Available online at www.sciencedirect.com
ScienceDirect
Available online at www.sciencedirect.com
Procedia Engineering 00 (2017) 000–000
ScienceDirect
www.elsevier.com/locate/procedia
Procedia Engineering 202 (2017) 88–108
4th International Colloquium "Transformer Research and Asset Management”
Bushing Failure- Investigation process & findings
Tarik Al Abria, Mohan Lala, Ibrahim Al Balushia, Mohammed Al Zedjalia
Oman Electricity Tranmission Company ,P.O BOX:1389 ,Al Khoud P.C:132 ,Omana
Abstract
Between 28 August 2011 and 8 September 2011, five 33kV OIP bushing failures occurred at JBB Ali Grid Station
belongs to Oman Electricity Transmission Company (OETC). Three failures resulted in a complete loss of supply
and around 130,000 consumers were reported with a lack of electricity. OETC assigned a third party investigator to
conduct a power failure investigation in order to obtain a possible root cause for these failures.
The investigation approach or process that were followed included the following phases:
•
Phase 1: Facts finding and preliminary analysis
•
Phase 2: Destructive dismantling and laboratory investigation through.
•
Phase 3: Root cause analysis.
From the investigation and buildup of hypothesis, no single most likely cause of failure could be identified. Metal
migration and semi-conductive copper sulphide migration in the oil that led to surface contamination of insulation in
the stress zone is considered causal factor for bushings insulation degradation. Copper sulphide in bushing was
detected caused due to corrosive Sulphur in oil. There has to be another factor influencing the weakened insulation
breakdown. The source of this factor can be found in the suspected frequently occurring earth faults in the 33 kV
line feeders, leading to a line voltage to ground of 33 kV and transient over voltages due to switching off the faulted
feeder.
* Tarik Al-Abri. Tel.: +96899501136; fax: +96822309046.
E-mail address: tarik.alabri@omangrid.com
1877-7058 © 2017 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the organizing committee of ICTRAM 2017.
1877-7058 © 2017 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the organizing committee of ICTRAM 2017.
10.1016/j.proeng.2017.09.697
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
2
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
2
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
89
All the OIP bushings were supplied in the same patch filled with the same oil and operated in identical service
All
the OIPsobushings
wereofsupplied
in the
patch
filled with
theconsidered
same oil and
operatedstate.
in identical
service
conditions
degradation
insulation
as same
chemical
process
can be
in similar
Electrically
all
conditions
so degradation
insulation
as chemical
processstresses
can beasconsidered
in are
similar
state.inElectrically
all
bushings were
subjected to of
identical
voltage
and load related
transformers
operated
parallel. Thus
bushings
wereofsubjected
identical
and as
load
related
stresses
as transformers
areofoperated
parallel. Thus
combinations
these twotofactors
are voltage
considered
causal
factor
for frequent
breakdown
multipleinbushings.
combinations of these two factors are considered as causal factor for frequent breakdown of multiple bushings.
© 2017 The Authors. Published by Elsevier Ltd.
© 2017 The Authors. Published by Elsevier Ltd.
Peer-review
under responsibility
responsibility
of Elsevier
the organizing
organizing
committee of
of ICTRAM
ICTRAM 2017.
2017.
Peer-review
of
the
© 2017 The under
Authors.
Published by
Ltd. committee
Peer-review under responsibility of the organizing committee of ICTRAM 2017.
Keywords: Bushing; tan δ; Bushing failure; contamination; deposits; transients; JBB Ali; Mudhairib; Investigation; Stress Zone; ESM/EDS
Keywords: Bushing; tan δ; Bushing failure; contamination; deposits; transients; JBB Ali; Mudhairib; Investigation; Stress Zone; ESM/EDS
1- Introduction
1- Introduction
The OETC network voltages comprise of 400KV, 220kV and 132 kV, and include 132kV/33kV substations feeding
The
network
voltagesofcomprise
of 400KV,
220kV and
132 kV,
132kV/33kV
into OETC
the primary
substations
the distribution
companies
at 33kV.
Oneand
of include
the substations
amongsubstations
the OETCfeeding
132kV
into
the primary
substations
of the
distribution
companies
at 33kV.
One of ittheis substations
thesubstation.
OETC 132kV
network
is JBB Ali
(132/33kV).
Since
this substation
is a radial
substation
considered among
a crucial
This
network is is
JBB
Ali (132/33kV).
Since this substation
a radial
substation itIncoming
is considered
a crucial
substation
equipped
with two 132/33kV,
125 MVAispower
transformers.
feeders
to the substation.
transformerThis
are
substation
is equipped
with
two
132/33kV,
125
MVAGeneration
power transformers.
Incoming feeders to the transformer are
from the 132kV
overhead
line
from
Al-Kamil
Power
plant.
from
the 132kVfive
overhead
lineoccurred
from Al-Kamil
Power
Generation
plant.
The following
incidents
in 132/33
kV substation
JBB
Ali:
following
five incidents
occurred
132/33
substation JBB
Ali:
2011 at 19:10
hrs,in125
MVAkV
Transformer
T2 tripped.
•The On
28th August
th
August 2011
at 19:10
MVA
Transformer
•• On
On 28
3rd September
at 19:25
hrs,hrs,
125125
MVA
Transformer
T1T2
at tripped.
JBB Ali substation tripped due to failure of 33kV
•
On 3rd September
at 19:25 of
hrs,
125transformers
MVA Transformer
T1 at JBB
due tothat
failure
of 33kV
Y-phase
bushing. Outages
both
at the substation
ledAli
to substation
a black-outtripped
in the zone
was fed
from
Y-phase
Outages
of both transformers
at the
substation
led to amaintenance
black-out inteam
the zone
that was
from
the JBB bushing.
Ali substation.
Following
this black-out
situation,
an OETC
swapped
the fed
R phase
the
JBB of
AliT1substation.
Following
black-out
situation,
an OETC
maintenance
swapped2011
the R
September
at phase
13.26
bushing
transformer
to the Ythis
phase
of T2 and
energised
transformer
T2 on 4thteam
th
bushing
of T1 transformer to the Y phase of T2 and energised transformer T2 on 4 September 2011 at 13.26
hrs.
•
•
hrs.
This replaced bushing failed however after 18 hours, on 5th September 2011 at 07.36 hrs. This caused second
This
replaced
bushing
18 hours,inonthe
5thshortest
September
at 07.36
hrs. This
caused second
subsequent
zone
black failed
out. Tohowever
restore after
this black-out
time 2011
possible,
an OETC
maintenance
team
subsequent
black
out. To
restore from
this black-out
the shortest
time
possible,
OETC
maintenance
team
removed
onezone
set of
identical
bushings
Mudhairibinsubstation
and
replaced
the an
failed
bushing
of Y phase
at
th
removed
set of identical
bushings energised
from Mudhairib
replaced
failed
of Ybushings,
phase at
Septemberand
2011
at 3.58thehrs.
Thebushing
other two
T2. This one
transformer
was eventually
on 6 substation
th
September
at 3.58
hrs. The T1
other
bushings,
T2. This from
transformer
wassubstation,
eventuallywere
energised
on 6on
received
Mudhairib
assembled
T1 at Y and2011
B phase.
Transformer
wastwo
energised
on
th
from at
Mudhairib
September
1.57 hrs. substation, were assembled on T1 at Y and B phase. Transformer T1 was energised on
7received
•
•
th
September
at 1.57
hrs. hrs T2 tripped again due to failure of B phase bushing. Following this failure OETC
7On
7th September
at 18.33
th
On 7 September
at 18.33
hrs T2from
tripped
failure
of B phase
bushing.
this failureT2
OETC
replaced
all the three
bushings
T2 again
with due
threeto new
bushings
received
fromFollowing
the manufacturer.
was
•
•
replaced all
three bushings
T2 bushing
with three
new bushings
received
the manufacturer. T2 was
18.33from
hrs after
replacement
and was
kept atfrom
no load.
energised
on the
8th September
th
energised on 8 September 18.33 hrs after bushing replacement and was kept at no load.
On 8th September at 19.48 hrs T1 tripped due to failure of R phase bushing (5th bushing from the lot of 6 at this
at 19.48
T1 tripped
due to failure
R phase
bushing
(5th
bushing
from the
lot from
of 6 atT2this
On
8th September
substation).
Following
this hrs
failure
OETC replaced
the Rofphase
bushing
of T2
with
the bushing
taken
Y
substation).
Following
OETCsubstation.
replaced the R phase bushing of T2 with the bushing taken from T2 Y
phase,
the one
brought this
fromfailure
Mudhairib
phase, the one brought from Mudhairib substation.
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
90
3
Figure 1 shows the schematic representation of the failures. Following these occurrences, a third party was assigned
by OETC to conduct a power failure investigation.
Figure 1: Schematic representation of the incidents
Nomenclature
OETC
JBB Ali
SEM
EDS
LV
PD
Oman Electricity Transmission Company
Jaalan Bani Bu Ali
Scanning Electron Microscope
Energy Dispersive Spectrometer
Low Voltage
Partial Discharge
2- Approach of Failure Investigation
In this section, the adopted approach towards the investigation of failure incidents is described. The failure
investigation was conducted in three subsequent phases:
2.1 Phase 1: Data collection and preliminary analysis
Data collection and preliminary analysis of facts and reporting with recommended plan of action for investigation
are part of phase 1.
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
91
4
2.1.1 Data collection
Data collection is the key process of the investigation. The status of the failure site was maintained till visit the site
and collects necessary evidence and data. All possible available data relevant to the incident were collected via the
following channels:
− Data request / questionnaire
− Interviews with client / stakeholders staff
− Visit to failure site
− Physical inspection of the available damaged components
− Photographs taken during investigative inspection.
The following data had been collected:
•
Failure incidents sequence as shown previously in Figure 1
•
SCADA event data
•
Loading data
•
Normal operation philosophy
The single line diagram of JBB Ali grid station is shown in Figure 2. It can be seen the both 125MVA transformers
T1 & T2 are fed through direct connection of 132KV feeders (Al-Kamil 1 & 2). The SLD shows the boundary
between OETC and Distribution Company as both transformers are connected to 33KV bus bars with 33KV
outgoing OHL feeders which are belong to the distribution company.
Figure 2: Single line diagram of JBB Ali Grid Station
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
92
5
The transformers are equipped with oil impregnated paper LV bushings of type COT 170-2500. The nameplate
details and failed bushings serial numbers were also collected.
2.1.1.1 Pre-service history of bushings
Six transformers of same type were supplied in a lot against an order to OETC. Two of these six transformers were
installed at JBB Ali, another two at Mudhairib grid station and two at Sur grid station. All six transformers were
fitted with 33kV bushings. Bushings were routine tested at the manufacturer facility. Thereafter, from the review of
available transformer factory test reports and site pre-commissioning test reports it can be stated that bushings were
not independently tested for any electrical tests i.e. tan Delta or capacitance test.
2.1.1.2 Post service history
Transformer T1 and T2 were commissioned in June 2002. Transformers’ routine maintenance is carried out on two
yearly bases. No information could be gathered on last date of bushing visual inspection. There was no record
available for off line visual inspection of these bushings. On 11 January 2011 on-line partial discharge (PD)
measurements and Infrared measurements were carried out on the transformers. The UHF method was applied to
measure PD. No PD activity was detected inside the transformers. Infrared measurements indicated no abnormality.
2.1.1.3 (Recent) maintenance/diagnostics
Following the third failure of a bushing, and prior to replacement of the bushing from Mudhairib grid station, OETC
maintenance team carried out tan delta and capacitance testing of the R & B bushings at transformer T2 and B
bushing at transformer T1. Results are abstracted in tables below.
Table 1: Capacitance & Tan delta test results for R & B phase bushings from T2
Table 2: Capacitance & Tan delta test results for B phase bushing from T1
In addition to these measurements on 6 September 2011 transformer T1, T2 turn ratio test was carried out,
Transformer oil BDV test was performed, 33kV cable insulation resistance was checked, neutral grounding
resistance was tested for its value. All results are within limits.
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
93
6
2.1.1.4 Protection recordings
Protection logs of three trip incidents were only available and were collected during site investigation. Following
three files were found:
− TX diff-2 (failure 28-08-2011 19:10:29) refers to Trip incident 1, Differential current: 0.72A (Y-phase)
− TX diff-1 (failure 03-09-2011 19:25:11) refers to Trip incident 2, Differential current: 2.2A (Y-phase)
- TX diff-1 (failure 08-09-2011 19:48:53) refers to Trip incident 5, Differential current: 0.389A (R-phase)
From the analysis it appears that differential protection relay operated as expected.
2.1.1.5 Site Inspection of Failed Bushings
All five failed bushings had been inspected at site. Most of the bushing had exploded oil conservator or top and
bottom oil seal damage. In addition, the 2nd failed bushing’s porcelain insulator exploded and others were having
broken sheds. In most of the bushings, the copper part inside the conservator was blackened as well as a small part
of the copper above oil conservator. The healthy bushing was inspected visually and it is normal. Two oil samples
was collected, one from a failed bushing that was not exploded and other sample from T2.
The healthy bushing (Figure 3 (f)) was tested for tan delta at 10KV test voltage and the results was 0.299% which is
within the acceptable limits.
(a)
1st Incident
(b)
2nd Incident
(c)
3nd Incident
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
94
(f)
(d)
4th Incident
(e)
5nd Incident
7
Healthy Bushing
Figure 3: Inspected failed & healthy bushings
2.1.1.6 Action Plan Following Site Investigation
The preliminary observations are that there are number of similarities between the failures:
−
Bushings are identical of same make, type and model.
−
Bushings were installed on the transformers installed at one grid station, same make and model.
−
Bushings failure took place at same grid station though at different transformers operated in parallel.
−
Both transformer loading patterns were identical and both were not loaded to full capacity.
−
There was pressure build up in the bushing caused due to internal fault.
−
Three out of the five failures occurred on Y phase bushings.
−
Four out of five failures occurred during evening hours (First incident at 19.10; second at 19.25, fourth at
18.33 and fifth at 19.48, only the third incident happened at 07.36)
It was recommended to carry out detailed dismantling at OEM factory of all the failed bushings followed by a
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
95
8
destructively dismantling of the un-failed bushing removed from service from JBB Ali substation. During
dismantling, samples will be collected of the failed component for a laboratory inspection based on the observations
made during the dismantling.
The possible tests are:
−
Testing and analysis of paper, chemical and microscopic (chemical analysis of paper i.e. degree of
polymerisation, scanning electron microscope (SEM))
−
Testing and analysis of copper
−
Testing and analysis of oil
−
Oil sample from a failed bushing has been collected for testing, DGA, Oil chemical and physical
properties.
−
Carry out routine testing of un-failed bushing.
−
Oil sample from un-failed bushing taken for testing prior to destructive dismantling.
2.2 Phase 2: Destructive dismantling and laboratory investigation
Dismantling of failed equipment or its components and laboratory investigations of failed components are
performed during this phase according to the action plan. All five failed bushings and the healthy bushing were
destructively dismantled. Black rings were noticed at different locations of the healthy bushing’s winding.
Photographs were allowed to be taken by the manufacturer for the failed bushings only. The following samples were
collected during bushings’ dismantling (see Table 3).
Table 3: Collected samples during dismantling
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
96
9
2.3 Phase 3: Root Cause Analysis
2.3.1 Analysis of observations on bushings (field and factory inspection)
From the review of observations during dismantling and analysis made at five failed bushings it can be stated that:
−
Four out of five bushings had sudden pressure surge due to fault and caused explosion of oil
conservator and oil seal burst.
−
Four out of five bushings had an exploded winding on the air side of the bushing.
−
Three out of five failed bushings have the fault location on the upper edge of the ground layer i.e.
starting point of upper axial stress zone (see Figure 4).
−
At one of the bushings, the fault location lies in the upper radial stress zone approaching the radial
stress zone.
−
One of the bushings had a fault in the radial stress zone. However this bushing also showed signs of
dielectric breakdown at the location starting point of upper axial stress zone.
−
No failure was found on the oil side of the bushing.
From this basic analysis it can be stated that bushing faults have occurred or initiated at starting point of upper axial
stress zone as it is demonstrated in Figure 5.
Figure 4: Radial and axial stress zones in a bushing
Tarik Al
Abri
et al. / Procedia
Engineering
202000–000
(2017) 88–108
Tarik Al-Abri
et al.
/ Procedia
Engineering
00 (2017)
97
10
Figure 5: Five Failed Bushings arcing locations
Bushing which was not failed but was taken out of service on dated 8 September 2011 by OETC was dismantled at
the factory. Evident black rings were observed at the bushing winding, the location of rings was found to match with
the concentric zone of radial stress. The prominent rings were found, formed at 540mm in the air side and even on
the oil side; at both locations matching with the edge location of the ground layer, while other non-prominent
deposits in ring shape were aligned to inner foil edge (see Figure 6). The black ring formation appeared as a deposit
in appearance with naked eye and paper sample from that location was later investigated at the lab through EDS.
From this basic analysis it can be stated that black ring formation at the outer layer of winding matches the starting
point of upper and lower axial stress zone.
2.3.2 Analysis of material test results
2.3.2.1 Oil analysis
During the investigation the following oil samples have been taken:
−
Transformer T2
−
Failed bushing
−
From healthy bushing prior the destructive dismantling of the bushing. Samples were collected before
and after the repetition of the routine tests.
Al Abri
al. / Procedia
Engineering
202 (2017)
88–108
TarikTarik
Al-Abri
et al.et/ Procedia
Engineering
00 (2017)
000–000
98
11
Figure 6: Black rings deposits on healthy bushing
The following parameters has been tested on all the oil samples:
-
Dissolved gas analysis (DGA)
Corrosive Sulphur content
Dibenzyl Disulphide (DBDS)
Bushing oil samples were also tested for
-
Dissolved metal in oil
Dielectric Dissipation factor (DDF)
Acidity
No specific observations were made on an oil sample of the transformer oil. Oil from the failed bushing indicated
high energy electric discharges. Water content was found, while breakdown voltage was low. The DDF
measurement at 90oC can be stated high. Oil was found non corrosive. Small copper and silicon concentration was
detected.
Oil from unfailed bushing was analysed and considered oil dissolved gas analysis results as reasonable for 10 years
old bushings in service. The results of the investigation show no discharges and/or thermal faults in the bushing. No
significant dissolved metal in oil has been found. Oil was detected as non-corrosive and no DBDS was found. DDF
test was carried out at 90oC before testing and the value was 0.0561 with 24 GΩ resistivity. After testing value was
0.0450 with 21 GΩ resistivity.
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
99
12
2.3.2.2 Paper analysis
Five paper samples were collected from different failed bushings. The following analysis was carried out for the
collected samples.
- Paper DP testing
Degree of Polymerisation (DP) test on the paper sample from the un-failed bushing was performed. Three samples
were collected.
−
−
−
One from top layer
One from middle and
One from bottom layer
DP was tested and appeared to be 497, 480 and 512 respectively. One sample of the new kraft paper collected during
visit to the factory and was also tested with DP value of 1339. The DP vale of paper indicates normal aging of the
paper insulation in the bushing.
- SEM/EDS Testing
Several Scanning Electron Microscopic (SEM)/ Energy Dispersive Spectrometer (EDS) analysis are performed on
the paper samples of the failed and unfailed bushings. Paper samples were collected during destructive dismantling.
The areas investigated were blackened rings around the paper insulation of the good bushing, near the hole of the
paper where the test tap comes out of the bushing, identified particles on the paper and paper near the fault location.
Several black fibres are identified in the paper samples. From the stereo microscope this appeared to be a (paper)
fibre coming out of the surface of the paper (see Figure 7).
Figure 7: Black paper fibre
EM/EDS analysis on the deposits found around the hole where the (test tap ground) connection is made from the
first capacitive layer to the test tap. Several elements were found in the vicinity of the black hole, such as: Carbon,
Oxygen, Chlorine, Zinc, Tin and Lead. Also small amounts of sulphur and calcium are found. The traces of Zinc and
Chlorine are most likely to come from the use of soldering water for realizing the soldering of the connection of the
test tap. Tin and lead are likely to come from the soldering itself. The origin of carbon is expected from the paper
while sulphur is expected from oil or from sealing gaskets. The presence of calcium is unclear, only expected as
100
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
13
remains from the environment during manufacturing process.
Figure 8: Microscopic and SME/EDS picture of paper at Test tap hole
Black rings that are available in the healthy bushing (see Figure 9) was analysed through SME/EDS (see Figure 10).
An EDS analysis showed that the deposit contains the following major elements, being copper, sulphur, oxygen,
zinc and carbon. Incidentally iron have been also found between the paper fibres.
(a)
(b)
Figure 9: (a) Black deposit rings observed (b) SEM/EDS picture – black deposit on paper, elements analysis
Another sample of paper with aluminium foil was analysed from a failed bushing which is shown in Figure 10 (a).
Since the area near the edge of the aluminium foils is critical, it was decided to analyse this sample. Besides that the
damage shows some discharge or treeing like effect on the aluminium foil, see Figure 10 (b). The results of this
analysis shows that the majority of elements found are Carbon, Oxygen and aluminium. Negligible amounts are
found of Silicon, Sulphur, Calcium and Iron. In this sample no traces of copper is found.
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
(a)
101
14
(b)
Figure 10: Paper sample with aluminium foil
To have larger picture, one more sample of paper from the fault location of one of the failed bushings had been
analysed as shown in Figure 11 (a).
(a)
Figure 11: Paper sample from fault location
(b)
The sample taken from one of the failed bushings shown above has been divided into three samples, corresponding
to the three burnt paper holes. See Figure 11 (a). The major hole, where the actual fault from the conductor to
ground took place, shows the presence of copper (from the conductor), Aluminium (from the Aluminium layers) and
Calcium. Other minor elements found are Sulphur and Magnesium. The analysis of the lowest hole with the crack
and the most right fault location showed the presence of Carbon, Oxygen, Chlorine, Aluminium and Calcium. Trace
amounts of Iron, Sulphur and Copper have been found here. Besides that the most right location also showed the
presence of silver particles, indicated as number 2 of the Figure 11 (b). Point number 1 is an Aluminium particle.
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
102
15
One more sample of Darkened paper from the top of a failed bushing was analysed, see Figure 12. The analysis of
the deposit showed the presence of sulphur, copper, Carbon and Oxygen. Traces of Aluminium, Zinc, Silicon and
Calcium were also found.
Figure 12: Blackened paper sample
2.3.2.3 Copper analysis
The copper conductor of the healthy bushing has been analysed, since it had been found black deposits around the
copper and wanted to know the content of these deposits. The white line in Figure 13 (a) shows the tested copper
sample.
(a)
(b)
Figure 13: Copper stud from Healthy bushing
A detailed view and element analysis see Figure 13 (b), shows that sulphur has been detected in combination with a
majority of copper indicating a likely deposit of copper sulphide. Other elements found are, Carbon, Oxygen and
trace amounts of Nitrogen, Zinc and Calcium. A sample of failed bushing was investigated and shows
approximately the deposits as above.
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
103
16
2.3.2.4 Healthy Bushing Routine Tests at the factory
The healthy bushing was tested for routine tests at the factory with 85% withstand voltage as per IEC 60137-2008
standard. The bushing passed the test successfully. In addition, the bushing was tested for tan delta and capacitance
for the bushing as well as for the test tap. The results were compared with factory tests of 18bushings supplied at the
same time. Comparative test result analysis for tan delta of all 18 bushings produced in one lot exhibit variation of
over 35% while the capacitance value variation is up to 7.37%. The variation in measured values of the test tap
capacitance and tan delta is of the order of 20% and 376%, respectively. From this it can be stated that comparative
assessment for condition of bushings cannot be implied with test results of tan delta and capacitance values of same
family bushings even produced in one lot.
While comparing the test results of the healthy bushing from the December 2011 results to 2001 results, the bushing
tan delta value was found 40% higher at 10 kV while capacitance value can be stated quite comparative. From the
comparison of capacitance values bushing condition can be stated good. Change of over 5% in capacitance value is
recommended to take the bushing out of service. The variation in measured values of the test tap tan delta and
capacitance is of the order of (-) 75% and (-) 13%. It cannot be used a criterion for condition assessment.
2.3.3 Failure Hypothesis
The followed approach of the failure investigation, different failure hypotheses are formulated (including less
probable ones) in order to determine the root cause of the failure and to exclude less likely causes. Each failure
hypothesis result is weighted on a scale of 0 to 4 for causal factor of failure.
Weightage of scale is as following:
0 – Highly unlikely
1 – Unlikely
2 – Possible
3 – Likely
4 – Highly likely
The following possible failure scenarios are evaluated to determine the root cause of failure of the bushings:
1- Pre-service related factors
No evidence was found of any concerns that can be related to pre-service related factors. Though there were some
observations on the paper where some black linings were seen however appeared as paper fibre of that colour.
Installation of oil filled bushings with oil level view glass in an enclosed cable box is considered as a matter of
concern though could not be related to failure incidents. Pre-service factors are rated at “0” on the scale of 0 to 4 for
causal factor of failure, so is considered a highly unlikely cause of failure.
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
104
17
2- Crack in insulation body
Cracks in the porcelain may lead to loss of oil while cracks in epoxy insulation may lead to mixing of bushing oil
with transformer oil. There were no observations or evidence found on the failed bushings related to a crack in the
insulation body. Cracks in the insulation body is rated at “0” on the scale of 0 to 4 for causal factor of failure, so
considered as a highly unlikely cause of failure.
3- Leakage of oil
The leakage of bushing oil can take place during service life of the bushing due to failure or aging of gaskets, seals
or axial movement of the porcelain. Bushings are fitted with oil level observation glass and are designed for
installation outdoor in such a way that regular monitoring of the oil level can be done. It was noticed that bushings
are installed in a covered cable box where such monitoring is not possible under on line condition of the transformer
and is feasible only when the cable box cover are removed and level inspected. No documented record of such
inspection was available. Subsequent to the first two failures after each incident the oil level of the surviving
bushings was checked by OETC and found in order. To have a situation where the oil was lost in the relatively short
period between restoration and failure would have required a major oil leak and there was no evidence to suggest
that. Last failed bushing still had oil intact.
There was no evident sign of low oil level in any of the failed bushings so leakage of oil is rated at “0” on the scale
of 0 to 4 for causal factor of failure, so considered highly unlikely cause of failure.
4- Disconnection of earthing link of test tap
A loosely threaded cap can lead to poor contact of the ground connection and can lead to electrical discharges at the
test tap location. This will eventually lead to failure of the bushing. In none of the failed bushings such evidence is
found. Hence, this cause is rated at “0” on the scale of 0 to 4 for causal factor of failure, so considered as a highly
unlikely cause of failure.
5- Loose connections at current carrying joints
Loose connections at current carrying joint become a source of heat. This will lead to a gradual oxidation of the
contact point and subsequent progressive increase in contact resistance. This process continues and a thermal
runaway can originate. There was no evidence of a thermal runaway so this factor is rated at “0” on the scale of 0 to
4 for causal factor of failure, so is considered as highly unlikely.
6- Issues related to insulation oil
•
Oil contamination
From the available site test records on two out of five failed bushings and from the testing of the unfailed bushing at
factory, no such evidence could be found. This factor is rated at “0” on the scale of 0 to 4 for causal factor of failure,
so considered as a highly unlikely cause of failure.
•
Deterioration in oil quality while in service
Dielectric dissipation factor values as per IEC 60422 were found in acceptable range. Change of DDF with
temperature is expected though there is an observation that oil DDF adopted a sharp rise with temperature that gives
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
105
18
an indication of possibilities of presence of polar compounds. Further, the presence of oil decomposition products
such as sludge was noted during EDS investigation of the copper stem from un-failed bushing. Despite the fact that
the sample was degreased, still considerable amounts of carbon were present. With above evidences this factor can
be considered as a contributing factor to a limited extent so rated at “1” on the scale of 0 to 4 for causal factor of
failure, so considered as an unlikely cause of failure.
•
Corrosive sulphur
Bushings were filled with a type of mineral oil. Oil sample from the bushing was tested negative following IEC
62535 to determine whether oil is corrosive. DiBenzyl DiSulphide (DBDS) content was also found below detectable
limit of 5 mg/kg. DBDS is the most common compound identified out of the identified few corrosive sulphur
compounds so is normally tested along with corrosive test. Following test standards oil was found non corrosive.
It is known that no single corrosive sulphur compound is responsible for corrosive Sulphur issues that are present in
transformer oil. There can be number of different sulphur compounds present in the oil as sulphur compounds are
present in crude oil. Few compounds/species are known to degrade from stable into reactive. Responsible factors
identified for this process are service time, oxygen deficient environment (bushings under investigation are nitrogen
blanketed) and temperature.
SEM/EDS investigation of deposits over the copper stud (blackened area) gives ‘positive identification’ of copper
sulphide (Cu2S), it is important to notice that the amount of sulphur is often equal or sometimes even higher than
the amount of copper.
Presence of copper and sulphur at paper samples was also found and presence of copper sulphide could not be ruled
out. From above forensic investigation evidences it can be stated that bushing was affected with corrosive sulphur
though the oil was tested “non-corrosive” following standard laboratory tests.
With all mentioned facts taken together, corrosive sulphur is considered as a contributing factor in the failure. It can
therefore be rated as “2” on the scale of 0 to 4 for causal factor of failure, so classified as possible cause of failure.
•
Metal migration in oil
SEM/ EDS investigation of paper samples indicated that the black ring formed at the winding mainly contains
−
Copper
−
Sulphur
−
Zinc
Some particles containing iron, calcium, magnesium could also be found. Traces of silicon and aluminium were also
found at a few locations. The black ring is the accumulation of semi-conductive, conductive and polar particles. All
these particles deposited on the surface of the winding paper during the service life of the bushing. The migration of
metal particles took place with in the oil. It was noticed that deposition of these insoluble products occurred in areas
of high electrical stress. Four out of five failed bushings’ fault locations matches with the location of deposits found
on the winding at the unfailed bushing. The surface contamination can reduce the electrical strength of the insulation
system. The rate of deposition in all the bushings operating in similar conditions at a substation can hypothetically
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
106
19
be considered same. This applies to the state of deposits within the bushings prior to failure, which could not be
established. Deposition appeared to follow a pattern like treeing over the paper surface.
The Oil DGA result could be stated normal for such bushing with no indications of partial discharges. Conclusion
could be made that even if bushing had visible evidence of electrical treeing or presence of deposition of copper or
sulphur, oil DGA may not indicate any gases related to partial discharge.
These particles deposition formed a progressive close/semi close ring across the winding during the service life of
10 years.
The deposition of metal and (semi)conductive particles initiated partial discharges under the influence of voltage.
The fact remains that PD intensity during testing of the un-failed bushing at the factory was recorded below
detection limit of 2pC so hypothesis shall only be considered with other contributing factors. The known fact is that
five bushing failed with in a span of ten days, the logic of chemical process taking place in all the bushings
operating in similar conditions at a substation can only support this hypothesis.
The premature failure at OETC bushing have been caused due to deposition of metal, (semi)conductive particles and
oil decomposition products at the active core of the bushing in the stress zone. The migration of metal and semi
conductive particles has taken place in oil during the service life of the bushing. From this hypothesis, migration of
metal is rated as “3” on the scale of 0 to 4 for causal factor of failure, so classified as likely cause of failure. Though,
it shall be considered with other contributing factors.
•
Failure due to network transients
Lightning or switching impulses can lead to bushing insulation failure. The day of the incidents was without any
thunders / lightning and none of the bushing appeared to be failed at the time of reported switching. The known fact
is that all the 33kV feeders from the substation are through overhead lines with short parts of cables used for road
crossings. It could be observed that there have been instances of 33kV overhead lines conductor falling, jumper
broken, cable lugs burnt that leads to ground faults. Single phase ground fault with resistive grounding will increase
the phase-to-ground voltage of the remaining two phases to phase to phase voltage.
Review of protection log indicates changes in HV currents along with measurable changes at 33 kV side, that makes
us to develop an opinion that first fault occurred just after a fault occurred somewhere else (deeper in) the 33kV
network though no feeder trip recorded at that instance.
On 28th August 2011 at 15:04.45 there was a reported incident at Al-Kamil feeder 7L5 feeder tripping at distance
protection that is about 4 hours prior to first bushing failure. Though it was reported to be an incident of jumper
broken, phase of that jumper was not reported. It is likely that an over voltage affected the bushing at this incident
and initiated the process of electrical discharges leading to first incident of failure, knowing the fact that bushings
insulation were affected due to deposition of semi-conductive and conductive particles.
In case of earth faults on 33kV feeders, if considered frequent, all the bushings will be stressed with a root 3 higher
voltage as transformers are operated in parallel. Bushings with degraded insulation may cause a set of bushings to
fail in a short period as experienced at OETC. Assumption is made that number of ground faults has significantly
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
107
20
increased in a period prior to the first fault. Assumption is made based on limited data received from the distribution
company.
From this hypothesis, network transients are considered as a likely factor under the prevailing insulation condition
of the bushings so is rated as “3” on the scale of 0 to 4 for causal factor of failure.
•
Harmonics
JBB Ali substation feeds BBC (British Broad Casting) relay station, a relay station contains power electronics and
therefore the load will contain harmonics. No record of harmonic content was available. The load of BBC feeder
was stated to be of the order of 7 – 8 MW. Load comparing to the capacity of the bushing is low. An important fact
remains that out of the eighteen bushings supplied and installed at three different substations, only bushings installed
at JBB Ali substation failed at the same period of time.
The factor cannot be ruled out completely so rated at “1” on the scale of 0 to 4 for causal factor of failure, hence
considered as an unlikely cause of failure.
The resume of the evaluations of the failure hypotheses as discussed and elaborated above is presented in Table 4 as
below.
Table 4: Failure hypotheses
Tarik Al Abri et al. / Procedia Engineering 202 (2017) 88–108
Tarik Al-Abri et al. / Procedia Engineering 00 (2017) 000–000
108
21
3- Conclusion
In conclusion, five bushing failures were taken place in 125MVA transformer at JBB Ali grid station belong to
OETC. To find the root cause analysis of the failure, an investigation was initiated. The investigation approach or
process that were followed included the following phases:
•
Phase 1: Facts finding and preliminary analysis
•
Phase 2: Destructive dismantling and laboratory investigation.
•
Phase 3: Root cause analysis.
From the investigation and developed hypotheses, no single most likely cause of the failures could be identified.
Despite that a likely cause of failure is migration of metal in the oil leading to surface contamination of insulation in
the stress zone it was also contributed by the presence of corrosive sulphur products i.e. copper sulphide. All the
bushings can be considered in same state of insulation degradation due to this chemical process occurred in identical
service condition and life time of the bushings.
Since the above mentioned mechanism will probably not lead to a direct breakdown of five bushings within a span
of two weeks, there has to be another factor influencing the failure of weakened insulation. The source of this factor
can be found in the frequently occurring earth faults in the 33 kV line feeders, subjecting all the bushings to a line
voltage to ground of 33 kV and transient over voltages due to switching off the faulted feeder, knowing that both
transformers are operated in parallel during the service life. The combination with the surface contamination in the
bushings deteriorated the dielectric withstand capability of the bushings leading to a frequent breakdown of multiple
bushings.
Acknowledgements
We would like to thank Dr. Adil Al-Busaidi, Sr. manager-Asset Management & Planning Department and Ahmed
Al-Rahbi, Manager-Asset Performance Management who motivated us to participate and contribute to Cigre.
References
[1] KEMA, Power failure investigation: Transformer bushings failure at JBB Ali Substation, 2nd Edition, 2012.