Journal of Power and Energy Engineering, 2015, 3, 1-6
Published Online April 2015 in SciRes. http://www.scirp.org/journal/jpee
http://dx.doi.org/10.4236/jpee.2015.34001
How to cite this paper: Geng, L.Q. and Zhang, N. (2015) Simulation and Analysis on Winding Deformation of a Power
Transformer in Current Transformer Connecting Manner. Journal of Power and Energy Engineering, 3, 1-6.
http://dx.doi.org/10.4236/jpee.2015.34001
Simulation and Analysis on Winding
Deformation of a Power Transformer in
Current Transformer Connecting Manner
Lanqin Geng1, Ning Zhang2
1Section of Fundamental Theory, Hengshui College of Vocational Technology, Hengshui, China
2Departmen t of Electrical Engineering, North China Electric Power University, Baoding, China
Received Sep te mber 2014
Abstract
Transformer winding deformation is one of the main types of transformer faults. To check if a
power transformer is being under winding deformation, the transformer can be connected in a
current transformer during its testing. A transformer winding simulating model is set up under
this connecting manner. Then the simulation has been performed with current source which is the
frequency sweep power. The simulation results show that the winding def or ma tio n can be re-
flected effectively with current source method. This method lays the foundati on for the realization
of on-line monitoring and diagnosis of the transformer neutral directly grounded side winding.
Keywords
Winding Deformation, Frequency Response Method, Current Source
1. Introduction
Power transformer is one of the core equipment of power system. Its safe operation is crucial to guarantee the
security of the power system [1] [2]. Transformer winding deformation is one of the main types of transformer
faults [3] [4]. The maintenance period of transformer is more than half a year with high cost and wide influence
[5]. The realization of on-line monitoring and diagnosis for transformer winding deformation will benefit for
real-time record of winding running state.
Frequency response method is one of the main methods for the detection of transformer winding deformation
at home and abroad, which can detect the winding short circuit impedance change of 0.2% or axial size change
of 0.3% [6]. The winding is equivalent to a passive linear network when the excitation signal frequency is higher
than 1 kHz. The frequency response characteristic is the outstanding property of passive linear network. The
frequency response characteristic is the only for a given network. The winding equivalent parameters will be
changed with the winding deformation. The frequency response curves will change reaction to the winding fre-
quency response characteristic.
Current source is the frequency sweep power in the simulation. For the transformer neutral directly grounded

L. Q. Geng, N. Zhang
2
side winding which is in the online running state, current source sweep signal is applie d to the grounding line
through the current transformer. And the high frequency current signal is obtained in the outlet side through the
current transformer. Then the frequency response curves are drawn to determine winding deformation through
the comparative analysis of deformation response curves [7] (hereinafter referred as the current source met h od ).
The method will benefit for real-time record of winding running state and high sensitivity.
2. Establishment of Simulation Model
The core permeability and air permeability are almost the same when transformer is in the high frequency signal
excitation (Usually more than 1 kH z) . The transformer winding is equivalent to a passive linear network which
consists of distributed inductance and capacitance [8] [9]. As shown in F igure 1, the winding equivalent model
has 6 levels, and its distribution parameters are uniform distribution parameters for hypothesis.
Where, L is transformer winding inductance; K is transformer winding longitudinal capacitance; C is tran s-
former winding ground capacitance; R1 is input matching resistor; R2 is equivalent resistance of output mea-
surement loop [10]. The specific parameters of simulation model are shown in Table 1.
PSPICE software was used for simulation. Sweep frequency mode is order of magnitude sweep. Scanning
range is from 1 kHz to 1 MHz. Scanning point is 200 in each order of magnitude. Simulation result of normal
transformer winding is shown in Figure 2.
Figure 1. Transformer winding equivalent diagram.
10000100000 1000000
-500
-400
-300
-200
-100
0
Frequency/ Hz
Figure 2. Frequency response curve.
Table 1. Simulation parameters.
Parameter Li Ki Ci R1 R2
Value 40.426 mH 19.5 pF 1213.27 pF 1 Ω 50 Ω

L. Q. Geng, N. Zhang
3
Figure 2 shows that frequency response curve is an oscillating curve with troughs and peaks. There is a peak
in the frequency response curve when series resonance occurs in the winding. And there is a trough when paral-
lel resonance occurs. T he change of model parameters will cause the change of corresponding peaks and trough s.
Winding deformation can be distinguished by comparing frequency response curves before and after the change
of parameters [11] [12].
3. Simulation with Current Source Method
All parameters are increased of 25% respectively to simulate different parts of winding defo rmation. The results
show that the curves are the same when the same winding deformations occur in the winding symmetrically. So
the repetitive curves are not shown. The abscissa of curve is freque ncy which is measured in Hz. The ordinate is
frequency response amplitude. Dotted line is the image under normal conditions, and solid line is the image after
chan gi ng para meters.
3.1. Simulation Results after Changing L
Increase L and the frequency response curves are shown in F igure 3.
Figure 3 shows that peaks are offset to lower frequency and the response amplitudes are changed when the
winding inductance is increased in turn. The trough a mplitude in high frequency is decreased. There is a new
trough in the left of the original trough. Peaks that change obviously are different when changing inductance
parameters of different positions. Change L1and peaks 1, 2, 3, 4 change large. Change L2 and peaks 1, 3, 4, 5
chan ge la r ge. Change L3 and peaks 2, 3, 4, 5 change large.
10000100000 1000000
-500
-400
-300
-200
-100
0
100002000030000 40000
-60
-40
-20
0
(a)
10000100000 1000000
-500
-400
-300
-200
-100
0
100002000030000 40000
-40
-20
10000100000 1000000
-500
-400
-300
-200
-100
0
100002000030000 40000
-60
-40
-20
0
(b) (c)
Figure 3. Simulation results after increasing L. (a) Increase L1 for 25%, (b) Increase L2 for 25%, (c) Increase L3 for 25%.

L. Q. Geng, N. Zhang
4
Similarly frequency response curves that L is decreased of 25% can also be made. How the change of L in-
fluenc e s frequency response curves is summarized in Table 2.
3.2. Simulation Results after Changing K
Increase K and the frequency response curves are shown in Fig ure 4.
Figure 4 shows that peaks are not offse t basicall y when longitudinal capacitance is increased in turn. The
trough amplitude in high frequency is decreased. And there is a new trough in the left of the original trough.
Peaks that change obviously are different when changing longitudinal capacitance of different positions. Cha nge
K1and peaks 4, 5 have a slight cha n ge. C ha nge K2 and peaks 4, 5 change lar ge. Change K3 and peak 4 has a
slight cha nge , peak 5 changes la rge .
Similarly frequency response curves that K is decreased of 25% can also be made. How the change of K in-
fluenc e s frequency response curves is summarized in Table 3.
3.3. Simulation Results after Changing C
Increase C and the frequency response curves are shown in Figure 5.
Figure 5 shows that peaks are not offset basically when C1 is changed, but they are offset to lower frequency
when C2 or C3 is changed. Peaks that change obviously are different when changing C of different positions.
Chan ge C1 a nd peaks are unchanged basically. C han ge C2 and peaks 2, 3, 4, 5 change l ar ge . Change C3 and
peaks 2, 4, 5 change large. The change of peak 5 is especially large when C2 or C3 change s.
Similarly frequency response curves that C is decreased of 25% can also be made. How the change of C in-
fluenc e s frequency response curves is summarized in Table 4.
The above simulation analysis shows that current source metho d can reflect the winding deformation type ef-
fectively.
Table 2. Effect of L on the frequency response curve.
Change of L Differences Similarities
Increase 1) Peaks are offset to lower frequency. 2) A new trough appears in the left of the
original trough.
1) Changes in low frequency are more obvious.
2) The trough amplitude in high frequency is decreased.
3) Peaks that change obviously are different when changing
inductance parameters of different positions. But peaks change the
same when changing inductance parameters of same positions,
whether increase or decrease.
4) The curves are sensitive to the change of L.
Decrease 1) Peaks are offset to higher frequency. 2) A new trough appears in the right of
the original trough.
Table 3. Effect of K on the frequency response curve.
Change of K Differences Similarities
Increase A new trough appears in the
left of the original trough. 1) Peaks are not offset basically. 2) Peaks that change obviously are different when changing longitudinal
capacitance of different positions. But peaks change the same when changing
longitudinal capacitance of same positions, whether increase or decrease.
3) The trough amplitude in high frequency is decreased.
4) Changes in medium frequency are more obvious.
5) The curves are less sensitive to the change of K.
Decrease A new trough appears in the
right of the original trough.
Table 4. Effect of C on the frequency response curve.
Change of C Differences Similarities
Increase Peaks are offset to lower frequency when C2 or C3
changes. 1) Peaks are not offset basically when C1 is changed
2) Frequency response curves are sensitive to the change
of C.
Decrease Peaks are offset to higher frequency when C2 or C3
changes.

L. Q. Geng, N. Zhang
5
(a)
(b) (c)
Figure 4. Simulation results after increasing K. (a) Increase K1 for 25%, (b) Increase K2 for 25%, (c)
Increase K3 for 25%.
10000100000 1000000
-500
-400
-300
-200
-100
0
600000800000 1000000
-240
-220
(a)
10000100000 1000000
-500
-400
-300
-200
-100
0
100002000030000 40000
-60
-40
-20
10000100000 1000000
-500
-400
-300
-200
-100
0
100002000030000 40000
0
(b) (c)
Figure 5. Simulation results after increasing C. (a) Increase C1 for 25%. (b) Increase C2 for 25%. (c)
Increase C3 for 25%.

L. Q. Geng, N. Zhang
6
4. Conclusions
1) The simulation results show that current so ur ce method can reflect the winding deformation types ef fe c-
tively, which lays the foundation for the realization of on-line monitoring and diagnosis of the transformer neu-
tral directly grounded side winding.
2) Current so urc e me tho d can’t apply to on-line monitoring and diagnosis of the transformer delta connection
winding. This method needs further study and discussion.
References
[1] Xie, Y.-P. and Zhang, C.-L. (2000) The Research for Checking the Deformation of the Transformer Coils by Using the
Frequency Response Analysis Method. Inner Mongolia Electric Power, 18, 4-5.
[2] Tan g, G. (2007) Frequency Response Analysis for Winding Deformation of Power Transformer. Xihua University.
[3] Arri, E. and Mo eei, F. (19 94) Status Assessment of Transformer Wingdings Via On-line Stray Inductance Measure-
ment. IEEE Instrum. Meas. Technolgy. Conf IMTC94, Hamamatsu, May 1994, 1171-1174.
[4] Sun, G.-F. (2011 ) Frequency Response Mother On-Line Monitoring Transforme’s Winding Deformation. Electric
Switchgear, 44, 13-15.
[5] He, Y.-Y. (1996) Discussion on the Domestic Power Transformer Ability to Withstand Short Circuit Problems. Tran s-
former, 33, 2-7.
[6] Wan g, Y., Li, Y.-M. and Zhang C.-L. (1997) The Comparison Study on LVI and FRA Method of Detecting Winding
Deformation Faults in Transformers. High Voltage Engineering, 23, 13 -15.
[7] V aessen, P.T. and Hanique, E.A. (1992) New Frequency Response Analysis Method for Power Transformers. IEEE
Transaction on Power Delivery, 7, 384-391.
[8] Li u , H.-L. and Lei, X. (2011) A Detecting System for Transformer Winding Slight Deformation Based on Impulse
Response Analysis. Electrical Measurement & Instrumentation, 48, 10-13.
[9] Isla m, S.M. (2001) Detection of Shorted Turns and Winding Movements in Large Power Transformers Using Fre-
quency Response Analysis. IEEE Engineering Society Winter Meeting, Singapore, 1.
[10] Mor eau, O. and Gullloty, B. (199 9) FRA Diagnosis Method Sunulation Applied to Feasibility Study and Effi ci e n c y
Improvement. IEEE Engineering Society Winter Meeting, New York, 31 January-4 February 1999, 1059-1064.
[11] Zhao , S.-S. (2008) Analysis of the Deformation Detecting Test of Power Transformer Winding and Its Simulation
Technology. Zhejiang University.
[12] Dick, E.P. and Eve n, C.C. (1978) Transformer Diagnostic Testing by Frequency Analysis. IEEE Transaction on Power
Apparatus and Systems, 97, 2144-2153.