Condition Assessment of Power Transformer

Bushing Using SFRA and DGA as Auxiliary Tools

Bahar Mohseni (student Member IEEE), Naser Hashemnia, Syed Islam (senior member IEEE)
Department of Electrical & Computer Engineering, Curtin University
Perth, Western Australia
[email protected]

Abstract—Dielectric insulation of a transformer bushing Current detection methods include, capacitance and dissipation
deteriorates as a function of temperature, oxidation, and moisture. factor measurement, partial discharge and oil sampling [6-8]. In
This causes accelerated aging of oil and cellulosic solid insulation, addition FRA is a trenchant tool for monitoring transformers to
generating fault gases within bushing oil and eventual permanent diagnose mechanical faults such as axial and radial
failure. To prevent such failures, effective analyses and diagnoses displacement [9, 10], disk space variation [11] and inter-turn
need to be investigated. Dissolved Gas Analysis (DGA) can give the
indication of internal abnormalities inside the transformer bushing.
short circuit [12] at an initial level. However, little attention has
In addition, Frequency response analysis (FRA) is a widely accepted been given to the impact of bushing faults on the FRA signature.
tool for mechanical deformation diagnosis within power In this paper, FEM is used to simulate the physical
transformers. Although a large number of studies have been geometrical dimension of a transformer bushing, and
conducted on the detection of transformer winding deformation by investigate how different levels of fault will affect equivalent
FRA technique, the impact of bushing faults on the transformer circuit parameters and FRA signature. To attempt this concept,
FRA signature has not been sufficiently investigated. It is the goal of DGA and FRA have been utilized as auxiliary tools to detect
this paper to propose precise simulation as well as practical analyses transformer bushing faults prior to failure.
demonstrating the impact of bushing faults on the FRA signature. A
real transformer bushing geometry is modelled through 3D finite
element analysis (FEM) on which different bushing faults are
emulated. To verify the derived simulation results, DGA of
transformer oil as well as FRA are performed on a three-phase, 132
kV, 315 MVA power transformer. It can be observed clearly from the
results, that bushing faults have an impact on the FRA signature and
DGA of the power transformer.
Index Terms- condition monitoring, power transformer bushing,
FRA, DGA, finite-element

I. INTRODUCTION
Power transformers are unquestionably one of the most
crucial pieces of equipment in electrical power systems. As a
result of the continuing advancement in electric power
transmission and distribution systems and the increasing
worldwide transformer population age, the likelihood of
transformer catastrophic failures, which in turn leads to power Figure 1. 3D model of a power transformer
outages, is increasing [1]. Hence, the importance of
developing a standard diagnosis tool to detect potential
mechanical problems, including displacements or II. FINITE ELEMENT AND ELECTRICAL MODEL OF A
deformations [2-4]. TRANSFORMER BUSHING
Bushings are critical components of power transformers and Failure of a power transformer results in serious damage to
also one of the major initiators of transformer failures [5]. assets and reduced reliability of the power system. A failed high
According to CIGRE, transformer bushings contribute to voltage (HV) bushing is one of the major transformer
around 17% of all transformer failures, are the 3rd most common components responsible for forced outages of a power
cause of breakdowns, and are the most common cause of transformer. There are two common types of transformer
transformer fires and explosions. Therefore, establishing state- bushings, namely solid porcelain bushing, which is used for low
of-the-art diagnostic tools to help reduce the risks of failures voltage (LV) transformers, and oil impregnated paper (OIP)
and serious fires is of paramount importance. bushing, for HV transformers. According to CIGRE, close to
By detecting the defect or degradation of bushings caused by 50% of serious transformer fires are initiated by Oil-
the stresses acting on them, bushing failures can be avoided. Impregnated-Paper (OIP) bushings and they are the most

978-1-4673-8848-1/16/$31.00 ©2016 IEEE

common cause for transformer fires. The insulation body of an • Step 2: The permittivity of oil and paper in the bushing are
OIP transformer bushing is comprised of several cylindrical changed within properties material option to simulate the
layers of insulation material wound around a central core. The moisture content and insulation degradation. The resultant is
insulation layers are usually filled with an insulating fluid such solved in electrostatic solver to calculate the variation of the
as oil whose ageing and moisture content have noticeable capacitances for different levels of fault.
impacts on bushing dielectric properties. In order to achieve a • Step 3: The electrical parameters are extracted for normal and
high degree of uniformity in the electric field, conductive layers faulty conditions.
are placed between paper layers [13]. The outermost conductive • Step 4: The FRA signature is plotted as the gaining transfer
layer is connected to the flange of the bushing which is function TFdB = 20 log10 |V /V | for healthy and faulty
grounded while intermediate conductive layers remain conditions.
insulated. The above described construct acts as a set of
capacitors connected in series, creating a capacitive voltage
splitter.
The equivalent electrical model of the bushing connected to
a one-phase transformer is shown in Figure 3. The capacitance
Cbn of each layer can be calculated using equation (1), where Ɛr
is the relative dielectric constant of dielectric material, ln is the
length of nth bushing layer and rn is the radius of nth bushing
layer. Volume resistance Rsn can be obtained using equation
(2), where ρ is the resistivity and Rpn is a resistance in the range
of MΩ which represents surface resistance of bushing layers
and bushing external insulation dielectric.

2πεrε 0l1 2πεrε 0l2 2πεrε 0ln Figure 2. 3D model of Bushing solved in electrostatic FEM solver.
Cbn = = = ... =
r r r (1)
ln 1 ln 2 ln n III. BUSHING FAULT DETECTION
r0 r1 rn−1 As a result of moisture ingress inside the bushing dielectric
insulation, the capacitive components, C1 and C2 change. In
FEA, this effect can be simulated through altering the
rn
dr ρ r permittivity or conductivity of the dielectric insulation.
Rsn = ρ  = ln n (2) Moisture content variation inside the bushing insulation is
rn−1
2πrln 2πln rn−1 emulated by increasing the oil-paper permittivity from 1% to
5% for the model under study. Table 1 shows the percentage
of corresponding alteration of the bushing capacitances C1 and
The main insulation system is known as the C1, which is the C2 due to the permittivity variation inside the bushing dielectric
equivalent capacitance of the inner insulation layers. The last insulation. It is clear that for a particular percentage increase in
few layers near ground are known as the C2 insulation. dielectric permittivity, C1 varies more significantly in
As a result of insulation degradation, which occurs mainly comparison with C2, e.g. a 2% permittivity rise leads to a 9.5%
due to heat, oxidation, acidity, and moisture ingress, the relative increase in C1 while the value of C2 increases by a mere 4%.
dielectric constant of the bushing, Ɛr varies, thereby changing
the bushing capacitance. Bushings are designed to have a TABLE 1. CAPACITANCE VARIATION OF BUSHING MODEL
constant dielectric capacitance over the asset operational life
[14] and therefore, a 3% to 5% variation in bushing capacitance Dielectric Insulation Permitivity Cpacitance
can indicate potential problems within bushings [15]. To Variation(%) Variation(%)
accurately identify the effect of bushing faults on the electrical C1 C2
parameters of the bushing, in Figure 2, the physical geometrical 1 8 2
dimension of a 200 KVA bushing of a three-phase transformer 2 9.5 4
shown in Figure 1 is simulated using FEM software (Ansys) for
3 11 5.5
healthy and faulty conditions. In order to emulate a fault in the
4 13 6.8
bushing, its insulation complex permittivity is altered. The
5 15 7.5
electrical parameters for normal and faulty conditions are
extracted using FEM based on the following 4 steps:
Figure 4 shows the impact of moisture ingress in bushing
• Step 1: The Transformer bushing is modeled using FEA with insulation on the HV winding FRA signature. The variation of
properties material in electrostatic, magnetostatic and DC high voltage FRA signature for rises of 2% and 4% in the
conduction solvers to obtain the capacitances, conductance permittivity of the transformer bushing oil and paper have been
and inductance for healthy condition. investigated in the simulation.
 Cg-L /2 Cg-L
G
G s-L

M ij-L
Cbn R s-L Cs-L
Rsn Ls-L
R pn
G CHL M HL
Ls-H

R s-H C
s-H
M ij-H

G s-H
G
Cg-H /2 Cg-H

Figure 3. Electrical model of transformer winding and bushing

As can be seen, moisture content inside the dielectric DGA methods to diagnose bushing faults. The results of
insulation can affect the FRA signature at high frequency DGA test are presented in Table 2.
range over 700 kHz where the resonance and anti-resonance
frequencies decline in comparison with the healthy TABLE 2. DGA TEST DATA
condition. The impact can be more predominant if more
moisture penetrates within bushing insulation. To simulate Test Date Dissolved Gas (ppm) Moisture
content
the worst conditions, the permittivity of the dielectric (ppm)
insulation can be further varied in FEM. Dec 2010 C2H4 CH4 CO CO2 10
Feb 2011 0.1 2.25 135 250.43 12.5
March2012 4943 10556 4195 1547 24
July 2013 2940 7790 3100 4650 26
June2014 3650 5965 4800 5900 35
Nov2015 4500 65022 8718 6380.6 45

According to Table 2, the total dissolved gas content has

increased continuously over the 5-year period. Carbon
monoxide (CO) and Carbon dioxide (CO2) concentration
level increased rapidly from 135 and 8715 to 250.43 and
6380.6, respectively. According to IEC standard, the
increment of CO and CO2 level in transformer oil can be
used as an indication of dielectric insulation aging and
decomposition. This degradation can be due to the increase
in moisture content inside the oil and then penetration of the
paper insulation which leads to the decomposition of
Figure 4. Effect of moisture content in bushing insulation on HV winding cellulose [16]. Table 2 shows that the moisture content
FRA signature. increased from 10 ppm to 45 ppm over the period, which
leads to a higher degradation rate of the bushing insulation,
III. PRACTICAL MEASURMENTS hence significantly affecting the dielectric insulation relative
In this study, a 200 MVA step up transformer has been permittivity (Ɛr).
taken as a case in order to evaluate the capability of FRA and Detailed investigation of the level and the change of
combustible gases in the insulating oil revealed that the fault
is developing in nature and adversely affecting the dielectric observed in SFRA signature to diagnose bushing faults
insulation inside the bushing. at initial stages.
To validate the DGA results, the transformer HV winding
FRA signature is plotted. Figure 5 shows the impact of Appendix
bushing failure on the transformer FRA signature. FRA Table A.2. Dielectric properties of bushing
measurements were performed as the magnitude of the Dielectric properties Oil Paper layer porcelain
transfer function (TF) in dB, 20log10(Vo/Vi), against ɛ 2.4 2.5-16 6.5
frequency. In this case, the FRA measurement is at two Table A.4. Bushing Model Parameters (T Model)
different time intervals. The responses shows that as the Electrical parameters C1 C2 LS RS
condition of the dielectric insulation deteriorates, resonances Value 2000pF 450pF 2.5µH 0.2Ω
and anti-resonance frequencies are shift to the left. The
variation is seen in the high frequency range after 800 kHz.
REFERENCES
-20 Curve Info [1] Rahimpour, E., et al., Transfer Function Method to Diagnose Axial
Healthy Displacement and Radial Deformation of Transformer Winding. IEEE
Poor condition Power Engineering Review, 2002. 22(8): p. 70-70.
-40 Worst Condition [2] Ahn, H.M., et al., Finite-Element Analysis of Short-Circuit
Electromagnetic Force in Power Transformer. IEEE Transactions on
Industry Applications, 2011. 47(3): p. 1267-1272.
-60
[3] Bigdeli, M., et al. Transformer winding diagnosis using comparison of
transfer function coefficients. in Electrical Engineering/Electronics,
T F (d B )

Computer, Telecommunications and Information Technology (ECTI-

-80 CON), 2011 8th International Conference on. 2011.
[4] Abu-Siada, A., et al., Understanding power transformer frequency
-100 response analysis signatures. IEEE Electrical Insulation Magazine, 2013.
29(3): p. 48-56.
[5] Hashemnia, N., A. Abu-Siada, and S. Islam, Detection of power
-120 transformer bushing faults and oil degradation using frequency response
analysis. IEEE Transactions on Dielectrics and Electrical Insulation, 2016.
23(1): p. 222-229.
-140 [6] Almendros-Ibanez, J.A., J.C. Burgos, and B. Garcia, Transformer Field
0 1 10 100 1000 3000
Frequency (kHz) Drying Procedures: A Theoretical Analysis. IEEE Transactions on Power
Delivery, 2009. 24(4): p. 1978-1986.
Figure 5. Practical phase-A FRA signature with 3% moisture content in the [7] Ekanayake, C., et al., Frequency response of oil impregnated pressboard
bushing oil and paper samples for estimating moisture in transformer insulation. IEEE
Transactions on Power Delivery, 2006. 21(3): p. 1309-1317.
[8] Du, Y., et al., Moisture equilibrium in transformer paper-oil systems.
IV. CONCLUSION IEEE Electrical Insulation Magazine, 1999. 15(1): p. 11-20.
[9] Hashemnia, N., A. Abu-Siada, and S. Islam, Improved power
This paper presents an improved method using transformer winding fault detection using FRA diagnostics ??? part 1: axial
comprehensive finite element analysis to analyse the effect displacement simulation. IEEE Transactions on Dielectrics and Electrical
of dielectric insulation failure of HV transformer bushing on Insulation, 2015. 22(1): p. 556-563.
the FRA signature. Based on simulation results and practical [10] Hashemnia, N., A. Abu-Siada, and S. Islam, Improved power
tests, the following conclusions can be drawn: transformer winding fault detection using FRA diagnostics ??? part 2: radial
deformation simulation. IEEE Transactions on Dielectrics and Electrical
Insulation, 2015. 22(1): p. 564-570.
• Simulation and practical results also reveal that [11] Hashemnia, N., et al. Offline to online mechanical deformation
moisture content inside the dielectric insulation of a diagnosis for power transformers. in Power Engineering Conference
transformer bushing can be observed in the high (AUPEC), 2014 Australasian Universities. 2014.
frequency range of the transformer FRA signature as it [12] Behjat, V., et al., Diagnosing Shorted Turns on the Windings of
changes the magnitude and resonance frequencies Power Transformers Based Upon Online FRA Using Capacitive and
Inductive Couplings. IEEE Transactions on Power Delivery, 2011. 26(4):
position. p. 2123-2133.
• By considering the change in capacitances of the [13] Melo, M.A.C., M.E.G. Alves, and G.A. Moura. Experience with on-
bushing model, failures could be detected through the line monitoring of capacitance and tangent delta of condensive bushings. in
frequency response analysis (FRA) signature of the Transmission and Distribution Conference and Exposition: Latin America,
model, and the FRA signature gained from the 2008 IEEE/PES. 2008.
simulation is validated by the FRA signature derived [14] Goff, M.B. and A.H. Eltom. Oil filled bushing secrets revealed. in
Industry Applications Society Annual Meeting, 2013 IEEE. 2013.
from the practical measurement.
[15] Peng, H.d., et al. Insulation diagnosis for 220kV oil-immersed current
• During operation, minor faults developing inside the transformer by frequency dielectric spectroscopy. in 2011 Electrical
transformer bushing will develop into major faults and Insulation Conference (EIC). 2011.
further result in catastrophic failures. To avoid [16] Duval, M. and A. dePabla, Interpretation of gas-in-oil analysis using
permanent failure, DGA and SFRA diagnostic new IEC publication 60599 and IEC TC 10 databases. IEEE Electrical
techniques can be used as auxiliary tools which relate Insulation Magazine, 2001. 17(2): p. 31-41.
fault gases’ concentration level with the resonance shift