Transformers are the vital components in a power network. So, to the stakeholder, in spite of its age, an existing transformer is far more preferable than a new one. This is mainly because of its high monitory value and the difficulties in replacing it at appropriate times. But to preserve a transformer effectively it is always necessary to know its remaining life, as condition based maintenance solutions can be prescribed, based on its age. This activity is called RLA or Remnant Life Assessment. There are two methods of performing it - either by analyzing the Furan content in the transformer oil or by assessing the DP of the insulation paper. This paper explains the effectiveness of DP technique under different circumstances, also weighing its pros and cones, with the help of two case studies.
I INTRODUCTION
In a switchyard, transformer is a very crucial component. In the world about 80% [1] of the transformers are already over 40 years of age. In such a scenario, the stakeholder must certainly know, for how long his aged transformer will render service, so that he can make proactive decisions for refurbishment or replacement of the transformer. The activity which is carried out for estimating the remaining service life of the transformer is known as Remnant Life Assessment or RLA. The scope of RLA is not limited to estimating the remnant life of a transformer alone, but it also explains, in what state the transformer is. There are two methods of performing RLA. The first method is the analysis of furan content in the liquid insulation and the second, is the estimation of DP of the insulation paper. From the experience of a number of transformers, CGL has found that DP estimation is a promising technique for RLA. Hence the paper discusses two of such case studies where DP was used as a vital tool for RLA studies.
II INSULATION AGEING AND REMNANT LIFE
The insulation material used in transformer consists of cellulose paper, pressboard, wood and impregnating mineral oil. When the paper insulation is overstressed by high temperatures or electrical discharges, the chemical bonds in its molecules break creating new ones. These irreversible reactions cause ageing of the transformer, with decrease in the electrical and mechanical strengths of its insulation [2], [5]. These reactions are further catalyzed by the presence of moisture leading to a chain reaction inside the paper. The byproducts, so generated are called Furans which get dissolved in the transformer oil. These Furanic compounds keep on accumulating in the oil as age passes and therefore by determining the concentration of furans in oil, it is possible to find out the remnant life of the transformer.
During ageing, along with furans, another phenomenon that happens is the rupture of large cellulose chains or polymers into smaller ones with less no. of glucose molecules in them. The number of glucose units in a cellulose polymer of the paper insulation, is called the Degree of Polymerization (DP) of the paper. Fresh paper insulation may have a DP value between 1000 and 1500, where as a very aged paper insulation may have a DP value of
200-250. Therefore DP analysis predicts remnant life of the transformer by counting the numbers of glucose units in the cellulose paper. DP is a very reliable method for life estimation, as the life estimate is predicted from the paper samples directly, unlike furan analysis. But this requires a shutdown to open up the transformer for paper sampling. Hence DP is an offline test, which is its disadvantage. Moreover the places of paper sampling and rewinding these places with fresh paper require high level of knowledge and expertise, which is generally available with an OEM. However, in spite of the labour
1
TRAFOTECH 2010 Eighth international Conference on Transformers
involved in the paper sampling, the end user is benefitted as he gets the most accurate results of his transformer's age. This has been practically verified in the following case studies, which demonstrates the credibility of DP as a promising remnant life estimation tool.
III CASE STUDIES
Case Study 1 - 60 MVA, 110kV 1976 make
Sub-station transformer
The above mentioned transformer was one of the key transformers, supplying power to the city where it was installed. Since it had completed 32 years of service, the responsible utility felt the necessity of RLA and the job was awarded to CGL in 2008 for the same. CGL followed its self-generated, knowledge based diagnosis steps and it initiated the activity with an examination of the history of the transformer. The history highlighted the following events:
In 1985, oil pumps were commissioned along with 2 additional radiators and the rating was upgraded from 60MVA to 72.5MVA
Oil was replaced in the year 1993
In 1997, tap changers of all phases were repaired or replaced.
In 2004, COPS was installed.
Oil filtration was done in the year 2004
Following, the DGA results of this transformer from 2006-08 were examined. These are shown in the Table.1 below.
Table.1 - DGA results
with past data. Table.2 shows the furan results from 2006 to 2008.
Table.2 - Furan analysis results
Date of
Sampling
2
FAL
5
HM2F
2
FOL
2
ACF
5
M2F
Total
ppm
03/09/06
11.27
0.43
0.13
0.14
0.29
12.26
16/01/08
13.08
0.28
0.26
0.11
0.31
14.04
Siteofpapersampling
ObservedDP
Values
Remarks
HV'B' phaseBottom
Leadouterlayer
473
Accelerated
Ageing
HV 'Y' phase Bottom
Leadouterlayer
343
Accelerated
Ageing
HV'R' phasecentre leadouterlayer
515
Normal
Ageing
OldbarrierboarbetY
&B phasecentre
603
Normal
Ageing
Newbarrierboarbet
Y&B phaseBottom
719
Normal
Ageing
OldbarrierboarbetY
&B phaseTop
1244
Normal
Ageing
New barrier board between
Y&B phaseBottom
1017
Normal
Ageing
HV'B' phaseTop
Leadouterlayer
456
Accelerated
Ageing
HV'B' phaseCentre
Leadouterlayer
479
Accelerated
Ageing
HV'R' phaseTop
Leadouterlayer
767
Normal
Ageing
YPhasewinding , BottomLeadinnermost layer
868
Normal
Ageing
BPhasewinding , bottom lead innermost layer
890
Normal
Ageing
B Phase winding bottomLeadInnermost
764
Normal
Ageing
R Phase winding , Centre,innermost
821
Normal
Ageing
From furan test results it seemed that the life of the transformer was over as the furan content had reached to ppm level. But in order to have a final confirmation, DP test was performed on the transformer. Table.3 below shows the DP values observed in paper samples collected from various places inside the transformer.
Table.3 - Results of DP values
Sampling
Date
H2
ppm
C2H6
ppm
C2H2
ppm
C0
ppm
CO2
ppm
CO2/
CO
2006
482
23
0
290
3032
10.5
2007
336
38
0
14
148
10.6
2008
260
10.5
ND
488
14011
28.7
Moving on to the next step, fresh oil samples were collected from the transformer and evaluated. The moisture analysis of the oil indicated 44.8 ppm, corresponding to 4% wetness of solid insulation. This was too high for any transformer. Furan analysis was also done on the same oil sample and compared
The DP values of Kraft paper ranged between
343 and 890, while the DP values of pressboard insulation were well above 600. From DP test it
was identified that about 40% of life was remaining in the transformer. Thus a
transformer which was found as failed in furan
2
TRAFOTECH 2010 Eighth international Conference on Transformers
analysis indicated another 40% life in the DP analysis, which in any case was beneficial to its owner. However one observation that was made from DGA and furan tests was that, the CO2/CO ratio was increasing along with the furan content. This was indicative of heating up of cellulose [3] [6] under high moisture content, which could have lead to the accelerated ageing of the insulation paper and rise in furan content.
Case Study 2 - 315 MVA, 3 phase,
400/230 kV ICT
The above mentioned transformer was referred to CGL for RLA due to presence of high level of furan in the oil. This transformer had experienced two major faults in its history, which are as follows:
• First fault on 20/1/1989 when the W-phase winding succumbed to internal short circuit and
was replaced with a spare healthy winding
• Second fault on 20/7/1991, when the V- phase winding was replaced due to internal
short circuit. The following Table.4 shows the
furan content in oil from 2004 to 2007. The colour of oil when last sampled was brown.
TABLE.4 - Results of furan values
Siteofpaper sampling
ObservedDP Values
Remarks
HV'U' phasemain lead(outerlayers)
539
NormalAgeing
HV'V' phase
Canopy
719
NormalAgeing
HV'W'phasemain lead(Stressshield
canopy)
719
NormalAgeing
LVmiddlephase
695
NormalAgeing
IV 'Y' phase
539
NormalAgeing
Neutrallead
761
NormalAgeing
HV'V' phase, Barrierboard-top
954
NormalAgeing
HV'V' phase, Barrierboard- bottom
826
NormalAgeing
IV 'W'phase, Barrierboard
973
NormalAgeing
was used for interpretation [4]. It is also an indicator of overheating/normal aging of cellulose insulation. However since two of the windings were replaced, it was recommended that the RLA should be conducted using DP test. Therefore impregnated insulation paper and pressboard samples were carefully removed for DP from various locations of the transformer on 12/01/2007 and were replaced with the new material to avoid further damage. On analysis in the lab, the average DP value of thin insulation was found to be 662 and that of thick insulation was 918. Pressboard insulation was in good condition with minimum life consumption among all the solid insulations. Table.5 shows the DP values of impregnated paper collected from various parts of the transformer.
TABLE.5 Results of DP values
Dateof
Sampling
2
FAL
5
HM2F
2
FOL
2
ACF
5
M2F
TOTAL
ppb
15/09/04
1290
56
0
0
0
1346
27/12/04
1480
82
0
0
0
1562
28/09/05
1488
52
0
0
0
1540
28/07/06
1705
0
0
0
0
1705
27/12/06
1682
48
0
0
0
1730
12/03/07
1430
50
0
0
0
1480
The table indicates that the values of 2-FAL were high. The transformer has a history of two faults, when a considerable amount of impregnated cellulose material from the old windings would have degraded and liberated furans into the oil. The same oil was used till the RLA process. Even though oil processing was done in course of its service, unlike gases and moisture which could be removed by efficient vacuum filtration of oil, the furans couldn't be removed as they were dissolved in the oil. 2- FAL Furaldehyde is a more stable compound among all five furan compounds and hence it
Assuming that the insulation inside the winding
would experience a higher temperature than the insulation in bulk oil and winding leads, a DP
value less than the least value indicated in
Table.5 was considered for RLA. This value corresponded to a remnant life of 57%. If the absolute value of 2-FAL, which is 1682 ppb, was considered, then it would correspond to a DP value of 370, which in-turn would give a remnant life value of about 44%. However, this could not be considered as the actual estimate, owing to the replacements done. Therefore, the transformer which was initially considered for replacement due to continuous increase in
3
TRAFOTECH 2010 Eighth international Conference on Transformers
furan concentration was given a service clearance and it is still working well at its respective site.
IV CONCLUSION
DP is a highly reliable tool for RLA. The two cases studies mentioned in sec. III clearly indicate the significance of DP in estimating the remaining life of the transformer. Weighing the pros and cons of furan analysis and DP, the latter greatly outshines the former, when the history of the transformer is unclear. But furan too has its advantage being an online method. Therefore the choice of the method for RLA depends upon situation and the requirement of the end user. The end user can decide between two methods - one which yields better results, but requires a shut down and the other which is performed online, but misleads under special circumstances. However it is it is always advisable to employ DP for remnant life estimation, if the history of the transformer is unclear.
