www.ietdl.

org
Published in IET Generation, Transmission & Distribution
Received on 18th March 2012
Revised on 22nd August 2012
Accepted on 5th September 2012
doi: 10.1049/iet-gtd.2012.0166

ISSN 1751-8687

Influence of sympathetic inrush on voltage dips

caused by transformer energisation
Jinsheng Peng1, Haiyu Li1, Zhongdong Wang1, Foroozan Ghassemi2, Paul Jarman2
1
School of Electrical and Electronic Engineering, University of Manchester, Manchester M13 9PL, UK
2
National Grid, Warwick CV34 6DA, UK
E-mail: [email protected]

Abstract: Energising a transformer with other adjacent transformers in service may induce sympathetic inrush that could lead to
long duration voltage dips. This study addresses this issue through analysing voltage dip events observed on a 400/132 kV grid
when energisations of several generator step-up transformers were performed. With the help of a number of eld measurements of
voltage dips involving sympathetic inrush, a detailed network model in alternative transients program/electromagnetic transients
program has been developed and validated. Based on the network model, comprehensive simulation studies have been performed.
In addition, because of the nature of multiple controlling parameters, sensitivity assessment has been carried out to identify the
key inuential parameters. It is found that the sympathetic inrush, although does not affect voltage dip magnitude, can prolong
voltage dip duration by 143% under the worst energisation condition. The sympathetic inrush occurring in substation transformers
gives rise to a network-wide sympathetic interaction, resulting in further prolonged voltage dips on the 132 kV grid. Finally, it is
proven that application of static var compensator is able to speed up the voltage dip recovery.

1 Introduction magnitudes; the dependence of voltage dip magnitudes on

network congurations was studied in [9]; the impacts of
Power quality is a growing concern for power system switching angle, residual ux and system loading condition
operators as well as customers in view of increasing on voltage dip magnitudes were assessed in [10]. Also,
utilisation of sensitive loads, such as variable speed drives voltage dips caused by energisation of medium-voltage
and microprocessors [13]. One of the main power quality wind turbine transformers were studied in [1114] to check
issues is voltage dip (sag) which is dened in the IEC grid code compliance. With detailed considerations of
61000-4-30 as temporary reduction of the voltage transformer saturation and upstream network characteristics,
magnitude at a point in the electrical system below a all those studies utilised Electro-Magnetic transients program
threshold. The possible causes of voltage dips can be (EMTP) simulation to carry out comprehensive assessments
short-circuit faults, motor starting or transformer [814]. However, most previous studies were devoted to
energisation. Voltage dips caused by short circuits and evaluating the magnitude of voltage dips rather than its
motor starting were systematically evaluated [46]. During recovery which could be more important when the
recent years, the voltage dips induced by transformer transformer inrush is coupled by a phenomenon called
energisation are gaining more attention [714]. In UK, a sympathetic inrush. A transformer could arouse and engage
3% threshold is normally applied to the voltage dips sympathetic inrush when it is energised into a system where
caused by transformer energisation, and this limit could be there are adjacent transformers in service. The mechanism
further tightened if the interval between consecutive of this phenomenon was interpreted in [16] and the key
transformer energisations is short, according to Engineering inuential factors were evaluated in [17, 18]. With most of
Recommendation P28 (ER-P28) [15]. the previous studies focusing on its impact on differential
The characteristics of voltage dips caused by transformer protection, recent studies also raised the concerns over its
energisation were revealed in [7] that they are inuences on voltage dips [19, 20].
non-rectangular (voltage dips recover gradually) and Based on an existing 400 kV transmission grid, this paper
non-symmetrical (each phase has different dip magnitude investigates the inuence of sympathetic inrush on voltage
because of the different degree of saturation). Voltage dips dips that occurred when energisation of large GSU
because of energisation of large generator step-up (GSU) transformers was carried out. Measurements of voltage dip
transformers for connecting conventional power plants were events accompanied by sympathetic interaction between
reported in [810]. Specically, study in [8] investigated 400 kV GSU transformers are presented. A network model
voltage dips because of energising a 315 MVA GSU is developed and validated against the eld measurement
transformer from a 138 kV network and suggested results. It is then used to assess the network-wide impacts
segregated point-on-wave method to minimise voltage dip of sympathetic interaction on voltage dips and perform

IET Gener. Transm. Distrib., pp. 112 1

doi: 10.1049/iet-gtd.2012.0166 & The Institution of Engineering and Technology 2013
www.ietdl.org
sensitivity studies to identify the key inuential parameters. should be noted that in this paper, the current drawn by the
Application of static var compensator (SVC) is also energised transformer is named as inrush current and the
evaluated to show its ability to speed up voltage recovery. induced current in adjacent connected transformers is
named as sympathetic inrush current). The measured
waveforms are shown in Fig. 2 which includes: the
2 System under study and field three-phase currents observed at power feeder 1 in terms of
measurements instantaneous value (Fig. 2a) and root mean square (RMS)
value (Fig. 2b), the voltages measured at substation K in
A large power plant needs to be connected to a 400 kV term of RMS value (Fig. 2c), and three-phase currents
transmission grid. As shown in Fig. 1, the grid consists of measured at the feeder IK which links substation I and K
11 substations, most of which accommodates a number of (as labelled in Fig. 1) in term of instantaneous value
autotransformers (400/132/13 kV, 240 MVA, YNa0d11). (Fig. 2d ). It should be noted that the calculation of RMS is
Substations are linked by 400 kV double circuit over one cycle and refreshed every half a cycle, which is
transmission lines with lengths ranging from 21 to 97.54 in accordance with IEC 61000-4-30. As can be seen in
km. There are capacitor banks located at substation G, H Fig. 2a, the no-load magnetising currents become gradually
and B and SVC devices connected at substation I, E and B. larger after the energisation, which indicate the initiation of
The power plant is located near substation K. The coupling sympathetic inrush in GSU transformers T2 and T3. Fig. 2b
circuit breaker CB1 is normally closed. Through power shows that the sympathetic inrush currents took about 2 s to
feeder 1, GSU transformers T2 (345 MVA, 400/19 kV, reach their maximum magnitudes (the largest maximum
YNd1) and T3 (415 MVA, 400/21 kV, YNd1) are magnitude appeared in phase C, which is about 120 A in
connected to substation busbar via CB2; through power RMS) and the decay of the sympathetic inrush lasted more
feeder 2, GSU transformer T1 (345 MVA, 400/19 kV, than 25 s. In Fig. 2c, it can be observed that the system
YNd1) is connected via CB3. took a long time for a full recovery of voltage for
To connect the power plant, three steps were conducted: three-phases. The currents measured at feeder IK (shown
the rst step was to energise the GSU transformers to in Fig. 2d ) only contain system load currents before
gain power supply from the main grid; the external power transformer energisation; after energisation, they are
supply was then used to start the auxiliary equipments superimposed by a portion of transformer inrush currents.
which are necessary for heating up the gas and steam The second measurement considered energising GSU
turbines; nally the generators were synchronised. transformers T2 and T3 simultaneously, with transformer
However, the energisation of GSU transformers caused long T1 already energised. The observation point of voltage dips
duration voltage dips which were reported by downstream is the same with that in the rst measurement. It was also
distribution utilities. Field measurements were therefore found in the second measurement that the inrush currents
conducted for further investigations. Here, two decayed very slowly. The recorded maximum dip
measurement cases involving sympathetic inrush between magnitude at substation K was about 32.5 kV on the most
GSU transformers are illustrated in details. affected phase (7.85% of the initial voltage). The voltage
The rst measurement was carried out during the dip disturbance lasted about 20 s with the onerous dips
energisation of GSU transformer T1, with T2 and T3 lasting about 10 s. It triggered low voltage alarms and led
already energised but with secondary side unloaded (it to responses of reactive power compensation devices.

Fig. 1 Schematic diagram of the system under study

2 IET Gener. Transm. Distrib., pp. 112

& The Institution of Engineering and Technology 2013 doi: 10.1049/iet-gtd.2012.0166
 www.ietdl.org

Fig. 2 Three-phase current and voltage waveforms of the rst measurement

a Instantaneous sympathetic inrush currents
b RMS sympathetic inrush currents
c RMS voltage dips
d Instantaneous currents of feeder IK

IET Gener. Transm. Distrib., pp. 112 3

doi: 10.1049/iet-gtd.2012.0166 & The Institution of Engineering and Technology 2013
www.ietdl.org
3 Sympathetic inrush and voltage dip amplitude of source voltage and is the energisation angle;
Vc(t) is the voltage of common busbar; is(t), i1(t) and i2(t)
The network shown in Fig. 1 can be generically described in are the currents owing through supply, TE1 and TE2,
Fig. 3a by a Thevenin equivalent circuit seen from any of two respectively.
parallel transformers TE1 and TE2 within the system. The Applying Kirchhoffs voltage law and Kirchhoffs circuit
line impedance between TE1 and TE2 is neglected; R1 and law, the system can be described by
R2 represent transformer winding resistances; L1 and L2
represent transformer leakage inductances; L1m and L2m
are the magnetisation inductances of TE1 and TE2, dis dl
respectively. Vs(t) is equal to Vmsin(t + ), where Vm is the Rs is + Ls + R1 i1 + 1 = Vm sin(vt + u)
dt dt

Fig. 3 Graphical representation

a Generic power system involving sympathetic interaction between transformers connected in parallel
b Flux-linkages of TE1 and TE2
c Sympathetic inrush currents i1 and inrush current i2
d RMS voltage dip at common busbar

4 IET Gener. Transm. Distrib., pp. 112

& The Institution of Engineering and Technology 2013 doi: 10.1049/iet-gtd.2012.0166
 www.ietdl.org
dl1 dl while the DC components in 1 are of opposite polarity and
R1 i1 + = R2 i2 + 2 the time constant of the rst DC component 1[ = (L + 2Ls)/
dt dt
(R + 2Rs)] is smaller than that of the second DC component
is = i1 + i2 (1) 2[ = L/R], so i1(t) will gradually reach the maximum peak,
and gradually decay afterwards.
where 1, 2 are the ux-linkages of TE1 and TE2, The simplied analytical analysis shows in a general way
respectively, and 1 = i1(L1 + L1m), 2 = i2(L2 + L2m). the variation of ux-linkages in TE1 and TE2 which
Owing to the non-linearity of core magnetisation depends on the time constants formed by the inductances
inductances, mathematical solution for (1) cannot be readily and resistances of the circuit branches. In reality, the core
obtained. To qualitatively show the sympathetic inrush inductance is non-linear and therefore the time constants
process, an analysis is made by assuming L1m and L2m to cannot be so readily determined. More accurate estimation
remain constant [21, 22]. Assuming TE1 is identical to of the sympathetic inrush transient can be achieved
TE2, we can obtain R1 = R2 = R, L1 + L1m = L2 + L2m = L. by using Alternative Transients Program (ATP)/EMTP
Supposing TE1 has already been energised, energisation of simulation. Based on the network equivalent parameters and
TE2 would induce change of 1 and 2 as a function of by applying a non-linear saturation curve to magnetisation
time, which can be expressed by inductances L1m and L2m, it is possible to simulate the
transformer uxes (1 and 2), the currents (i1 and i2) and
L the busbar voltage Vc, as shown in Figs. 3b, c and d,
l1 (t) = V sin(vt + u a)
Z m respectively. It can be seen that energisation of TE2 causes
1
 asymmetrical distribution of ux-linkages of TE1 and TE2,
l1 (0) l2 (0) e((R+2Rs )/(L+2Ls ))t resulting in inrush current i2 and sympathetic inrush current
2
1
 i1; and correspondingly, a signicant dip of Vc with slow
+ l1 (0) l2 (0) e((R/L))t (2) recovery.
2
L
l2 (t) = V sin(vt + u a) 4 Model description and validation
Z m
1

l1 (0) l2 (0) e((R+2Rs )/(L+2Ls ))t
Field measurement results indicate that the system would
2 experience unacceptable long duration voltage dips during
1
 energisation of GSU transformers, therefore further studies
l1 (0) l2 (0) e((R/L))t (3) are required to understand the characteristics of this
2
transient phenomenon. Owing to the difculty in carrying
where Z = [(R + 2Rs)2 + (L + 2Ls)2]1/2, = arctan[(L + 2Ls)/ out more eld measurements and the limitations of generic
(R + 2Rs)]; 1(0) and 2(0) are the initial ux of TE1 and analytical analysis (e.g. the assumption of linear
the residual ux of TE2, respectively. magnetisation inductance), detailed studies can be achieved
From (2) and (3), it can be seen that both 1(t) and 2(t) through computer simulation. This section describes the
consist of one sinusoidal component and two exponential development of a network model for simulation studies and
DC components. The AC component and the rst DC its validation against the eld measurement results.
component are the same, but the second DC component in
1 is opposite to that in 2, therefore i1(t) and i2(t) are 4.1 Description of network model
opposite to each other and appear alternately. Also, because
the DC components in 2 are negative, the maximum peak The network model was developed using ATP/EMTP based
of i2(t) would appear right after the energisation of TE2, on the 400 kV grid shown in Fig. 1. The sources S1 and S2

Fig. 4 Estimated core saturation curves of 415 MVA and 345 MVA GSU transformers

IET Gener. Transm. Distrib., pp. 112 5

doi: 10.1049/iet-gtd.2012.0166 & The Institution of Engineering and Technology 2013
www.ietdl.org

Fig. 5 Comparison of currents and voltage dips between measured and simulated results
a Instantaneous sympathetic inrush currents
b RMS sympathetic inrush currents
c RMS voltage dip
d Instantaneous currents of feeder IK

6 IET Gener. Transm. Distrib., pp. 112

& The Institution of Engineering and Technology 2013 doi: 10.1049/iet-gtd.2012.0166
 www.ietdl.org
Table 1 Comparison of three-phase voltage dip magnitudes adopted here is that air-core inductance is equal to twice the
transformer short-circuit inductance [29]. This assumption
Phase Dip magnitude, kV Deviation, %
is reasonable, because, when an unloaded two winding core
Field test Simulation type transformer is energised from high voltage (HV)
winding side (usually the outer winding), the cross-section
A 4.23 4.35 2.8 area of the air-core cylinder enclosed by the HV winding
B 9.28 9.08 2.1 where the ux goes through under deep saturation is
C 9.65 9.6 0.5 normally about twice the cross-section area of the gap
between HV and low voltage (LV) winding where the ux
goes through during the short-circuit test. These curves
are represented by ideal voltage source. Equivalent source were implemented into type-96 non-linear inductors which
impedances to represent source strength are derived from were then connected in delta and located at the LV terminal
the short-circuit levels at substation A and B which are to form the core representation [23].
7.1 and 6.4 GVA, respectively. All the double-circuit Based on transformer test report data, the approach used to
transmission lines linking the substations are modelled by model substation transformers is similar to that applied to
Bergeron routine, with line dimension, length and modelling GSU transformer T1, T2 and T3, that is, using
transposing scheme taken into account. System loading BCTRAN with externally connected core representation.
connected to each substation transformer is modelled by
constant lumped impedances connected at 132 kV side, 4.2 Model validation
with the power factors of local demands considered.
Regarding reactive power compensation devices, all the Model validation was made according to the rst
capacitor banks are modelled by constant capacitor banks measurement of energisation. The model parameters were
rated at 60 MVar; the SVC devices located at substation E obtained as follows: the source voltages were assigned in
and B are modelled by Thyristor controlled reactors rated at accordance with the eld measurement; source impedances
225 MVar in parallel with xed capacitor banks rated at were derived from corresponding system short-circuit level;
150 MVar; the SVC connected at substation I is modelled the amount of system loading at each substation is half of
by one tertiary connected Thyristor-switched capacitor with the rated capacity of the substation transformers as stated by
a rating of 60 MVar. The details of the SVC models are not the network operator; energisation angles were obtained
included in this paper, because the model is mainly for from the measured voltage waveforms in which three
validation purpose and for most case studies in following phases were energised simultaneously at 80 (angles are
sections, the SVC are not considered, unless otherwise relative to the positive-going zero crossing of phase A
specied. line-to-ground voltage); transformer air-core inductances are
The modelling of transformer is a main concern. To equal to twice of transformer short circuit inductances;
calculate inrush transients using ATP/EMTP (with because of lack of de-energisation voltage waveform, residual
ATPDraw as the graphical pre-processor), two types of uxes were assumed zero. Based on these parameter settings,
standard available models are often used, one is BCTRAN the simulation results were simultaneously generated by the
and the other is Hybrid Transformer [23, 24]. Owing to the developed network model.
lack of core design data and the need of initialising residual Fig. 5 gives the comparison between simulation and
uxes, BCTRAN was selected to model GSU and measurement results. In Fig. 5a, instantaneous wave shapes
substation transformers in this work. Data obtained from of three-phase sympathetic inrush currents were compared,
test report were used to derive the parameters in transformer with focus on the initiation of sympathetic inrush. It can
model. Based on short-circuit test data, transformer leakage be seen that the simulation results can replicate the
network was represented by RL 1 matrix. To approximate double-peak patterns, as well as growing trend and peak
core saturation characteristic, open circuit test data were magnitudes. The build-up and decay of sympathetic inrush
converted by using the algorithm presented in [25, 26] and were compared in Fig. 5b between the simulated RMS
curve-tted to form i curves, based on which hysteresis sympathetic inrush currents and the eld measurement
curves were then derived by using a subroutine called results. Good agreement is achieved in terms of the
HYSDAT [27]. Fig. 4 shows the lower half hysteresis initiation, the peak instants and magnitudes as well as
curves for GSU transformers. The saturation points used to the decay. Fig. 5c illustrates the comparison regarding the
dene the major hysteretic loop were appointed; beyond the long duration RMS voltage dips, with the dip magnitudes
saturation points, the nal slopes of saturation curves were particularly compared in Table 1. As can be seen, the
quantied by /i. Each nal slope was determined by simulated voltage dip recovery traces are similar to eld
the corresponding saturation inductance deduced from measurement results; the biggest deviation of dip magnitude
transformer air-core inductance, with the effect of winding is < 2.9%. The comparison in Fig. 5d is the currents
leakage inductance considered [28]. One assumption measured at the feeder IK. Good agreement can be seen in

Table 2 Energisation cases and results of phase C voltage dips

Case Sympathetic inrush GSU transformers being energised Adjacent GSU transformers in service RMS voltage dip

Magnitude, % Duration, s

1 no T1 no 9.6 2.670
2 yes T1 T2 and T3 9.6 6.479
3 no T2 and T3 no 18.1 3.376
4 yes T2 and T3 T1 18.1 7.718

IET Gener. Transm. Distrib., pp. 112 7

doi: 10.1049/iet-gtd.2012.0166 & The Institution of Engineering and Technology 2013
www.ietdl.org

Fig. 6 Voltage dip patterns and recovery traces on 400 kV and 132 kV side
a Patterns of voltage dip magnitudes across the whole network substations
b Patterns of voltage dip duration
c Voltage dip recovery traces observed at 400 and 132 kV busbars of substation I

8 IET Gener. Transm. Distrib., pp. 112

& The Institution of Engineering and Technology 2013 doi: 10.1049/iet-gtd.2012.0166
 www.ietdl.org
the range of steady stage (i.e. prior to energisation), which the comparison between Cases 3 and 4. By comparing
conrms the correct modelling of system loading. In the Cases 1 and 3, it can be seen that energising transformers
range of transient stage (i.e. after energisation), simulation T2 and T3 together would cause voltage dip with longer
results also show good agreement with eld test results. It duration and with magnitude almost twice than that caused
should be noted that the initial part of voltage recovery was by energising T1 alone. However, by comparing Cases 2
affected by the response of SVC, which has also been and 3, it is shown that the duration of Case 2, which
correctly replicated in the simulation results. Regarding this involves energising T1 alone but is inuenced by
SVC effect, more details will be described in Section 7. sympathetic inrush, is signicantly longer than that of Case 3.
In summary, the comparisons show that the network model The above case studies show that the sympathetic inrush
is capable to replicate measurement results and that the does not affect the voltage dip magnitude, which is obvious
BCTRAN model, although with simple core representation because at the initial stage of the energisation transient,
connected externally, is able to simulate the sympathetic voltage dip is mainly determined by the inrush current in
inrush transients between transformers. Therefore the the energised transformers, whereas at this time the
network model was employed in following simulation case sympathetic inrush drawn by the adjacent transformer has
studies. yet to build up. With the increase of the sympathetic inrush
current, the decay of the inrush current in the energised
transformer is slowed down and as a consequence the dip
5 Effect of sympathetic inrush between GSU duration is prolonged. The comparisons made between the
transformers case studies indicate that this prolonging effect can be very
signicant.
Using the validated model and the same system initial
conditions, simulation study was carried out regarding the
energisation cases with and without sympathetic inrush 6 Effect of sympathetic inrush attributed to
between GSU transformers (as shown in Table 2). For all substation transformers
cases, the theoretically worst energisation condition was set
as follows: three phases were simultaneously energised at Based on Case 4, further attempts were made to investigate
the positive-going zero crossing of phase C line-to-ground the sympathetic inrush of substation transformers and the
voltage; residual ux in phase C is 0.8 pu of peak rated consequent impacts on voltage dip performance. Voltage
ux (in the direction of ux build-up); residual ux in dip magnitude and duration at 400 and 132 kV substation
phase A is 0.8 pu of peak rated ux; and phase B retains busbars were observed across the whole network. Fig. 6a
zero residual ux [12]. With such energisation shows the patterns of voltage dip magnitudes of phase C on
conditions, the highest inrush current and biggest voltage both 400 and 132 kV side. It can be seen that the two
dip would appear in phase C. Therefore phase C voltage patterns are almost identical, which indicates that, as far as
dips observed at the 400 kV busbar of substation I are dip magnitude is concerned, the propagation of voltage dip
shown in Table 2, in terms of dip magnitude and duration. is not much affected by the substation transformers in the
It should also be noted that in the following studies, 3% is system studied. The dip magnitude observed at each
taken as the beginning and end threshold for quantifying substation is found to be related to the distance between the
the duration. substation and the supply source and also the distance
Comparing Cases 1 and 2, it can be seen that the voltage between the substation and the energised transformers. For
dip magnitudes are the same, however, the dip duration can those substations (including H, I and J ) located in the
be prolonged by 143%. Similar observation can be made in proximity of the energised transformer and relatively far

Fig. 7 Patterns of voltage dip duration at 400 kV side for various SVC locations

IET Gener. Transm. Distrib., pp. 112 9

doi: 10.1049/iet-gtd.2012.0166 & The Institution of Engineering and Technology 2013
www.ietdl.org
away from supply source, the observed dip magnitudes are side starts to be affected by sympathetic inrush about two
relatively large; for those substations (including D, E, F and cycles after the energisation and continues to be affected for
G) located relatively far away from the energised more than 9 s. This effect is found to be related to the
transformer and close to the supply source, the observed dip distance between substation and the power plant where
magnitudes are relatively small. GSU transformers being energised. For instance, the
Fig. 6b shows the patterns of voltage dip duration observed substation transformers located at substation I and J have
on 400 and 132 kV side. It can be seen that the dip durations been found the most affected, because of their electrical
on 132 kV side are longer. The prolonged voltage dip distances to the GSU transformers are the shortest. The
duration at 132 kV side is attributed to the sympathetic above ndings show that, under the worst energisation
inrush of substation transformers which is further illustrated condition, energisation of large GSU transformers in such
by voltage dip traces observed at substation I shown in Fig. systems (characterised by long transmission lines between
6c. It can be seen that both sides witness the same amount the supply source and the energised transformers) may
of voltage dip magnitude, however, the recovery at 132 kV trigger a network-wide sympathetic inrush.

Fig. 8 Results of sensitivity assessment

a Impacts of parameter variation on voltage dip magnitude
b Impacts of parameter variation on voltage dip duration

10 IET Gener. Transm. Distrib., pp. 112

& The Institution of Engineering and Technology 2013 doi: 10.1049/iet-gtd.2012.0166
 www.ietdl.org
7 Effect of SVC network studied here is characterised by long transmission
lines between supply source and the transformers being
The voltage dip recovery performance is further studied when energised, and therefore the effect of source strength, to
SVC is involved. Here, a set of 300 MVar SVC was applied at some extent, has been offset by the impedances of long
different substations and the corresponding voltage dip transmission lines; in those systems where the supply
recovery performance at 400 kV side are shown in Fig. 7. source is located closer to the transformers being energised,
The results suggest that the application of SVC can help the impact of source strength variation would be more
speed up the voltage recovery. This effect is inuenced by signicant. The variation of core saturation inductance
the SVC location. The substations nearest to the SVC directly inuences the magnitude of inrush currents and
location will benet the most; therefore it is desirable to thus produces great impact on the dip magnitude, but it
locate SVC close to the substation where the transformers only slightly affects the decay time constant of inrush
are energised. transients (because of the relatively small saturation
inductance compared with network impedance) and
therefore shows little impact on the dip duration. However,
8 Sensitivity assessment most transformers up to date are only tested at factory up to
110%; open circuit test at higher voltages is required for
The voltage dips involving sympathetic interaction is more accurate estimation of core saturation inductance. The
inuenced by multiple parameters. Sensitivity assessment signicant impact of transformer load losses on dip duration
was carried out to identify the most inuential parameters. is because the decay of sympathetic inrush is highly
Here, the concerning parameters include the system source determined by the losses of the connection between the
strength, system loading, GSU transformer load losses and GSU transformers and their own losses. Owing to the short
GSU transformer core saturation inductance. Switching electrical distance between the GSU transformers, the
angle and residual ux were not taken into account in the amount of load losses of the GSU transformers is the key
sensitivity study because they were treated as deterministic contributor to the damping of sympathetic inrush and
parameters forming the worst energisation condition. therefore imposes the greatest impact on dip duration.
The sensitivity assessment addresses variations of the
concerning parameters between + 100 and 50%, utilising
Case 4 (presented in Table 2) as the base case. Their 9 Conclusions
impacts on 400 kV side voltage dip magnitude and duration
are illustrated in Figs. 8a and b, respectively. In both This paper conducts a comprehensive study on the inuence
gures, each column represents the deviation caused by of sympathetic inrush on voltage dips caused by
variation of a certain parameter. transformer energisation based on an existing 400/132 kV
Fig. 8a illustrates that change of core saturation inductance network. A number of eld measurements of voltage dips
presents the biggest impact on dip magnitude. The biggest involving sympathetic inrush have been carried out,
deviation of dip magnitude appears in the case of increasing indicating the sympathetic inrush currents can induce long
the core saturation inductance by 100%. Source strength is duration voltage dips. A network model was developed in
the second most inuential one. In fact, in the reduced ATP/EMTP and validated against the eld measurement
cases, the impact of source strength variation on voltage dip results. It is shown that the BCTRAN model, although with
magnitude is comparable with that caused by variation of simple externally connected core representation, is able to
core saturation inductance. The variation of system loading simulate the sympathetic inrush transients between
only results in minor impact on dip magnitude. The transformers.
variation of transformer load losses barely impacts the dip Under the worst energisation condition, the validated
magnitude. network model was used to evaluate the degrees of
In Fig. 8b, it can be observed that change of transformer sympathetic interactions between large GSU transformers
load losses dominates the impact on voltage dip duration (e. and between GSU and substation transformers. It is shown
g. 50% reduction of transformer load losses could increase that the sympathetic inrush, although does not affect voltage
dip duration by 88%); the cases of reduced full-load losses dip magnitude, can prolong voltage dip duration by 143%.
have more impact than those of increased. The second most The prolonged duration could be more pronounced on 132
inuential parameter is again the source impedance. 50% kV side, because the sympathetic phenomena extend to
reduction of source strength will increase the duration by substation transformers. It is also shown that application of
36.5%. Variation of system loading only shows minor SVC devices can contribute to speeding up the voltage
impact on voltage dip duration, which is similar to its recovery, and it is more effective by locating SVC closer to
impact on dip magnitude. The variation of core saturation the transformers being energised.
inductance, which exhibits great impact on dip magnitude, Considering sympathetic inrush and voltage dips are
however, shows very little impact on voltage dip duration. controlled by multiple parameters, sensitivity assessment
It should be noted that, although the above analysis focuses was carried out to identify the most inuential parameters
on phase C, the other two phases also exhibit the same trend. on dip magnitude and duration. The amount of transformer
The minor impact of system loading variation suggests that load losses has been proven as the key inuential parameter
the system loading condition is of less concern when carrying to determine the duration of sympathetic inrush and voltage
out transformer energisation in the network studied, which dip. Transformer air-core inductance is found to be most
can be attributed to the low ratio of maximum substation inuential on dip magnitude; this parameter, which is
loading to the source strength (which is <0.09), that is, the normally not available from factory test report, can be
source strength is too strong relative to system loading. estimated using analytical or eld calculation or both based
However, it is noteworthy that for those systems having a on transformer design data; therefore it is suggested that
higher ratio of system loading to source strength, the effect transformer manufacturers provide it in the test report so as
of system loading might become more inuential. The to improve the modelling of transformer deep saturation.

IET Gener. Transm. Distrib., pp. 112 11

doi: 10.1049/iet-gtd.2012.0166 & The Institution of Engineering and Technology 2013
www.ietdl.org
10 Acknowledgment 20 Peng, J.S., Ang, S.P., Li, H.Y., Wang, Z.D.: Comparisons of normal
and sympathetic inrush and their implications toward system voltage
depression. Int. Universities Power Engineering Conf. (UPEC),
The authors would like to express their sincere gratitude September 2010, pp. 15
to National Grid for the nancial and technical support. 21 Bi, D.Q., Wang, X.H., Li, D.J., Yu, G.W., Wang, Z.J., Wang, W.J.:
The author, Mr. Jinsheng Peng, would like to thank Your Theory analysis of the sympathetic inrush in operating transformers,
Manchester Fund at the University of Manchester for Autom. Electr. Power Syst., 2005, 29, (6), pp. 18
22 Hayward, C.D.: Prolonged inrush currents with parallel transformers
providing Alumni Research Impact Scholarship which affect differential relaying, Trans. Am. Inst. Electr. Eng., 1941, 60,
partially funds his PhD study. (12), pp. 10961101
23 Prikler, L., Bnfai, G., Bn, G., Becker, P.: Reducing the magnetizing
inrush current by means of controlled energization and de-energization
11 References of large power transformers, Electr. Power Syst. Res., 2006, 76, (8),
pp. 642649
24 Mork, B.A., Gonzalez, F., Ishchenko, D., Stuehm, D.L., Mitra, J.:
1 Sarmiento, H.G., Estrada, E.: A voltage sag study in an industry with
Hybrid transformer model for transient simulation-part
adjustable speed drives, IEEE Ind. Appl. Mag., 1996, 2, (1), pp. 1619
I. development and parameters, IEEE Trans. Power Deliv., 2007, 22,
2 Kazibwe, W.E., Ringlee, R.J., Woodzell, G.W., Sendaula, H.M.: Power
(1), pp. 248255
quality: a review, IEEE Comput. Appl. Power, 1990, 3, (1), pp. 3942
25 Neves, W.L.A., Dommel, H.W.: Saturation curves of delta-connected
3 Lamoree, J., Mueller, D., Vinett, P., Jones, W., Samotyj, M.: Voltage
transformers from measurements, IEEE Trans. Power Deliv., 1995,
sag analysis case studies, IEEE Trans. Ind. Appl., 1994, 30, (4), pp.
10, (3), pp. 14321437
10831089
26 Chiesa, N., Hoidalen, H.K.: Analytical algorithm for the calculation of
4 Heine, P., Lehtonen, M.: Voltage sag distributions caused by power
magnetization and loss curves of delta-connected transformers, IEEE
system faults, IEEE Trans. Power Syst., 2003, 18, (4), pp. 13671373
Trans. Power Deliv., 2011, 25, (3), pp. 16201628
5 Milanovic, J.V., Aung, M.T., Gupta, C.P.: The inuence of fault
27 ATP rule book (Leuven EMTP Center, Leuven, Belgium, 1987)
distribution on stochastic prediction of voltage sags, IEEE Trans.
28 Chiesa, N., Mork, B.A., Hoidalen, H.K.: Transformer model for inrush
Power Deliv., 2005, 20, (1), pp. 278285
current calculations: simulations, measurements and sensitivity
6 Williams, A.J., Grifth, M.S.: Evaluating the effects of motor starting
analysis, IEEE Trans. Power Deliv., 2011, 25, (4), pp. 25992608
on industrial and commercial power systems, IEEE Trans. Ind. Appl.,
29 Povh, D., Schultz, W.: Analysis of overvoltages caused by transformer
1978, IA-14, (4), pp. 292305
magnetizing inrush current, IEEE Trans. Power Appar. Syst., 1978,
7 Styvaktakis, E., Bollen, M.H.J.: Signatures of voltage dips: transformer
PAS-97, (4), pp. 13551365
saturation and multistage dips, IEEE Trans. Power Deliv., 2003, 18, (1),
pp. 265270
8 Nagpal, M., Martinich, T.G., Moshref, A., Morison, K., Kundur, P.:
Assessing and limiting impact of transformer inrush current on power 12 Appendix
quality, IEEE Trans. Power Deliv., 2006, 21, (2), pp. 890896
9 Pors, A., Browne, N.: Modelling the energisation of a generator step-up
transformer from the high voltage network. Australasian Universities The details of GSU transformers rated at 345 and 415 MVA
Power Engineering Conf., Syndey, Australia, December 2008, pp. 15 from test reports are given in the following Tables 3 and 4,
10 Seo, H.C., Kim, C.H.: The analysis of power quality effects from the respectively.
transformer inrush current: a case study of the Jeju power system,
Korea. IEEE PES General Meeting, Pittsburgh, USA, July 2008, pp.
16
11 Ma, T., Cadmore, A.: System studies of voltage dips resulting from
energisation of MV wind turbine transformers. Int. Conf. and Table 3 GSU transformer test report (T1 and T2, 345 MVA)
Exhibition on Electricity Distribution, Turin, Italy, 2005, pp. 15
12 Smith, K.S.: Transformer inrush studies for wind farm grid Main data [kV] [MVA] [A] Coupling
connections. Int. Conf. Power Systems Transients, Montreal, Canada,
June 2005 HV 400 345 498 YN
13 Arana, I., Holbll, J., Srensen, T., Nielsen, A.H., Holmstrm, O., LV 19 345 10483 d1
Srensen, P.: Voltage dip caused by the sequential energization of open-circuit Eo, kV,% [MVA] Io, % Po, kW
wind turbine transformers. Proc. Conf. on European Wind Energy LV 17.1 (90) 345 0.049 126.2
Association, Marseille, France, March 2009 19 (100) 345 0.103 174.1
14 Rioual, M., Reveret, J.C.: Energization of step-up transformers for 20.9 (110) 345 0.350 239.9
wind-farms: modeling and its validation by tests performed on a 10 short-circuit [kV] [MVA] Z, % Ps, kW
MW site. IEEE PES General Meeting, Calgary, Alberta, Canada, HV/LV 400/19 345 17.8 838.4
2009, pp. 17
15 Engineering Recommendation P28: Planning limits for voltage
uctuations caused by industrial, commercial and domestic equipment
in the UK (The Electricity Council, 1989)
16 Bronzeado, H., Yacamini, R.: Phenomenon of sympathetic interaction
between transformers caused by inrush transients, IEE Proc., Sci. Table 4 GSU transformer test report (T3, 415 MVA)
Meas. Technol., 1995, 142, (4), pp. 323329
17 Kumbhar, G.B., Kulkarni, S.V.: Analysis of sympathetic inrush Main data [kV] [MVA] [A] Coupling
phenomena in transformers using coupled eld-circuit approach.
IEEE PES General Meeting, Tampa, Florida, USA, 2007, pp. 16 HV 400 415 599 YN
18 Wang, Y.G., Yin, X.G., You, D.H., Xu, T.Q.: Analysis on the LV 21 415 11410 d1
inuencing factors of transformer sympathetic inrush current. IEEE open-circuit Eo, [kV,%] [MVA] Io, [%] Po, [kW]
PES General Meeting, Pittsburgh, USA, July 2008, pp. 18 LV 18.9(90) 415 0.05 156.8
19 Pontt, J., Rodriguez, J., Martin, J.S., Aguilera, R.: Mitigation of 21(100) 415 0.069 211.1
sympathetic interaction between power transformers fed by long 23.1(110) 415 0.179 290.8
overhead lines caused by inrush transient currents. IEEE-IAS Annual short-circuit [kV] [MVA] Z, % Ps, kW
Meeting, New Orleans, Louisiana, USA, September 2007, HV/LV 400/21 415 17.12 924.7
pp. 13601363

12 IET Gener. Transm. Distrib., pp. 112

& The Institution of Engineering and Technology 2013 doi: 10.1049/iet-gtd.2012.0166