DGA Whitepaper2
DGA Whitepaper2
DGA Whitepaper2
imagination at work
The Transition to Next-Generation Online DGA Monitoring Technologies Utilizing Photo-Acoustic Spectroscopy
Abstract
Abbreviations
PAS: Photo-Acoustic Spectroscopy
DGA: Dissolved Gas Analysis
The gases that are associated with specific fault types are
Hydrogen (H2), Carbon Dioxide (CO2), Carbon Monoxide (CO),
Ethane (C2H6), Methane (CH4), Ethylene (C2H4) and Acetylene
(C2H2). They are known collectively as the diagnostic gases.
Analysis of these gases allows for the diagnosis of developing
faults using a variety of methods, including Duvals Triangle, Key
Gas Method etc
Traditionally DGA was limited to a laboratory environment because
of the complexity of the equipment required to first extract gases
from the oil sample taken and then measure these gases, often
at quantities as low as one part-per-million (ppm). Historically gas
extraction from a sample was performed using a strong vacuum
pump called a Toepler pump apparatus (1). More recently, IEC,
ASTM and others have published a new gas extraction technique,
Headspace Gas Extraction, which is becoming standard in
laboratories due to its convenience, excellent repeatability and
perhaps more importantly its inherent suitability for automation.
Historically the DGA technique comprised of an extraction
process (such as the aforementioned headspace) and Gas
Due to the time and costs involved (site visit, logistics and
laboratory analysis), DGA analysis of an oil sample would usually
be restricted to once per year, with repeated or more frequent
samples collected and tested only if significant fault gases were
detected in the routine annual sample. As many types of faults
can progress significantly in less than one year, this approach
often resulted in missed diagnostic opportunities. Faults could
occur and progress for up to 12 months before being detected
and significant damage would be caused to the transformer in the
mean-time as a result.
H 3200
4000
2800
2300 2100
C-H
3000
C
N
Triples
2,380
CO 2
1800
O
C
C
1500
Doubles
Finger print
Singles
1460, 1380
nujol
2000
Parabolic
Mirror
Radiation
Source
Chopper
Wheel
Wavelength Selection
(Filter wheel)
PAS Technology
Photo Acoustic Spectroscopy (PAS) works along the following
principle: A gas substance absorbs light energy following local
heating by an IR light and transforms it into kinetic energy (by
the energy exchange process). Regularly interrupting this process
causes a series of pressure waves (sound) that can be detected by
microphones. By measuring the sound at different wavelengths,
cm-1
1000
Sample IN
Analysis
Chamber
Microphone
Microphone
Sample OUT
<1
<1
Methane (CH4)
<1
Ethane (C2H6)
<1
Ethylene (C2H4)
<1
Acetylene (C2H2)
<0.2
The Transition to Next-Generation Online DGA Monitoring Technologies Utilizing Photo-Acoustic Spectroscopy
Application of PAS
Technology to DGA
Standard Gas-in-Oil
Samples
Gas ID:
Methodology
The Transition to Next-Generation Online DGA Monitoring Technologies Utilizing Photo-Acoustic Spectroscopy
Results
Gas
Range :
LOW
MEDIUM
HIGH
UNIT :
(PPM)
(PPM)
(PPM)
Ethylene
17,4
35,1
70,2
Ethane
16,0
32,1
64,2
Methane
18,1
36,3
72,6
Carbon Monoxide
18,3
36,6
73,2
Carbon Dioxide
17,9
35,8
71,6
Acetylene
17,4
34,7
69,4
Quantification
Diff.
(%)
Std
Dev
(ppm)
65
-10,9
1,87
59
63
-6,28
1,60
64
70
70
-8,17
3,27
59
57
62
58
-19,7
1,94
143
165
162
111
142
103
19,4
58
60
58
62
61
-13,8
1,60
Std.
Test
1
Test
2
Test
3
Test
4
Test
5
Test
6
Ethylene
70,2
62
61
63
60
64
Ethane
64,2
59
60
61
59
Methane
72,6
68
62
66
Carbon
Monoxide
73,2
60
57
Carbon
Dioxide
71,6
151
Acetylene
69,4
60
Actual
Value
(ppm)
Mean
Max
Std Dev
Ethylene
511.0
510.9
512.4
514.1
1.1
Ethane
509.0
513.8
514.3
514.7
0.4
Methane
493.0
495.0
495.3
495.7
0.3
Acetylene
499.0
504.2
504.5
504.8
0.2
Carbon Dioxide
10000.0
10088.7
10113.1
10137.5
24.4
Carbon
Monoxide
506.0
509.7
510.4
510.8
0.4
Quantification
Diff.
(%)
Std
Dev
(ppm)
35
-5,98
1,41
33
+0,73
2,66
36
38
-4,96
2,43
29
31
31
-17,6
0,98
131
121
122
+256
7,12
30
32
32
-11,1
1,17
Std.
Test
1
Test
2
Test
3
Test
4
Test
5
Test
6
Ethylene
35,1
31
33
33
32
34
Ethane
32,1
29
31
32
32
37
Methane
36,3
31
34
35
33
Carbon
Monoxide
36,6
29
30
31
Carbon
Dioxide
(CO2)
35,8
130
139
122
Acetylene
34,7 gas-in-oil
29
31 results
31
Table
5. Medium
Gas ID:
Quantification
Diff.
(%)
Std
Dev
(ppm)
16
-3,26
0,41
15
16
+2,08
1,21
19
18
18
+1,29
0,82
16
14
16
15
-16,2
0,82
131
124
105
111
+542
10,14
16
15
15
-9,96
0,52
Std.
Test
1
Test
2
Test
3
Test
4
Test
5
Test
6
Ethylene
17,4
17
17
17
17
17
Ethane
16,0
17
17
18
15
Methane
18,1
19
19
17
Carbon
Monoxide
18,3
15
16
Carbon
Dioxide
17,9
110
109
Acetylene
16
16
16
Table
6. Low17,4
gas-in-oil
results
Quantification
AVE
(PPM)
Diff. (%)
Test 1
Test 2
Test 3
Ethylene
(C2H4)
145
145
144
144,7
0,58
0,39
Ethane
(C2H6)
165
164
164
164,3
0,58
0,35
Methane
(CH4)
63
65
64
64
1,56
Carbon
Monoxide
(CO)
32
31
33
32
3,13
Carbon
Dioxide
(CO2)
334
312
331
325,7
11,9
3,66
Acetylene
(C2H2)
118
119
119
118,7
0,58
0,49
Quantification
AVE
(PPM)
Diff. (%)
Test 1
Test 2
Test 3
Ethylene
161
161
164
162
1,73
1,07
Ethane
174
177
180
177
1,7
Methane
68
68
71
69
1,73
2,5
Carbon
Monoxide
38
41
44
41
7,3
Carbon
Dioxide
181
182
187
183,3
3,2
1,75
Acetylene
130
156
147
144,3
13,2
9,1
The Transition to Next-Generation Online DGA Monitoring Technologies Utilizing Photo-Acoustic Spectroscopy
Gas ID:
Field Experiences
Estimates put the number of installed online DGA instruments at
more than 80,000, of which the vast majority are composite gas
monitors employed as alarms to detect developing faults. More
than five thousand PAS based DGA instruments are now installed
globally, in more than 120 countries. They are normally deployed
on critical or large transformers where their cost benefit is more
obvious.
Many examples exist of where DGA has prevented further
damage occurring to a transformer by notifying asset managers
of a developing faults through increasing gas levels .
The Transition to Next-Generation Online DGA Monitoring Technologies Utilizing Photo-Acoustic Spectroscopy
PAS based systems are now providing this capability all over the
world through the provision of high quality DGA results.
Example of successful
diagnosis of developing
fault
A GE Transfix monitor was installed on a transformer in Asia.
Results observed over several months identified a thermal
problem in the transformer evidenced by:
Figure 4. Best match (86% pattern match) to the Key gas diagnosis:
Arcing from load current
Conclusion
PAS based DGA instruments have been developed with the express
purpose of addressing the shortcomings of online GC based
instruments. They provide a real alternative to GC by matching
their performance and operating successfully in the field.
Utilising a technology historically designed for online application,
PAS instruments are very stable and repeatable monitoring
instruments suited for the tough environmental and operational
demands associated with remote transformer monitoring. PAS is
the new high-end standard for monitoring critical transformers.
References
[1] IEC 60567 standard
[2] McIlroy, C. D. PhotoAcoustic Spectroscopy. A New Technique
for Dissolved Gas Analysis in Oil. EPRI Substation Equipment
Diagnostics Conference, New Orleans, LA, USA. February 23 26,
2003
[3] McIlroy, C. D. A Comparison of PhotoAcoustic Spectrometer
and Gas Chromatograph Techniques for Dissolved Gas Analysis
of Transformer Oil. EPRI Substation Equipment Diagnostics
Conference, New Orleans, LA, USA. February 15 18, 2004