Effect of Moisture on Transformer Insulation and Cases of Power

Transformer Failure due to the Entry of Free water

Oommen P. Joshua Lissy Augustine
Transformers and Electricals Kerala Ltd (TELK), Angamally

1.0 Introduction 2.0 Sources of moisture

The insulation structure of any oil filled power The major sources of moisture build up in a power
transformer is a combination of an insulating liquid transformer can be broadly divided into three viz.
and a solid insulation impregnated with the same residual moisture especially in thicker insulation
liquid. Generally mineral insulating oil is the after drying , ingress from atmosphere due to
insulating liquid while Kraft paper, press board exposure to humid conditions, breathing etc. and as
and Densified wood forms the major solid a byproduct of decomposition .
insulation. All these solid dielectrics are made of
Cellulose fibre. Cellulose are Hydrophilic as well 2.1 Residual moisture

as Hygroscopic. Hydrophilic means water loving ,

readily absorbing or dissolving in water. Transformers are usually dried in the factory until

Hygroscopic means the ability to attract the average moisture content in the cellulosic

water molecules from the surrounding insulation is 0.5%. This limit may be lowered or

environment through either absorption which is a raised depending on the kV rating. A more rigorous

process involving the entire volume of the material drying out operation is necessary for EHV and

or adsorption which is a surface-based process UHV transformers. But moisture higher than the

where a film of adsorbate is created on the surface. above average can remain in bulk insulation at the

Even though hygroscopic materials are used as end of drying particularly in wood and resin

major insulation in transformer, water even in impregnated materials which are used as lead

minute quantities, is harmful to transformer supports, winding supports and OLTC support

because during operation it will be attracted to the structure. This is mainly because the drying time

places of greatest electrical stress and results in required is higher for thicker material. For

dielectric breakdown of the equipment. Water laminated boards which are glued together, the

accelerates the deterioration of both the insulating adhesive layer practically prevents any drying

oil and the paper insulation, liberating more water perpendicular to the lamination. Drying therefore

in the process. Once the paper insulation has been takes place only parallel to the lamination and thus

degraded, unlike the oil, it can never be returned to needs more time. High density material also takes

its original condition. Therefore water is the most more time for drying.

unwanted substance in transformer insulation During service this local excessive moisture will
come out of insulation, driven by heat, and will

1
become redistributed spatially. This will occur over
many temperature cycles with oil serving as the
carrier for moisture.

2.2 Ingress moisture

i) By Exposure to high humid conditions

At factory , after drying the transformer , it is

Fig 1 Moisture absorption of 3mm High Density
taken out from the oven for retightening. Time Press Board at 23O C at various humidity’s of
taken for this work varies from a few hours to ambient air[3]
several hours depending upon the method If the continuous exposure is more than 8 hours
employed by different manufacturers. Some do this particularly in high humid conditions, possibility of
work before first oil filling and others do this after moisture going deep in to the insulation can not be
initial oil filling. It is always convenient to do the ruled out.
retightening before oil filling. But chances of Part of this is removed by subsequent vacuum and
higher moisture ingress will be there if the HOC treatments at factory as well as at site. But
insulation parts are opened to atmosphere before many transformers that are dried at factory to a
initial oil filling. After retightening, the transformer moisture level of 0.5% or less, reported to have a
is again exposed at the time of lead taping and much higher moisture level at the time of
connection work. Transformer is again exposed commissioning itself which may be the cumulative
during despatch preparation. During transport also effect of the exposure to humid conditions at
the gas sealing integrity is affected in some cases. different stages.
At site the transformer is opened to atmospheric air
during lead connection , bushing mounting etc. for ii) Due to the breathing action through the
long duration due to various facility limitations. silicagel breather and conservator.
Many a times, at factory as well as at site these
works are not carried out in controlled atmosphere As per CIGRE report, the rate of water
resulting in re-entry of moisture in to the insulation contamination for transformers with membrane-
parts. With four days exposure of a 3 mm thick sealed conservator preservation systems is about
pressboard to air of 50 % RH, the water content in 0.03 to 0.06 % water in the cellulosic materials per
the insulation will increase by 5 %. At 97% RH it year. But in case of transformer with open
will be up by 15%.So the Relative Humidity of the breathing arrangement the yearly increase of
air and duration of exposure will be the critical moisture in the cellulose is 0.2%.
factors that decide the reentry of moisture in to the
insulation. iii). Ingress by molecular flow through very small
ducts and tubes

Moisture can enter into the tank by molecular

(Knudsen) flow due to the difference in water
vapour pressure in the atmosphere and the
2
transformer gas space or oil. The term Knudsen or include free glucose molecule, water, CO and CO2.
molecular flow describes the diffusion of gases Presence of oxygen promotes oxidation of cellulose
through very small ducts and tubes (e.g. molecules and water is a byproduct from this.
gaskets).Ingress of moisture by this method is Hydrolytic aging is mainly due to the moisture and
practically negligible except during the vacuum acid particles in oil. Water dissolves the oxygen
process. bridge between the glucose rings. For every
molecule of water consumed, three molecules of
iv) Ingress of moisture and entry of free water H2O are released in subsequent dehydration. Thus
through poor seals above conservator levels the process of degradation of cellulose insulation
by pyrolitic, oxidative and hydraulic mechanisms
Defects like improper top sealing arrangement of produce water as a byproduct.
draw lead bushing , loose hand hole cover at the
conservator top, ruptured explosion diaphragm etc ii) Oil degradation also produces water.
are another group of potential water entry points in
transformer. These bad seals will not cause oil leak Oxidation of oil due to moisture and oxygen in the
when they are above conservator oil level and thus presence of heat produces acids. The process is
goes unnoticed also. But this will result in viscous cumulative as acid coupled moisture further
flow of wet air into the transformer during the decompose the oil giving rise to more acids and
cooling cycle of transformer and large amount of moisture.
rain water can be sucked in during rainy season
3.0 Distribution of moisture
2.3 As a by product of ageing process

The water entered in to the transformer in various

i) Decomposition of cellulosic insulation material.
ways will remain in solid insulation as
i) Absorbed to surface
Cellulose fibre consists of cellulose molecules that
ii) In vapor form
lay side by side. Cellulose is a chain molecule
iii) Free water in capillaries
made up of glucose rings linked by OH bridges.
iv) As embedded free water
The no of glucose molecules forming the cellulose
chain determine the length of cellulose molecule.
In the oil, the moisture will appear as
Higher the no of glucose units per cellulose
i) Tightly bound to oil molecules - seen
molecule, higher will be the strength of cellulose.
particularly in aged oil
ii) Dissolved
iii) Emulsified water – ie water supersaturated
in oil but has not yet totally separated from it.
It usually gives oil a milky appearance.
Fig 2 Molecular Structure of cellulose [3]
iv) Free water suspended in the oil as water
Three most common degradation factors of
droplets , when moisture in oil is much above
cellulose are heat, oxygen and water. When
saturation value
cellulose is exposed to heat up to 2000C the glucose
rings disintegrates. By products of this reaction

3
The water inside the transformer is not shared paper insulation system. The dielectric properties
uniformly between cellulose and oil as cellulose of oil paper insulation system are good when the
has several hundred times affinity for moisture than paper is dried and impregnated and oil is dry.
oil. Generally, 99 % of total water content inside a
transformer will be in solid insulation and balance
only will be in oil. Moreover, water in insulation
will not be uniformly dispersed also.

It is also important to note that Moisture

distribution between Cellulose and oil is not
constant. Depending upon temperature, moisture
migrate from one element to the other. When
temperature increases moisture from paper goes to
oil and vice versa during cooling. Returning of Fig 3 Mineral oil Breakdown voltage at room
moisture from oil to paper is always slower temperature as a function of water content of the
oil [6]
process. The moisture distribution therefore is a
function of moisture content , temperature
distribution , masses and dimension of cellulose
insulation structure

Based on the water contamination in transformer

insulation, one of the classifications adopted is
given below..

Class Water content by weight

in Insulation
Fig 4 Dielectric strength of solid insulation with
Good Dry transformer - 0.5 ~ 1.0%
different moisture content at various temperature [7]
Fair - 1.0 ~ 2.0%
But dielectric strength of oil as well as solid
Wet Insulation - 2.0 ~ 4.0% insulation will be significantly reduced as shown in

Excessively wet Insulation - > 4.5% fig(3) and fig(4) when the water content is high in
oil and / or insulation. Higher moisture content will
result in low partial discharge inception voltage
4.0 Effect of moisture also.

Moisture in transformer has many direct as well as 4.2 Accelerates paper ageing
indirect adverse effect on its performance and life.
The ageing rate of cellulosic insulation is
A few of them are narrated below to bring out the
dependent upon temperature, water content and
importance of keeping the transformer dry.
oxygen. Generally temperature is the primary
4.1 Reduction in Break down strength factor that is affecting aging. But water in higher
quantities also have an influence on degradation of
Electric strength of oil is the controlling factor for
the power frequency strength requirement of oil

4
cellulose. The effect of water on ageing is generally concentration increase.” As per other published
considered as directly proportional. A humidity of data, for dry transformers (ie transformer with
4% can shorten a transformer’s operational lifetime <0.5% moisture in paper) hot spot up to 180oC may
by a factor of 40 (i.e., by 97.5%) from that of a dry be possible with little risk of bubble formation.
transformer. Thus moisture in insulation has a great bearing on
temperature at which bubble formation takes place
depending upon which overloading capability of
the transformer is decided.

Fig 5 Expected life for solid insulation and its

dependence upon moisture and temperature.[5]

4.3 Reduction in the overloading capability of

the transformer

Water travels through micro capillaries of cellulose

when temperature and vapour pressure are low
enough. During overloading , due to the rapid rise Fig 6 Bubble evolution temperature with different
levels of moisture content in insulation [8]
of temperature, rapid evaporation of absorbed
water and rapid rise of vapour pressure occurs that 4.4 Appearance of free water on sudden cooling
will force oil out of micro capillaries resulting in
In case of a transformer with higher moisture
appearance of vapour filled cavities on insulation
content, when the transformer is loaded and
surface. This phenomena is called bubbling. Since
operated at a high temperature for considerable
the dielectric strength of the bubbles is significantly
time, large quantity of moisture might have
less than the insulation system, their formation can
migrated from insulation to oil. If this transformer
result in discharge events starting from partial
is tripped or switched off on a cold day or on a
discharge to flash over. So the temperature at
rainy day , it will start to cool down very quickly.
which bubble forms put limitation in overloading
When the oil temperature goes down the solubility
capability of a transformer.
of water in oil reduces and water will start to go
IEC standard 60076-7 which is the loading guide back to cellulose. But returning of water to paper is
for the transformer clearly states“… bubbles are a slow process and free water may appear inside
likely to occur when the hot spot temperature the transformer . If transformer is restarted after
o
exceeds 140 C for a transformer with winding such a cool down but before moisture returning to
insulation moisture content of about 2%.This cellulose , a very critical situation can arise. If
critical temperature will decrease as the moisture transformer is sufficiently dry the moisture coming

5
in to the oil will be less during loading and
saturation condition may not appear while cooling.
So dry transformers are much less susceptible to
this type of problem.

Therefore it can be concluded that eventhough

increase in moisture may not result in instant
failures, it will be a silent killer of the transformer.

5.0 Measurement techniques

Even though moisture in solid insulation of a Fig 8 Griffin curves for water equilibrium in
cellulose mineral oil system [4]
transformer is so critical, direct measurement of it
is very difficult. The estimation of moisture level is Similarly Dew point-Moisture plots published in
done based on published Equilibrium curves for oil 1983 by Dr .TV Oommen shown in fig 9 is a very
filled environment or gaseous environment. In both convenient tool to find out moisture level in
the cases it is assumed that when the active part of insulation. If the gas phase is at a slight positive
the transformer is placed in a medium of dry oil or pressure the dew point will be slightly higher.
dry gas for sufficiently long period the moisture
from solid insulation will migrate to the medium
and will attain an equilibrium stage. By measuring
the moisture content of the medium and then using
the equilibrium curves, the moisture in solid
insulation is estimated. Moisture content for oil
medium is measured by Karl fischer titration
method and that of gaseous medium is found out by
Dew point measurement. Oommen curve published
in 1983 and Griffin curves published in 1988 are
the two widely used moisture equilibrium curves
for water equilibrium in cellulose mineral oil
system.

Fig 9 Chart for moisture estimation from Dew

Point measurement. Gas phase at one
atm pressure [2]

The limitations of the equilibrium models is that

the time taken to attain the equilibrium condition
may vary from a few hours to several days
depending upon temperature and is also dependent
upon direction of migration. So steady state
equilibrium may never be attained completely due
Fig 7 Moisture equilibrium curve by Dr.T.V
Oommen for paper/oil system [2] to continuous variations of load and ambient

6
temperature. Hence repeated measurement over a was opened inter turn failure was seen near HV line
period of time and averaging out the value can be a lead on a few discs below the middle of winding.
solution to reduce the error. As abnormal rust was seen inside the tank, it was
assumed that free water entry was there in to the
In recent years, various Dielectric Response
transformer. Each and every component at tank top
Analysis are also used to determine the moisture
was carefully examined. Finally some colour
content in solid insulation of power transformer.
difference was seen at the top of HV bushing. On
Recovery voltage Measurement (RVM), Dielectric
careful examination it was noticed that a (damaged)
spectroscopy in time domain ie measurement of
cork gasket was used between the terminal cap and
polarisation and depolarisation currents (PDC) and
the expansion chamber where as the bushing was
Dielectric frequency domain spectroscopy (FDS)
supplied with Nitrile rubber O ring gasket for this
are the most popular techniques available in this
joint. When terminal was removed clear colour
category. However the measurement by these
difference was noticed at the inside surface of the
methods are influenced by material properties,
tube through which draw through lead was coming
insulation geometry, Oil condition especially oil
out. On further investigation following points were
conductivity etc.
brought out.

6.0 Case studies Transformer was erected by customer and due to

some problem they have changed the gasket
As narrated above, excess moisture always has an
between the cap and expansion chamber. The new
harmful effect on transformer life. But direct entry
gasket used was not properly seating the joint
of free water can result in immediate failure of
resulting in ingress of free water .Part of the water
transformer. Recently we had two problems due to
entered in to the transformer soaked the HV line
the ingress of free water through Bushing terminal.
lead and has gone to HV winding resulting in inter
Both occurred after erecting / maintenance by the
turn failure. Some of the water particles was fallen
customer.
in to the tank and it might have travelled to one of
6.1 Case 1 the corners of the tank may be because of slope of

Transformer after erection by the customer was the tank in that direction. This free water that got

commissioned in the presence of manufacturer’s accumulated has caused the rusting of the tank.

testing Engineer. Transformer worked satisfactorily Commissioning Engineer from manufacturer has

for a few months . During the first rainy season it not gone to the top of the bushing to verify the

tripped on differential. On inspection after draining intactness of the terminal connection. As bushing

the oil, abnormal rusting of a large area was seen terminal was above conservator level when

inside the tank. . It was a transformer with HV line connection was not tight there was no leak, but

lead at the middle of the winding. When winding water could get in during rainy days resulting in the
problem .

7
 Fig 10 Rust / Water drops seen inside the tank

Fig 11 With Correct gasket at terminal fig 12 With Wrong gasket

6.2 Case 2 and a retainer over this. A cap is then provided over
this with O ring sealing to avoid moisture entry in
Higher absolute value and an increasing trend in
to the pipe. On inspection it was found that four
tangent delta was reported for a 400kV bushing
disc springs were found used in this particular case
from one of the sites. The bushing was brought
and a hair line crack also was seen on expansion
back to the factory for investigation and
chamber. With four disc springs the cap could not
rectification. Water entry marks were seen at the
position properly and there was gap between
top of bushing in the preliminary inspection. On
bottom of cap and expansion chamber cover. To
careful examination following defects were found
reduce this gap over tightening might have carried
out.
out resulting in the crack on expansion chamber.
The construction of 400kV bushing is such that the Water might have crept in through this gap /
top of draw through lead of Bushing is fixed in a through the crack resulting in higher tangent delta
slot at the cover of expansion chamber. The lead is for the bushing. This mistake / gap also went
kept tight and locked using three nos disc springs

8
undetected at different stages including will only go to the top of the bushing and
pre-commissioning test. However as the bushing everybody rely on their expertise. As HV Bushing
top is generally above conservator level, any loose
connection at the top can result in direct entry of
free water into the transformer and the result can be
catastrophic. Therefore well trained and reliable
expert people only shall be used for such works.
Any modifications at top shall be with the approval
of manufacturer and shall be verified by well
qualified persons.

7.0 Conclusion

“Water is not Friend, but Enemy No.1 for

transformers. Water is not an insulator, but a
conductor, and water absorbed in the paper and oil
could cause even an instant failure of a transformer.
So every effort to be taken to keep water out of it”.
Fig 13 Gap seen with wrong disc spring assembly Marginally higher values of moisture need not
result in the immediate failure of the transformer
was taken out when higher tangent delta was
and it gives a wrong confidence to many to handle
reported , great danger of bushing failure /
the transformer with out enough caution towards
transformer failure was averted.
entry of moisture. But, a transformer with low

6.3 Learning moisture will always have enhanced life and higher
overloading capacity. So every effort to keep water
Direct entry of free water in to the transformer can out of the cellulose insulation and oil, at all stages
cause immediate failure of the transformer. In of manufacturing , installation and operation is
many cases , such direct entries can go unnoticed as essential to ensure a long and trouble free service
they will not be revealed in any tests. Generally of the transformer.
utilities believe that if transformer is commissioned
in the presence of manufacturer’s representative all References
the mistakes till that point will be taken care off.
1.0 F.M Clarke “Insulating Materials for Design
But as commissioning is mainly an electrical
and Engineering Practice”
testing oriented activity, points that are not revealed
in any electrical testing will not be detected during 2.0 T.V Oommen ”Moisture Equilibrium Charts
pre-commissioning activities. Bushing terminal for Transformer Insulation Drying practice”
connection is one of them. In India, generally IEEE Transaction on Power Apparatus and
engineers will not go to the top of the bushing to Systems Vol PAS -103 No 10 October 1984
verify the gasket sealing or connections as suitable
climbing facilities will not be available at most of
the sites. People who can climb over the porcelain

9
3.0 CIGRE Brochure 349.”Moisture Equilibrium
and moisture migration with in transformer
insulation systems”, June 2008

4.0 Y.Du, M.Zahn, B.C Lesieutre, A.V Mamishev Name : Lissy Augustine

“Moisture Equilibrium in Transformer Paper- Qualification :

Oil Systems” IEEE Electrical Insulation B Tech in Electrical Engineering from Kerala
magazine January/February 1999-Vol 15 No.1 university in 1989

Present organization : TELK Angamally

5.0 Lars E. Lundgaard, Walter Hansen, Dag
Linhjell, and Terence J. Painter “Aging of Oil- Designation: Senior Manager (EHV transformer
Design)
Impregnated Paper in Power Transformers”
IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 19, NO. Experience :
1, JANUARY 2004
About 20 years experience in GCB design and
Power Transformer design
6.0 HP MOSER / V Dahinden “ Transformer
Board II” published by M/s Weidmann

7.0 Thomas A. Prevost , Weidman “Degradation

of cellulose insulation in liquid filled power
transformer” 2005 Fourth Annual Technical
Conference

8.0 T.V Oommen, Stan Lindgren ”Bubble

Evolution from Transformer Over load”
Transmision and Distribution conference 2001
IEEE/PAS

Authors

Name : Oommen P.Joshua

Qualification :

Degree in Electrical Engineering from Kerala

university in 1979 and M Tech from Cochin
university in 1985

Present organization : TELK Angamally

Designation: General Manager (Technical)

Experience :

More than 30 years experience in EHV Power

Transformer Design and testing

10