transformer are dependent on material properties and geometries. By completing a frequency response
on a unit, variations in results will relate to variations in health. This technique is commonly used after
delivery of a transformer. This can typically be carried out periodically (every 5 or 10 years) during the life
of the transformer to understand the transformer asset health.
This is an offline test and will require the transformer to be switched out. This would not normally be
suitable to be carried out more often than during scheduled maintenance windows but due to the nature of
the TASS scheme, this could be implemented more frequently (i.e. when the transformer is switched out
for TASS as such there will be no requirement to organise an outage).
SFRA could be combined with power factor, tan delta and capacitance tests to gain a fuller understand of
the transformer asset health.
A summary of the proposed techniques and the associated risk that they are designed to monitor is
provided in Table 4.1. It should be noted that for a number of the transformer condition monitoring
methods, in order to identify the activity of a specific risk, a detailed understanding of the network
configuration at the point of switching is required.
Required
transformer Able to detect Able to detect
outage for mechanical oil/insulation
Method Advantage Disadvantage test damage degradation
information
DGA with local Real time Requires the No Partially (only Yes
testing information installation of on- some effects of
line equipment mechanical
May not be damage)
possible on all
transformers
Furfuraldehyde A powerful tool Additional No Partially (only Yes
when used in information some effects of
combination with required to mechanical
DGA identify faults damage)
Partial discharge Extremely sensitive De-noising in a No Partially (only Yes
sensors are substation some effects of
available environment is mechanical
challenging, real damage)
time analysis
equipment is
expensive
Current and Potentially use Large volumes of No Will provide No
Voltage existing SCADA data will be indication of the
measurement system to acquire generated, forces on the
data. As such this generating transformer
will be low cost and relationships
provide valuable between trends in
data on inrush data and asset
current limiting health will take
technologies. time
Acoustic / Vibration Potentially provide Need to identify No Yes No
online monitoring of an independent
mechanical analysis system
condition that is
commercially
available
Sweep frequency Detailed diagnostic It is only available Yes Yes No
response analysis of the core condition offline.
(SFRA) available. Can be Each test requires
combined with a testing team on
Power factor/tan site and a planned
delta and outage.
capacitance test.
The TASS scheme
provides the
opportunity to carry
out frequent offline
tests.
4.4 Conclusion
There does not appear to be a single condition monitoring technology that will provide comprehensive data
required to fully assess if TASS is damaging a transformer. As such, we must consider using two or more
condition monitoring technologies to assess the health of the transformer. We will consider the condition
monitoring equipment required to assess each of the major identified risks.
For tests detecting if physical damage has resulted from magnetising inrush current, perhaps the first thing
to consider is if high levels of magnetising inrush current will be experienced by the transformer. Where
inrush limiting measures (such as dedicated relays, point on wave switching or pre-insertion resistors) are
applied; the need to monitor the for this physical damage can be mitigated by calculations showing the
inrush currents will be limited and onsite data showing that the actual inrush currents are limited.
If the inrush currents are not limited, then techniques of acoustic monitoring and sweep frequency
response analysis (SFRA) have been assessed as two potentially applicable asset health monitoring
technologies. Of these, only SFRA is widely applied in the transmission and distribution industry.
To carry out SFRA will require a transformer outage and an on-site testing team. This may be feasible for
a test of a transformer with TASS, as TASS will require the transformer to be switched out. The frequency
of testing will require further consideration. SFRA may be combined with power factor/tan delta and
capacitance tests.
For detecting the migration of moisture between paper and oil as well as moisture into the oil from external
sources, the most effective test would be oil sampling with tests for moisture content and dielectric
strength.
The moisture content testing could be carried out in a lab (from frequent oil samples) or an online moisture
content monitoring device could provide the information. The choice between the two options would
depend on; if retrofitting online moisture content monitoring is feasible, the frequency of testing chosen and
if visits to site to sample oil could be combined with other activities.
In addition it may be advantageous to carry out oil sampling with tests for moisture content before applying
TASS to any transformer so that transformer with high moisture content detected could be discounted.
Alternative network topology (ANT) is intended to ensure that the security of supply is maintained for
customers where the TASS scheme is applied at a substation.
In particular where TASS is applied at a two transformers substation, the purpose of ANT is to ensure
security of supply when the primary (non-TASS) transformer fails and the TASS transformer is out of
service.
The simplest application of ANT is to allow for interconnection between the substation with TASS and
adjacent substations.
Figure 5.1: Standard operating arrangement for primary distribution substations with an interconnecting circuit
The normally open point allows load to be moved between the two substations.
Where TASS is applied at one of these substations; during periods of low load one transformer will be
switched out of service. This is shown in Figure 5.2
In this case, if the primary transformer at substation A fails, all the load at substation will be lost until the
TASS transformer is switched on.
In order to ensure security of supply it is proposed that for ANT the interconnection is closed between
substation A and substation B. This is shown in Figure 5.3.
To limit fault levels, this interconnection could be closed only when the TASS transformer is about to be
switched out.
In addition, the interconnection need not be rated for the full load of Substation A. Firstly, because the
TASS transformer will only be switched out at low loads (normally less than 50% of the substation firm
rating). Secondly, the ANT scheme would normally be operated with automatic switching in of the TASS
transformer. As such the interconnector would only be expected to carry the full load of Substation A for a
short period of time.
These and other issue will require consideration for each substation where TASS could be applied. The
issues selected for further consideration are detailed in the initial risk assessment presented here:
6 Conclusion
This report provides a technical analysis of a number of issues related to the application of the LEAN
project including:
Consideration of required transformer energisation procedures to ensure compliance with the relevant
codes and practices.
Identify potential failure modes in the transformers with TASS applied and select condition monitoring
procedures to allow management of the risk of failure.
Identifying the risks to the security of supply from applying TASS and ANT. Find potential ways of
ensuring the security of supply is maintained, and identify risks in applying these methods
Transformer Energisation
From the studies documented in this report, we can make the following conclusion; for the majority of
installations, uncontrolled switching of 33/11kV transformers in the SEPD network using no additional
equipment will not comply with the requirements for normal routine switching operations given in the
Distribution Code.
The use of pre-insertion resistors can provide a distribution code compliant solution but will require the
installation of site specific primary equipment for each application of the TASS scheme.
A controlled switching relay design applicable for use with a ganged operation circuit breaker such as
SynchroTeq MV may provide a distribution code compliant solution. This would not require the installation
of any additional primary equipment (other than measuring devices such as VTs if these are not presently
installed). However, as SynchroTeq MV does not yet have an extensive operational experience within the
power industry in the UK, it is recommended that this conclusion be verified through discussions with the
SynchroTeq manufacturer - Vizimax Inc and through on-site testing.
To monitor physical damage due to inrush it is recommended to first monitor the magnitude of inrush
currents directly to ascertain the severity of the risk. If repeated large inrush currents are detected it is
recommended to explore the possibility of regular offline sweep frequency response analysis tests on the
transformer to detect if any physical damage has occurred.
For detecting the migration of moisture between paper and oil it is recommended to carry out regular oil
sampling with tests for moisture content and dielectric strength.
We have carried out an initial review and risk assessment of the issues related to the application of TASS
and ANT. This is provided in Figure 5.4 and addressing these issues will form a key part of the work
package three scope of the LEAN project.
Appendices
PSCAD-EMTDC is a time domain based simulation package. It applies numerical techniques to solve the
system differential equations using a time-stepping integration algorithm. This allows the complex
interaction that takes place between each part of the system to be included in the model. In the time
domain, the frequency dependant characteristics of transmission circuits can be represented. The model
is therefore much more complex than a load-flow, short-circuit, or transient stability program. In
“conventional” power system analysis packages, the electrical characteristics of the system are
represented by a single, per phase, equivalent circuit the parameters of which are selected at rated
frequency. Such programs cannot be used to analyse switching transients – for that task an
electromagnetic transients program such as PSCAD-EMTDC is required. For the studies reported, the
most recent release of PSCAD-EMTDC (X4) has been used, in conjunction with the Intel FORTRAN 11
compiler. Post-processing of the transient time domain wave shapes produced by PSCAD-EMTDC has
been performed using Livewire.
According to the TASS Method, as a minimum, there will be at least one Energised Primary Transformer
and one Switched Primary Transformer (TASS transformer). The energised primary transformer will
remain energised throughout the TASS method while the Switched Primary Transformer will be energised
or de-energised depending on the substation’s varying load throughout the day [7]. Figure A.1 shows a
general configuration of a typical primary distribution substation.
The following discusses the PSCAD-EMTDC modelling of the main elements of a TASS network:
The winding resistance and the leakage inductance (from the short-circuit test) define the series
impedance of the transformer.
At present there are a few developed theories used to model the magnetization of a transformer core
accurately. These models are generally based on the observed behaviour of real materials, and include
magnetic hysteresis effects. Two of the best-known and most widely used empirical methods to model
hysteresis are the Jiles-Atherton and Preisach models, which both provide an accurate representation of
the hysteresis loop.
Jiles and Atherton used a soft magnetic material for their investigation and modelled the saturation
characteristic using a modified Langevin function. Preisach theory is based on the fact that a
ferromagnetic material is made up of dipoles that can be represented as having coercivity, a magnetic
characteristic with two statistically distributed parameters and a field due to neighbouring dipoles.
For this study, the Jiles-Atherthon Core Hysteresis Model will be used.
The Jiles-Atherton theory describes the relationship between M (the magnetic moment), and H (the
magnetic field intensity). Electrical engineers are more accustomed to dealing with a B–H relationship in
the form of a hysteresis loop. Conversion from a B-H to an M-H loop is straightforward using the
relationship:
The PSCAD-EMTDC “classical” transformer model with the Jiles-Atherton Core Hysteresis Model
calculates changes in magnetizing current at every time step based on the changes in magnetizing flux.
The magnetizing flux is also calculated by integrating the magnetizing branch voltage.
Equation 2: Jiles-Atherton
Since no information on the type and geometry of the transformer cores has been made available, it is
assumed that the transformer core is made of typical good quality grain-oriented steel with a flux density in
the steady state of the order 1.6 – 1.7 T [3]. The key factor as far as inrush is concerned is the saturated
air core inductance of the winding. The air core reactance value Xs is varied until the peak inrush current
returned by the model, for worst case switching angle and residual flux combination, corresponds to the
values given by [9]] presented in Figure A.2 below.
The worst case assumed residual flux is 80%, 0% and -80% in the three limbs of the transformer [9]. This
is pre-set in the model as 0.8, 0 and -0.8 pu remnant flux at transformer core 1, 2 and 3 respectively.
Figure A.2: Non-linear Saturation Characteristic - Flux Linkage Plotted against Magnetising Current derived from
the PSCAD-EMTDC classical transformer model with the Jiles-Atherton core hysteresis model.
Figure A.3: Relative Magnetising Inrush Currents for Various Transformer Connections [3]
Equation 3 is used to calculate the possible worst case peak of the instantaneous inrush current. The
calculated inrush currents are shown in Figure A.3 above.
𝑀𝑉𝐴×106 ×𝐼𝑛𝑟𝑢𝑠ℎ𝐹𝑎𝑐𝑡𝑜𝑟
𝐼𝑃𝑒𝑎𝑘 =
√3×𝑉1 ×103
The modelling of mutual coupling between the phase winding will require more detailed transformer design
data. This data are not available for the transformers considered in the LEAN study as these transformers
were constructed 10-50 years ago. Based on [10], with respect to the inrush simulations, the transformer
model without the magnetic coupling between phases presented higher inrush current peaks than those
obtained from the model with magnetic coupling. As such more pessimistic results will be observed.
Transformer winding stray capacitances are calculated based on the recommendation from [3] and
estimated HV winding capacitance, CH to earth taken from [6].
The stray capacitance of the LV winding (CL) is three times the stray capacitance of the HV winding (CH).
The Stray capacitance between the HV and LV winding (CLV) is four times the stray capacitance of the HV
winding.
The 33 kV system is represented by a Thevenin equivalent source and the network system impedance is
calculated from SEPD minimum fault level and X/R ratio. The impedance will be calculated as follows;
𝑉𝑠𝑦𝑠𝑡𝑒𝑚2
𝑍𝑔𝑟𝑖𝑑 = ∗ (cos(atan(𝑋⁄𝑅 ) + 𝑗𝑠𝑖𝑛(atan(𝑋⁄𝑅))
𝑆𝑦𝑠𝑡𝑒𝑚 𝑓𝑎𝑢𝑙𝑡 𝑙𝑒𝑣𝑒𝑙
The interconnector cable is represented by pi-sections. The zero sequence impedance of the cable is
approximated by taking the ratio of the positive sequence/zero sequence impedance from a published
datasheet for a typical 33 kV cable.
The cables were modelled as cascaded pi-sections which effectively model the travelling wave effect but
neglect the frequency dependency of the cable resistance, resulting in slightly pessimistic results.
The number of pi-sections required for the correct representation of the lines depends mainly on the
expected frequency of oscillation. The highest frequency that can be attained by a pi-section
representation is the natural frequency of one individual element representing a fraction of the total line
length. This is given by:
1
𝑓 𝑚𝑎𝑥 = 𝐶′
[Hz]
2𝜋𝑠√(𝐿′. )
2
Where:
Table B.3: 11 kV load at the LV side of the 33/11 kV Primary-transformer associated with the cross-over point (load
assumed 0.95 lagging power factor)
Substation S (MVA) P (MW) Q(MVAR)
Rose Green 8.85 8.41 2.76
Hedge End 6.68 6.35 2.07
Hedge End 33 kV substation represents a good example of an SEPD standard primary distribution
substation that enables the TASS method to be employed using the existing switchgear. A single line
diagram of the Hedge End substation is shown in Figure C.1.
A PSCAD-EMTDC model of Hedge End 33 kV substation is shown in Figure C.2. This has been
developed to investigate the impact of transients during transformer switching operation. The following
modelling assumptions have been considered:
The 33 kV network is represented at the Hedge End 33 kV bus based on the minimum system fault
level [system minimum fault level]
C1MT is the Primary-Transformer which will remain online during the analysis. The transformer is
represented by its short circuit impedance. The transformer core saturation and winding stray
capacitances are also represented.
C2MT is the TASS-Transformer which will be switched in and out. The transformer is represented by
its short circuit impedance. The transformer core saturation and winding stray capacitances are also
represented.
The transformer feeder cable is assumed to be short and to therefore has a negligible impact on
results.
The load on the LV side of the primary-transformer corresponds to the cross-over point loads. For
Hedge End, the calculated cross over point load is 6.68 MVA. For the model the load is assumed to
have a 0.95 lagging power factor. The TASS-Transformer is assumed to be unloaded.
The circuit breaker is assumed to be a standard 33 kV gang-operated circuit breaker with a typical
chopping current level of 5 A.
Figure C.2: PSCAD-EMTDC model of the Hedge End 33 kV Primary Distribution Substation
TASS Switched Transformer
33 kV 2.16e-2 [uF]
5400e-6 [uF]
1.62e-2 [uF]
0.0
34.5
V Ph
200 [ohm]
5400e-6 [uF]
1.62e-2 [uF]
P+jQ
Scenario 1 – The 15 MVA 33/11 kV TASS-Transformer is energised against the 33 kV system at minimum
fault level without the availability of the Primary-Transformer.
Scenario 2 – The 15 MVA 33/11 kV TASS-Transformer is energised against the 33 kV system at minimum
fault level, with the Primary-Transformer online. The transformer peak inrush current, system voltage drop,
transformer terminal voltage waveforms, and impact of sympathetic inrush are analysed and reported.
For all the scenarios considered, the energisation is performed by closing the associated circuit breaker
simultaneously over one power-frequency electrical AC cycle (0-360 deg) with an incremented step of
5 deg using the PSCAD-EMTDC multi-run component. This is to calculate the least favourable switching
angle that yields the worst case condition.
Scenario 1 investigates the energisation of the Hedge End 15 MVA 33/11 kV TASS-Transformer without
the availability of the Primary-Transformer. The study assumes that the 15 MVA transformer is unloaded,
with the worst case residual flux distribution of ±80% in its iron core.
Figure C.3 shows the variations of the system retained RMS voltage at the 33 kV busbar as a function of
switching angle over one power frequency electrical AC cycle. Figure C.4 shows variations of the
corresponding transformer maximum magnetising peak inrush current.
Figure C.3 shows that at any switching angle (0 to 360 deg), assuming the worst case transformer residual
flux distributions, the voltage drop at the POI will in most cases exceed the 3% voltage drop limit specified
for the study.
The instantaneous peak magnetising transformer inrush currents are shown in Figure C.6. The maximum
inrush current calculated is 1.763 kA which is about 6.7 times the transformer full load current.
The results presented here are associated with the least favourable switching angle which yields the worst
case voltage drop at the 33 kV grid energising bus. The maximum voltage drop calculated is 9.3%. Note
that the maximum voltage drop is very likely to occur on only one phase due to the unsymmetrical
saturation characteristic by different switching instants; the different point-on-the-voltage-wave at the
instant of the switching/energisation; and the residual flux distributions.
The calculated voltage drop is consistent with the prospective voltage drop estimated from Figure 3.2.
Figure C.3: Retained RMS Voltage at the 33 kV Bus Figure C.4: Peak Transient Inrush Current
Figure C.5: Transformer Core residual flux associated Figure C.6: Peak Instantaneous Inrush Current
with the worst case voltage drop associated with the worst case voltage drop
Figure C.7: Worst case retained RMS voltage at the 33 Figure C.8: Instantaneous phase to ground voltage
kV Bus
Scenario 2 investigates the energisation of the Hedge End 15 MVA 33/11 kV TASS-Transformer with the
availability of the Primary-Transformer. The study assumes that the 15 MVA transformer is unloaded, with
the worst case residual flux distribution of ±80% in its iron core.
Figure C.9 shows the variations of the system retained RMS voltage at the Point of Common Coupling
(PCC) as a function of switching angle over one power frequency electrical AC cycle. Figure C.10 shows
variations of the corresponding transformer maximum magnetising peak inrush current.
Figure C.9 shows that at any switching angle (0 to 360 deg), assuming the worst case transformer residual
flux distributions, the voltage drop at the energising bus will in most cases exceed the 3% voltage drop limit
specified for the study. The maximum voltage drop 7.55 % as shown in Figure C.13.
The instantaneous transformer residual flux and peak magnetising transformer inrush currents are shown
in Figure C.11 and in Figure C.12. The maximum inrush current calculated is 1.8 kA which is 6.9 times the
transformer full load current.
The impact of sympathetic inrush can be seen in the longer duration of the transient inrush current as well
as in the voltage drop recovery. However as shown in Figure C.14, no sustained voltage disturbance was
observed at the energising bus throughout the study. The simulation shows that employing the TASS
method at the Hedge End 33 kV substation will not be affected by sympathetic inrush. However,
mitigations are needed to reduce the inrush current and its associated voltage drop at the energising bus
to be within the allowable limit.
Figure C.9: Retained RMS Voltage at the 33 kV Bus Figure C.10: Peak Transient Inrush Current
Figure C.11: Transformer Core residual flux associated Figure C.12: Peak Instantaneous Inrush Current
with the worst case voltage drop associated with the worst case voltage drop
Figure C.13: Worst case retained RMS voltage at the 33 Figure C.14: Instantaneous phase to ground voltage
kV Bus
Rose Green 33 kV substation is an example of an SEPD standard primary distribution substation where
the transformers are energised remotely. The 33/11 kV 20 MVA transformers at the Rose Green
Substation are energised from the Hunston 33 kV substation via a 5 km feeder cable circuit as shown in
the network single line diagram below (Figure C.15).
When energising a transformer together with a significant capacitance presence (i.e. long cables/overhead
lines) the best case scenario anticipated is that the core of the transformer is fully demagnetised and the
cable is discharged due to the interchanging of energy absorption between the transformer inductance and
the cable capacitance. This phenomenon is referred to as the transient’s ring-down effect. However, there
is always a possibility that the core will not be fully de-magnetised and that there will be residual flux in the
transformer iron core after it has been de-energised. This scenario is therefore considered for the
transformer energisation analysis at Rose Green 33 kV substation.
A PSCAD-EMTDC model of Rose Green 33 kV substation, as shown in Figure C.16, has been developed
to investigate the impact of transient during transformer switching operation. The following modelling
assumptions have been considered:
The 33 kV network is represented at the Hunston 33 kV bus based on the minimum system fault level
C1MT is the primary-transformer which will remain online during the analysis. The transformer is
represented by its short circuit impedance. The transformer core hysteresis and winding stray
capacitances are also represented.
C2MT is the TASS-Transformer which will be switched in and out. The transformer is represented by
its short circuit impedance. The transformer core saturations and winding stray capacitances are also
represented.
The transformer feeder cable is represented by two cascaded pi-sections with minimum length of
2.5 km to represent the cable response correctly up to 5 kHz.
The load at the LV side of the primary-transformer corresponds to the cross-over point loads. For
Rose Green, the calculated cross over point load is 8.85 MVA. For the model; the load is assumed to
have 0.95 lagging power factor. The TASS-Transformer is assumed to be unloaded.
The circuit breaker is assumed to be a standard 33 kV gang-operated circuit breaker with a typical
chopping current level of 5 A.
The Rose Green 33 kV substation is assumed to be earthed by a 1000 A earthing transformer. The
transformer is placed at the 33 kV bus.
Scenario 1 – The 20 MVA 33/11 kV TASS-Transformer, with no residual flux or trapped charge in the
cable, is energised against the 33 kV system at minimum fault level with the Primary-Transformer online
Scenario 1 investigates the energisation of the 20 MVA 33/11 kV TASS-Transformer at the Rose Green 33
kV substation. The study assumes that the 20 MVA transformer is unloaded and the Primary-Transformer
is online.
Figure C.17 shows the variations of the system retained RMS voltage at the Point of Common Coupling
(PCC) as a function of switching angle over one power frequency electrical AC cycle. Figure C.18 shows
variations of the corresponding transformer maximum magnetising peak inrush current.
Figure C.17 shows that at any switching angle (0 to 360 deg) - assuming the worst case transformer
residual flux distributions - the voltage drop at the POI will in most worst cases marginally exceed the 3%
voltage drop limit specified for the study.
The results presented here are associated with the least favourable switching angle which yields the worst
case voltage drop at the 33 kV grid energising bus. This is the best case scenario as it is assumed that
there is no residual flux in the transformer iron core prior to the energisation.
The maximum voltage drop calculated is 3% on Phase A as shown in Figure C.21. This is calculated
30 ms after the instant of the circuit breaker closing.
The instantaneous peak magnetising transformer inrush currents are shown in Figure C.20. The maximum
inrush current calculated is 656 A. The transformer full load current is approximately 2.5 times the
maximum inrush current.
Assuming the best case condition (i.e. no residual flux in the transformer iron core) the TASS method can
be employed at the Rose Green 33 kV substation using the existing circuit breaker while maintaining the
voltage drop at the energising bus within the allowable limit.
Figure C.17: Retained RMS Voltage at the 33 kV Bus Figure C.18: Peak Transient Inrush Current
Figure C.19: Transformer Core residual flux associated Figure C.20: Peak Instantaneous Inrush Current
with the worst case voltage drop associated with the worst case voltage drop
Figure C.21: Worst case retained RMS voltage at the 33 Figure C.22: Instantaneous phase to ground voltage
kV Bus
Scenario 2 simulates the ring-down transients when the transformer is being de-energised to calculate the
probable residual flux in the iron core. The transformer is energised then de-energised to simulate the ring-
down transients. In the best case scenario the residual flux in the iron core will be near to zero. This
condition is almost ideal as the ring-down transients depend on many uncertainties, for example the
resonance frequency between the capacitance and the inductance.
Figure C.23 shows the variations of the system retained RMS voltage at the Point of Common Coupling
(PCC) as a function of switching angle over one power frequency electrical AC cycle. Figure C.24 shows
variations of the corresponding transformer maximum magnetising peak inrush current.
The results presented here are associated with the least favourable switching angle which yields the worst
case voltage drop at the 33 kV grid energising bus. The instantaneous peak magnetising transformer
inrush currents are shown in Figure C.26. The maximum inrush current calculated is 2.47 kA which is
about 7.1 times the transformer full load current. The maximum voltage drop calculated is 10 % on Phase
B as shown in Figure C.27.
The simulation shows that there will be some residual flux in the transformer iron core when it is being de-
energised. In the worst case scenario these residual fluxes may reach up to 80% of the nominal flux. For
remotely controlled substations/transformers, mitigations are still needed to reduce transient inrush and to
maintain the voltage drop at the energising bus to keep it within the acceptable limit.
Figure C.28 shows the instantaneous phase to ground voltage at the energising bus during the
energisation. The instantaneous transient overvoltage of 35 kV-peak can be observed at the instant of the
energisation. The overvoltage quickly decays.
The worst case voltage drop calculated in both of the scenarios is less than the prospective voltage drop
calculated in Figure 3.2. This because the simulations carried out to produce the curve did not take any
damping elements into consideration other than the network impedance. The simulations show that the
curve in Figure 3.2 can only be used to estimate system voltage drop when considering a transformer that
is locally switched.
Figure C.23: Retained RMS Voltage at the 33 kV Bus Figure C.24: Peak Transient Inrush Current
Figure C.25: Transformer Core residual flux associated Figure C.26: Peak Instantaneous Inrush Current
with the worst case voltage drop associated with the worst case voltage drop
Figure C.27: Worst case retained RMS voltage at the 33 Figure C.28: Instantaneous phase to ground voltage
kV Bus
The following results shows the system response calculated at the Rose Green and Hunston 33 kV
substation during the energisation of the Rose green 20 MVA 33/11 kV transformer from the Hunston
33 kV substation. The energisation assumes that the switching angle is deduced from the transformer
residual flux requisition method.
Figure D.1 shows the prospective optimal switching angle when switching the transformer at the Rose
Green 33 kV substation remotely from the Hunston 33 kV switching bus. At this switching angle, the
transformer residual flux is equivalent to the prospective core flux of the transformer. Figure D.2 shows the
peak Instantaneous inrush current across the 33 kV circuit breaker at the Hunston 33 kV substation during
the energisation of the 20 MVA 33/11 kV transformer. The inrush current is mainly the component of the
cable capacitive current, which is shown by the rapid frequency characteristic at the instant of the
energisation. Figure D.3 shows the retained RMS voltage at the Hunston 33 kV energising bus. No
evidence of excessive voltage drop was observed.
Figure D.4 and Figure D.5 shows the maximum instantaneous phase to neutral voltage at the Hunston 33
kV energising bus as well as the Rose Green 33 kV. No excessive overvoltages were observed.
Figure D.1: Optimal switching angle (point-on-wave) for energising the 20 MVA 33/11 kV transformer at the Rose
green substation (blue plots – system phase to earth voltage ‘00000 V) (green and maroon plots – measured and
prospective residual flux – kwB-N)
Figure D.2: Peak Instantaneous inrush current across the 33 kV circuit breaker at the Hunston 33 kV substation
during the energisation of the 20 MVA 33/11 kV Transformer at the Rose Green 33 kV Substation
Figure D.4: Instantaneous phase to neutral voltage at the Rose Green 33 kV bus during the energisation of the
20 MVA 33/11 kV transformer from Hunston 33 kV substation
Figure D.5: Instantaneous phase to neutral voltage at the Hunston 33 kV energising bus