2016 International Conference on Computation of Power, Energy Information and Communication (ICCPEIC)
Study of Dissolved Gas Analysis for Pre­
Determination of Faults in Transformers
Nikhileshwar P. Adhikari
Amit Gupta
Triparna Nandy
PG Scholar
Assistant Professor
GGCT, Jabalpur
Assistant Professor
LNCT Group Of College
GGCT,Jabalpur
India
India
nandytriparnal90lgmail.com
Bhopal, India
nikhil adhikari0lrediffmail.com
amit.hvps0lgmail.com
One of the best methods of detecting certain
This de-gradation can lead to pre-mature failure of
problems which can lead to failure of the transformer is
the transformer. The two important functions of the
Abstract -
Dissolved Gas Analysis (DGA). For diagnosing fault
oil are to provide cooling and electrical insulation.
condition on oil filled insulation, one of the most recent
techniques developed is Dissolved Gas Analysis.
The
chemical analysis of the generated gases is known as
Dissolved Gas Analysis or DGA. For DGA, oil sample
from the transformer has to be removed which can be
Oil used in transformer is a mixture of
Hydro-carbons. Lower order of Hydro-carbons like
Methane,
Acetylene,
Ethane,
Ethylene
etc.
are
produced
with some permanent gases like Carbon
done without de-energization of the transformer easily.
Di-oxide, Carbon Mon-oxide and hydrogen during
Then
oil
the process of degradation. Failures are inevitable in
paper
oil filled equipment such as power transformer if
presents complete DGA study of 12.5 MVA 132/33KV,
proper care is not taken. The mixture of hydrocarbons
by
sample
using
is
gas
analyzed
chromatography
in
the
technique,
laboratory.
This
EMCO makes Transformer: at 132 KV SIS Srinagar
(M.P.). The data are analyzed over a period of 2006 to
2012. The DGA results by different standard methods
are compared.
and permanent gases is sealed environment can cause
an
explosion.
In
the
running
condition,
proper
monitoring is needed on the concentration of these
explosive gases [7] . Initially the generated gases will
Keywords - Dissolved Gas Analysis (DGA), Parts per
dissolve
in
the
oil.
When
the
generated
gases
million (PPM), Gas Chromatography (GC), Rogers Ratio
increases in its volume, more of it will dissolve into
Method.
the oil. Eventually a stage will come when the oil will
be completely saturated with the dissolved gases and
I.
Oil
is
used
gas in the oil so it will come out as free gas. By
insulation purpose. Mineral oil de-gradation in oil
evaluating the amount of generated gas present and
filled transformers is a major concern. Analysis of oil
the rate of gas generation, abnormal condition can be
sample
detected.
maintenance
transformers
transformer
tool
for
is
for
any further increase cannot be contained as dissolve
and
of
in
INTRODUCTION
useful,
cooling
predictive
determining
health
and
of
transformer. Ageing of the oil, overheat, overvoltage,
II. METHEDOLOGY
environmental conditions and numerous unknown
DGA: It is an effective as well as practical method
factors are responsible for the oil degradation [2] .
which helps in the detection of incipient faults along
[Type text]
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
634
Tripama Nandy et af: Study of Dissolved Gas Analysis for Pre-Determination of Faults in Transformers
with its order of severity. After commissioning of the
transformer oil using vacuum extraction method
transformer, DGA shall be repeated once in a month
For
extraction
of
gases
whereas in case of transformer which has been
extraction,
repaired, DGA should be carried out in a week soon
extraction can be used. Disadvantages of this method
after decommissioning and then after about 3 months.
are it is very expensive and also time taking. Due to
DGA is very helpful as faults are identified and
limitations
necessary actions are taken before any catastrophic
hydrogen
failure occurs. [6]
spectroscopy have been developed. Both
headspace
of
GC,
on-line
two
from
the
sampling
new
monitoring
oil,
and
techniques
and
stripper
vacuum
namely
photo acoustic
For the improvement in reliability and power
of these methods requires less time as compared to
availability, DGA plays an important role. On the
Gc. For analyzing dissolved gases in insulating oil,
basis of results obtained, investigation has become
GC has been used for the last 60 years. After IEC,
easier and problems thus are resolved in less time [5] .
IEEE and ASTM published relevant guidelines, this
Many industries are now aware of this tool and using
method became more popular.
it for reduced maintenance and reduced cost of
methods, GC is accepted as the best method for
repairs of power transformers.
measuring the concentration of gases dissolved in
Gas chromatography: One of the most widely used
transformer oil. In the laboratory environment, GC
methods
analysis is successfully conducted [7] .
is
Gas
Chromatography
(GC)
. Gas
Among the three
Chromatography apparatus of Chernito Company is
used in MPPTCL. For the separation of different
complex mixtures, GC is used. From the operating
transformers, oil samples are collected and taken to
the laboratory for the gas extraction and analysis. Oil
sample is taken from the transformers and transported
to the laboratory using syringes, flexible metal cans,
and special glass bottles or in the calibrated stainless
III. ANALYSIS METHODS
steel cylinders. From the bottom of the tank, oil
Rogers Ratio Method: In this method, A 4 digit ratio
samples are usually taken.
code is generated from the five fault gases namely
This method for DGA is used once in a year only as
hydrogen, methane, ethane, ethylene and acetylene
it is a costly process.
for determining 15 diagnosis rules for transformer
conditions.
Table 1 Roger's ratio gas
Fig.1
Extraction of dissolved gases from
RATIO
CODE
CH4/H2
1
C2HJCH4
2
C2HJC2H6
3
C2H2/C2H4
4
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
635
2016 International Conference on Computation of Power, Energy Information and Communication (ICCPEIC)
1. If all the four ratios are equal to zero then it
remaining ratios equal to zero,
indicates no fault.
energy discharge (arc with power flow through).
2. If code 2=5, in other words when the ratio of
9. If code 3 is equal to zero, in other words when the
ethane to methane is equal to 5 and remaining three
ratio of ethylene to ethane is equal to 1 with
codes are zero then it indicates partial discharge of
remaining ratios equal to zero, it indicates insulated
low energy density or hydrolysis.
conductor overheating.
3. If code 1=1 and code 2=5, in other words when the
lO.lf code 3 is equal to one and code 4 also equal to
ratio of methane to hydrogen is equal to one and that
1, in other words when the ratio of ethylene to ethane
of ethane to methane is equal to 5 respectively with
is 1 and ratio of acetylene to ethylene is also equal to
remaining two ratios equal to zero, it indicates partial
1 with remaining two ratios equal to zero, it indicates
discharge of high density which is possible with
complex thermal hotspot and conductor overheating.
tracking.
11.1f code 1 is equal to one and code 4 is equal to
4. If code 2=5 and code 3=1, in other words when the
one, in other words when the ratio of methane to
ratio of ethane to methane is equal to 5 and ratio of
hydrogen is 1 and ratio of acetylene to ethylene is
ethylene to ethane is equal to 1 with remaining two
also 1 with remaining two ratios equal to zero, it
ratios zero,
indicates
then it indicates coincidental partial
coincidental
thermal
it indicates high
hotspot
and
low
discharges and conductor overheating.
energy discharge.
5. If code 2=5 and code 4=1, in other words when the
12.1f code 2 is equal to one, in other words when the
ratio of ethane to methane is equal to 5 and that of
ratio of ethane to methane is equal to 1 with the
acetylene to ethylene is equal to 1 with remaining
remaining three ratios equal to zero, it indicates
two ratios equal to zero, it then indicates partial
thermal fault of low temperature range less than 150
discharges of increasing energy density.
degree Celsius.
6. If code 1 is greater than 1 but less than 2, in other
13. If code 2 is greater than zero but less than 2 and
words when the ratio of methane to hydrogen is
code 4 is equal to l,in other words when the ratio of
greater than 1 but less than 2 and remaining ratios are
ethane to methane is greater than zero but less than 2
zero,
and ratio of acetylene to ethylene is 1 with remaining
it
then
indicates
low
energy
discharge
(flashover without power flow through).
two ratios equal to zero, it indicates thermal fault of
7. If code 1 is greater than 1 but less than 2 and code
temperature range from 100 degree Celsius to 200
3 is equal to 1, in other words when the ratio of
degree Celsius.
methane to hydrogen is greater than 1 but less than 2
14. If code 2 is equal to one and code 3 is also equal
and ratio of ethylene to ethane is equal to 1 with
to one, in other words when the ratio of ethane to
remaining ratios equal to zero, it indicates low energy
methane is equal to 1 and ratio of ethylene to ethane
discharge (continuous sparking to floating potential).
is 1 with remaining two ratios equal to zero, it
8. If code 1 is greater than 1 but less than 2 and code
indicates thermal fault of temperature range from 150
3 is equal to 2, in other words when the ratio of
degree Celsius to 300 degree Celsius. It occurs due to
methane to hydrogen is greater than 1 but less than 2
overheating of copper due to eddy currents.
and ratio of ethylene to ethane is equal to 2 with
[Type text]
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
636
Tripama Nandy et af: Study of Dissolved Gas Analysis for Pre-Determination of Faults in Transformers
15. If code 2 is greater than 1 but less than 2 and code
When the ratio of acetylene to ethylene is not
3 is equal to 2, in other words when the ratio of
significant (code 1=not significant), ratio of methane
ethane to methane is greater than 1 but less than 2
to hydrogen is greater than 1 (code 2 >1) and ratio of
and ratio of
with
ethylene to ethane is less than 1 (code 3 <1) it
it indicates
indicates thermal fault of temperature less than 300
ethylene to ethane is equal to 2
remaining two ratios equal to zero,
thermal fault of temperature range from 300 degree
degree Celsius.
Celsius to 700 degree Celsius. It occurs due to bad
When the ratio of acetylene to ethylene is less than
contacts/joints. [3]
0.1 (code 1<0.1), ratio of methane to hydrogen is
IEC Basic Ratio Method: This method has originated
greater than 1 (code 2>1) and ratio of ethylene to
from the Rogers Ratio Method, except that the ratio
ethane is in the range of 1 to 4 (1<code 3 <4) it
of ethane to methane is not included in this method as
indicates thermal fault of temperature between 300
it only indicated a limited temperature range of
degree Celsius to 700 degree Celsius.
decomposition.
When the ratio of acetylene to ethylene is less than
0.22 (code 1 <0.22), ratio of methane to hydrogen is
greater than 1 (code 2>1) and ratio of ethylene to
Table 2 IEC basic ratio gas
Ratio
Code
ethane is greater than 4 (code 3>4) it indicates
C2H2/C2H.
1
thermal fault for temperature greater than 700 degree
CH./H2
2
C2HJC2H.
3
When the ratio of acetylene to ethylene is not
significant (code 1=not significant), ratio of methane
to hydrogen is less than 0.1 (code 2<0.1) and ratio of
ethylene to ethane is less than 0. 2 (code 3 <0.3) it
Celsius.
Increasing value of acetylene may indicate that the
hot spot temperature is higher than 1000 degree
Celsius. [4]
Key Gas Method:
Table 3 key gas parameter
Gas
Normal
Abnormal
Tnterpretation
H2
<150 ppm
>1000 ppm
Arcing, Corona
CH.
<25 ppm
>80 ppm
Sparking
C2H.
<10 ppm
>35 ppm
Local Overheating
C2H.
<20 ppm
>100 ppm
Severe Overheating
CO
<500 ppm
>1000 ppm
Severe Overheating
When the ratio of acetylene to ethylene is in the
CO2
<1000 ppm
>15000 ppm
Severe Overheating
range of 0.6 to 2. 5 (0.6<code 1>2.5), ratio of methane
N2
1-10%
N.A
to hydrogen is in the range of 0.1 to 1 (0.1 <code 2<1)
O2
0.2-3.5%
N.A>0.5%
indicates partial discharges.
When the ratio of acetylene to ethylene is greater
than 1(code 1>1), ratio of methane to hydrogen is in
the range of 0.1 to 0.5 (0.1 <code 2<0.5) and ratio of
ethylene to ethane is greater than 1 (code 3 <1) it
indicates discharges of low energy.
and ratio of ethylene to ethane is greater than 2 (code
Combustibles
IV. DATA OBTAINED FROM MPPTCL
3 <2) it indicates discharges of high energy.
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
637
2016 International Conference on Computation of Power, Energy Information and Communication (ICCPEIC)
Table 4 data collected from MPPTCL
Hz(in
Date of
Testing
parts per
million
ppm)
Oz(in parts
per million
ppm)
COz(in
parts per
million
ppm)
CzH.(in parts
per million
ppm)
CzH. (in
parts per
million
ppm)
CzHz(in parts
per million
ppm)
CO(in
CH.(in
parts per
million
parts per
million
ppm)
ppm)
p+p (in
parts
per
million
ppm)
TOTAL
(in parts per
million ppm)
02/01106
6
19559
1352
2
I
NIL
17
NIL
2
20939
30111106
2.8625
0
795.3932
0.6463
0.2897
0.0022717
64.096
1.6356
0.3746
865.5007
14/02/07
2.6905
5305.125
1401.0163
0.6675
0.7763
0
29.1343
1.2301
1.126
674.7665
12/06/07
0.9062
11030.57
651.7298
0.6249
0.2312
0.020202
37.841
0.7889
0.2481
11722.955
05/03/08
1.4004
10354.9902
552.5699
0.00016324
0.075920
0.0043099
35.1829
0.6540
0.3462
10945.2239
01104/09
0.7966
8760.1555
261.1016
0.000024859
0.0005662
0.076823
9.4063
0.3298
0.1128
9031.9740
30/07/10
2.9941
2676.6658
560.195
0.6411
0.1154
0.00092215
8.6342
0.4447
0.2975
3249.9895
17/02/11
0.5005
3776.0119
490.1201
0.9172
0.093243
0.0062040
0.7331
0.2446
0.5729
4269.1997
19112/11
87.9727
--
0.0
1978.380
0.9659
0.00
168.0650
9.1488
0.00
2244.5324
05110/12
0.8443
--
1169.000
0.8917
0.4800
0.0797
135.1660
1.3775
0.2769
1308.1169
Table 5 the Result Obtained according To ROGER'S RA no METHOD
Dates
Code 1
Code 2
Code 3
Code 4
02.01.06
NiIl6
IINil
211=2
NiIl2
30.11.06
1.6356/2.8625=0.57
0.2897/1.6356=0.177
0.6463/0.2897=2.2309
0.00227717/0.6463=0.035
14.02.07
1.2301/2.6905=0.457
0.7763/1.2301=0.631
0.6675/0.7763=0.8598
0/0.6675=0
12.06.07
0.7889/0.9062=0.87
0.2312/0.7889=0.293
0.6249/0.2312=2.702
0.020202/0.6249=0.03
05.03.08
0.6540/1.4004=0.467
0.075920/0.6540=0.116
0.00016324/0.075920=0.02
0.0043099/0.00016324=26.40
01.04.09
0.3298/0.7966=0.414
0.0005662/0.3298=0.017
0.000024859/0.0005662=0.043
0.076823/0.000024859=3090.349
30.07.10
0.4447/2.9941=0.148
0.1154/0.4447=0.259
0.641110.1154=5.555
0.00092215/0.6411=0.00143
17.02.11
0.2446/0.5005=0.4887
0.093243/0.2446=0.38
0.9172/0.093243=0.98366
0.0062040/0.9172=0.0067
19.12.11
9.1488/87.9727=0.1039
0.9659/9.1488=0.1055
1978.380/0.9659=2048.22
0.00/1978.380=0
05.10.12
1.3775/0.8443=1.63
0.4800/1.3775=0.348
0.8917/0.4800=1.8577
0.0797/0.8917=0.8937
[Type text]
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
638
Triparna Nandy et af: Study of Dissolved Gas Analysis for Pre-Determination of Faults in Transformers
V. TEST RESULTS AND ANALYSIS
H2
Result Obtained according To Roger's Ratio Method:
collected
data
of
failed
transformer
is
checked
according to roger's ratio method. It is found that on
100
I
5�
1\,
'" .. .. .. .. .. i
o..-ioooooo..-i..-i
NO.,tNu-i..-iO"aiu-i
date 05.10.12, it indicates high energy discharge
-
H2
0('1) ........00('1)
..
......
..
0
..
which was the cause of failure of the transformer.
Fig 3. Graph for H2 from 2006 to 2012.
The results are presented in table 5.
Result Obtained According To IEC Basic Ratio
Method:
Collected
data
of
failed
transformer
is
checked according to IEC basic ratio method: It is
found that on date 19.12.11, it indicates high energy
discharge which was the cause of abnormality in the
transformer. The results are presented in table 6
The above graph indicates that H2 is almost constant
over a period of 4 years , it increases in 2011 due to
internal
arcing
and
corona
discharge.
During
overhauling of transformer some loose connection
were found and the same were rectified by soldering
and
tightning
of
bushing
to
internal
winding
connection.
hereunder
Table 6 Collected data of failed transformer is
checked
Dates
Code 1
Code 2
Code 3
02.01.06
0
0
2
30.1l.06
0.035
0. 57
2.2309
14.02.07
0
0.457
0. 8598
12.06.07
0.03
0. 87
2.702
05.03. 08
26.40
0.467
0.02
01.04.09
3090.349
0.414
0.043
30.07.10
0.00143
0.148
5.55
17.02.11
0.0067
0.4887
0.98366
19.12.11
0
0.1039
2048.22
05.10.12
0. 8937
l.63
l. 8577
02
40000
�------
20000 +t-----­
o +-��r_����
-
02
Fig 4. Graph for O2 from 2006 to 2012.
The graph shows the content of oxygen decreases due
to combustion and ingress of particle impurity. The
oil flirtation was done in November 2006 and the
oxygen increase, it further continuously decreases
over a period from 2007 to 2012 . The oil filtration is
recommended at present.
C02
Result Obtained According To Key Gas Method:
2000
According to key gas method , The cause of failure of
1000
this transformer was severe overheating as C2H4 gas
o
-
CO2
is much more than normal value which is 100 parts
per million(ppm).
Fig 5. Graph for CO2 from 2006 to 2012.
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
639
2016 International Conference on Computation of Power, Energy Information and Communication (ICCPEIC)
The above graph indicates that COz has decreased in
November 2006 due to oil
filtration.But
due to
overloading in February 2007, it has again increased
0. 1
causing severe overheating. In June 2007, the problem
0.05
of overloading was solved and COz decreased.But
o
again
load
increased
in
October
2012
I
A
causing
overheating as well increase in concentation of COz.
Fig 8. Graph for CzHz from 2006 to 2012.
due to overloading its concentation increased causing
4000
severe overheating. Oil filtration is recommended at
2000
o
i
i
i
i
i
i
i
ZS
i
- CzH4
co
Fig 6. Graph for CZH4 from 2006 to 2012.
This graph shows that CZH4 was almost constant
from January 2006 to February 2011.Suddenly in
December
2011
concentration
present.
increased
200 ,----------------10
�
I¢;
"7
•
•
t,
- co
causing
severe overheating. Remedial actions were then taken
Figure 9. Graph for CO from 2006 to 2012.
to decrease its concentration.
This graph indicates that concentration of CO has
increased from January 2006 to November 2006 as a
result of overloading.Due to some remedial actions it
2
decreased to some extent. But again load increased in
I
� S?S; ..
December 2011 as a result its concentarion also
C,
increased causing severe overheating.
..
,
0 ""';000000 ""'; ""';
NO.,tNu-i ""';o,,_: criu-i
om ...... ...... 0 om ...... ...... o
Fig 7. Graph for CZH6 from 2006 to 2012.
This graph indicates that CZH6 decreased initially
from January 2006 to November 2006 due to oil
10
: 1�""",i,i6"
T--------------��-
,....; ,....;N�rti.,t.._: NNO
0 ...... 000000 ...... ......
NO.,tNu-i ""';o,,_: criu-i
filtration and then slightly increased in February 2007
causing local overheating. Due to remedial actions
taken in time its concentration decreased.
This graph shows that initially concentation of CzHz
was under normal operating limits but
om ...... ...... oom ...... ...... 0
Fig 10. Graph for CH4 from 2006 to 2012.
This graph indicates that CH4 increased in December
2011.
[Type text]
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
640
Triparna Nandy et af: Study of Dissolved Gas Analysis for Pre-Determination of Faults in Transformers
The cause behind this was sparking. Prior to 2011 it
severe overheating etc are predicted by DGA. During
was under normal operating limits.
ovehauling of transformer, loose connections should
be
p+p
3
2
1
0
rectified by soldering and tightning of bushing to
internal
.-I
.-I
connection,
addition
of
DBPC
Powder, oil reclaimation and other remedial actions
I�,�,
� � r-- r-- 00 (j) 0 .-I
0 0 0 0 C! 0 .-I .-I
winding
should be taken to decrease the concentration of
gases in oil. It will increase the active life span of the
-p+p
transformer.
N
.-I
.-i .-i N <.0 M � r' N N ci
C! .-I 0 0 C! C! 0 C! .-I .-I
N 0 � N LJ") .-I 0 r-- (j) LJ")
0 M
.-I
.-I 0 0 M
.-I
.-I 0
VII. REFERENCES
Figure 11. Graph for P+P from 2006 to 2012.
This graph shows that concentration of P+P i. e
[1] Rahul Pandey and M. T. Deshpande, "Dissolved
propylene and propane has decreased over the years.
Gas
Analysis
(DGA)
of
Mineral
Oil
Used
in
Transformer." International Journal of Application or
Innovation in Engineering & Management (IJAIEM)
VI. CONCLUSION
Volume 1, Issue 2, October 2012.
The rate of decomposition and the type of gases
[2] Ali Saeed Alghamdi, Nor Asiah Muhamad and
change during defective operation, which could be a
Abubakar A. Suleiman, "DGA Interpretation of Oil
result of thermal overloading and/or electrical faults
Filled
has been analysed for 12.
Transactions
5 MVA 132 /33 kv
Transformer
on
Diagnosis".
Condition
electrical
and
electronic
Transformer . It is shown that using DGA , based on
materials. Vol. 13,No.5,pp.229-232,October 25,2012.
the quantity or type of fault gases, the gas increase
[3] Er. Lee Wai Meng, "Dissolved Gas Analysis
rates and the proportions between the gases, the type
(DGA) of Mineral Oil Used in Transformer."(2009).
of failure can be pre -deduced. Partial discharges with
[4]
lower
Stefan Tenbohlen, "Improvement of interpretation of
energy
mainly
lead
to
the
formation
of
Jackelyn
Aragon-Patil,
Markus
Fischer
And
hydrogen and methane, as well as small quantities of
dissolved gas analysis for power transformers".
ethane. Electrical
[5] M.Duval, "Dissolved gas analysis: It can save
discharges)
acetylene,
discharges
cause
contionusly
of
and
spark
hydrogen
and
as well as methane and ethylene.
thermal-oxidative
quantities
separation
(arcs
of
CO
cellulose
and
conducting
degradation,
C02 are
the
larger
formed.
DGA
Test
By
By
and
your
transformer,"
IEEE
Electrical
Insulation
Magazine, Vol 5, No. 6,pp.22-27,1989.
[6] Norazhar Abu Bakar, A.Abu- Siada and S.Islam ,
"A Review of Dissolved Gas Analysis Measurement
and
Interpretation
Techniques"
IEEE
Electrical
monitoring the results , the internal Faults can be Pre­
Insulation Magazine, Vol 30, No.3, May/June.
determined in EHV Class Power Transformers, hence
[7] Joseph B. DiGiorgio, "Dissolved Gas Analysis of
to prevent transformer failure
Mineral Oil Insulating Fluids," Northern Technology
major power interruption.
which further causes
Effective steps can be
& Testing.
taken when internal arcing and corona discharge ,
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
641
2016 International Conference on Computation of Power, Energy Information and Communication (ICCPEIC)
[8] Sung-wook Kim, Sung-jik Kim, Hwang-dong
Seo, Jae-ryong Jung, Hang-jun Yang and Michel
Duval, "New Methods of DGA Diagnosis using IEC
TC lO and Related Databases Part-I: Application of
Gas-ratio
Combinations",
IEEE
Transactions
on
Dielectrics and Electrical Insulation, Vo1.20, No.2,
April,2013.
[Type text]
978-1-5090-0901-5/16/$31.00 ©2016 IEEE
642