Vous êtes sur la page 1sur 353


(TRAFOSEM - 2008)
" Modernization & Upgradation of
Transformer Technology on Global Horizon
and Challenges for Indian Power Sector"
Organized by :
Indian Transformer Manufacturers Association (ITMA)
Central Electricity
Bureau of Energy
Central Power
Research Institute
,II!iJ Power Finance
<-v Corporation Ltd.
CREe Rural Electrification
w. ....... ,..... Ltd.

NTPC Eloctrlc Supply
f l.
Power Grid Corporation
Compeny Llmltd of India Limited
ff tI z1 IIIf
National Power
Bhakra Beas
Training Institute
Management Board
r:l'!Ti1 PHD Chamber of The Associated Chambers
_ Commerce and Industry 'lA')' of Commerce & Industry of India
India Energy Forum
PHDCCI z ...
11th It 12th November 2008
Convention Centre, India Habitat Centre, Gate No.3, Lodhi Road, New Delhi - 110003.
A. s.
Power is one of the most essential infrastructural requirements
for the overall development of the country's economy. Providing
adequate and affordable electric power of acceptable reliability
and quality at reasonable prices to all sectors of the economy is
essential for sustainable economic development, human welfare
and higher standard of living.
Indian economy is at a crucial juncture of its evolvement wherein
the reform process initiated in the early nineties has begun to
fructify. Therefore, making available power on demand to all
is one of the top most priorities of the Government. Making
power available to the consumers, be it industries, commercial
or domestic consumers or farmers for their use, involves an
overall development of the entire chain of the power system
network from power generation to its transmission and ultimately
its distribution to the point of consumption in the most optimum
and efficient manner.
Since India has taken rapid strides in the
development of the Power Sector both in terms of enhancing
power generation and in making available power to widely
distributed geographical boundaries. At the beginning of the
10,h Plan (April 2002) 80% of villages were electrified whereas
in June 2008 82.3% of villages have been electrified. Still a
large number of people in our country have no access to
electricity and a large number of villages are still out of reach
of electrical power.
Power Sector was perceived to be riddled with some
fundamental weaknesses which necessitated the need to initiate
the reform process in the Sector. Major weaknesses identified
were as follows:
Low priority accorded to long term viability of SEBs and
therefore these utilities facing huge losses
Mounting energy and peak power shortages
High AT &C losses
Skewed tariffs and therefore uncovered subsidies
State Power Utilities unable to raise resources for
Rural Electrification very slow
With a view to tackle the above problems, need was felt to
introduce innovative changes in the Power Sector and therefore
the reform process was initiated during the early nineties.
During 1990's the power sector reforms were initiated and the
Liberalized Policy was announced encouraging privatization
of generation. The States were asked to unbundle their SEBs
into small manageable and efficient corporations which would
attract private investment to meet the power sector nee<Js. To
help the State Governments in the reform process, the Union
Poliry nnirl".linf"<: from time to time. The
Electricity Regulatory Commission Act was enacted in 1998
for setting up of Regulatory Commissions at the Centre and in
the States. Electricity Conservation Act 200 I was enacted to
promote Energy Efficiency in consumption and Demand Side
Management measures. The enactment of the Electricity Act
in June 2003 was a major milestone which paved the way for
development of the Power Sector within a competitive and
liberal framework while protecting the interests of the consumers
as well as creating a conducive environment for attracting
investments in the sector. The National Electricity Policy and
the National Tariff Policy were finalized by the Government
to provide direction to evolution of the Power.Sector within the
ambit of the Act. The Integrated Energy Policy was announced
by the Planning Commission in 2006.
National Electricity Policy aims at laying gt.:idelines for
accelerated development of the Power Secto:-. The Policy shall
provide direction to the evolution of the Power Sector within
the ambit of the Electricity Act, with per capita power to be
increased to over 1000 units by 2012 & Power to all by 2011-
These policy and. legislative level changes will significantly
impact the logistical decisions, planning and'pricing strategies
of the power producers, fuel suppliers, transmission &
distribution companies and the ultimate consumers. However
even though these legislative changes have attempted to address
and streamline most of the issues involved, some trepidation is
being felt on the transition of the energy sector from a regulated
to a deregulated environment. It is vital to capture the essence
of the transition process. The transition period could be turbulent
and it cO'lld take some time to translate the provisions of the
Act into ground realities and remove policy disconnects.
Focused and united effort on the part of the Government and
all the stake holders would be required to minimize the transition
period as also the uncertainties during this time, in order to
give the required boost to the power sector on the path of fast
Our country has significant potential for diversified resources
of power generation. There is a need to widen the energy
resource base for electricity generation and have an optimal
mix of different energy sources ie., Coal, lignite, gas, LNG,
Naphtha, Liquid fuel, Hydro and Nuclear. Renewable sources
of energy such as Wind, Small Hydro, Bio mass and Solar
Energy have the duel advantage of being economical at remote
locations as well as nonpolluting and therefore need to be
encouraged to the maximum extent possible. New emerging
fuels and technologies also need to be developed towards
stainable development of power. Limited availability offossil
els like coal and gas has further highlighted the importance
power from these sources to supplement power from
mventional Sources.
)al is expected to remain as the main fuel for power
neration in the foreseeable future. Therefore, it is crucial
It policies are driven by a forward technology vision to create
obust and managerially strong coal industry as also to ensure
:1cient extraction of domestic coal with strong stress on cost
tting and quality improvements. Also important is to promote
Jre efficient and clean coal based power generation
:hnologies. Our country's coal reserves are estimated to be
out 240 Billion Tonnes within the depth range of 0-1200
:ter for seams 0.9 meters and above in thickness. Of this the
wen category constitutes about 90 BT (37 %). The total coal
)duction in the country during 2007-08 was 424.47 MT and
o MT in 2011-12. Out of this about 70% coal is allocated to
wer sector.
~ n i t is available at limited locations in India such as Neyveli
Tamil Nadu, Surat, Acrimota in Gujarat and Barsingsar,
lana, Bithnok in Rajasthan. Over 88% of the resources are
:ated in the State of Tamil Nadu alone, whereas the rest 12%'
; distributed in various other States. The geological reserves
lignite have been estimated to be 30 Billion Tonnes. Since
snite is prone to catch fire if transported over long distances,
use for power generation at pithead stations is therefore found
be attractive.
e estimated reserves of natural gas in the country are about
2 Billion cubic metres (Be urn). Natural gas from Mumbai
gh Offshore and Gujarat oill gas fields is transported by GAIL
ough Hazira-Bijapur-Jagdispur pipe line for use primarily in
tilizer and power plants in combined cycle modes in Western
j Northern Regions. Natural gas from Krishna - Godavari
tds and Cauvery fields meets the local requirement for use as
uel in combined cycle plants in A.P. & T.N. Natural gas in
sam & Tripura is used in power plants in these states.
ass production of natural gas, which was almost negligible
he time of independence, was about 30 Bcum during 2001-
The Government has opened up the sector (related to gas
110ration and exploitation) to private participation, and
liance Industries have discovered 7 Trillion Cubic feet (TCF)
gas, the production planning of which is in progress.
clear energy is an inevitable option in the long term to meet
projected electricity demand through a diversified resource
,e and the need for environmentally sustainable development
h energy security. India is endowed with limited uranium
ources. However the thorium resources, amounting to nearly
: third of entire world's thorium resources, are found in India.
mium is the only naturally occurring fissile element that can
used in a nuclear reactor to produce energy through fission.
lrium is a fertile element, which can be converted into fissile
lerial i.e. uraniunl233, in the reactor. The strategy for large-
Ie deployment of nuclear energy in the long term is focused
towards utilizing thorium. Judicious utilization of limited
indigenous uranium to ultimately tapping the vast thorium
resources has been one of the key objectives of the programme.
Therefore, as a long-term strategy, the nuclear power programme
has been based on three-stages linking the fuel cycles of these
Our country has vast potential of hydropower totaling to about
84,000 MW at 60% LF (about 150 000 MW installed
capacity) with associated energy of600 TWh. Out ofthis about
17.6 % has been developed and about 6.4 % is under
development. 2.7 % of the hydro potential has been cleared by
CEA for further development. Thus about 75 % of the hydro
potential is yet to be harnessed.
The total installed capacity at the beginning of the 10lh Plan i.e.
1.4.2002 was 1,05,046 MW comprising 26,269 MW hydro,
74,429 MW thermal (including gas and diesel), 2,720 MW
nuclear and 1628 MW wind-based power plants. In addition to
above, 1,697 MW of capacity from remaining Renewable
Energy Sources was connected to the Grid.
Figure No.1
1697 MW
I Total - 1,05,046 MW I
74,429 MW
The power supply position in the country at the beginning of
1 DIll Plan period was of shortages both in tenns of demand met
during peak time and overall energy supply. On an All India
basis, the peaking shortage was 12.6% and the energy shortage
For the 10
Plan, against a need based capacity addition of
57,000 MW, a target of 41,110 MW comprising 14,393 MW
hydro, 25,417 MW thermal (including Coal Lignite & Gas)
and 1,300 MW nuclear was fixed based on preparedness of
projects and availability of resources. The sector wise, type wise
details are indicated below:
Figure No.2
SECTOR, 41,110 MW
The target set for capacity addition during the 10th Plan was
41,110 MW. Even though stringent monitoring of projects was
done, the capacity addition achieved during 10th Plan was
21,180 MW.
Projects totalling to a capacity of 21,281 MW, comprising
14,554 MW thermal, 6,507 MW hydro and 220 MW Nuclear,
included in the 10th Plan target of 41,llO MW have either
slipped or have been dropped .. However, due to consistent
efforts, certain additional projects not included in the 10th Plan
target were identified for benefit during lOth Plan by expediting
the process of project implementation and compression of the
construction schedules. These additional projects totalling to
1,351 MW,comprising 1,251 MWthermaland 100 MWnuclear
power projects, were considered to offset to some extent the
shortfall in capacity addition due to non-materializing of targeted
projects during 10th Plan. .
With the addition of new projects and the projects slipped!
dropped from the 10th Plan, the capacity addition during 10th
Plan was 21,180 MW.
Sector wise and type wise details are given in the Table-l below:
Ser.tor Hydro Thermal Nuclear Total %
Central 4495 7330 1180 13005 56.9
State 2691 3553.6 0 6244.6 55.9
Private 700 1230.6 0 1930.6 27.1
Total 7886 12114.2 1180 21180.2 51.6
% 54.8 47.6 90.8 51.6
From the above, it may be noted that the capacity addition in
the lOth Plan is 51.6% of the target. The contribution of the
Central and the State Sector has been about 56% whereas the
Private Sector contribution has been as low as 27%. The main
reasons for slippage of projects during 10th Plan were delay in

) Q
placement of Main Plant and Balance of Plant (BoP) equipment,
delay in award of works and investment decision, E&F
clearances, lack of preparedness of projects, delay in transfer
of technology arrangements by BHEL & TEC/PIB clearance of
Hydro projects.
The total Installed Capacity at the end of 10th Plan i.e. as on
31.03.2007 was 1,32,330 MW comprising 34,654 MW hydro,
86,015 MW thermal (including Coal ,Lignite Gas & Diesel)
3,900 MW nuclear based power plants and 7,761 MW from
renewable energy sources including wind.
Figure - 3
(As On 31.03.2007)
7,761 MW

3,900 M

86,015 MW

At the end of 10th Plan, the country faced power shortages in
terms of Peaking power (MW) and Energy availability in (M Us)
The details of shortages are given in Table -2 as follows:
Peak (MW) Energy (MU)
Requirement 6,93,057
Availability 87,105 6,24,716
(-)Shortage/(+) Surplus 13,610 68,341
(%) (-) 13.5 (-) 9.9
The Government's Mission is to provide reliable and quality
power to all at reasonable rates by 2012. It has been assessed
that the peak demand by 2012 is likely to be about 1,52,000
MW and the energy requirement is likely to be about 1038 BU
by 2012. A realistic assessment of feasible projects for the II th
Plan has been done to meet the above demand and a well planned
Strategy adopted while finalizing the list of II 'h Plan projects.
Targetfor 11th Plan
A capacity addition target of 78,577 MW was set for the II th
Plan as per the Eleventh Plan document prepared by Planning

. --t)

Commission. The break. up of this capacity is given in
Table -3 as follows:
Table -3
CENTRAL 9685 26800 3380 39865
STATE 3605 24347 0
_ . .,,--
PRIVATE 3263 7497 0 10760
TOTAL 16553 58644 3380 78577
The Planning Commission has subsequently agreed to some
changes in the hydro/thennal capacity and therefore a revised
capacity addition of 78,700 MW is proposed for the Eleventh
plan Period.
Feasible Capacity Addition and Status during Ilth Plan
A capacity of 16,335 MW was proposed to be commissioned
during 2007-08 which was reviewed and was fixed at 12,039
MW. The actual capacity addition during the Annual Plan 2007-
08 was 9,263 MW against the revised target of 12,039 MW as
illustrated below:
Total Prg.: 12039MW
Achiev .: 9263MW
Thermal Nuclear
. :. Programme. Achievement:
Out of the II th Plan revised target of 78,700 MW it has been
assessed that a capacity of 4280 MW is likely to slip whereas
an additional capacity of about 4594 MW has been ordered for
commissioning during the 11 th Plan. Therefore as per our
latest estimates, a capacity of 79,190 MW is expected to
materialize during the 11th Plan as detailed below. In
addition a capacity of about 4210 MW is expected in the
11 th Plan on Best Efforts basis.
Figure -5
26563 MW
Additional capacity
34190 MW
: New Renewables - 14,000 MW
: Captive - 12,000 MW
About 11,333 MW has already been commissioned during
2007-08 and 2008-09 ( till date) comprising of 2,673 MW hydro,
8,440 MW thermal and 220 MW nuclear power plants. Further
a capacity of 67,857 MW is under construction.
Fuel Requirement by End of lldr Plan (2011-12)
The likely fuel requirement for 11'" Plan capacity addition of
about 80,000 MW during II'" Plan has been estimated and the
requirement of coal is about 550 MT. The total coal
availability from domestic sources is expected to be about 482
MT per annum by 2011-12. Accordingly, imported coal of the
order of 40 MT, equivalent to 68 MT of Indian coal, may have
to be imported. This quantity may reduce provided production
of domestic coal is increased. The requirement of gas is
expected to be 89 MMSCMD of gas requirement at 90% PLF
in 2011-12. At present, the availability of gas is of the order of
36 MMSCMD and therefore is not sufficient to meet the
requirement of even existing plants.
Table 5
Fuel Requirement (2011-12)
Domestic Coal* 550 MT
Lignite 33 MT
GaslLNG ** 89 MMSCMD
Other Measures during 11th Plan
In order to bridge the gap between demand & supply of power,
specially in the context oflimited financial resources available,
it has become imperative to look for other options which are
not as capital intensive as new capacity addition and which could
be implemented in a comparatively shorter time frame. In this
regard optimum utilization of existing installed capacity in the
country to maximize generation through Renovation &
modernization (R&M) of existing power plants had been

r C
. --l)
considered to be most cost effective option. Besides, steps are
being taken to increase availability by reduction in T &D losses,
Efficiency & Demand Side Management, Captive plants,
renewable energy sources and Technology Up gradation
Through Research and Development.
Optimum utilization of captive power plant (CPP) capacity,
existing as well as planned, is an important issue which needs
careful planning and detailed deliberations. A large number of
captive plants including co-generation power plants ot varied
type and sizes exist in the country which are either utilized in
the process industry or used for in house power consumption.
The Installed capacity ofCPPs has increased from 588 MW in
1950 to about 22,335 MW at present generating about 82 billion
units of electricity. More than 5200 MU of surplus power is
being exported to the grid. It has also been estimated that captive
plants with an estimated capacity of about 12,000 MW would
be added during the II th Plan period. Captive plants therefore
playa supplementary role, though an extremely useful one, in
meeting the country's power demand.
The vibrant economy of India is largely dependant on the
availability of quality supply of electricity to consumers at
competitive rates. Numerous challenges are likely to be faced
oy the Power Sector in the new millennium. The Electricity Act
2003, National Electricity Policy and the National Tariff Policy
have created an open, competitive environment for the growth
of the Sector. However, having created a conducive
environment, the main task before us is to realize and
implement the goals set before us and to translate the Plans
into reality on a priority basis.
ABSTRACT: This paper discusses the issues and concerns as
regards use of HTLS conductors for overhead lines, distributed
power generation, reliabilit and power quality as they pose
challenges in Indian power sector. The constraints and problems
are identified and some solutions are suggested.
ACSR has been used as conductor material for overhead
transmission lines for the past many decades. The increase in
power demand and power sector deregulation has placed new
demands on power transfer and delivery systems. Everywhere
the rate of addition of new transmission lines in much less than
the rate of increase of electricity demand. Therefore power
system engineers have to constantly seek new innovations to
increase pOWer transfer capacity over the existing lines. One of
these innovations is the development and use of HTLS (high
temperature low sag) conductors for overhead transmission lines.
These conductors can be operated a temperature of around 2000C
thus doubling the power transfer capacity. Some of these
conductors are:
Aluminium conductor steel supported
Aluminium strands are formed to produce
trapezoidal wave thus increasing the
aluminium cross sectional area for the same
Special zirconium alloy aluminium
conductor invar steel reinforced.
Gapped TAL alloy Aluminium conductor
steel reinforced.
Aluminium conductor composite reinforced.
Aluminium conductor strands over a low
thermal elongation polymer matrix
composite core.
The t;Se of these conductors for reconductoring of existing lines
as well as for construction of new lines calls for special attention
and more so for reconductoring of existing lines. Some of these
considerations are:
I. The present day load flow studies, stability studies are done
by assuming constant resistance of conductor. However
the change in resistance from about 300 C to 2000 C is very
appreciable. This will affect the factors like accuracy, rate
of convergence etc.
2. The resistance of conductor is, many times assumed to be
negligible in short circuit studies. When HTLS conductors
are used this assumption may not be valid.
3. Significant increase of active power results in significant
iii';;';-';;".>';; ~ ."",,.;.'" !'VWCi ai:.u. Tilis will call for
strengthening of shunt compensation equipment at the
receiving end.
4. When a line with HTLS conductors is operating in parallel
with other lines, the loss of this line would result in extra
congestion on the other lines.
5. Planning for use of HTLS conductors has to be done by
keeping the design parameters, facilities for stringing of
these conductors and other required changes in view.
Distributed power generation where in electricity is generated
at consumer end therebi avoiding the transmission and
distribution costs would playa very important role in electric
power systems in future. There would be a significant increase
in distributed gel!eration connected to regional grids in the next
The controls, communications and power electronics have
enabled small size DO systems to provide many services at a
low cost. In addition to low cost, the consumers benefit from
lower frequency and duration of outages and better power quality.
The advantages of DO can -be enumerated as under:
a) Utility perspective: Transmission capacity relief,
Distribution capacity relief, Lower costs, Easier and
more effective management.
b) Consumer perspective: Improved reliability, Better
power quality, Lower costs, Incentives from utility for
providing capacity reserve.
c) Commercial Power Producer perspective: Increased
business opportunities in power generation, New
business opportunities to provide services like reactive
power support, reserve capacity etc.
Applications: Voltage support, Frequency responsive hot and
cold reserve, Power quality correction (surges, harmonies etc.),
Reactive power management, Unit Commitment and dispatch,
Emergency control, Elimination oflocal stability problems, Local
congestion management, Peak shaving.
The development of Power Electronics interfaces has made the
conventional sources non-competitive. Solar energy, Wind
energy, Biomass and Fuel cells are likely to form the bulk of
future DO systems. Typical size will be in the range of I - 5
IEEE std 1547 (2003) titled 'Standard for interconnecting
Distributed Resources with Electric Power Systems' provides
the functional technical requiremento required to ensure proper
installation and operation of the interconnection. This standard
has 4 parts.
The inter connection issues have not been addressed in India as
yet. It is high time some standards as suited to Indian grid are
Indian electricity act 2003 encourages generation.
As per the provisions of this act generation has been liberalized.
In rural areas one can set up a stand-alone system for generation
and distribution. This can also be done in urban areas after getting
clearance from SERe. Every power producer shall have the
right to open access to the grid provided capacity is available
and wheeling charges are paid. It is expected that the above
provisions will encourage the entrepreneurs in our country to
set up distributed generation systems.
The output voltage of a DG systems may be dc ( in case of
photovoltaic syStems and fuel cells) or ac ( in case of wind energy
systems). However it is mostly a variable voltage and has to'be
made compatible with grid voltage. Therefore a power electronic,
interface is always needed between a DG system and grid.
The importance of reliability of electric supply can not be over
emphftsized. Inspite of tremendous growth of power sector, the
reliability continues to be a mater of great concern. Many critical
loads like digital systems, power electronic systems, variable
speed drives etc., require a very high reliability level.
Reliability of electric supply can be seen from qualitative as
well as quantitative aspect. Qualitative aspects which affect
reliability are:
1. Power generation, transmission and distribution facilities
and their expansion to take into account increase in demand.
2. Finance
3. Fuels
4. Management and administration
5. Legislative and regulative affairs.
All the above aspects are interlinked in more than one way. Two
distinctive quantitative reliability assessments methodologies are:
I. . Collection and analysis of system outage data.
2. Predictive reliability assessment i.e combining historical
data and theoretical models to estimate the reliability of
future systems or future expansion plans.
Generation, transmission and distribution systems in advanced
countries are designed and operated for certain predetermined
reliability indices. Commonly used indices are:
ASAI i.e. average service availability index
= Total customer hours for which supply is available
Total customer hours
CAIDI i.e. customer average interruption duration index. This
index gives the average duration of one interruption.
SAIDI i.e. System average interruption duration index.
SAIFI i.e. system averagejoterruption frequency
The figures for the above indices for a typical power system in
USA for the year 2005 are as under:
146.60 minutes
31.51 minutes
0.21 interruptions per customer
Some efforts have been made in our country for collection of
outage data. However the authenticiiy of Gala :;uppiicu uy
different utilities is still to be verified. Moreover this data has
. .
not been use forany systematic reliability studies.
The unreliability of electric supply has compelled a very large
number of domestic, commercial and industrial consumers to
make alternative arrangements for emergency purpose. A very
significant number of domestic consumers use inverters. The
use of an inverter means a dual conversion of energy with an
overall efficiency of above 75%. Even ifit is assumed that only
about 10% of domestic consumers use inverters for their
minimum emergency requirement i.e. one tube light, one TV
and one fan, the ,lose of energy on this account comes to as high
as 108 kWh per year for the whole country.
The commercial and industrial consumers use DG sets which
use the Petroleum products and the loss on this account is still
It was expected that with the unbundling of power sector the
situation as regards reliability would improve. However the
situation remains the same. It is necessary that government
agencies, public sector undertakings, research organizations and
industries use cognitive skills and analytical processes to
adequately face the reliability problem. Comprehensive planning
and engineering excellence are necessary. Piecemeal solutions
are not enough.
Customers demand both reliable and high quality electric service.
Reliability refers to continuity of electric supply. Power quality
describes characteristics like momentary interruptions, voltage
sags, voltage surges, harmonic distortion. Modern equipments
have electronic and Microprocessor components which are very
sensitive to power quality. Power electronic equipment is used
to convert 50 Hz ac into variable frequency ac, variable voltage
ac and variable voltage dc. The current consists of fundamental
component and higher order frequencies which are integral
multiples of 50 Hz fundamental frequency.
Many standards have been proposed by different organizations
as regards the maximum permissible magnitude of harmonic
currents. Some of these standards are:
l. IEC Norm 555 - 3 prepared by International Electrical
2. West German Standard VDE0838 for household appliances,
VDE0712 for fluorescent lamp ballasts and VDEOl60 for
3. EN5906 (European standard) The limitation of disturbances
in electric supply networks caused by domestic and similar
appliances equipped with electronic devices.
4. IEEE Guide 'Hannunic control and reactive compensation
of static power converters'. ANSI I IEEE standard 519
of 1992.
Hannonics can be reduced substantially by using a shunt active
filter. Shunt active filter is connected to ac mains through
coupling reactor .It removes current hannonics by injecting equal
but opposite harmonic current. Therefore, hannonic component
ofload current is cancelled and the-current drain from the mains .
becomes sinusoidal with unity power factor. The combination
of non-linear load and active filter becomes an ideal resistor as
seen from the source. The non-linear load draws load current
having rectangular wave shape. The active filter injects a current
whose wave shape is so as to make ovemll wave shape sinusoidal.
Shunt active filters may be single phase or three phase. They
may also be connected in cascade for high power non-linear
loads. It requires a proper control scheme to calculate the current
reference waveform for each phase, maintain de voltage constant
and generate inverter gate signals. The reference current
compensates the load current harmonics.
1. HTLS conductors can be opemted at about 2000c. Thus
the power transfer capability of a line having HTLS
conductors is double that of a line with ACR conductor.
However the use of HTLS conductor requires detailed
system study.
2. Distributed power generation has a number of sped
advantages. All steps should be taken to encourage th
form of generation.
3. Inspite of power sector reforms in many states, reliabilil
or electric supply continues to be poor. Comprehensio
planning and analysis is required to improve reliability.
4. Modem equipments using electronic components an
microprocessors required high quality supply. Harrnoni
content can be reduced by using snUllI active filler.
1. B.R.Gupta, 'Power System Analysis and Design', S.Chan
& Co., New Delhi, 2008.
2. B.R.Gupta, 'Genemtion of Electrical Energy', S.Chand
Co., New Delhi, 2008.
3. EPR! ,'HTLS Tmnsmission conductor', report 1001811 Jun
4. GlGRE WGB 2.12, 'Conductors for Uprating 0
Tmnsmission lines', Tech. Brochure 244, 2004.
5. Vandana Singhal & B.R.Gupta, 'Distributed Powe
Generation - A Leap into the Future', Internationa
Conference on 'Green Power 5: Development an,
Management of Resources and Energy Security' Centra
Power Utilities, 2 - 3 February 2006.
6. EPR! 'Integration of distributed resources in electric powe
systems: current interconnection practice and unifiec
approach'. Tech. Ref. TR-11489, 1998 .
. --
Anil Kr. Aggarwal
Availability of power is one of the key components of economic
development. Transformer is the major and indispensable
machine for generation, transmission and distribution of
electricity in the country. We have a strong base over 500
transformer industries in India with an overall manufacturing
capacity of more than 800,000 MVA per annum. The
contribution towards domestic utilities is over 500,000 MVA,
the rest are exporting to more than 50 countries.
Today our industries are fully geared up with proven technology
and with . enough capacity to undertake manufacturing wide
range of power and distribution Transformers and other special
transformers, like welding, instrument, traction, mining, furnace
etc. About 95% of the transformers installed in the Indian power
network are of indigenous origin. Energy efficient transformers
with low loss and low noise level are also made to suite
international requirement. Besides, the transformer industries
in India are well versed and matured enough to supply all types
of transformers and can meet the country's demand for
transformers of all voltage levels including 800 KV.
Some of the transformer manufacturers in India have
successfully developed in-house technology and cost
effectiveness in their design and manufacture of transformers.
With the help of strong R&D facilities in the organization and
infrastructure available at national level, the transformer
industries are continuously upgrading the product performance.
With the progressive and pragmatic policies ofthe government
of India, the transformer industries are on fast track from the
past couple of years, having impressive growth rate of 19%
during 2007-08 and a cumulative average growth rate of 9%
over past 5 years. The continued high growth of power sector is
vital for GDP, which is expected to touch between 8% to 9%.
The demand for smaller rating distribution transformers,
especially 10 KVA to 100 KVA have grown in a very big way,
particularly due to the thrust given by central and state
government for rural electrification under RGGVYN schemes.
However transformers ranging from 250 KVA to 1000 KVA
have not registered upward trend, but maintained the normal
In respect of power transformers, there is very good prospect
for transformers of 400 KV level while 220 KV segments
maintains its normal growth. So far 800 KV level transformers
are concerned, the demand has started peaking up due to capacity
addition of 100,000 MW generation in 10th and lIth plan
periods in aggressive manner with the announcement of NEP
and NTPC for attracting investors both in National and
international market. The transformers at UHV level shall
assume priority due to upcoming national grid and interlinking
of trunk lines for inter - transfer of bulk power from surplus to
deficit states.
All SEBS/ Utilities/ Discoms/ Transcos under govemment
initiative have taken up improvement <;>f transmission and
distribution system in a big way with funds being provided by
PEC, REC organizations. PSU, like PGCIL, NTPC have g()ne
in for major expansion plans. All these have shown increased
business potential for the industry. By and large the order book
position for all manufacturers are fairly healthy and margins
have improved.
Transformer industry and power sector of the country are closel y
associated and in a way are inter dependent as electricity cannot
reach the end users without inserting transformers en-route for
transformation of various voltage levels. The major customers
are SEBS/ Utilities/ Discoms/ Transcos and heavy and small
industries of the country.
The current installed generating capacity is around 132,000 MW
as against a targeted generation of 210,000 MW at the end of
11 th plan prior, which means, we are required to generate
another 78,000 MW by the end of2012. Besides, the Ministry
of power has undertaken to complete 7 Ultra Mega Power
projects of 4000 MW each at a cost over Rs. 100,000 Crores in
the next 3 to 4 years in the state of Madhya Pradesh,
Chhattisgarh, Gujarat, Maharashtra, Karnataka, Orissa etc.
With such ambitious programmes of government of India for
accelerating the growing economy and for providing electricity
to all, and electrifying remaining 100,000 villages by 2012,
the future of electrical manufacturing industries, suppliers
of transformers and accessories, lamination and Stampings ,
Switch gear and Control gears, Insulating materials, Insulating
oils, Insulators, Windings wires, Conductors and Accessories,
Instruments, Measuring, Indicating and testing equipments,
Cables, Capacitors, Transmission line towers, Switch-yard, Sub-
station equipments and Software Services assume a lot of
importance in the power sector. Therefore, in all, over 78,000
mw power shall be added within coming 3 to 4 years.
As a matter of fact, the ratio of power generation to that of
transformation capacity is 1:8 i.e. We need transformers
equivalent to 78,000 x 8 = 624, 000 MW or 840, 000 MVA.
We therefore, need 840,000 MVA transformers to distribute a
power of 78,000 MW from the point of generation upto the
distribution centre. The following paragraphs will help us to
identify the approx. no. of power and distribution transformers
needed to distribute a power equivalent to 840, 000 MVA.
Let us assume that the distribution transformers take half of
the MVAand balance 50 % goes to power transformers account.
This assumption leads us to conclude that we need distribution
transformers equivalent to 420,000 MVA and power
transformers equivalent to also 420, 000 MVA.
Most of the transformers being used in rural areas are ranging
from IO KVA to 100 KVA. while in urban areas, it is between
100 KVA of 400 KVA. Metro cities and industrial units consume
transformers ranging from 630 KVA to 2000 KVA or above.
63 K VA would have been the ideal average rating of distribution
Therefore, no. of transformers required for coming 4 years
_ f A .... I\ IV'" ... '1'\" I 'l
= 66,66,666 nos.
and per year requirements = (66,66,666 I 4 ) = 16, 66, 666 nos.
Similarly, the requirements of no. of power transformers per
year can be calculated.
With the induction of 800 KV system, the country need to
add 50 to 60 units of 800 KV.
1000 units of 400 KV, 2000 units of 220 KV and thousands of
units of capacity varying from 10 MVA to 3l5'!v!.VA. These are
anticipated rough figures and indicate enormous scope of
transformers in the country.
I f we take 10 MVA is the average rating of power transformers,
the no. of Power transformers required = (420,000 110) = 42,000
unit:,. Therefore per year requirements of power transformers=
(42,000 14) = 10,500 units.
From the above, we may conclude that the fmure requirements
of transformers are enormous and the Indian manufacturers have
to gear themselves up bOth mentally and financially to this
daunting job for the prosperity of the country.
Increase of raw material cost, especially copper, aluminum, oil,
electrical steel, insulating materials etc. are very very on high
side and need to be stable at one point. Further, availability of
electrical steel in the international market in not healthy. CRGO
silicon steel is not manufactured in India and is being imported
at 2.5 times the prevailing cost due to its shortage all over the
world. The price of copper as indicated by major manufacturers,
like Sterlite industries (I) Ltd., Hindustan Copper Ltd, Birla
Copper Lto, has gone by over 100 %.
In today's globally liberalized economy, Indian industries are
expected to be internationally competitive in terms of price and
quality and have to meet challenges posed by China and South
Asian Countries.
The prevalent policies of Central and state government have to
be suitably modified with incentives for infusing efficiency and
creating level playing fields in order to evolve a common market
for meeting national I international competition.
The foremost action for procurement of CRGO steel and
reduction in cost of copper (by way oflevies on both items) has
to be initiated by the government of India. This is extremely
necessary to succeed in providing sufficient, reliable and quality
power at affordable cost besides ensuring commercial viability
of power industries by the end of 11th Five Year Plan Period.
As already estimated, aH India addition (production of
transformers) of MVA capacity upto 1212 expected to be in the
order of 840,000 MVA.
This capacity is required to be added by installation of millions
of transformers in the system for power sector reforms,
launched particularly under APDRPI AREPIRGGVYN schemes.
It will omy be successful if the transformers industries are
facilitated to over come the problems being faced by them. In
fact the matter has already been brought to the notice of the
Ministry of Power, the Ministry of steel and the Ministry of
Commerce through ITMA forum, but little has been done.
Let me add a few more paragraphs on 'inflation', 'Raw material
prices' and 'Electrical industries' .
The Indian Electrical Industries, during the past two to three
years, have been facing a daunting challenge on account of the
global phenomena of rising prices of raw materials like ferrous
and non ferrous metals, oil and others. Un-precedenting inflation
has added further, fuel to the fire with its impact on prices and
On looking back, one realize just how much the electrical
industry has been hit by price increases. During the period April
2005 to April 2008, the price of copper has increased by 105%,
aluminum by 37% and steel by 48%. And finally the price of
crude oil, which impacts the prices of transformer oil base stock,
furnace oils, plastics etc. which was about 30 to 40 dollars a
barrel a couple of years back, is now hovering around 100
dollars a barrel.
As is universally known, the electrical industry is highly sensitive
to increase in the prices of raw materials, since the cost of raw
materials constitutes more than 70% of the prices of finished
products. Any erratic behavior in the price of raw materials
therefore puts tremendous pressure not only on the prices of
the finished products, but also on deliveries since suppliers are
unable to absorb the entire price rise and seek intervention of
the buyers in remedial measures.
As price rise is a global phenomena, nothing much can be done
to control them. However, completion of projects on time in
equally important for the growth of our country. The best option
available for both the suppliers and the buyers is to ensure that
the projects are completed on time to avoid additional cost
overruns and more importantly, delayed use of the infrastructure
To avoid such delay, it would be prudent for suppliers and buyers
to address the problems together and arrive at reasonable and
realistic increase in the prices of the finished products agreeing
to absorb some part of the increase in input prices.
And to avoid such situation of conflicts, disputes and reSUlting
delays, the buyers may consider stipulating the use of price
variation clause while awarding contracts.
Many of the manufacturers are exporting transformers to various
part of the world, including even to western countries and some
parts of USA. The price realization has improved. though not
upto the desired level.
The input cost of basic raw materials as well as accessories has
gone up substantially, but without commensurate increase in
market price, resulting in reduction of profit margins. It is
incumbent upon the industry to invest more in manufacturing
technology in order become a world class exporters of the quality
However, aesthetic look of Indian products in found to be
inferior in comparison to global standards. CPRI certification
for Indian products are not accepted in many countries. This
requires Indian manufactures to conform to the specifications
of other agencies like CESI (Italy) or KEMA (Netherland) etc.
The export and import policies of the government of India on
transformer and various other raw materials and components
need to the reviewed further for the benefits of Indian
transformer industry. The export market has been doing better
than the import for last 2 years. The future looks brighter
provided the policies are reviewed favorably and availability
of electrical steel is ensured.
Barring a few transformers manufacturers, the average quality
ofIndian distribution transformers catering to home market are
far below vis-a-vis international standard. This is mainly due to
lack of adequate levels of automation. Aesthetics of Indian
products is found to be wanting in comparison to the global
A paragraph may be added here on 'BEE Rating Distribution
TransfQrmers'. BEE has come out three categories of
transformers as 3- STAR, 4- STAR and 5-STAR. The gradations
have been done in respect of total losses. The specification has
called for maximum total loss at 100% and at 50% load. The
manufactures are required to design the component losses (i.e.
no- load and load loss) with such numericals so that the total
losses at 50% load as well as at 100% load should not more
than the guaranteed values mentioned in the following tables.
Table 1.
KVI customer -HSEB Customer- MSEB
Max. total Max. total Max. total Max. total
loss at 50% loss at 100% loss at 50% loss at 100%
loading loading loading loading
(watt) (watt) (watt) (watt)
1611113 135 440 150 480
25/11/3 190 635 210 695
63/11/3 340 1140 380 1250
10011113 475 1650 520 1800
160111/3 - - 720 2200
200/1113 - - 890 2700
The following tables may be consulted to arrive at the
component losses for CRGO steel transformers.
Table 2.
KVI customer-HSEB customer- MSEB
No Load Load Loss Max. Total Max. total
loss (watt) loss 50% loss at 100010
(watt) loading loading
(watt) (watt)
16/1113 60 300 135 360
25/1113 85 420 190 505
63/1113 155 700 340 895
100/11/3 220 1020 475 1240
Table 3
KVI Customer-HSEB Customer- MSEB
No Load Load Loss Max. Total Max. total
loss (watt) loss 50% loss at 100%
(watt) loading loading
(watt) (watt)
16/1113 60 360 150 420
25/11/3 85 500 210 585
63/1113 155 900 380 1055
100111/3 220 1200 520 1440
16011113 300 1680 720 1980
200111/3 350 2160 890 2510
The concept of star rating transformers is quite a new subject in
the Indian power Scenario. Manufacturers should get used to
develop such transformers and get them tested at any
government test house. The tests should include short-circuit,
impulse and loss verification tests.
With the application of more and more star rating transformers
in the distribution net work., the line losses will come down
The paper briefly discussed the targeted generation of power
by the end of 11 th Five Year Plan and the State of transformer
industries in India. The future power position of India looks to
be quite favorable for the Indian transformer industries
especially after the unclear agreement with us. It is high time
for all the power industries to go for capacity enhancement in
respect of quantity as well as <J.dding infrastructure to undertake
manufacture of EHV transformers.
The paper has briefly touched the requirement of losses for star
rating transformers. This subject is quite new to our industries.
We feel there is huge scope of discussion on this aspect. ITMA
should take leadership to use it's forum to organize many such
deliberations on star rating transformers .
. --
A. K. Saha
India is going through a crucial phase in her quest to increase
Power Generation and Distribution from the present capacity
of 260GW to 960GW by 2031-32. The investment will
approximatelv be 200 Kcrores. In this DIan RGGVY will invest
5(f,OOO crore and APDRP 40,000 crore for electrifying the
villages. Bureau of Energy Efficiency is working on the
modalities to make transformers more efficient by giving star
ratings depending on maximum losses at 50% and 100010 ratings.
The paper will discuss in details applicability of the methods
suggested from national perspectives.
Voltage kVA 3 Star
Ratio Rating
11000/433-250 V Max. Max.
Losses Losses
at 50% at 100%
(Watts) (Watts)
11000/433-250 V 16 150 480
11000/433-250 V 25 210 695
1l000/433-250 V 63 380 1250
11000/433-250 V 100 520 1800
11000/433-250 V 160 770 2200
11000/433-250 V 200 890 2700
In our persuit and quest for excellence we must find a way to
reduce energy loss in totality from the point of generation to
the end point of consumer terminal. Transformers playa
considerable part in energy loss, but they cannot be discussed
in isolation because distribution system losses are considered
significant ifnot higher.
As per Central Electricity Dt:partment the maximum allowable
losses at rated voltage and rated frequency permitted at 75C
for 111.433 KV transformers can be chosen by utilities from the
values of 3 star, 4 star or 5 star rating for transformers upto
ratings 200 kV A as indicated in Table I.
4 Star 5 Star
Max Max Max. Max.
Losses Losses Losses Losses
at 50% at 100% at 50% at 100%
(Watts) (Watts) (Watts) (Watts)
135 440 120 400
185 655 160 615
330 1170 280 1100
440 1700 360 1600
670 1950 570 1700
780 2400 670 2100
Table 1
The maximum loss of a transformer at any given rating is the
sum-total of no-load loss and 12R loss if other stray losses are
Therefore if transformer loss at 100% rating= a+b then
transformer loss at 50% rating= a+ b/4
where a = no-load loss
b = J2R loss for 100% rating at 75C
Voltage Ratio kVA 3 Star rated
From the table for a 16 kVA 3 star rated transformer a + b = 480
watts and a + b/4 = 150 watts Solving the equation a = 40 watts
and b = 440 watts
In case of 16 k VA 5 star rated transformer a + b = 400 watts and
a + b/4 = 120 watts Solving the equation a = 27 watts and b =
373 watts
Similarly the no load loss and, the load loss for 3 star and 5 star
transformers are obtained for other ratings as given in Table-2
5 Star rated
N L Loss in watts Load loss in watts N L Loss in watts Load loss in watts
11000/433-250 V 16 40
11000/433-250 V 25 48
11000/433-250 V 63 90
11000/433-250 V 100 93
11000/433-250 V 160 23
11000/433-250 V 200 283
It is evident that for a particular design of 16 kV A transformer
3 star, 4 star or 5 star rating may be achievable but for other
ratings a particular design will not be suitable. Hence different
designs for a particular rating are to be worked out to satisfy
the need for 3 star and 5 star ratings. The designs may have to
use different materials, hybrid design for core as well as windings
with varied construction and manufacturing process.
Even if we assume that the figures for 3 star or 5 star transformers
are achievable, the transformer of a particular rating will have
many variants to suit the losses. The idea can not be generalised
for RGGVY or APDRP transformers because specifications are
to be drawn with national perspective and use. For special need
a transformer may be designed to suit the special requirement
of a customer, but if the need is to satisfy the national
requirement, special attention of the authorities is needed to
draw the specifications.
Categorisation of a transformer under different load conditions
may give wrong signals to different customers mainly the
utilities. With so many variants in materials, design and
construction the interchangeability of a transformer is lost. A
particular rated transformer will not be generally suitable in a
cost effective way for a replacement if required.
Moreover there will be no standardisation of products because
in their quest to achieve star marks manufacturer shall not only
use different materials, different designs and different
manufacturing processes but also a combination of all those
that are available. It will not be possible for the utilities to
standardise the products and .distribute the power effectively
and economically.
With the application of total owning cost (sum-total of
capitalized cost of losses and initial landed cost) utilities are
trying to economise distribution by reducing energy losses
during the life of transformer. The no load loss and load loss
are the two main constituent parts of capitalisation. Initial landed
cost is the price a purchaser has to pay to get the materials.
While capitalized cost of no load loss is constant for life of
transformer with fixed anergy cost, capitalised cost of load loss
depends on energy cost, load factor and hours of operation.
Moreover to get a realistic value of capitalized cost, life of
transformer, running cost, load growth, outages as well as
failures are to be taken into account. The failure rate of
distribution transf6rmer in India is of the order of 10 - 14% and
cost offailure sometimes exceed the landed cost of transformer.
Added to this the depletion value of money, makes the matter
worse to achieve realistic value of capitalisation during its life.
With so may variants in the playing field the cost of capitalization
will loose its meaning unless the losses and the manufacturing
materials are standarised.
RGGVY scheme stipulates capitalization cost of Rs. 1.44 lacs
for each KW of Iron Loss and Rs. 0.64 lacs for each KW of
Load Loss with penal provision of2oo% for each KW of excess
loss without any benefit for the reduced losses measured during
t ~ t s APDRP specification however stipulates capitalization
cost or Rs. 0.98 lacs for each KW of Iron Loss and Rs. 0.20
lacs for each K W of load loss without any penalty provision for
exceeding losses. The whole exercise of virtual cost seems to
be futile unless the variants are realistically reduced.
The RE Projects under RGGVY scheme stipulates maximum
top oil temperature rise at 35C and maximum winding
temperature rise at 40C for 10, 16 & 25 k VA transformers. I t is
an established fact that the capacity of any electrical equipmE:mt
is determined when thermal equilibrium is achieved while
working at full capacity. Reduction of the thermal limits of
temperature rise will directly increase the rating of transformer
and hence the cost. Moreover we are not using the'materials to
their operating limits thereby adding costs and defeating the
t::.::;;;:;;; ;;: t:::;:!uct being economically efficient. Moreover If
we take into consideration that most of the materials needed for
manufacture of transformers are import oriented it will be a
huge national waste if we do not consider it now.
APDRP specification also stipulates II KV internal HV fuse
inside the oil cooled transformer. With each make and break of
II KV switch an arc wili develo'p and the oil will quench the
arc without any arcing schute or a separate oil tight enclosure.
With the load shedding and outages that are so rampant in our
system, it is doubtful whether oil quality can be maintained and
consequently the life of the transformer.
At present we loose 35 to 40% of power generated. In other
words 60 to 65% of power generated reaches the consumer
terminals. Transmission loss and. transformation loss are
therefore equally important to develop a healthy system. The
target is to reduce the losses to 32% during 2011-12. We
therefore have to proceed in small steps and determine its
compatibility with the system at each step for long term benefit.
The idea on which REC standardised the electrical, equipment
has not outlived its purpose. Simplifying the specification with
less number of variables and stringent application of the
specification will benefit us in the long run to achieve energy
Manufacture and utilization of Energy Efficient Transfonner is
the call of the day. A copper rich transformer is highly efficient
across a wide range while the iron rich transformer may be more
economical if the transformer is lightly loaded during service.
Therefore specifications are to be drawn keeping the national
purpose and the material objective in mind after due
consideration to the above mentioned factors. As the factors
are widely variant a compromise is to be worked out for each of
the following while drawing the specification of Energy Efficient
I. Core Material
2. Winding Material (primary & secondary)
3. No load loss
4. Load loss at 75C
5. Temperature rise oil/winding
With simplification of standards and reducing the number of
variables products will be simplified, interchangeable and
standardized. With a simplified governing specification it will
be easy to evaluate the cost and performance at any given time
under any given circumstances. It is the need of the hour and it
is time to act now.
Prasenjit Paul, K. Bheema Prakash, Somnath Chattopadhyay, Subrata Dutta
The inrush currents are high magnitude hannonic rich currents
generated when transfonner cores are driven into saturation at
the time of switching ull. T!.cy ilVUI.1U!!, U; .,!1V1"
usually of the order of milIi seconds.
Ever since the phenomenon of transient transfonner inrush
currents was first established by Fleming in 1892, there are
continuous efforts to analyze this phenomenon and find
solutions-till 1988, reduction of the same by installing pre-
insertion resistors, 1988 point on wave controlled switching
and later controlled switching taking into account the residual
An attempt is made in this paper to consolidate data available
in the literature on various aspects of inrush current phenomenon
of transfonners. The inrush current phenomenon is explained
along with its nature and with the effect of different cases of
switching instants.
The effect of inrush magnetizing currents are compared with
respect to short circuit currents.
The factors on which inrush currents are dependent are listed.
The shape, magnitude and duration of the inrush current depend
on several factors such as size of transfonner, impedance of
source, magnetic properties of materials, residual flux, instant
of switching in, the method of transfonner connection etc
Different fonnulae available for estimating peak inrush currents
are used and range of values are calculated for various
transformers. The difference values is attributed to the
assumptions made in the equations Proper care should be taken
for setting protection relays.
A case study of 25M VA 132 K V transfonner has been presented
and it was established that the frequent tripping during switching
was attributed to inrush magnetizing current and being very
nearer to incoming substation, the source impedance had vital
role in this phenomenon. The setting of the protective relay had
to be delayed to off set this problem.
There are various methods to mitigate problems caused by
magnetic inrush currents such as controlled switching,
providing, hannonic filters, analyzing residual magnetism, use
of resistors etc.
The shape, magnitude, and duration of the inrush current depend
on several factors such as size of transformer, impedance of
source, magnetic properties of materials, residual flux, instant
of switching in , the method of transformer connection Star,
delta, etc
Finally it is concluded that selection of suitable transformer
with optimum operating flux density depending on various
parameters indicated above is crucial to mitigate problems oj
undesirable tripping due to inrush magnetizing currents.
The paper contains the description of the aspects indicated above
along with Annexure 1
Indicating different limiting conditions for switching, typical
wave shape, problems related to protective relaying, factors
affecting inrush current
Annexure 2 contains various formulae and results of estimated
inrush current values
And annexure 3 contains case study of description of substation
along with transient wave shapes recorded on the transformer.
1.0 Introduction Magnetizing inrush current
Inrush current is a fonn of over current that occurs during
energization of a transfOrmer and is a large transient current
which is caused by part cycle saturation of the magnetic core of
the transformer. The inrush currents are high magnitude
harmonic rich currents generated when transformer cores are
driven into saturation at the time of switching on. They are
normally of short duration usually of the order of milli seconds.
Ever since the phenomenon of transient transformer inrush
currents was first established by Fleming in 1892,there are
continuous efforts to analyze this phenomenon (Ref: 1-27), and
find solutions-till 1988, reduction of the same by installing pre-
insertion resistors, 1988 point on wave controlled switching
and later controlled switching taking into account the residual
flux Transformers are essential elements which facilitate the
transmission of electric power at suitable voltages over long
distances and transfonner energisation is a common occurrence.
The magnetizing inrush phenomenon produces current input to
the primary winding, which has no equivalent on the secondary
side. When a power transformer is energized, there is an
important transient inrush of current that it is necessary in order
to establish the magnetic field of the transformer. Some
transformers exhibit peak current demand up to approximately
. ten times the nominal value. In addition, during the first few
cycles high values of the homopolar components are also
requested by the transfonner. The inrush current depends on
number of parameters including instant of switching on, type
and construction of transformer, the source impedance, distance
from input substation, protection systems employed etc. These
details are briefed below and a case study on a 132kV
transformer is presented
2.0 Explaining transformer inrush current
. Since it is very unlikey that transformer is switched on to AC
voltage at maximum value, it is almost always the case that iron
core will experience saturation. Because of magnetic
characteristics of iron core is highly non linear, saturation will
produce higher harmonics
Although usually considered as a result of energizing a
transformer, the magnetizing inrush may also be caused by
occurrence of external fault
Voltage recovery after clearing external fault
change of the character of the fault[ for example when phase
to ground fault evolves into a phase to phase to ground
fault and
out of phase synchronizing of connected generator
2.1 Different cases of switching instants
If it were possible to switch on a transformer at an instant of a
voltage wave which corresponds to the actual flux density in
the core at that instant, it would not have resulted in any transient.
In actual practice, a transient phenomenon in the form of inrush
is unavoidable since the instant of switching cannot be easily
controlled and the instant of switching favourable to one phase
is not favourable to other phases.
There are six limiting conditions to consider and these are
tabulated in Annexure I, Table 1
2.2 The nature of inrush current
The initial value of the current taken on no-load by the
transformer at the instant of switching in , is principally
determined by the point ofthe voltage wave at which switching
in occurs, but it is also partly dependent on the magnitude and
polarity of the residual magnetism which may be left in the core
after previously switching out
The basic circuit of transformer can be treated as combination
of predominantly inductance and with resistance. As per the
constant flux linkage theorem ,magnetic flux in an inductive
circuit cannot change suddenly, the flux just after closing the
switch must remain equal to the flux just before closing the
switch. Based on the different cases indicated in Table I, the
flux density could reach peak positive value of Bresidual+ 2
Bmp, maximum valu"e. For sinusoidal input voltage, when linear
magnetic characteristics are assumed flux wave has a transient
DC component which decays at the rate determined by the ratio
of resistance to inductance of primary winding and with a steady
state ac component. For a phase switched on at the most
unfavourable instant, at zero crossing of applied voltage, the
typical inrush current wave form which is completely offset in
first few cycles with wiping out alternative half cycles is
observed because the flux density is below saturation value for
these half cycles. Resulting in very small current value. Hence
inrush current is highly asymmetrical and has a predominant
second harmonic component Typical inrush current
waveforms are indicated in, Annexure 1, fig 1
The time constant LlR of the circuit involving transformer is
not constant, the value of L changes depending on the extent of
core saturation. During first few cycles, saturation is high and
inductance is low. Hence initial rate of inrush current is quite
high The total phenomenon lasts for a few cycles.
3.1 Efreet of inrush magnetizing currents
These asymmetrical high magnitude inrush currents have an
ample harmonic spectrum including even and odd components.
These currents may cause a variety of undesirable effects in the
system such as
improper operation of protective relays and fuses
1.. ____ = _____ _ __ , ...
.... uJ. ... J.v .... ~ ... "" .... v .. u .......... .....
mechanical damage to the windings
deterioration of the insulation and
voltage sags.
oscillatory torque in motors
current transformer saturation
Since the magnetizing branch representing the core appears as
a shunt element in the transformer equivalent circuit, the
magnetizing current upsets the balance between the currents at
the transformer terminals and is therefore experienced by the
differential relay as a false differential current.
From the stand point oftransformer life time, tripping out during
inrush conditions, it is very undesirable situation, breaking a
current of pure inductive nature, generates high over voltage
that may jeopardize the insulation of a transformer and be an
indirect cause of an internal fault.
3.2 Short circuit currents
Abnormal currents may occur in the primary windings under
certain adverse conditions when switching in a transformer on
no-load, but much heavier currents may flow in both primary
and secondary windings when transformer momentarily supplies
its heaviest load, when a short circuit occurs across the secondary
Four di,stinct conditions to which transformer may be subjected
a r ~ __ "
Condition Remarks
Transielll switching, no-load Danger to the windings
current inrush
Steady no-load current Normal thermal effects
Steady normal load current Normal thermal effects
Transient short circuit current Danger to the windings,
special precaution required
as it results in severe
mechanical stresses in the
The operating conditions of power transformers warrant great
importance to minimize the frequency and duration of unwanted
outages and make relaying task very difficult.
The problems related to protective relaying of power
transformers is listed in Annexure t Table 2
The inrush currents resulting different cases are listed in Table
1 above.
The shape, magnitude and duration of the inrush current depend
on several factors such as size of transformer, impedance of
source, magnetic properties of materials, residual flux, instant
of switching in, the method of transformer connection etc. and
these are listed and described in Annexure 1, Table 3
The transformer user is generally interested in knowing the
maximum value of inrush current and the rate of decay of inrush
current If the saturation flux density of the core material is
2.03 teals, The flux of magnitude 2.03X Ai is contained in the
core where Ai is the net core area.
The formulae are derived in reference [ 4] for single phase basis.
Depending on connection of three phases, Star or delta, , the
maximum inrush current is two thirds that for single phase in
case of bank of three single phase transfolmers. And for three
phase three limb, three phases are magnetically interlinked and
can be treated as consisting of three independent single phase
The different formulae used in practice by designers are given
in Annexure 2, A to D, and estimated peak inrush magnetic
currents are tabulated in TABLE 4 Annexure 2
The difference in values obtained is due to the approximation
and assumptions made for different equations. The calculation
accounts for a number of important transformer and system
parameters that significantly impact the magnitudes and wave
form of the inrush current. Inrush current parameters, peak,
second harmonic and duration of modern power transformers
differ from older designs due to use of higher grain oriented
core steels, steplap core joint type and higher rated design core
induction value, which have significant impact on proper
relaying protection of the transformer.
First of all it should be underlined that the transient of current
that is required to establish the magnetic field of the transformer
during the magnetization, cannot be considered a fault condition
so it should not cause protective relays to operate. The duration
and amplitude of the inrush current is a function of two sets of
parameters. The first one considers parameters which belongs
to the transformer and the second includes parameters from the
power system.
7.1 Factors to be considered for PROTECTION
N{)minal power of the transformer
material used to build the core
residual flux just before the connection to the powel
short circuit power at the common coupling point
distance between the bus of the substation and the powel
The transformer design and station installation parameters affect
the magnitude and nature of the inrush currents significantly. [Ref
I J Ueslgn ana system parameters [winding Connection, short-
circuit network capability] etc
In designing high speed protection system the fast discrimination
of magnetic inrush current is also indispensable to prevent false
tripping of relays
The conventional method of inrush current detection for
transformer protection recognizes the ratio of second harmonic
component of the differential current, new approach based on
nonlinear behaviour of transformer magnetization
characteristics. The second harmonic detection method as
problem of detection within one cycle, For effective protection,
it is important to identify inrush current from other events such
as load switching, capacitor switching, single phase to ground
fault, etc
In order to extract ifirush current and internal fault, many
schemes are used-some schemes make use of the information
obtained only from incoming currents of transfonner such as
method based on second harmonics restraint principle,. Care
has to .be taken to take into account considerable second
harmonics generated by wide spread use of underground cables
and increasing nonlinear loads . Also second harmonic
components of magnetizing current in the magnetizing inrush
current tend to be relatively small in new transformer because
of improvement in transformer core materials. Some schemes
make use of variation of transformer terminal voltages, based
on voltage restraint principle,
7.2 Different ways to solve inrush problem
Different ways to solve the inrush problem are
Design measure on transformers
oversized fuses
limiting with resistors on the primary side while start, then
bridge the resistor
switching on with solid state relays with random
characteristic or peak switching characteristic
use a dimmer to ramp up the voltage is expensive not sure
combinations of methods above
And limiting with nonlinear resistors etc
Each method has its own advantages and limitations and details
are available in literature
Different ways to solve the inrush problem is affected by
different aspects -Lower inrush brings higher weight, higher
costs. Higher fusing brings dange: of overheating. The
mufacturer and user should find optimum solution based on
eraU requirement .
19le line diagram including a medium power transformer
depicted in Fig 2
lDexure 3. The transformer is connected to I32kV Substation
lich is very close [less than l00meters] from substation. 132k V
lDsformers. The connection is made through 132kV SF6
"Cuit breaker. During switching in of transformer, the supply
j breaker at Substation used to trip because of earth fault
)ping due to high inrush neutral current On detailed analysis
> tripping was attributed to huge transient phenomenon of
gnetic inrush current [REF 21]
Jical wave shapes were recorded using FAST ACTING
ANSIENT RECORDER available with customer. The wave
pe is damping in nature and stabilizes after a few cycles.
e initial peak was causing nuisance tripping. The
responding relay setting time was increased to offset the
ial peaks, this avoided nuisance tripping. No abnormality of
transformer was observed after conducting confirmatory tests
The transient phenomenon of Inrush magnetizing currents
is briefed in this write up .. This cannot be avoided during
switching in of transformer. Adequate care is taken in
design and manufacturing oftransforrner to limit the peak
value of inrush current..
Due to phase displacement in three phases, unbalance in
inrush current is Unavoidable.
Setting of protection circuit for delayed operation and
setting for higher values are common and established
methods to take care of inrush currents and unbalances
Advances made in transformer technology and design in
the recent past have changed the characteristics of the
transformer inrush current and have introduced
modification in operations in the existing harmonic restraint
relays during energisation
Incase of magnetic core saturation which sometimes
happen during switching in transfolll)er, the inrush current
is limited by primary winding along with the power line
system impedance in the circuit This system impedance
and distance from the source substation influence the
magnitude of inrush current.
Special care and consideration is required while
selecting transformers to be located very Dear to
incoming substations aDd with high short circuit
capacity, Dux density in transformen should be limited
to lower tbannormal values
(vii) When multi pulse solid state controlled power systems are
in use such as 12 pulse or 24 pulse thyristor/solid state
device controlled systems, the harmonics in the system
are mote. Examine the possibility of using transformers
with phase shift connections To reduce the harmonics with
typical connections such as extended delta and zig-zag.
10.0 List of references
Ref I: BOOK Transformer e n g i n e ~ r i n ~ T F Rhm"., r.
Camilli,ABoyajaian, VM Montsinger Johnwileyand
sons, LONDON]
Ref 2: The J&P Transformer book, Martin J Heatcote, Newnes
Ref 3: Transformers, Book BHEL Tata Mcgraw hill publishing
company New Delhi
Ref 4: Transformer engineering Design and Practice:, BOOK
SV Kulkarni, SA Kbarpade, liT MUMBAl, Marcel
Dekker publication
Ref:5: Effects of magnetizing inrush current on power quality
and distributed generation Manana MEtal ETSIIT
University of Can tab ria Spain
Ref 6: Transformer protection against inrush currents Vinaya
K Thakare, Electrical India Vol 47, nO. 8Aug 2007
Ref 7: Power transformer characteristic and their effect on
protective relays", Fahrudin Mekic, Ramsis girgis,
Zoran Gajic, Ed te Nyenhuis, ABB, 33
protective relay conference Oct 17-2006
Ref8: Bogdan Kasztenny, Ara kulidjian, 53
conference for protective relay Engineers april 11-13,
Ref9: Transient inrush currents due to closing time and
residual flux measurement-deviations if controlled
switching is used Andreas Ebner, ETH Zurich
Ref 10: AR Sedihi, MR Haghifarn, Tehran, Electrical power
and energy systems, 27, 2005 PP 361-370" Detection
of inrush current in distribution transformer using
wavelet transform
Ref 11: Effects of magneitising inrush current on power quality
and distributed Generation Manana M Eguiluz, LL
Ortiz, G Renedo: C PEREZ S University of Cantabria,
Ref 12 : Avoiding of transformer inrush current peaks with a
TSR Dip Ing (FH) Michael Konstanzer Manana M
Eguiluz, LL Ortiz, G Renedo, C PEREZ S University
of Cantabria, Spain
Ref 13: High voltage engineering network protection,
Lappeenanta university of technology
Ref 14: Inrush current control in transformers, Seshanna
Panthala Faculty of Engineering, Assumption
UniversityBangkok, Thailand
Ref 15: implementation of transformer static differential relay
with harmonic blocking Mohammed El Gamal.Ahmed
Lofty. Hussien Dessouki, Alexandria Egypt
Ref 16: Sami G Abdulsalam, Wilsun xu,IEEE, international
conference on power systems transients IPST05 in
Montreal Canada june 19-23 2005 Using neutral
Ref 17: Large transformers for power electronic loads, OW
Andersen, Norwegian institute of technology
Trondheim, Norway
Ref 18: A new technique for mitigation of transformer inrush
current, Balachandran DP, .. Sreerama kumar R,
Shimnamol P ,19TH International conference on
electricity distribution Vienna 21-24 may 2007 CIRED
Paper No 0694
Ref 19: Jose C Oliveira, Carlos Tavares, Roberto Aplolonio et
alTransformer controlled switching to eliminate inrush
current, 2006 IEEE PES T&d conference and
exposition Latin America Venezuela
Ref 20: Impact of inrush currt,nts on the mechanical stress of
. high voltage power transformer coils Michael Steurer
Ref2I: Internal report and discussions with MIS M
Vijayakumaran, Ranjit C Das and D Ghosh May 2004
Ref 22: Electrical engineering hand book Siemens Aktien
gesellschaft, Erlangen, BOOK Wiley eastern LTD
Ref 23: Alternating current machines Book AF Puchstein, TC
Lloyd, AG Konrad ASIA publishing house Bombay
Ref24: Switch gear protection Book Sunil S Rao, Khanna
publishers, Delhi 1908
Ref:25: Transformer controlled switching to eliminate inrush
current Part II On a l00MVA three phase transformer,
Heriveltos, bronzeado, et aI, .2006 IEEE PES
Transmission and distribution conference and
exposition Latin America, Venezuela
Ref 26: Elimination vf ill;",.;!. ill pwu!!",,!
by sequential phase energization KP Basu, Ali asghar,
VOL 4 NOS, PP 147-152, Malaysia
Ref27: Transformer design for reducing the inrush current by
asymmetrical windingChih ku cheng, Shin der chen,et
aI, Journal of Chinese institute of engineers, Vol 28,
No.5, pp 811-818, 2005
The authors are grateful to the management of AREVA T&D
INDIA LTD., NAINI works for the permission to present this
paper. The authors are thankful to Sri A K Sinha (Unit
management Director - DTI) for the constant encouragement
and Sri G K Gupta (Unit Managing Director - PTI) for the
support and Sri M Vijaykumaran for the guidance and
discussions in preparing this paper. The authors are thankful to
all those involved during measurement and analysis of this
transient phenomenon.
The authors are obliged to organizers of "TRAFOSEM 2008,
opportunity to present this.
Different Hmiting conditions for switching in of transformer
1 Switching in at zero voltage,No residual Due to hysterisis and nonlinear B-H curve of core steel, the transient
magnetism flux may reach to approximately twice the normal flux
current waves may be extremely peaked and contain prominent
2 Switching in at zero voltage-with maximum The flux wave starts at a value corresponding to the polarity and
residual magnetism having polarity opposite to magnitude of the residual magnetism in the core, theoretically flux
that to which the flux would normally attain can reach a value 3 times normalThe magnitude of inrush current
under equivalent normal voltage conditions. will be higher than case I and takes longer time to reach steady state.
3 Switching in at Zero voltage with maximum Compared to case 2, results in a diminution of the initial maximum
residual magnetism having the same polarity as value of the flux and of the inrush current. If the residual magnetism
that due to which the flux would normally attain is lower than the corresponding maximum, initial flux waves are
under equivalent normal voltage conditions unsymmetrically disposed about zero axis
4 Switching in at maximum voltage, no residual Follows normal steady time distribution of the flux, the no-load
magnetism current pursues its Iformal course and does not exceed the magnitude
of normal load current.
Switching in at maximum voltage-with Residual magnetism introduces transient components, initial flux
maximum residual magnetism having a polarity waves unsymmetrically disposed about zero axis, high initial
opposite to that which the flux would normally maximum flux values are attained, current inrush may reach a value
attain under approximately to twice the normal flux density
Switching in at maximum voltage with The initial flux will be unsymmetrically disposed about the zero axis,
maximum residual magnetism having the same as in case 5 , above, but the flux and current waves would initially he
polarity as that to which the flux would normally disposed on the opposite side of the zero axis
attain under equivalent normal voltage
.,:> ........ c. ___

f\ f\ f\
C\ f'.... '.,.,. ,.-.....
......... 0- .......... cr..1: i.'C"' ... " ... s'l. ....
Ie:. tbr .. e- pb-.-..c-__
Accurate estimation of
the second and the 5
harmonics takes arounq
one cycle Off nominal
frequencies create extra
measuring errors in
harmonic ratio
The measured currents
display enormous rate
of change and are often
significantly distorted
In modern power
transformers due to the
magnetic properties of the
core , the second harmonic
during inrush and the 5
harmonic during over
excitation may be very low,
jeopardizing relay security
The presence of
higher harmonics It actually takes one full j
does not indicate cycle to reject the I
necessarily inrush, magnetising inrush and I
the harmonics may stationary over excitation I
block a relay during hypotheses, if an internal '
severe internal faults fault is not severe enough
due to saturation of to be tripped by the
__________ -+-=th::e=-=cr=rs:'-_---':--_--I unrestrained differential
The 5"' harmonic may elements
External fault current when
combined with ratio
mismatch may generate a
false differential signal. The
CI's when saturated during
external faults may produce
an extra differential signal
The internal fault current
may be as low as few
percent of the rated value.
Attempts to cover such
jeopardize relay
be present in internal
fault currents due to
saturation of the CI's
and due to rotor.
asymmetry of
generators and lor
power electronic
All the means of
preventing false
tripping during
to a certain extent the
dependability of the
The security demands
under inrush
,overexcitation and
external faults may
limit relay
The means of restraining
the relay from tripping
during external faults may
limit the relay speed of
The means of restraining
the relay from tripping
during inrush, over
excitation and external
faults may limit the relay
speed of operation
Table 3: Factors affeding inrush current
Factor Description
Size of transformer The peak values of the magnetizing inrush current are higher for smaller[ < 1 OOk VA transfonners
with time constant for decaying current is in the range of 0.1 of a second. for large transfonners
duration is longer and decaying time in the range 1 second
Impedance of the system The inrush current is higher when the transfonner is energized from a powerful system Trnnsfonnl'r<;
from which a located closer to the generating plants display inrush currents lasting much longer than transfonners
transformer is energised installed away from the generators
Magnetic properties of The magnetizing inrush is more severe when the saturation flux density of the core is low. Peak
the core material inrush current increases as the design induction level increases, % second hannonic decreases as
the design induction level increases. Step lap joint have higher magnitudes of inrush and lower
second hannonic
Remanence of the core Under most unfavourable combination ofthe voltage phase and the sign of the remnant flux, higher
remnant flux results in higher inrush currents, Modern core HIB has lower remanence
Moment when a The magnitude of the inrush current is a random factor and depends on the point of the voltage
transformer is wave fonn at which switchgear closes as well as on the sign and value of the residual flux. It is
switched on approximated that every 5
or 6
energizing of a power transfonner results considerably in high
values of the inrush current
Way a transformer is The maximum inrush current is influenced by the cross-sectional area between the core and the
switched in winding which is enrgised. Higher values of inrush current are observed when inner[having smaller
diameter] windings energized first. In grain oriented steels, inrush current may reach 5 to 10 times
rated when energized from outer winding and 10 to 20 times when energized from inner winding.
Transformers with Switching in via resistance reduces the magnitude of inrush current and increases damping.
special switchgear Transfonner equipped with air type switch may result in successive halfcycles of the magnetizing
voltage of same polarity., The consecutive same polarity peaks cumulative the residual flux and
reflect in a more and more severe inrush current
Number of phases and Inrush current in 3 phase transfonners is equal to or less [by 3C-40%]than single phase depending
winding connection upon the winding connection, Delta, star, grounded star, Auto etc, residual flux is also influenced
by winding and other capacitances connected to transfonner
to = Point at which core saturates (function ofB
, Bs, B
, 0)
BR = Remnant flux density
AI General inrush current equation (Ref7J
f2V[ -(t-t ]
-- sin(wt-)-e 0 *sina *K *K
Zt r W S
U = Applied voltage
Zt = Total impedance under inrush, including system
to = Energization angle
Bs = Saturation flux e ~ s i t y
BN = Nonnal rated flux density
t = Time constant of transformer winding under inrush
a = Function of to
Kw = Accounts for 3 phase winding connection
K. = Accounts for short-circuit power of network
BI Inrush current calculation formula INHOUSE:
lomax = [V ..[ 2 / Xa ] * [ I + ] Amps, Where
V = RMS phase voltage applied to primary
= [Bn + Br - Bs ] I Bn
Bn = Working flux. density ,
Bs = Saturation flux. density
Br=0.8 xBn
Xa = Air core reactance of primary winding
= Xam + Xatap + Xm+ Xas
Xam = Reactance of main winding,
Xatap= Reactance of tap winding
Xm = Mutual inductance of main and tap,
l(as = System reactance
rush Current as per Blume et at :
*h* Ac*(Bres+2*Bmax-20.2)*2
x= ~ ~
\c = Cross section area in sq. cm
3max = Working flux. density in Kilo-lines/sq. cm
3res = Generally 60% of Bmax in Kilo-lines I sq. cm
>l = Number ofTums in Primary
\s = Effective cross section of air core magnetic field
within excited wdg in sq. cm
= Winding height in inches.
DJ IDrush CurreDt as per KulkarDi etal
I omax = (1- cosO)
Where ,
v = R M S value of applied alternating voltage ,K2 = correction
factor for the peak value = 1.15
Xs= ~ O ' N 2 * Aw *2*7r* f
N = Number oftums of the excited winding
Aw = Area inside the mean tum of excited winding
hw = Height of energized winding
0= KI*COS-I{BS- Bmp- Br}
Where, Bs = Saturation flux density, Bmp = Peak value of
designed steady-state flux density
Br = Residual flux. density = 0.8 * Bmp, =KI = Correction
factor for saturation angle = 0.9
MVA HV(kV) Connection Flux density i ..... (lph) i ..... (lph) i .... (lph)
( Tesla) [BlumeJ [Kulkarni) (AREVA)
(Amps ) (Amps) (Amps)
1 Il Delta 1.73 412 477 560
1 Il Delta 1.66 443 513 610
5 33 Delta 1.72 444 514 667
5 33 Delta 1.59 411 477 624
10 33 Delta 1.65 759 881 1018
10 33 Delta 1.68 772 897 1091
15 33 Delta 1.64 832 1076 1255
15 33 Delta 1.62 1108 1521 1555
20 132 Star 1.68 685 939 895
20 132 Star 1.44 775 1132 935
25 132 Star 1.53 842 1090 1074
25 132 Delta 1.69 535 425 621
r N 132111 kV 132111 kV 132111 kV
c----\J 16I2OMVA 16120MVA 20125MVA

l r N OTHER PROT 87T & 64 HV
_ c----\J &64LVNOTSHOWN
FIG 3A: Transient current Recorded in 25MVA 132/11KV

'_ ..

".. .
.... , .. \

" r



t, i
..... .1' ,,,->
Mayur Karmarkar,
The Indian centre of the ICA, New York
lCA is a non-profit, non-commercial organization
Leading organization for promotion and defense of
copper use worldwide
Headquartered in New York, regional offices in:
o Brussels, New York, Santiago, Singapore
31 offices in 24 countries on 6 continents
Active in over 60 countries
The United Nations Framework Convention on
Climate Change (UNFCCC) adopted in 1992
UNFCCC adopted Kyoto Protocol in 1997 -mandates
quantitative limits on GHG emissions from developed
countries in the 2008-2012 timeframe, and creates the
Clean Development Mechanism (CDM) framework
(Article 12)
Allows developed countries committed to reduce GHG
emissions to assist developing countries in achieving
sustainable development
Benefit from project activities result in Certified
Emission Reductions (CER's)- used by developed
CDM requires an approved methodology by UNFCCC.
.w < __ ___ __ ___ ._
I Project Note, . , 1
t r-litR

.... I
; l
Ii' QOE 1--+ '-----r---J
01 11.":.
/ __ '''"'6-.ttanon
Anqually DOE (Verification) --+ i UNFCCC ED ,. ] -+ CERs
DNA - DeslpIItd National Agency
DOE - D<sIpIItd Operldonal Eadly

Host ComtIy LeA P > 90""10
New rlethodo'09Y p < 50%
Implementation PrOject
issua1ce p 85% - but varies from low of 20%
(geothermal), (Iancfit)
to nigh of 118% (N20)
p - Success probability
Source - IR: Research, UNEP

_. ..
Utilize existing approved methodology (AMS-Il.A)
To facilitate higher usage of Star 3 OT
Develop revenue stream for utilities - Incentives
Monitor energy consumption reduction in a project
area as outlined in AMS-ILA
CERs generated after monitoring, validation and
oversight of CDM Executive Board (CDM-EB)
Revenue from sale of CERs used to service
investments-Estimated revenue/ kVA of Rs_ 6_6 per
Define project area - DISCOM based
Utility tender provir.ion Star 3 (BEE) with
Develop monitoring plan (Asset mgmt)
Preparation of Project Design Document (PDD)
Validation of PDD by DOE
PDD approval by DNA
DNA approved POD posed for final approval ofCDM-
Monitoring! validation commences as per AMS-II.A
under this framework
Programmatic COM to reduce individual project
tmnsaction costs for replicability
l t I nll.TAfUnl:1l:1 I

1 .. 1 I -r-" --
Rs approvalj Project approvall
lity (Project Proponent)
,esUDent in EEDT
IDstaII GSM - Meters
Prtpaft CPA - DD
:>IIi toring
set management
I Database I TC old &: new Dr

.. ---------,,_ ... --.- ... -.. -..
GSM AMR . , Project area
fw llof III., 'I" JSCOM
- ." ." II"
m!! " It
Ji] (' (( II ,," I I I .. I "
. < "'''@I'"'II&,,I
I: I ... I' r" I' ,"
". r ",, If PSG I.. ,
" I r' ., "I' If l"lf
0ad0a400. IIi t I,I
\) "', 111"1" II' h
'IIIi1sIoII factor I I." .. II " 1'1 "I -Project
!: nductIoa I I, I" ,,,, , Sample Group
I.i I r ,I ,I I
Cor It years 'I (II"

---_ .. ------_ ..._------..
Define project area - DISCOM based
:lase of DT wise test reports/ Service Hrs / loading
-\ssist in selection of Project sample group (PSG), as
-equired under AMS-II.A
:nstall DT meters (GSM based)
.nformation of EEDT installed
\sset management system & monitoring of DT
,ystem of 'Performance Testing' of a repaired DT
Estimation of technical distribution losses in the
electricity grid
Initial investment for the cost differential
Step 1: Determination of the project area (s)-Based on DISCOM
areas each with a maximum 60GWh (60 MUs)(around 1 million
Step 2: Establishment ofa project activity implementation plan
Step 3: Installation of measurement equipment
Step 4: Establishment of PSG
Step 5: Establishment of DT database, manage & monitor of
each DT installation & repair
Step 6: Monitoring of utilization hours, loading, Voltage &
frequency in the PSG
Step 7: Determination of the load factor & loss curve
Step 8: Calculation of the energy saving based on Test Reports
& loss curves
Step 9: Estimation of technical distribution losses in the
electricity grid
Step 10: Cross-check of monitoring results by random sampling
of DTs not included in the PSG .
Step 11: Calculation of emission reductions
Step 12: DT inventory management
:' AveragekVA-63kVA :' Project period - 10 years
',' Incremental cost each for Energy savings per
EEDT - Rs. 2O,<XXl EEDT -185 W X 6000 Hrs
"Annual Dr demand - :! National Annual Energy
3OO,<XXl Nos. Savings - 333,(KX) MWh
;, Debt cost - Interest rate 10%
PA for 10 years
;, ARR proposal with 70%
debt & 30% equity
:l Average energy cost at my
feeder level- Rs. 31 kWh
:\ 1 T CO,2, = ,0.85 MWh
(Source eM)
;, Return over equity - 14% :' 1 T CO
= 1 CER = Rs.500
PA (As per regulator's

(Amount in Rs. Crore) Debt Equity Total
Incremental capital investment 420 180 600
Annual capital serving cost proVlGea OJ ':JL.
in ARR for to years
Annual Energy savings due to BE 100
Annual CER Revell\lC due to BE 14
Total annual savings due to BE 114
Annual surplus due to EE 22
._+ n. r -0-,
11 (",I i, "f
.... ,. -.....

Any use of this presentation other than as a whole and
in conjunction with this disclaimer is forbidden.
This presentation may not be copied in whole or in
part or distributed to anyone.
This presentation and information and statements
herein are based in whole or in part on information
obtained from various sources. ICPC(I) makes no
assurances as to the accuracy of any such information
or any conclusions based thereon. ICPC(I) is not
responsible for typographical, pictorial or other
editorial errors. The presentation is provided AS IS.
No warranty, whether express or implied, including
the implied warranties of merchantability and fitness
for a particular purpose is given or made by ICPC(I)
in connection with this presentation.
You use this presentation at your own risk. ICPC(I) is
not liable for any damages of any kind attributable to
your use of this presentation .
. \
David W. Mlliure & Jerry Smith
National Appliance Conservation Act (NAECA) of
1987 gave US Department of Energy (DOE) specific authority
to set national standards for certain residential and commercial
appliances. Energy Policy Act (EPACT) of 1992 expanded
number of products DOE could set standards for while requiring
it to assess feasibility of Energy Conservation Standards for
Distribution Transformers.
In response to this act, National Electrical Manufacturers
Association (NEMA) developed a voluntary Guide called TP I
in 1996 and latter revised it in 2002. NEMA TPI efficiencies
for different products are shown in Tables below.
Reference Condition Temperawre % 01 Nameplale Load
Load Loss 55"C 50%
No Load Loss 2fr-C 50%
SlnglePhise Three Phase
kVA EnlCle kVA EInden
10 98.4 15 98.1
15 98.6 30 98.4
25 98.7 45 98.6
37.5 98.8 75 98.7
50 98.9 l1Z.5 98.8
75 99.0 150 9b.9
100 99.0 225 99.0
167 99.1 300 99.0
250 99.2 500 99.1
333 99.2 750 99,2
500 99.3 1000 99.2
667 99.4 1500 99.3
833 99.4 2000 99.4
2500 99.4
Medium Vollage
Low M.dium Voltage
97.7 976 91.6
\18.0 91.9 979
00.2 981 !l81
982 1/82
.. 8.5 98,4 [184
986 985 985
1'18.1 98.e 1'18.1
988 989 988
989 990 98.9
9!/,1 990
99.2 99.0
- 99.2 99.1
% of Na .... plat. Load
Thr .. Phue
Low Medium Volta"e
81l ..eolilV8tl
97.0 96.8 968
1175 97.3 97.3
91.7 976 91.6
98.0 97.9 97.9
98.2 001 98.1
98.3 98.2 982
!lB.5 98.4 98.4
986 1'18.6 98.5
118.7 \088 98.7
9a.8 98,9 98.8
9B.9 99.0 989
99.1 98.0
99.2 990
992 99,1
Even though a few states had made this standard as mandatory
minimum requirement especially for Dry type transformers, for
Liquid-Filled units it was mostly on a voluntary basis.
Mandated Study conducted by DOE determined that a new
standard could save energy and would result in Life Cycle Cost
(LCC) Savings. Advanced Notice of Proposed Rulemaking
(ANOPR) meeting was called in September 2004 to review
results of their findings with stakeholders. After getting inputs
from stakeholders, DOE issued Notice of Proposed Rulemaking
documents in September 2005 and held another
meeting of stakeholder in September 2006. The Final Ruling
was issues in October 2007.
The Final Ruling:
The final ruling sets minimum efficiency standard for Liquid-
filled and Medium voltage Dry type transformers up to 2500
kVA and is effective January 1, 2010 (see Table I.1 & Table
1.2). Standard for Low Voltage Dry type product was set at
TP 1 as a part of EPACT 2005 effective January 1, 2007.
10 ____ ._. __ ............ _ ......__ ..._ .. 98.62 15 _ ....... _. __ __ .... __ .._ ............................ ..
15-_ ..__ ..... , .......... , ... _ ..... _._ .._ ... 98.71i 30 ........... _ ....___ .._ .............. __ ..... ..
25..-__ . ___ ._ ............... ___ ....... __ 98.01 45 .......... _ .._. ___ ..... _._._._ .. _ .............. _ ...
31.5 .... ___ ..... _ ___ ........._ ... _ ............ . 99.01 15 ........,.... .._ ... _ ..... _ ... _._ ...... _ ........ _ ........ ..
50 __.... _ ....... __ ........... _._ ...._ .._ .......... ,_ .. . 99.08 !l2.5 ..... _ ..._ .. _ .................. _ .......... .
75 ........ "_ ....,, ........ _ ...................... .. 99.17 150 _ ............ ______ ... _ ......... , .... _ .................. ..
100, ........ __ ._ ........_ ....................................... , ...... .. 9923 22S ............ _ ................... , ........ _ ..
9925 ................ _ .... _ ..
99.32 500 .............................. , ....... " .... _ ..
167 ....... _._ ..... _ .._ ...... , .... _ .......... _ ..................... .
250 ......................... _ .......... " ..
333 ................................. _ ............... 99.11 750 ....................... ' ....................... .
500 ...... __ .._ ....__ .............. .. 99.42 Ioc() ............... ' ................................ _ ..
567 ........... _., ....... _ .................................. , ........... 9946 1500 .... , .... _ .... _ ............ , .. _ ..
S33 ......... _._._ ...__ ......................................... .. 99.49 2000 ...... _ ..... __ ... _,_._ ....... , ....... __ ....' ............ ..
2500 ...... _ ........ _ ......... _ .................................... ..

N .. 10CFRPart431.SuIr


2G-i5 kV 4&-t5kV
2O-.45kV 46-Q5 tV 196kV

IIIc:lerey elItleney .Idorey
1';1 (\1 1".,
15 ........ __ ......_ .... 98,10 91,86 15 11.50 97,18
25 ........................ _ ... 98.33 98.12 30 -_ ..... __._-_ .....-._., 97.90 97,63
31.5 ...... _ ... _ ........ __ . PB,49 .., 45 M.IO 97,86
50 ... , ... __ ....__ ...
98.60 M.42 15 ... -" .... _---- ... __.. 98.33 ga,12
/S, ......__ .......... _ ... 91.13 gaS?
112.5 .................. _ ....... M.4P 91.30
100 ......... __ .... _ ... .. 9882 98.67 aU3 150 ...................... _._ ... 98.60 91.42 "'M,
161 ......._ .. _ ... . aU6 ta.83 ".80 225 .............. , _ ... _ 98.13 9857 9853
250 .. " ..___ ......... , .. ".07 tags
'8.01 ... , ................. -....... 98.12 98,67 98.63
m ..... __ .......... _ ... .. 99,1.1
ta.89 50) ... , ........................ 98.16 9800
500 ............._ ..... _ ... . 8922 ",12 ".09
750 ........... ' .... __ ... !lI.07 98.91
fIit ........... _ ......... _ 9927 89 IS 10.15 1000 .. _ ................. _ ....... 19.14 9903 9899
83) ..... __ .... _
89.20 1500 ._ ......... _ .. __ .. 1922 ".12 ",09
2000 .. __ .. .. __._ ..._-- fl.P ".1. ".15
2500 ........ 99.11 89.23 99.20
Engineering Analysis:
Prior to setting these rulings; DOE had done an extensive
Engineering analysis of various ratings, considering all available
material options including Amorphous metal to arrive at
minimum Life Cycle Cost (LCC) under various operating
conditions and energy costs. Following is an overview of
methodology used in such analysis.
First DOE classified products range into ten Product class.
Initially, Product classes 3 & 4 were for Low Voltage Dry type
transformers. Since efficiencies of these products (Low Voltage
Dry type) were incorporated into EPACT 2005, DOE dropped
these product classes from further analysis hence missing from
following table.
TUlrW Pr*tCEtsadNllktlk\'A_
DimiRtiaI Iran PndtcIOm kHRDIt
1. I.iqIId-iumid. \JB!ltF 10-833 B
15-!500 14
5 )45t\'Bn.
6 .
!. Dry-type . . .\6.95 tVBR. l)-8},
Dty-type. thrtt-phalt. 46-95 tv Bn.
9. ,tV BII.
.... 0"
10. b-pIw. kY Bn.
As it would be impractical to analyze all kVA ratings in all
voltage class of these product classes, DOE set out to simplify
analysis, Product classes were further subcategorized by shape
of the tank used to make product. resulting in 10 Design lines.
Recognizing that many ratings used essentially same
construction techniques, one representative rating was chosen
from each design line for detailed analysis, Design lines and
representative rating are shown in table below,
Tablr 5.2.2 Dfslp LiIn (DLs) lid Rtpmtllatln tllirs for Alalrns
PC' D1
T Dktribllioa tl'A ItprtWltltil't r iii for ""
Traasrollllfr 1Ja&, EDliDHriDl DtYaal.ilf
50 k\'A. 65'('.IIIl!ltpbut. 6OHz.14400r priuwy.
ffCQllplll tmI; 1 ltCondary. rtclall[llW' bilk
1 1
25 k\'A. 65,(,. 60Hz. 1+100\' primuy.
IOIIIdlallk I!O 1-10\' ltCondarv. rouod tIIIk
IIIlIIImld. 150 .. 133
sao tH. WC 60Hz. I+lOO\' .
.imlllerlfdl!mflw 1>-500
150 tn. 65=(. iIm.phatt. 60Hz..
JIrlIII2IY. ltCoadary
5 750-!500
1500 i\'A 65'(' Ibrttplult. 60Hz.
!494OGrdY 1+100\'
300 k\'A. 150'C.Ibrtt-pbaIt. 60Hz. -I16OV
iplw. iprllllllY. 1m' 1tC.,..
Iky-l)'pt. 1Itd!Im-,olq(f. __
1500 tVA I)O'C dlfttiJlmt. 60Hz.
iiUt. sccoudaly. 4)1;Hll
JOO tVA. 1)O'C'.I!m-pblt. 60Hz. 1!470r prunary.
pIast. Bll .QOy.mv 1tcOIIdary'. 9ft\' Bll
Iky-l)'Pt. mtdium-ioluftI!m,
1500trA \5O'C'.1iIm-pbaIt. 60Hz. 1H10V
pIast. Bll triaw;, 410Y 95,\' Bll
10 IJ
!OOOt\'A IW<'. 1im.puw.6OHz.11470V
. .qoY2m ltCoadaty.l!5kV Bll
PC IDUlKprodatttlau(1ft C!upIrr 3 ofllle ISD) .... Dtp1DnIaIcaIutd .1IlaI)1InIHIt\b PC,. PCl.
mI PC, based OR dIt rnuII\ b ria dm-pIw Tht Dcpnnr cLd lIOIltltd aay rqmcauUI'C
from lilt "\t.lDtdillaHoI!aft prodatt cbucs (PeS, PC1 aad PO) lItmr.t of lin lot: uIH ,oIume.
Graphically, above table can be depicted as follows:
Tabif 5.l.3 lJqldd-laIwnId Dnip. i.iIlft and RtprtWllhaUvt l'Dils
Product Class 1
Product aass 2
Table 5.2.4 Dr,"-Typt. Yoltage, Siug]f'-Phase Lints
Tablt' 5.25 DQ'-T)]>e. :\lrdium-Yoltagt. Lines
PC 10
High Bll
known among transfonner design engineers that there is a
batlbellnatlLC81 relationship that exists between the kVA rating
physical size, cost and performance of the transfonners.
Size vs. Perfonnance relationship arises from fundamental
tmll8tilons describing transfonner's voltage and kVA rating.
used this methodology to extrapolate cost and perfonnance
of remaining ratings in the design lines.
Core Configuration:
IndustrY has used different core configurations to make different
design lines. Following describes what core configuration were
used for which design lines.
tllllr U.4 COft ( ourations rsttl in . Lilt
Stad:M. bun.lap or Joint
OLIO crucifOl1ll jomt
OL 11
Graphically, core constructions are shown below.
01 000
Figure 5.3.2 Gl'sphif of Single-Phase Core Configul1Itions

Figure 5.3.3 Graphic of Three-phase Core Configurations
Materials Considered:
Various combinations of materials were considered in the
analysis. Following is sample of material considered for one
design line.
TabJtll6 Combinations (or tk Rrpl'Mltltin UBit from Design Lint 2
CGftMatDl , mp-Voltlgt lollVoItage Cort DfsIgD T}')It
COHuctOf Cold1Kter
M6 AI (uire) Ai(stnp) Shen DG "'oun\! Core
M6 Cu(wire) Al()trip) Shen. DG Wound Core
M4 Al(wire) AI (strip) Shen DG WOIWd Core
M4 Cu(wire) AI ()trip) Shen DG Wound Core
M3 Cu(wire) Al(stripi Shell DG Wound Core
MJ Cu(wire) Cu(slripl Shtll DO Wound Core
Ml Cu(wire) A1(strip) Shtn DG Wound Core
Cu(wire\ Cu(strip) Shell 00 Wound Core
ZDMH (u(wire) Cu(;trip) Sjell DG Wound (ere
SAl (Amoqlhoul) (u(stripl -DG \\'oundCore
Material and Labor Costs:
The Department used a standard method of cost accounting to
determine the cost associated with manufacturing. Following
figure illustrates this methodology where production costs and
non-nroduction costs are comhined to determine the
FuU Produtlion COS1'


Design Database:

Hoo-Produttion Cost
........ ' .. .
... ,
. ....,. ....

All designs reSUlting from above mentioned combinations of
materials were stored in Design database. This data set was
then used to feed Life Cycle Cost (Ltc) Model to arrive at the
optimum solution for the nation as aggregate. Design data base
can be depicted in a scattered plot showing cost of the each unit
along with its efficiency for each combination of the material.
Following is one such scattered plot for one of the design line.
For reference, NEMA TPI efficiency is shown on this graph.
The scatter plot clearly shows cost (& hence price) of unit rising
as efficiency increases. It also shows for each material
combination, a range of efficiency that can be achieved. For
example use of M6 grade of core steel will at best achieve
NEMA TPI efficiency. To achieve better efficiency, better
grades of steel and/or use of copper is necessary. Amorphous
metal provides the best level of efficiency. In that respect it is
Best and only in the Class by itself
LCC Standard Levels:
The output ofLCC model yielded the most optimum efficiency
for each design line. DOE called this level as Trial Standards
Level 4 (TSL4). NEMA TPI was called TSLl as this was the
voluntary standard prior to the ruling. Best in Class Amorphous
metal units were assigned TSL6. Other intermediate levels were
equally spread between these three classes. As one might expect,
such assignments resulted in some discontinuity and discrepancy
of efficiency not only among k VA range but also between single
and three phase within a given Trial Standard Level.
Technically, Final ruling set at TSL4 would have theoretically
results in most optimum condition for the nation. However,
because of discontinuity in the efficiencies, the final ruling was
set between TSL 4 & 5 for small single phase and between TSL
2 & 3 for three phase ratings. Such adjustment pennitted smooth
transition going from k VA to K VA and pennitted Hannonization
of efficiency between single and three phase units.
Benefit of the Ruling:
The Department estimates the standard will save approximately
2.74 quads (quadrillion (lOIS) British Thennal Units (BTU)) of
energy over 29 years (2010-2038). This is equivalent to all
energy consumed by 27 miIlionAmerican households in a single
By 2038, DOE expects the energy savings from standards to
eliminate six 400 MW power plants and 238 million tons of
Carbon dioxide (C0
), Using 3% discount rate, the cost of
standard is $460 million per year in increased equipment and
installation cost while annualized benefits are $904 million per
year in reduced operating costs.
EgU Stryken, MSc. EE
Abstract - Increasing the efficiency of transformen will
involve reducing the transformer losses. Considerable
H:uu\:i.ivu VI Ui:)l' iUU,ivlI 'I ....... U me. :u.i.id in relation to
the common present practice is fully attainable. Some
concrete examples in this paper illustrate consequences of
different degrees of loss reductions on the transformer
Globally today 30 GT of CO
are released to the atmosphere
per year. The CO
content of the atmosphere has risen from a
pre-industrial value of 28()' ppm. to 379 ppm in the year 2005.
With the emerging economies consuming more energy and the
increasing global population, this trend is accelerating. The
electrical energy system contributes to the CO
release, as long
as electrical power is generated through non-renewable energy
sources such as coal, oil and natural gas fired plants.
The total global energy consumption was ill 2006 close to
127000 TWh. About 50000 TWh was used to generate 19000
TWh of electricity. Renewable sources globally stand for seven
percent of the electricity generated. About seventy percent of
the total global electrical energy is generated from fossil fuels.
As soon as primary energy is transformed to electrical energy
through a thermo-dynamic process, between 40 and 55% of the
primary energy is lost. In older power generation plants for
example the efficiency may be below 30%. The rest is heat
released to the atmosphere unused. Given the high percentage
of coal, oil and gas fired plants this low efficiency involves
waste of CO
emitting resources.
Today's existing power generation, transmission and distribution
systems efficiency is as low as 33% while it could be higher
than 50% with existing technologies. A huge reduction of CO
emission is possible.
Typical losses are about 3.5 percent in the transmission system
and 4.5 percent in the distribution system. The losses in power
generation in thermal power plants are higher. In addition these
plants are the direct sources CO
emission. The capturing of
would be the Hlost efficient way to reduce the content of
in the atmosphere, but the technology for such capturing
in a larger scale is not yet developed.
This article deals with the losses in the distribution transformers,
which represent a considerable part of the total distribution
losses. There is an essential potential for energy saving when
providing low loss transformers for new stations as well as by
replacing old transformers by new ones with lower losses. This
can be achieved by means of existing technology, and the process
can be started without any delay caused by technical reasons.
Transformers are designed to fulfil the users' specification,
which normally includes the technical requirements given in
national or international standards. In addition the design intends
to reflect an optimal balance between the spending of materials
in the transformer and acceptable electrical transformer losses
in service. Some users have preferred a lowest possible purchase
price. while others have specified capitalised values for the
losses. These capitalised loss values are assumed to reflect the
cost for the energy needed to cover the losses in service for a
number of years. Hence a pure economical consideration in an
attempt to minimize the total costs, which in this respect are the
sum of the purchase price and the costs of the losses in service
during a certain number of years.
Viewed in the light ofa vital necessity of reducing the emission
of CO
through reduced energy consumption, there might be a
basic change in the traditional way of optimizing the transform
er dimensions, that is, setting the minimum total t.ansfonner
costs as the only target. For standardized distribution
transformers this change may come as a result of low specific
maximum limits for permitted losses, or of specified minimum
limits for the transfonner efficiency, referred to a specific service
condition. Such limits may be established voluntarily by the
users, or by mandatory decisions from the authorities. Lower
transformer losses can be achieved indirectly by increasing the
capitalized loss values. Some concrete examples that illustrate
how different loss requirements and loss capitalisations
influence the transformer economy, are presented in this
document. 100 kVA and 630 kVA are the transformer power
ratings that are chosen for the examples.
A. 100 kVA Transformer
The specified losses are shown in TABLE 1.
No Load Losses [W] 260 210 180 145
Load Losses [W] 2150 1750 1475 1250
125 105 90
1060 900 770
The losses in the 4 left columns with numbers are taken from
the European Standard EN 50464-1:2007, [2], Table 2 and 3,
column D, C, B, A, where column D is assumed to be the most
frequently used loss figures at present. The 3 columns to the
right contain loss figures that are about 15 % lower than those
in the next foregoing column to the left. The total losses in the
last column to the right are about one third of the total losses in
the first number column to the left.
The transformer manufacturers' task is then to minimize the
material costs of the transformers with the seven different sets
of loss figures in TABLE I. The result is shown in the
diagram Fig. 1

"C S +-------------------------------
2 1,00-1,12_1,39 '
"j 1.....,--------.... -
"ij 0 +-A----.... - ..... --...... -.-

Losses P3'1\. [W]
Fig. 1 Relative material costs for transformen with different
levels of losses.
The material costs of the transfonner with P.,IP l = 26012150 W
is used as the reference = 1. The diagram shows that the material
costs increase with decreasing loss levels. The reason is
increased quantity of materials. For a transfonner with PgfP
901770 W, the material costs are 3.27 times as high as the
transformer with P.,IP L = 260/2150 W. In other words, to bring
down the total losses to one third of the losses that are assumed
to be typical today, increases the material quantity and costs by
a factor of 3.27 in this example. These high material costs will
in service be compensated by lower expenses for the energy
needed to cover the losses in operation, depending on energy
prices, the degree of loading of the trnnsfonner and the number
of years the transfonner will be in service.
The transfonners represented by the four left columns in Fig. 1
are designed with stacked core and conventional oriented laser
scribed steel, while the trnnsfonners represented by the three
right columns are designed with amorphous material and wound
core with 3 limbs. The low losses in the latter transformers could
also be achieved with stacked core and oriented steel, but the
material costs would then be higher, as indicated by the white
bars in the diagram. The efficiency of a transfonner is given by
the fonnula:
ll% == 00
ro + L
Po is the no load loss (W) at rated voltage
P L is the load loss (W) at rated current
is the active power (W) supplied to the load
m is the relative degree of loading.
At rated current m = 1.
is calculated according to:
/2 is the loading current on the secondary side
is the secondary voltage in no load condition
is the voltage drop through the transfonner
referred to the secondary side
cos , is the power factor of the loading current
The relative drop is calculated as:
m (U
cos cp+ u
' sin cph 1_..Jr-
-. (-lLr-' s-in-cp---u-
-. C-OS-CP-Y
u and u are the transformer short circuit resistance and short
circuit respectively in p.u ..
It appears that in addition to the losses the efficiency depends
on the short circuit impedance of the trans fonner, the power
factor of the loading current and the degree of loading.
The efficiency of the 100 k VA transformers with the losses in
TABLE I is shown on the horizontal axis in Fig. 2. m is
assumed = 0.5 and cosf = I. As will be seen, increasing the
efficiency by 1 % from 98.44 % to 99.44 % causes 3.27 times
higher material costs in this example.
.2,38 -
... 1,96

r+ 1,00-- 1,12
98,40 98,60 98,80 99,00 99,20 99,40 99,60
Efficiency %
Fig. 2 Relative material costs at ditt:erent efficiencies
B 630 kVA Transformer
The specified losses are shown in TABLE II.
I :! I = I': I:: 1 ::0 I:: 1 I
The losses in the 4 left columns containing numbers are mainly
taken from the European Standard EN 50464-1 :2007, [2], Table
2 ami 3, column D, C, B, A. where P/P
= 900/8400 W is
assumed to be the most frequently used loss figures at present.
1be 3 columns to the right contain loss figures that are about
20 to 14 % lower than those in the next foregoing column to the
left. The total losses in the last column to the right are about
.one third of the total losses in the first number column to

-. 6
4 ., 14
2.4tn .,n


I i

Losses P o/PL [W]
Fig. 3 Relative material costs for transfonners with different
levels of losses.
The material costs of the transformer with PiP L = 900/8400 W
is used as the reference = 1. The diagram shows that the material
costs increase with decreasing loss levels. For a transfonner
with P DIP L ;;;:; 290/2800 W, the material costs are 2.81 times as
high as the transformer with P DIP L = 900/8400 W. In other words,
to bring down the total losses to one third of the losses that are
assumed to be typical today, increases the material costs by a
factor of2.81 in this example. These high material costs will in
service be compensated by lower expenses for the energy needed
to cover the losses in operation, depending on energy prices,
the degree ofloading of the transformer and the number of years
the transformer will be in service.
The transformers represented by the four left columns in Fig. 3
are designed with stacked cores and conventional oriented laser
scribed steel, while the transformers represented by the three
right columns are designed with amorphous material and wound
core with 3 limbs. The low losses in the latter transformers could
also be achieved with stacked core and oriented steel, but the
material costs would then be higher with the present relation
between material prices, as indicated by the white bars in the
The efficiency of the 630 k VA transformers with the losses in
TABLE II is shown on the horizontal axis in Fig. 4. As will be
seen, increasing the efficiency by 0,64 % from 99.05 % to 99.69
% causes 2.81 times higher material costs in this example.
8 2,5

E 1,5

99,00 99,10 99,20 99,30 99,40 99,50 99,60 99,70 99,80
Fig. 4 Relative material costs at different efficiencies
Instead of specifying specific loss values for standard
distribution transformers, capitalized loss values may form an
essential part of the basis for the dimensioning of the
transformers. Capitalized loss values means that the cost of the
electric energy needed to cover the losses in a transformer in
service is taken into account.
The reasoning is:
A transformer can be designed to achieve a lowest possible
manufacturing cost as the only target. Such a transfonner has
relatively high losses, only limited by the cooling costs involved
to keep the transformer temperature within the permissible
limits. Consequently the costs in service will be high, but the
service reliability is not affected by the losses. Depending on
the number of years the transformer is going to operate, the
sum of the running expenses over a number of years may exceed
the acquisition price. The task then becomes to design a
transformer that minimizes the sum of the manufacturing costs
and the running costs. Spending more materials reduces the
losses and the running costs, but on the other hand increases
the manufacturing costs. An optimizing computer program is
used to find the lowest sum of the manu fact uring costs and the
running costs. This sum is often called the total transfonner
costs. To enable the manufacturer to run such a program, the
user must provide capitalized loss values.
To calculate capitalized loss values the following formulas may
be used:
Capitalized loss value for no load loss =
8760d (I+!f.;)" -,l,oO
I +..l!.)"
Capitalized loss value for load loss =
1;'"".8760d (I+!f.;lr" I
I+_i)" (I+_Z J2.(I+L)_1
100 100 100
(How these formulas are derived is explained in [1], p 88- 98,
of the bibliography).
In these formulas the meaning of the symbols are:
d the average energy price at the transformer location in
the power system the first year in operation [$/kWh]
the average annual inflation rate [%]
the average annual increase in energy cost rate [%]
the average increase in loading current per year [%]
Ieqlnlt is an equivalent loading current the first year in operation
that is defined as
Ieqinil =
where I, is a current within a small time interval ? t. The
year is divided into a large number of such intervals, each
interval being so small that the current can be regarded as
nearly constant within each interval.
n needs a little longer explanation. The design that gives the
lowest acquisition price is regarded as a basic price. To obtain
an alternative design with lower' losses, an additional price has
to be paid, which may be regarded as an additional investment.
n is the number of years that will elapse before the additional
investment is earned in the form oflower running costs. n is the
only parameter in the loss capitalisation formulas the user or
investor is free to choose. All the other parameters are subject
to estimat ion based on the best possible judgement of the future
A numerical example with the following assumptions:
d= 0,12 $/kWh
IIeqintt = 0,33 p.u.
p= 7,2
n= 10 years
This gives the following capitalized loss values:
No load losses: 12 $IW
Load losses: 2.4 $IW
These values comply well with those used in some countries.
(Any other than $ can of course be used in the loss
capitalization fonnulas).
Alternative calculations with n = 5, 10 and 15 years while
keeping the same value of the other parameters give the
following table:
n a 0/0
Capitalized no load loss Capitalized load loss
r$/W) ($/WI
5 22.1 5.5 0.76
10 12.2 12 2.4
15 9.0 20 6
The table contains a column for a, which is the annual interest
rate on the amount invested to achieve lower losses than the
losses of the transfonner having the lowest acquisition price. a
depends on n and on i according to the fonnula:
lOO( i)n
ao;., = --;; 1 + 100
It appears that the capitalized loss values increase when n
increases, while the annual interest of the additional investment
decrease. Or said in another way, the economical value of the
losses inaease the more years the purchaser is willing to wait
before the additional investment in lower losses is earned in the
form of saved running costs of the transformer. A reasonably
acceptable annual interest rate could also be used as the freely
chosen parameter to detennine the capitalized loss values.
A. 100 kVA Transformer
The results fnun thf" .. t;n" t" .. t..., ..
and relative material cost shown in TABLE IV:
n No load loss Load loss ReI. material
[W] [W] costs
5 168 1728 1.27
10 62 1193 2.04
15 80 629 3.46
It can be noticed that the optimized losses decrease when n
increases, which is logical. Further it appears that when n = 15
(corresponding to a = 9 %), the optimized losses are lower than
the alternative with the lowest specified losses in TABLE 1.
The material costs of the transfonner with P g1P L = 260/2150 W
is used as the reference = I. In the case when n = 5, laser scribed
oriented steel is chosen. The total costs (material cost plu::; the
economical value of the losses) are the same when using
amorphous core material and wound core. However, the material
costs are lowest with o.iented steel. In that situation purchasers
are inclined to choose the. alternative with the lowest acquisition
price. Amorphous material is used in the core in the two other
cases since it was found to be the most favourable solution under
the given conditions. Looking at the second line of TABLE IV
(n = 10), the material costs are doubled compared to the
reference transfonner in the first column of TABLE I, while the
losses are nearly halved. The accumulated expenses due to the
losses during 10 years have been calculated based on the same
values for d, p, z etc. as before, and the difference between
the reference transformer and the transformer based
on the capitalized losses with n = 10 and the result is illustrated
in Fig. 5.
S 7000



... >(
tIS a,)


o 2 3 4 5 6 7 8 9 10 11
After number of years
Fig. 5
The transformer based on capitalized losses has nearly 7000 $
less accumulated expenses due to losses after 10 years in service
compared to the reference transformer. The excess price for the
optimized transfonner based on loss capitalization might be
earned in fonn oflower expenses due to the losses in the course
of 6 - 7 years. After that increasing savings continues for the
rest of the lifetime of the transfonner.
B. 630 kVA transformer
The results from the optimization give the transformer loss es
and relative material cost shown in TABLE V:
n a% No load loss Load loss ReI. material
[W] [W] costs
5 22.1 220 5637 1.78
10 12.2 261 3738 2.25
15 9.0 295 2738 2.92
Also for this transformer the total losses decrease when n
increase. The most economical solution in all 3 cases is wound
cores with 3 limbs and amorphous material. The relative material
costs increase with increasing n. The basis for the relative
material costs is the transformer in the first column to the left in
TABLE II having Po = 900 Wand P
= 8400 W.
Results of calculations of the expenses that accumulate in service
because of the transformer losses are shown in Fig. 6.
--t- Capitalized,
- - - - - - - - - - - - - - - ~ ~
20000 0
-Ii- Capitalized,
- - - - - - - - - - - ~ ~ ~ ~ ~
-ir- Fixed, oriente
- - - - - ~ ~ ~ ~ ~ ~ ~
0 5 10 15
After number of years
Fig. 6
The calculations are made for three different 630 kVA
transformers and based on capitalized values 12 and 2.4 $!kW
for no load and load losses respectively:
1. One with laser scribed oriented steel in a stacked core,
based on 12 and 2.4 $IkW;
2. One with amorphous material in a wound core, based on
12 and 2.4 $IkW capitalized losses;
3. One with laser scribed oriented steel in a stacked core,
based on fixed losses, 900 W no load losses and 8400 W
load losses at nominal current. This is also the reference
transfonner with the relative material costs = 1.
It will be seen from Fig. 6 that the transformer with the fixed
losses (indent 3) after 10 years has accumulated running
expenses due to the losses that are nearly 38000 $. This
transfonner has also the lowest acquisition cost and relative
material costs = 1.
The transformer with stacked core (indent 1) has after 10 years
accumulated running expenses due to the losses of almost 22000
$ and has relative material costs of 1.82.
The transformer of indent 2 has the lowest accumulated running
costs, slightly below 15000 $ after 10 years operation. However,
it has the highest acquisition cost with its relative material costs
of 2,25.
From a technical point of view it is easy to make
transformers with considerable lower losses than those
which are frequently delivered today. 60 -70 % loss
reduction is well attainable;
The low loss transformers have 2 -3 times higher weight
and have larger total outer dimensions;
The acquisition price will be ill the order of 2 -3 times as
high as for usual transformers today if the loss reduction
shall be of significance;
The higher acquisition price will be compensated by
reduced running costs due to lower losses in the course of
5 - 10 years. After that the lower losses will give a net
revenue in form of lower expenses that is higher for the
low loss transformers compared to the transformers with
losses normally installed in the near past. By replacing old
high loss transformers with new low loss transfonners this
revenue becomes still higher;
Given that drastic energy reduction is imperative in order
to save the global climate, low loss transfonners have to
be chosen on other grounds than solely on short term
profitability aspects.
When acquiring transformers it is strongly recommended
to base the choice on the total transfonner cost, which is
the sum of the acquisition price and the expenses to cover
the losses in service. .
In order t<r-make conscious the importance of the losses, it
is recommended to base the enquiry on capitalisation of
the losses as described in clause V rather than specifying
fixed losses. Use at least n = 10 or preferably n = 15.
[1] ABB Transformer Handbook, 3rd edition.
[2] European Standard En 50464-1 :2007 Three-phase oil-
immersed distribution transfonners 50 Hz, from 50 kVA
to 2500 kVA with highest volt age for equipment not
exceeding 36 kV - Part 1: General requirements .
. --
Anjull Chandra
1. Introduction
Distribution transformers operated and owned by electricity
distribution companies are responsible for supplying 70010 of
electricity to final Users at low voltage level and represent almost
98.5% of the transformer stock (numbers) installed by the power
utilities in the country. There are about 3457318 No. of
distribution transformers with a total installed capacity of
26253 7MVA(as on 31.3.2007) in the power Distribution system.
The state wise details of Distribution transformers is given in
Annex-I. Although the transformers are considered to be
efficient devices still roughly 1-3 % of distribution transformers'
throughput electricity is lost due to their inefficiencies. It is
estimated that about 18 billion kWh of energy is lost in
distribution transformers every year equivalent to about 6-8 %
of the total loss from the system. The distribution transformers
are growing at a rate of about 9.5% per year and about 3.5 lakh
transformers are being added every year. The trend toward
generating electricity at sites close to the point of use and
adoption ofless LT High Voltage Distribution Systems (HVDS)
suggests an increasing installation of small capacity transformers
in the network. As a result, losses from distribution transformers
will represent an even larger share of electricity use.
2. Energy losses in Distribution transformers
The energy losses in distribution transformers fall into two
components: No-load losses or iron losses (resulting from
energising the iron core; this phenomenon occurs 24 hours per
day, 7 days per week, over the lifetime of the transformer, 20-
25 years in average) and load losses (arising when providing
power to a user, from the resistance of the coils when the
transformer is in use, and for eddy currents due to stray flux).
The magnitude of these significant losses attributed to
distribution transformers is a consequence of:
Inefficiencies due to the design, materials and technologies
used in distribution transformers;
Inadequate type and rating of transformers selected for a
particular application.
Harmonics in the distribution network
Although manufacturers use well-established technologies and
designs, the distribution transformers are being manufactured
in .relatively low-cost facilities/small scale industries also. For
this reason the manufacturing sector supplying transformers -
has always focused on being very price competitive and in the
process sometimes quality and safety considerations are
sidelined to remain competitive. This has been resulting in
inefficiencies in the design, material and technologies and supply
of poor quality and high loss transformers to the utilities.
Technical solutions exist to reduce transformer losses by 75%
at minimwn (when replacing with modem transformers).Energy-
efficiency can be improved with better transformer design
(selecting better, lower-core-Ioss steels; reducing flux density
in a specific core by increasing the core size; increasing
conductor cross-section to reduce current density; good
balancing between the relative quantities of iron and copper in
the core and coils; and so on.), or by tile arioplion or t1Hj I
amorphous steel transformers. High-efficiency transformers are
available and are already in limited used by the utility and
3.0 Benefits of energy efficient transformers
The main reasons for considering an initiative on energy-
efficient transformers are given below:
I) The large number of distribution transformers in use and
the fact that most of the electric power generated
continuously passes through them implies that even small
improvements in transformer efficiency can result in
substantial energy savings.
2) Despite high average efficiencies (from 95 to 99%),
distribution transformers have a significant environmental
impact because they continuously consume power. The
environmental benefits of energy-efficient transformers are
very high.
3) Reduced physical size (unit dimensions);
4) Reduced transformer heating, hence lower need for
additional cooling or insulation
5) Longer operating lifetime.
6) Low-loss transformers also better withstand electronic
(harmonic) loads.
Despite the perceived benefits of energy-efficient distribution
transformers, take-up of efficient. transformers has been low.
Producing energy-efficient transformers is not a technical
challenge; it is the commercial considerations which prevail in
most of the utilities which are difficult to penetrate. Higher
efficiency transformers increase up-front costs. Since energy
losses are directly charged to customers via tariffs, reduced
losses do not necessarily b e ~ e f t the investor. It is a challenge
to convince utilities that, although t h ~ initial price associated
with these transformers is higher, the overall cost of using them
can be lower.
India is facing significant growth in electricity demand and could
benefit greatly by installing energy-efficient transformers. The
savings potential from distribution transformers may appear too
small and too difficult to achieve at the individual level bur
when considered for the country ""as a whole it represents
considerable savings in losses. A 0.5 % increase in efficiency
of all the distribution transformers in the country would result
in an annual loss saving of 2 to 2.5 billion kwhr. This would
mean emission reduction of about 1800 thousand tones ofe0
per year. However it needs to be considered that Utility's
investment in energy-efficient distnoution transformers has an
economic payback time between 1.5 and 8 years, and an internal
rate of return in the range of 20% to 50% depending upon
whether it is replacing an old inefficient transformer or
procurement for new installation.
In the XI plan we are planning to add about 78,000 MW of
generation capacity in the utility sector and corresponding
distribution transformers capacity would require to be added.
lilt; uU\'t:mment of India introduced the Accelerated Power
Development and Reform Programme (APDRP) and the
RGG VY programme under which improvement in the sub-
transmission and distribution system in the country and
electrification of un-electrified villages and strengthening of
the Distribution systems in rural areas is being taken up.
Addition of Distribution transformers by the state power utilities
on a large scale is being undertaken under these two programmes
as well as with their own funds. It is estimated that Distribution
transformer capacity addition would be about 2,72, 000 MVA
in the 11 til Plan. This projected requirement is on the lower side
and it is expected that additions are going to be much higher. If
all these new procurements are energy efficient it would result
in significant energy savings for the country.
4. Change Requirements in Procurement policies/
procedure of utilities for promoting use of energy efficient
transformers .
The era of increased competition for capital has commenced.
As utilities enter an era of increased competition. they are trying
to find ways to cut costs. Market deregulation is forcing utilities
to reduce capital budgets. The state based regulators have
reduced incentives for over-investments, however, they still
request enhancement of quality of supply and improvements of
customer services. These. sometimes conflicting requirements
have forced the electrical distribution utilities to move focus
from improving economic efficiency to minimizing capital
expenditure in the short-run. It seems that "low initial cost"
method (without proper assessment of total life cycle costs)
appears to be much more attractive solution to the utilities for
selection of distribution equipment. Although power
transformers are generally purchased after considering losses
over the I ife cycle of the transformer, Distribution transformers
are generally purchased based on least cost basis.
4.1 MEPS for Distribution Transformers
The transformer market is extremely competitive and the sector
involves a limited number of professional buyers, already
reasonably aware of the arguments for energy efficiency but
need some sort of push to change their procurement policies
and procedures. The procurement policies of the utilities should
be aimed at absorption of energy efficient techniques by using
various economic and financial instruments. In India, unlike in
many countries around the world, there is no general mandatory
standard or even voluntary approach for energy efficiency of
distribution transformers. The main document which describes
losses in transformers is IS1180. Mandatory Minimum energy
performance standards can enable achievement of some measure
of energy efficiency with immediate effect depending on the
standards set. However the Minimum Efficiency Performance
specifications in our country (Maximum allowable losses)
dermed as per IS 1180 are not high enough to promote purchase
of energy efficient transformers and need revision. Today
transformers with much lower losses are available in the market
The Bureau of Energy efficiency has come up with a labeling
programme with energy star ratings of I to 5 for Distribution
transformers with star 1 having lowest efficiency.and star having
highest efficiency. The Energy Star Transformer Program is a
voluntary energy efficiency programme designed to encourage
utilities to purchase and install high-efficiency, cost effective
transformers in their distribution systems. The MEPS should
be set so as to remove the worst products from the market. It
would be desirable if MEPS in IS 1180 itself were brought to
star 3 rating to remove the worst products from the market and
the utilities can gradually move towards star 5 rating. This can
be effectively used where there is a desire to use energy efficient
transformers, but not the time to go into complex specifications
and tender evaluation process. North America, Canada,
Australia, USA, China, Japan have all set MEPS for their
Distribution transformers.
Central Electricity Authority has prepared Specifications for
three phase and single phase oil filled Distribution transformers
for providing guidance to the utilities in procuring energy
efficient and good quality Distribution transformers. In the
specifications the maximum allowable losses (MEPS) have been
set out for 3 phase transformers for ratings from 10k VA to
2500 kVA and for single phase from 6.3 kVA to 25 kVA. The
values ofMEPS for transformers of3 phase with ratings between
16 kVA to 200 kVA have been taken from the star ratings of
BEE and for other ratings have been set at efficiencies of star
three level. With these MEPS it is hoped that utilities will move
towards buying transfonners with lower losses.
However the MEPS method does not provide manufacturer with
an incentive to provide higher efficiencies than those specified
and these in most cases will not be set high enough to achieve
full economic and environmental benefits. For this the TOC
evaluation methodology, is used by some utilities as it provides
a balance between cost of purchase and cost of energy losses.
The total owning cost to a utility for any transformer is made
up of two major elements:
1. The cost of owning the transformer. This includes the
purchase price, installation cost, ~ s i d u l value or cost at
the end of the life, as well as the effects of interest,
depreciation, taxes and other factors.
2. The cost of losses incurred by operating the transformer
discounted to present value.
Electricity losses over the lifetime of the transformer cost two
to three times more than the original purchase price of the
transformer. The formula adopted for TOC calculations is the
REC formula which captures the hrs of usage, cost of energy
(electrical tarift). equipment life (years of expected service) and
cost of money (rate of return). Today this evaluation method is
not being used extensively for purchase of Distribution
transformers, due to uncertainty in values of various parameters.
However, well done TOC calculations can lead to the
procurement of more efficient transfonners with reduceO iosses
both in the cores of the transfonners and in the coils.
5.0 Quality assurance for Distribution transformers
All electrical equipment is subjet to in-service failures and a
distribution transfonner is no exception. The failure rate of
Distribution transfonners in our country is as high as 20- 25%
for some of the utilities which is matter of concern. The
Distribution Transformer plays a vital role in the process of
servicing electricity to consumers. It is one of the last
components in supply chain of electricity to the consumers and
it's perfonnance affects the reliability and quality of power
supply to the consumers to a large extent. The failure of DTs in
the system results nol only in disruption of supply to the
consumers but also a significant loss to the licensee. Its outage
causes great inconvenience in network management and
involves high expenditure on account of repair/replacement. In
fact Distribution transformers represent a significant cost to
electric utilities, both as a capital investment anQ as an ongoing
operating expense and can account for approxiniately 9 percent
to 20 percent of total distribution capital spending in a year.
Any Utility, therefore, has to take all possible measures to reduce
downtime / failure of transformers to a minimum and to extend
their lives, at the most economic cost.
It is, therefore, necessary to improve design, raw material and
manufacturing process and procurement practices of distribution
transformers to improve their quality and reliability. The utilities
have to take following measures:
(i) Continuous up gradation of specification of transformers
and bringing necessary stringency in it, based on service
(ii) Exercise regular stage inspection and testing of each and
every transformer before it leaves factory
(iii) Carry out acceptance tests at site
5.1 Design Stringencies / Special requirements:
Quality control activity starts from the stage of formulation of
specifications by the utility. Stringent specifications would
ensure that quality transformers are procured by the Power
Utilities. Some of the features which could be stipulated to make
the specifications more stringent are:
(i) Every material/component should have material
specification to ensure quality and reliability. The
specification should clearly indicate the characteristics and
the tests to be carried out for verification of the quality
and reliability of the selected raw material & component.
(ii) Keeping provision for occasional unpredictable
overloading without affecting the service life, oil and
winding temperature rise limits specified in the
specification are made more stringent. In the specifications
prepared by CEA it has been specified that the temperature
rise over ambient shall not exceed the limits given below:
- Top oil temperature rise measured by thermometer: 35C
- Winding temperature rise measured by resistance : 4()OC
(iii) Keep provision for over fluxing in the design. The
transfonners core shall be suitable for over fluxing (due to
combined effect of voltage and frequency) upto 12.5%
without injurious at full load conditions and shall
not get saturated.
(iv) Specifying use of transfonner oil conforming to IS 335
and ensuring that recycled oil is not used by specifying
that the specific resistance of the oil shall not be less than
2.5 * 10
ohm-em at 27C .when tested as per IS 6103.
(v) By specifying electrical grade insulation of epoxy dotted
Kraft Paper! Nomex
(vi) Specifying Temperature rise test, Short Circuit Test and
Impulse Voltage Withstand Tests, pressure relief device
test etc.
(vii) By directing bidder to submit Quality Assurance Plan
giving the following:
Statement giving list of important raw materials, names of
sub-suppliers for the raw materials, list of standards
according to which the raw materials are tested. List of
tests normally carried out on raw materials in the presence
of bidder's representative, copies of test certificates.
Information and copies of test certificates as above in
respect of bought out accessories.
List of manufacturing facilities available.
* Level of automation achieved and list of areas where
manual processing exists.
List of areas in manufacturing process, where stage
inspections are nonnally carried out for quality control and
details of such tests and inspection.
List oftesting equipment available with the bidder for final
testing of equipment along with valid calibration reports.
These shall be furnished with the bid. Manufacturer shall
posses 0.1 accuracy class instruments for measurement of
* Hold points for purchaser's inspection
5.2 Quality control of raw material
Higher quality of raw material means better quality product with
lower failure rates. With shortage all over the world and very
high cost ofCRGO, some eleplents are importing
defective material for the use of manufacture of distribution
transformers. This material being used for the manufacture of
distribution transformers is causing colossal amount of losses
per annum besides burning of distribution transformers.
To prevent the use of seconds material it needs to be ensured by
the purchaser i.e the utility that raw material such as core
stampings used have been supplied by standard manufacturers.
The utility has to ask the supplier to furnish the manufacturers'
test certificate as well as the proof of purchase from these
manufacturers (excise gate pass). The purchaser should furnish
following documents along with their offer in respect of the
raw materials viz core stampings, winding conductors, insulating
paper and oil :
i. Invoice of supplier.
11. Mill's certificate.
iii. Packing list.
iv. Bill of landing.
v. Bill of entry certificate by custom.
5.3 Ensll,ing f,",lily o/tTflnsfonners during manllfacture
To ensure about the quality of transformers during manufacture,
the utility has to carry out inspection at following two stages:-
(i) On line anytime during receipt of raw material and
manufacture/ assembly whenever the utility
(ii) At finished stage i.e. transformers are fully assembled and
are ready for despatch
Quality control at manufacturing stage has to be augmented
substantially by a well equipped laboratory capable of carrying
out various routine tests and type tests specified in various
national!' international Each and every transformer
has to be thoroughly tested befor.e it leaves the factory. At least
10% transformers of the lot (minimum of one) should
be subjected to routine/acceptance test in presence of
purchaser's representative at the place of manufacture before
dispatch without any extra charges.
5.4 Ensuring quality t"roug" tests at site
The purchaser should conduct all tests on transformers selected
randomly from a lot after arrival at site and the manufacturer
should guarantee test certificate figures under actual service
conditions. The utility can also get short circuit tests carried
out at established laboratories with CPR! / ERDA on randomly
selected distribution transformers after they are delivered to
stores, without any prior intimation to the manufacturer /repairer.
Since this is a costly proposition, such tests could be carried for
each manufacturer once in every 3 / 4 years.
6.0 Conclusion
Highly efficient, yet cost-effective and quality distribution
transformers, which arc fully optimised for the expected service
conditions are the right solution for power utilities. Energy
efficient transformers are not being purchased because of the
assumptions built into utility investment decisions, regulatory
practices and tender evaluation methodologies adopted for
purchases. It is high time that we revise our minimum efficiency
performance standards and progressively move towards total
owning concept for evaluation and purchase of Distribution
transformers. Quality control has to start right from the stage of
preparation of specifications, to inspection of raw materials as
well as during manufacture as well as proper testing before
despach and on receipt at site. Introduction and use of such
equipment presents a significant challenge for the electrical
distribution utilities in the country and right environment has
to be created for supporting them through regulatory and
economic instruments.
Annex- I
StatelV.Ts. Number Aggregate
Capacity (KVA)
Haryana 155609 10910461
Himachal Pradesh 16949 1260296
Jammu & Kashmir 5661 883910
Punjab 252222 16231744
Rajasthan 367111 16046719
Uttar Pradesh 355271 19394270
Uttarakhand 30493 1953804
Chandigarh 1416 458368
Delhi 28448 9350597
Central Sector(NR) 500 636113
Gujarat 289302 23696946
Madhya Pradesh 378351 30561008
Chhattisgarh 47236 3551974
Maharashtra 282555 32238708
Goa 3653 446997
D. & N. Haveli 836 135256
Daman & Diu 460 98444
Central Sector(WR) 24 35500
Andbra Pradesh 529400 27082389
Kamataka 272791 22139094
Kerala 40603 5208474
Tamil Nadu 173073 22197412
Lakshadweep 84 6849
Puducherry 2 4000
Central Sector(SR) 228 222150
Bihar 37208 3014718
Jharkhand 11472 1137732
Orissa 50349 3426695
West Bengal 88012 6077231
D.V.C. 146
A.& N. Islands
538 82684
Sikkim 1274 63337
Central Sector(ER) 166 259520
Assam 19086 1802080
Manipur 2882 326601
Meghalaya 5053 405553
Nagaland 1192 345819
Tripura 5105 436491
Pradesh 160-" 117946
Mizoram 997 115319
Central Sector(NER) 13 6350
Maximum allowable losses for 3 pbase transformers at rated voltage and rated frequency permitted at 75 C.
Voltage Ratio
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
110001433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
11 000/433 - 250 V
11000/433 - 250 V
11000/433 - 250 V
330001433 - 250 V
33000/433 - 250 V
Voltage Ratio
33000/433 - 250 V
330001433 - 250 V
33000/433 - 250 V
33000/433 - 250 V
33.000/433 - 250 V
Rating (Kva) 3 Star
Max. Losses Max. Losses
at 50% at 100010
loading loading
(Watts) (Watts)
16 150 480
25 210 695
63 380 1250
100 520 1800
160 "770 2200
200 890 2700
Rating Max. Losses Max. Losses
(kVA) at 50% at 100%
loading loading
(Watts) (Watts)
10 98 300
250 1050 3320
315 1100 3630
400 1450 4630
500 1600 5500
630 2000 6640
1000 3000 9800
1250 3600 12000
1600 4500 15000
2000 5400 18400
2500 6500 22500
100 560 1820
160 780 2580
Rating Max. Losses Max. Losses
(kVA) at 50% at 100%
loading loading
(Watts) (Watts)
200 900 3000
315 1300 4300
400 1520 5100
500 1950 6450
630 2300 7600
4 Star 5 Star
Max. Losses Max. Losses Max. Losses Max. Losses
8t50% at 100% at 50010 at 100%
loading loading loading . loading
(Watts) (Watts) (Watts) (Watts)
135 440 120 400
190 635 175 595
340 1140 300 1050
475 1650 435 1500
670 1950 570 1700
780 2300 670 2100
330001433 - 250 V 1000 3450 11350
33000/433 - 250 V 1250 4000 13250
33000/433 - 250 V 1600 4850 16000
330001433 - 250 V 2000 5700 18500
33000/433 - 250 V 2500 7050 23000
Maximum allowable losses for 1 phase transformers at rated
voltage and frequency and at 75 C
Voltage Ratio Rating Total losses Total losses
(kVA) at 500/0 at 100 %
loading in loading in
Watts (max) Watts (max)
at 75C at 75 C
11/.../3 kV 1230 V 6.3 51 130
111.../3 k V 1 230 V 7.5 57 15-5
111.../3 kV I 230 V 10 65 200
111.../3 kV I 230 V 16 100 250
11/.../3 kV 1 230 V 25 110 350
11 kV I 230V 6.3 51 130
11 kV I 230V 7.5 57 155
11 kV/230 V 10 65 200
11 kV/230V 16 ' 100 250
11 kV/230V 25 110 350
Roman Targosz
Intelligent Energy TI Europe
Item Methodology and SOUIC8
Utility: 30% of population from SEEDT, rest from MMA report [7.4] and own estimates. Dry transformers have been excluded
as marginal for utility -1 % population share in EU average
Dry type 100/0 from SEEDT, rest from market reports [7.7], [7.81
Transformer Industrial oil filled - 10% from SEEDT, rest resulting from energy balances of non utility energy use and own estimates
SOUICeS: SEEDT country questionnaires, country reporls, MMA report on Characterization of European Transformer
Populations, Frost SuHivan, Goulden Repotf, KEMA Enstgy saving in industrial transformers [1.9J
Generally units above 2500 kVA have not been oonsidered (it is estimated that these account for less than 0,2% -10.000
pes) of total population and less than 1% of installed MVA
Similar as fleet, in general for utility, rate of population increase (average for 5 years) has been considered, in case of dry,
Frost and Goulden (7.71, (7.8] reports have been considered, in case of industry oil, mixed approach
Poland, Hungary detailed country statistics, few data from SEEDT, 30% deep structure analysis (e.g. assumptions about
energy used in drives and electric arc heating on MV have been made), rest based on Eurostat IlEA I Sources: Euroelectric
statistics; EUROPROG 2005, Eurostat 1999-2004, lEA 2004, Euroelectric publications 1997 and 1999 on structure of energy
use in Europe. Polish Energy Statistics 2006, other statistical data
Abo'"'t 25% from SEEDT. rest computed as average. In general the feeling is that the model is rather conservative some
Rated losses
SEEDT reports present losses which seem average of last few years rather than full age spectrum losses. In some cases
additional statistics reports about country transfonner losses or proportion of no load and load losses have been used for
cross checks.
Computing load
Woodrow. BuJler{1] model has been used to calculate time of peak loss from time of peak load. This model is regarded as
rather overestimating 't but less sensitive to COScp and to variation in marginal values
All values; populations, market, losses, consumption when possible are referred to year 2004
Analysis covers EU 27 countries and Norway (covered by SEEDT). Norway has been exclu.:led from losses balances. EU 27
Other issues
represents -80% of final electricity use of overall European 43 countries (excluding Russia and Caucasus but including
former USSR European Republics)
For coscp va lue of 0,9 has been used for utlity part and 0,85 for industry part
I. {ll To calculate lime Of T based on previouslycalclMled broad scale required by SEEDT_ applied one 01 the models listed in Annex
Fleet etJ.27
pes WI>.
< 400 WA 73% 34%
sedor oil' 5830 kVA 23% 48"4
>630WA 3% 11%
"7U15 110515
Total cisHIuIIon sector
<400WA 62% 20%
industry oil kVA 23% 27%
>630kVA 16% 52%
Total industry oil fnG m
<400kVA 23% 9%
industry dry 5630 kVA 40% M
>630WA 38% 63%
114017 143104
T ataI industry dry
4652172 1383988
Market EtJ.27
pes tItIA
68% 26%
mil 45%
7% 29%
62% 20%
22% 28%
16% 52%
.01f 15m:
m 271(
16% 4%
34% 20%1
50'% 76% ,
fun 1447'

139628 57266 '
tVA .....
Fleet detailed
Losses EU27
lMtyoi lPoleet 1Wl3 IPoIIIIIteI 348
6000 l'Pkmarklt 172
IPkIeet/JJllalll 27,3% 33.0%
lndusttol IPoteM 5544 tpoMbl :lIM
IPkIMt 2161 tPlI: ..... 95
induIrydly IPoIeet
IPo ....... 7S8
IPkIeeC me IPkmarket 120
33402 1269
Losses detailed
w " E
Operating efficiency
.... ,,-....-................. -
I .. ."4' i
.. .------------; __ ...... ' ;... ---------1
- . __ a CI Ii. "M I. U' W I.Y .... JI) '" Pi IE eo 10
Rated losses referred to AC'

.. ..
No load losses/a e
,.. ...................... tc' ..... .

' .. +---..--...--......--...,...-.....---ll
.. ........ ----
c.di .................... t<'_.HD.,
... =:: II

__ llOiVA I
l ...

_ _ _ _ _ _ _ _ _ _ _ _ I'
Load losses I age
.... ....., ............ I'IIIClIA,HDGIt





- - - -- - --
.. .... ..,.,.I
'"" I
, ..

__ NIIk..,,,
' ...
__ 13Q"vA

I !
- - - -
.." ,m
Load losses I age
......... leM __ ....... 'A,HO_
''''1-. -.. -- -- .

1. +- --- - --- -

The extra losses caused by reactive power losses are
difficult to quantifY. They can be mitigated by adequate
adjustment of system reactive power. There is an empirical
indicator of energy equivalent of reactive power. In the
past it was assumed to be -0,1 kW/kVAr. Some
measurement and calculations [10] indicate that at worst
scenario the total amount of losses would be doubled.
A hove mentioned indicator for sure, can no longer be
specified at so high level; however some extra losses must
As far as extra losses due to harmonics are concerned
there are similarly no accurate calculations. This is partly
consequence of inadequate measurement of harmonics
in a system and partly dynamic situation in a system. The
KEMA report [9J indicates (based on 6 cases only) that
additional losses in industrial transformers would be
equivalent to 0,3% of electricity used. This is roughly
equivalent to about 2 TWh losses for industry only.
Hannonics losses certainly exist in utility transfonners.
Although in industry IT intensive environments are main
source of harmonics it is proven that harmonics in power
system are ger..erated by small electronic loads like TV
sets, CFL's etc. There are no serious calculations how
much are these losses in European fleet of utility
transfonners but definitely these extra losses are not zero.
Our conservative estimate is that all these extra losses
(due to reactive power loss and harmonics) are at least 5
TWh (or 15% of calculated total of no load and load
losses) yearly for all; utility and private distributioll
Technical conclusion
European d'lStrilution transformer fleet and market is dominated by traditional technologies.
These technologies have certain limits but enable substantial losses reduction when compared
Extra losses loaveragetransfOOllEfmalketeffdency/eve/.
The SEEDT calculation indicates level to total
distribution transformers losses in EU-27 at level
of about 33,5 TWh. This calculation has however
completely ignored two aspects:
Extra losses due to reactive power losses in
distribution transformers which have influence on
active power losses in the network
Extra losses due to harmonics (voltage and current
Within the techncbJ111XJS far deve/oped there are suffICient measures to produce cost
efficient transfooners m have both no-load aid load losses by about 30% lower than EU
average level AC' accoRfing to HD 428. T ransfonner manufacturers may now very
dynOOli:ally shape tansfooneI' to acoommodale life cycle optinum cost
We are a rlllle bit scep1i3 about superconducling technology in distribution transformers.
These sholdd be rather and robust machines requirill9 mininll!l mailtenance,
diagnostic etc. We are no! convinced thatei:iency gains justify price user has to pay for this
lechnologyand probably acHtional installation & maintenance. We are however impressed by
overall devek>pneIi !litis 1ec:Ilndogy, which in case of larger transformers may very well and
soon become very atIIadive softition.
Amorphous lransfonners - very promising ENDESA
Starting point Distribution transfonner population, Age distribution of transfonner losses of electricity
market sales and losses in 2004 - EU-27 distribution companies in EU-25, increasingly
Owner Number In IIeet Number oIpcs Noft.bd Load losses in
(millions) sold In market Iosseslntleet IIeet
(1WI!Iyear) (lWhIytar)
Electricity 3.7 85,. 16.0 6.0
Industrial- 0.8 38.000 5.5 12
Industrial - 0.2 16.000 2.6 1.1
Total 4.6 140,080 24.1 9.3
Total technical potentials in 2004
"Static" potential (SAT):
_______ Electricity distribution
i 25.000
companies: 57.4 %
_______ Industry-oil:61.8%
? 20.000
------- Industry-dry: 31.3 %
Total: 55.5%
(is.S TWhIyear)
EItliOty distrib\Jtion lNtustry-di IrWsIry-dry
11earidIy losses lSavin\lpoXe!tia
What part of potential can be realised by 2025?
} 4 energy efficiency scenarios by SEEDT project
Important assumptions for calculating
electricity saving potentia1s until 2025
Baselines for electricity system development
PRIMES-TRENDS: BAU development of the
European electricity system with further increase in
electricity demand,
PRIMES EEiRES: Development of the European
electricity system with strong increase in energy
efficiency and renewable energies
Baseline for invesbnent in transformers until 2025:
2004 market behaviour: frozen efficien(,)'
Replacing the oldest (worst) DTs first
I: __ __ __ __
! ........ ..l1li
1m ____ ____
j Cl'Jo t-------:1tIII-
j 301 t----=_=_
i 201
j 10'10
. .
It4S- 1950- Ie5$- I. 1965- 1915- ,. ,.. 1995- zro>. :
1M 19504 1959 19504 1& 1114 11189 1994 1999 2004 :
Energy efficiency potentials 2025 (GWh/year) -EU-27
bt scenario: ReplacenMnt of the liquid.filled transfonners by AoBk and of the d!y-type HD 538-1.
2nd scwrio: of lilt liquld-filled IrwfoI1'Atl1 by AIJA1 and of the dlY-type HD 538-1,
reduced by 10%.
3fd tcenario: the liquld..filled transformers by AMOi Bk increased by
"aod dry-type as per HD 531-1, reduced by M.
f 411! acenario: Replacement of the IiqIJid.flJed tranafotmers by OOT.ak and drytype as per HO 538-1,
General Scenario 1 Scenario 2 Scenario 3 Scenario 4
development of
electricity system
6,2 7,4 10,6 11.6
5,0 5,8 7,9 8,2
Further assumptions for calculating
electricity saving potentials until 2025
Number of distribution transformers (DTs) bought in year x
= replacement rate (= 1 I lifetime)
change in OTs due to increase/decrease in electricity consumption
change in OT s due to change in structure and operation of distribution
system (e.g. elimination of redundant DTs in some German areas)
+ change in DT s due to increase in decentralised generation,
with lifetimes (marketlfleet ratios in 2004):
Bectricity distribution companies: 40 years (market/fleet = ca. 3%)
Industry -oil: 25 years (martetlfleet = ca. 5%)
Industry - dry: 30 years (rnarketJfleet = ca. 10%)
Economic impact of scenarios
For the EU-27 average, in all scenarios net cost savings
expected compared to BAU development
Economic impact on electricity distribution companies
strongly depends on regulatory scheme (and interest
Recent price developments have led to amorphous
transformers being more competitive
Results crucially depend on assumptions about
transformer price relations, interest rate, assumed lifetimes
and expected price developments (steel, copper,
Environmental impact of scenarios in 2025 in EU-27
(Mio t C02eq./year)
Baseline: Frozen efficiency (2004 market losses)
Policies and measures leading to first savings in 2010
General Scenario 1 Scenario 2 Scenario 3
of electricity
1,1 2,2 3,2
1,5 1,7 2,4
Scenario 4
Static technical electricity saving potential in 2004: 18.5 TWhfyear
(55.5 % of current DTs' energy losses of 33.4 TWhlyear)
Highest relative potentials in industry-oil, highest absolute
potentials in electricity distribution companies
Electricity saving potentials until 2025 between 5.0 and 11.6
TWhlyear in 2025, depending on scenario chosen and on general
development of electricity system
reduction potentials until 2025 between 1,5 and 3.5 Mio t
Potentials are economical, but calculations extremely sensitive to
assumptions I price developments steel, copper)
Economic impact on electricity distribution companies largely
depends on regulatory scheme
Economic impact on industry and commerce largely depends on
assumptions with regard to electricity prices and interest rate chosen
Proposed Mechnisms
Energie .,.",.,,-.!!tM! -w"""" ILl",
F_ _ 0 ThisIobel.blsedonaHlld .... on!r.
_ 0 ..... ....
'-II.-IoIIIIIoesei(moM eIIi!.iert)"'0111libeled B-.
.' .... "IZ-.""""'"" ... t
1 t! r 31!1 ..,"""" ..... "1!!'II!ood
o This libel is bIsed on.contiRIIionafnoloailtndloail ..... II..". IcadiJ,j it.
. ,.",.,3 -. IobtIIIaHd!ll! IPPt!ljdRplClW "' ..... 1!I1H!I!a ..
o TdalloaesNoloadl.oaes'1I3l.c1dL-.
TherationaledtlUbIIIlIa .. ULI--------,
- =1'eu;.;:;:d=lOI=mt_f-----l :::::.,
350 :===-... 4,..,..,
.... !t...... lG:
...... 1Q:
..... - :
1lIo...",d ... _ ... ...," .. , ...... """"
m.tt= fI'(J)dx=+Sr'-AI-tBt'
o ooncidI1g,lIlellJlllalelficienaesforllle""'*warielyallOadingsfiomOlo 1 call
be t'l)llSled IS Nll.,I3LL fcmUa
Labelling - comparison
HO 428 or
EN 50464 tech speak Proposal 1 Proposal 2 Proposal 3
8+/80 8 8
iAoAk co C B
Bk CC'-30% CO c/o C
lAoCk c- OlE E
lAook C- F G
BoBk DO 0 0
BoOk 0- FIG G
CoAk E+ 0 C
CoBk CC' EO E 0
DoAk F+/G+ E 0
DoBk CB' F+/G+ F E
Dook BB' FO/GO G G
EoAk G+ F E
EoBk CN G+ G F
EoCk AA' G+ G G
Labelling through integration of losses:
Proposed similar to appliances
through Inta.nt!on
LETTER I Integral of losses to reference
P .. = (NLL+O.333-LL) I (C.+O.333-B.)
C I 0.82 < P", IREfn S 0.92
I 102<P IREF 112
,,- "'- ... ",,,$'l<; ... _ ""'.,.-.
co: Class of no load losses as per EN 50464
Class of load losses as per EN 50464
CoBk = CC' of HD 428
Efficiency standards ,
Will Europe catch up with the US and Japan?
-'e - --
.. DoE USA min. TOC - 8H15 s.n.. AmoqIIhouI
_euz,_ .... 1.1:5_
Mandatory efficiency standard
European DT manufacturers are not interested in a voluntary
A mandatory EU-27 minimum efficiency standard will remove
the worst DTs from the'market
Only feasible if, at the same time {I), regulation of electricity
distribution companies removes any disincentives
It can be designed in several ways (preferred standard in
bold, if no labelling will be introduced):
t maximum allowable no load and load losses (CoCk), or
minimum efficiency at particular loading, or
just removing the worst labelling classes from the market
Mandatory standard
.. -----..-..... -.... -.... -. ;- ..... -... .. -.... -... -: .. ---'.J
IOOIXI ...... 4-;:;;.;e .. -..,------- ..---------
.. .... .. .. t ........................................ .
22SCI) --.--.--.--.. ----.-----.--f-----'-
2faII -------
11510. _._ ................... _ ....... 4IlO
l!U1l .... _ ... _ ............. _ ..... __ ._... . ....................... ....630
IllCID ..--__ . __ ... ....: ......... .:..; ...-.:.:.:: ..... ::+. ...... ..... = ....... .. .. ..-.-=+--::::4i ... = ...... -.
7!GO .. *"" .. .. ... .,.,... ... ..,... __ ..........-.>:_ . = ..... ___ . ..,..._ ....--t. __ >-._.------b.._=_--_+-_-+._-4_--._+-_ -...= ..
2!GO ...................... -----.... - ..... , .-.-.-.
Policies and measures
Different market actors face different
barriers and obstacles
Large electricity distribution companies
Large industry
Smaii and medium eieclnclty QlstnOUllon compames
Small and medium industry and commerce
Engineering firms, ESCOs, energy consultants,
Transformer manufacturers (and their suppliers)
Regulation of electriCity distribution companies
Reporting on losses I benchmarking (e.g. by using
labelling scheme) .> largest potentials first
Deviations from loss target could be rewarded I penalised
Incentive scheme should allow sufficient payback period
for investment
Maybe specific energy efficiency investment budget
outside the cap
At least existing disincentives should be removed AND
direct financial or fiscal incentives during transition
period as long as incentives are not induded in the
regulation scheme

Regulation of electricity distribution
companies Chances for implementation?
> How to convince the who concentrate(s)
on other issues at the moment?
> How to set it on the agenda of CEER I ERGEG?
) Real chance for implementation?
> Chance if addressing network losses and
distribution system optimisation in a more general
R&D and AMDT pilot projects
Increased interest at least by
ENDESA (and EDF) into pilot
projects with amorphous
distribution transformers
, European AMDT pilot project
with support from European
Commission (-> Strategic
Energy Technology Plan;
European Investment Bank?)
R&D support: From efficient
grid components like
distribution transformers to
efficient distribution systems
Source: Endesa .
(Test of 10 amorphous distribution
transfonners in Mallorca in 2008)
Information, motivation, advice
programmes etc.
Buyers and users information:
Inclusion into general energy advice programmes and
sector-specific energy concepts
Inclusion into general information, communication and
qualification on energy efficiency
Inclusion into information and marketing by
manufacturers and ESCOs
SEEDT TLCalc calculation tool for buyers
Bamers towards implementing the
proposed pOlicies and measures
Ambitious policy instruments proposed
Low replacement rate of distribution transformers (long IWetime)
Low energy saving potentials compared to some other energy
efficiency technology
No complete lifecycle analysis yet: not yet fully ready for
Not much interest by European manufacturers and distribution
companies yet
, Continuously changing regulatory schemes: Any planning and
calculation of investment possible? Any interest or acceptance by
regulators I CEER I ERGEG expectable? - Position Paper
Political differences between EU Member States more severe than
differences between states in US (difficult for setting a standar.Ql
Pilot Project EFFITRAFO ENDESA (.January 2008) :
El 1 ANLMRDS: :\o-load LOSSl'S LeHls
Pilot Project
___ rnor-

eu-zTSavIna ...... (FtM_ .. ___ to_1ooNI
Emission reduction: Effitrafo vs
Spanish Standard 2008
1[(\1\01.0(;\ fT\() \'1 f\TAtS: [,olution of the eGO SteelTechDology
idll1e5 .oN 0.27
013 0.18
..... "'JIO ...... .
CGO: Cold-rolled
grain oriented silicon
HiB: High
permeability grain
oriented silicon steels

Stat d danain re'ng
Energy Savings: EFFITRAFO vs Spanish Standard
.Amo- Bk
- S' = results

. I
One way to reduce no-ioad losses is to reduce the thickness of
core laminations which reduces eddy current losses by reduced
resitance between laminations
This process bagan with introduction of the two stage cold
rolling process.
Five main grades have been introduced M2 to M6 in increasing
order of gauge from 7 to 14 mil thickness
In domain refining. by laser scribing the laminations are sliced
into strips parallel to the surface so they are oriented in the
direction of flux lines.
Finally eddy current losses are greatly reduced by constructing
laminations that are clean and have burr-free edges by laser
__ .-':"':::N/i
TEC}[\OLOGY Fl l'D.\'IE\TALS: Core I"U,U, Induction
Core losses IWlkgJ
1,8 .
AlIlorpbous (0.025
HiD (0,30 nun)
HiD (0,23 nun)
Laser (0,23
0.0- -------
._--...,.-, B TIl)
1.0 1.1 12 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Allied Signal in early 80's. Allied Signal projected that in 2000
amorphous transformers may save 47 TWh of annual energy in
11 ( H\lli 0(,\ n \Jl \ \lEY\ \l "': \[OIn!(:
eGO Ttehnalogy AIIDT Technology
The amorphous metal is not crystaline like silicon steel. Atoms
here are arranged in non crystaline random mannner.
This structure looks like glass and is sometimes reffered to as
glass metal or "metglass" the name of the famous amorphous
steel supplier
In this structure atoms may easily reorient to be aligned with
magnetic field lines.
The movement of atoms create molecular friction which is
basically responisble for hysteresis loss.
In amorphous structure this friction is largely minimized
"[( II\OLot;Y H \DA\IENTALS: Properlies (I)
Amorphous Metals exhibit:
Easier IIIIp(lizatiel (low toercivity .. d higb
Lower IIIpdit loss (low coercivity, permeability .ad higb
Faster 00 rennal (u. result of 10 .. magaetie
Venalile .. guetie properties resultiDg from post-fabriealioa.
Dynamic vlbntion due to magnetostriction
! -----...., ::
..:r JO ' :t:
"E 100 !
JI .11
1 J
10 '-____ :::::J._...I ::t
1.11 t.l U 1.6 1.11
................ m
. ...-----...,
1.0 U J.4 1.6 11 2.8
1 [CJI\OLOG \ fD 0,\\1[\1 ALII: Thumal prvpeltil',
Heat Spectrum Radiated
in Oriented Silicon Steel Core Amorphous Metal Core
PRODUCTION PROCESS: Amorpbous Distribution Transformers
"kd Treatment Flow Chart
proccs, 10 obtain ,1rUClllfl'
(!Ii" ''('/, ;1.''1ulfCd ute f)f >' {
IJae Diagram Core Making

I'f{OI)I C r 10'0 "IHI(. '" -\lllUrph()u\ Distribution Transformers
Line Diagn.m Core Coil Assembly
N .... _......- ("._ tltae"l.vol .. dl
RIIIINMo ...... _ppctl,
to _"'""" ... c ...... joi.L
JeIac .......... p for c"",,,eillISKRlbly .ad .... ro-Ioftd.
.leialCIUI ......... d .... ......... li .. REPAIRING proce:..
0- pIoce __ ..... for..., -Wr. 5
Economic impact on electricity distribution companies largely
depends on regulatory scheme -> separate financial or fiscal
incentives might be needed for transition period
Barriets and obstacles different between market actors =>
Bundle of policy Instruments needed on EU and national level
Several barriers towards implementation of proposed policies
Some chances for implementing policy instruments:
Next steps of changes in (national) regulatory schemes
EU ActIon Plan on Energy Efficiency: Measures 10 reduce grid losses in
EuP Directive: combined with other transformers (> 200,000 pcs/year)
New CO2 reduction targets published in EU package d 23 Jan 2008
Promotion of AllOT pilot projects can increase competition in the
market for energy-effident transformers
is r

Vikrant Joshi, E. Ramaswamy, Shekhar Vora
Abstract: Calculationfor leakage impedance for magnetically
separate but electrically connected circuit has always been a
challenge and requires a non conventional way of /reatment.
Configuration like boosters transformer, zigzag connection etc.
belong to this category. This paper aims at calculating the
impedance of side limb taps Auto transformer which also has
similar arrangement. In case of single phase 3 limb
an'angement, tapping winding can be separated from main limb
and placed on return legs. So that mean diameter of the winding
is reduced & losses as well. Transformer with height limitations
makes it very difficult to achieve required impedance in
conventional transformer. In such cases placing tap winding
on return limb can be very good idea to achieve impedance.
This provides high degree of freedom and flexibility for the
designer; nevertheless an easy way of calculating impedance
is furnished. Three winding theory of transformer can be used
as an effective tool in deducing the impedances of for side limb
taps configuration & this be the major content of paper. The
values thus deduced are compared with a circuit approach based
software & results matched with reasonable accuracy.
Introduction: Leakage impedance of a transformer is an
important performance parameter that has to be guaranteed to
customer with defined tolerances by standards. Although its
estimation is quite straightforward for standard geometries
through flux linkage method, its estimation becomes quite
involved when non-standard designs are proposed, for reliability,
economy and feasibility. Iftwo windings on different magnetic
circuit are electrically connected, the leakage impedance offered
can be calculated in relatively easier way with three winding
theory.. .
In case of transformer where tapping winding is placed on return
legs of magnetic circuits, for design convenience, the discussed
methodology can be used to deduce the values of impedance
for any testing connections. To make this paper self contained a
brief introduction to construction of side limb taps transformer
and design approach for the same is also discussed followed by
the deduction of impedance expression for IV-HV, LV-HV, LV-
IV pairs at any taps usin. three winding theory approach. It
will be very useful to mention variable flux design philosophy
& three winding theory for completeness of this paper.
Variable flux design: Variable flux designs refers to special
cases of design where the volts I tum and hence the operating
flux density is not constant and varies with tap position. Designs
where taps are used at neutral ends in. star-connected auto-
transfonner (to facilitate the use of simple 3 phase tap changer
in one column) for HV variation, both the HV & IV turns are
changing for every tap positions. Since IV voltage is expected
to remain constant over the complete tap range, volts I turn
changes at each tap resulting in a variable flux design and
demands for a special approach to decide tap turns / step.
Theoretically speaking, although tap winding demands for
variable turns I step for maintaining IV voltage constant, due to
practical reasons tap turns I step is chosen to be constant so that
the ratio error between target voltages and achieved voltages
with the provided turns are weB within the specified tolerance.
Being introduced to variable flux concept, let us try to deduce
an expression to find additional turns, to be added to HV I IV
winding and p.u. V IT, for a neutral end taps autotransfonner
for HV variation.
HV principal tap volts = VI
HV principal tap turns = N 1
IV principal tap volts = V2
IV principal tap turns = N2
Tapping factor = p = 1 + % tap / 100
Let the turns added in HV & IV at tapping factor p he x
The variable flux operation shows
& =
~ pV\N
+ p ~ x = V
N, + V
PV;N2 -V,N,
x = -
- PVl
when p> I x is -ve
when p< I x is +ve
when p= I x is zero
Now, the ratio of V IT at the tapping factor "p" to that at the
principal tap is the p.u. V IT.
- V
+x) V _ N2
Hence p.u. - . or p.u. I r - ---
, V
N2 +x
Substituting the value ofx from (A) and simplifying, we get,
= ~ -V2
~ -V2
The above expression being deduced let us now summarize the
steps for variable flux design.
Calculate turns to be added in HV & IV at tapping
factor "p" using (J)
Calculate theoretical HV & I V turns at all tap positions
as N,+x, Nz+x respectively
Calculate theoretical turns I step required for tap
winding for all tap positions .
Calculate uniform tap turns I step for all tap positions
until the ratio error is within limits
Recalculate Actual HV & IV turns, volts I tum & volts
/ tapping step for all tap positions.
fhese complete the discussion on variable flux design and we
;an now proceed towards auxiliary transformer; it's need and
urns ratio selection.
fhe multi-circuit theory for three winding transformers:
the impedance analysis of the design under consideration
lemands for the basic understanding of multi-circuit theory, let
IS proceed with the explanation ofthe same, without proof. For
)roof, refer (MIT)
fhe leakage impedance characteristics of a three-circuit
ransformer can be represent completely by assuming that each
:ircuit has an individual leakage impedance and that no mutual
mpedance effects are produced by the load currents other than
vhat may automatically results from these individual leakage
mpedance drops.
n case of a transformer provided with three different windings
IV, LV & Tap (say), and if the magnetizing currents are
leglected, the transformer can be represented by the so called
II1IT), as shown below.
Star Eqt Ckt for 3 Wdg Transformer
'he above equivalent circuit represents a single phase
'ansfom1er or one phase of a three phase transformer.
'he terminals H, L & T are the terminals of KV, LV & TAP
vindings, where either the voltage sources or the loads can be
onnected for the analysis. The impedances of the equivalent
etwork can be evaluated as follows.
.et ZHL, ZLT & ZTH be the per-unit impedances between the
llffixed pairs of the windings at a common kVA base with the
winding left open-circuited. Also let
T = Z - ZHL
The resistive components of the impedances and reactive
components of the impedances can be independently calculated
using equations analogous to (B) with the letter Z replaced by
letter' R' and letter' X' respectively. It should be noted that the
'R's used here are not the ohmic values of winding resistances,
but are the 'per-unit resistance voltage drops'.
For example. RHL = load loss in kW between HV & LV I Rated
It should be noted that one of the terms RH or RL or RT may
tum out to be negative. So also. one of the terms XH or XL or
XT may tum out to be negative. In these cases, the negative
sign of the R or the X must be retained in the equivalent circuit
for correct analysis of the circuit. To understand the physical
significance of the equivalent circuit and meaning of negative
reactance the reader is recommended to refer Blume.
Design Approach for Auto transformer with neutral end
taps on side limbs fed by LV winding through Exciting
The design is divided into two parts
(I) Design of main limb
(2) Design of side limb
Prror to discuss design, circuit constraints must be laid down
with help of following figure.
U 2.1
SERI[S W'Nptr;
ct)!t!!'tf WINDING
L--__ ....;1"".3'--__ '"'\3
LV W'NDlr;
HVR Rated HV voltage at a particular TAP i.e. across terminals
1.1 & 1.2
IV Rated IV voltage i.e. across tenninals 2.1 & 1.2
SV No load voltage across series winding i.e. across
terminals 1.1 & 2.1
CV No load voltage across common winding i.e. across
terminals 2.1 & 1.3
No load voltage across LV winding i.e. across terminals
3.1 & 3.2
No load voltage across connected portion of tap winding
i.e. across tenninals }.3 & 1.2
No load voltage across exCiting winding i.e. across
tenninals 3.1 & 3.2
IV current
Series winding current
Common winding current
LV winding current
I RV Tap winding current
lEV Exciting winding current
NfV Turns of IV winding
N sv Turns of Series winding
N cv Turns of Common winding
Nl.v Turns of LV winding
Turns of connected portion Tap winding.
Turns of Exciting winding
By virtue of connections and physical positions of windings,
--=-- & --=--
N n' = 2 x N L'l ..... {with assumption that side limb
cross sectional area is half of main limb cross sectional
lEV = -IL"
EV =2x LV
6) ILvNLv+Is"NS\,+IC\,NCl/=O &
Il\VNRV + I c v N c ~ =0
7) When taps are not in circuit
I RV = 0 => lEV = 0 => I LV = 0 => N sv = - N cv
Here note that tertiary winding carries current at all taps except
one position.
The main limb design is based on variable flux design approach,
since the neutral end auto belongs to this category.
For side limb design the exciting winding & tap winding has to
be designed. Since LV winding is feeding exiting winding, their
voltages must be equal. But volts per tum are half in side limb,
the number of turns have to be double of that in LV winding.
The tap turns are frozen from previous step & since it is
connected in series with common winding its current is same as
current of common winding. The height of side limb windings
can be slight more than Main winding as requirement of end
insulations is not same as main winding.
Impedance expression:
I A = =
be called auto factor.
5, C, L be branch impedances of star connected equivalent
network of 3 winding transformer having series, common &
LV windings at nameplate VA of transformer, which is taken as
B be per unit impedance of booster nameplate MVA and
LV voltage as base. (I)
I H be rated HV winding current at nameplate MVA, hence
I H = I/HVR Amp (1)
I I be rated IV winding current at nameplate MVA, hence
I H = (1/ IV) - (1/ HVR) Amp (2)
I L be rated HV winding current at nameplate MVA, hence
IH = I/LV Amp (3)
Now suppose this transformer is subjected to impedance
measurement between HV & IV, by connecting terminals 2.1 &
1.2 together and applying a voltage across 1.1 & 1.2 of such a
magnitude that rated current I H = I/HVR Amp flows in the
HV terminal 1.1.
The connection of the circuit is shown in figure 2 below
C,.,.,.,., W!NMMi
3.1 -...------.
In this figure it can be noted that absolute voltage drops in
common and tap winding are equal and opposite. This is because
common and tap winding are shorted together & their voltage
sum must be electrically zero irrespective ofvoltage induced in
them. Also voltage source is as good as connected across 1.1
and 2.1 i.e. across series winding.
The above (fig 2) setup of the transfonner can be solved by
using the star equivalent network of 3 winding transfonner
consisting of series. common and LV windine as discussed here
c s
= [SV _ ev . (HVR - IV)]
1 [ ev . HVR ]
.IL = -- SV - + ev
Now absolute voltage drop across tap winding is
V; = VI X [HVR - (SV + ev)]
and absolute voltage drop across common winding is
V' = V x ev
c c
But equation (6) + equation (7) = O.
:. VI X [HVR - (SV + ev)] + Vc x ev = 0 (8)
. _ [(SV + ev) - HVR]
.. Vc - V
Now applying KVL to loop consisting of '(<; & V
we get,
Calling expression in RHS of equation (10):lS Kl,
we write, Vs - VI = Kl. (II)
Now applying KVL to loop consisting of \ZC: & V
we get,
SV CV . (HVR - IV)
V - V = -- . S + x C (12)
Calling expression in RHS of equation (12) as K2
and using equation (9) we write,
V - V x [(SV + CV) - HVR] = K2
S L ev .
Substituting V
= Vs - Kl from equation (ll) in equation
(13) above, we get
y) [(SV + ev) - HVR]
Vs - Vs - Kl x
K2 (14)
(HVR - (sv + CV)}] 2 (HVR - (SV + CV))
t-;;x 1+ CV =K +Klx CV
But HVR - (SV + ev) is nothing but RV.
K2 + Kl x -
K2 . ev + Kl . RV
[ev + RV]
Substituting back the values of Kl & K2 we get,
N. SV CV'ENR) xc]

S.-x(HVR-SV)+-x HVR-RV--- xRV+Cx----
HVR-RV--- xRV 2
Vs= HVR S'SV+(L+B)x IV +C{rv) sv (21)
V -V; =-5+-- SV+CV----
5V (L + B) [ CV HVR
(10) Now the voltage drop in series winding(which is

but the applied voltage) is given as
v: ==V
equation (21) & equation (22)
HVR-RV--- xRV "
f [ [ CV.HVR]]
;v;- SSV+(L+B)X IV IV sv (23)
. V' V'
f fi . z- s
However, pu Z 0 trans onner IS HVR or pu - HVR .
S V 2 [ . (R V )2 ( C V )2 ]
PUZ=--2 S+(L+B)x - +c-
This is the expression for pu impedance of an auto with neutral
end taps, with the lV feeding a sort of buster. If this buster is
not present and the tap winding is present on main leg only, the
value of" L " shall be taken as the branch impedance (pu value)
of tap winding in the star connected equivalent network of three
winding transfonner consisting of series, common and tap
Calculation of %Z between HV & LV.
Figure (4) & (5) above show the equivalent circuits when testing
the HV-lV impedance by shorting lV and supplying HV such
that rated current flows through it.
From figure (4) by applying Kel,
= (CV + SC)
:. V = -- C + 1 - -- L
CV ( RV)
Absolute voltage drop
, CV
v( RV)
Vc = --C+C 1---L
On the same line absolute voltage drop V; is,
. ( RV)
= -- S + SV 1 - -- L
Now from figure (5) by applying KVL,
V = --B
The absolute voltage drop V; is,
V' = -- B
Now, the pu HV-LV impedance is nothing but the sum of
absolute voltage drops and V;, expressed as a fraction
of HVR.
Hence, , , ,
SV {RVi} RV }
pu z = - C + C 1 - _.. + - S + S 1 - -- + - B
pu ZHL = [CV C + SV S + (SV + c0
L + RV B)
HVt\ .
Calculation of %Z between IV & LV:
Figure (6) & (7) above show the equivalent circuits when testing
the IV-LV impedance by shorting LV and supplying IV such
that rated current flows through it. HV is left open.
From figure (6) by KVL,
Vc = CV (C + L)
The absolute voltage drop ~ is equal to Vc CV. Hence,
~ = - (C + L)
From figure (7) by KVL,
V =-B
The absolute voltage drop V; is equal to V
RV. Hence,
V' = -- B
Now the pu IV-LV impedance is nothing but the sum ofabsolute
voltage drops ~ and V;, expressed as a fraction of IV.
pu Zn = - -- (C + L) + - B
1 [CV2 RV2 ]
= ~ rev2 (e + L) + RV2 B]
IV- r
The formulae derived in this paper are specific for auto
transformer with neutral end taps on side limb, fed by LV through
an exciting winding, for HV variation.
Results: The table below shows comparison of results obtained
by the method discussed above and TOK software which is based
on circuit approach.
Analytical method TOKResult % Error
1 V-LV 36.84 38.24 3.80
LV-HV 55.06 55.86 1.45
HV-IV 14.35 14.51 1.11
Conclusion: This paper discussed methodology of calculating
leakage impedance for an auto transformer with the tap winding
on return legs of magnetic circuit which operate with 50% volts!
turn. The expression thus deduced is used to calculate the LV-
HV, LV-IV, IV-HV impedances of an auto transformer and
compared with circuit approach based software & result matched
with 4 % accuracy. The approach discussed can be used for any
configuration with magnetically separated & electrically
connected windings to calculate impedance at any taps, thus
conveying generality and versatility.
theory, operation and application of transformer, by Late
L. F. Blume, A. Boyajian, G Camilli, T. C. Lennox, S.
Minneci, V. N. Montsinger. .
[2] Transformer Engineering Design & Practice by S. V.
These are especially designed of transformers for suppression
of electrical noises. Such Transformers can be designed with
different level of Isolation. Ultra Isolation Transformers
eliminate all types of electrical noise.
The use of Faraday s Unique Box shielding and special
.techniques totally blocks the transfer of electrical noise.
These transformers are designed with better regulation,
which do not generate common mode impedance coupling
Such Transformers protects against strong Lightning
impulse noise. Busbar or over head transmission lines
accidental short-circuits etc.
Such Transformers are also capable to protect
sophisticated equipments against electrical noise (Spikes)
due to accidental discharge of capacitors.
It protects the Micro processor based machines (CNC
Machines), Digital control devices and other valuable
equipments such as computers, Printers, Medical
Equipments etc. etc.
The Electrical Noises are generated due to:
Switching of electrical utilities like Capacitors, MCCB's,
ACB's ... etc. Larger the inductance of the system and larger
the current change during switching larger the magnitude
of electrical noise.
The inductive loads like big Motors, Compressors,
Overhead Cranes, Elevators, Presses etc. also generates
substantial switching noise.
Switching equipments like Invertors, Converters, SMPS
etc. generated electrical noise due to switching of
Thyristors, Transistors, Relays etc.
Welding systems pollutes earthing systems and add notches
in the wave form, generating harmonics.
Lightening, precipitation ofthe static charges and electrical
discharges in the atmosphere are the natural cause of
generation of various electrical noises.
Electrical Noise can be classified as under:
A) Common Mode: Occurs between ground and the current
carrying conductors including neutral. Being high
frequency noises, they don't get attenuated due to distance
and therefore reach farther. They are common source of
B) Normal Mode: Occurs between tWo current carrying
conductors like line to neutral or line to line. These are
generated and created due to switching actions of utilities
& equipment. They are generally of high magnitude,
causing device failures.
Electrical noises are observed to occur over a wide band
of frequency ranging from 1 KHz to 100 MHz and above.
In magnitude also observed to be as high as 4 - 3000 Volts
on 3 phase supply system.
The high frequency noise can interfere with digital
electronics equipments causing untraceable data entry,
change of programme, loss of memory, erratic behaviour
The high voltage spikes can cause the failure of Thyristors
or Transistor, Micro Processors and other sensiti ve devi ces.
The radiated noise can interfere in operation of remote
control equipment like Cranes, Digital controls or
Telecommunication Equipment.
Noise Suppression Transformers are specially designed
with split winding construction & bifilar Connections to
reduce over all Capacitance of the winding. Split
construction can be Axial or Radial.
The use of high insulating materials and special
techniques result in attenuating common mode and
minimizing transverse mode noise.
the core and its magnetic properties are properly selected
to exhibit sufficient leakage inductance to provide greatest
possible attenuation of normal mode power line noise,
consistent with the Transfer of Fundamental power
frequency. But al1 higher frequencies are blocked.
The use of Faraday's Unique Box shielding and special
shielding techniques totally block the transfer of electric
noise even by static chaFge, by way of proximity or R.F.
Due to better regulation, it does not generate common mode
impedance coupling effect. i.e. CP= Coupling Capacitance
impedance of important factors of control of noise electrical
reduction transformers.
All the types of electrical noise, predominantly common
mode noises can be eliminated by Noise Suppression
Transformers. Since it isolates primary and secondary or
separates neutral to ground bond on the secondary side,
can be used to create separately derived source to combat
current loops.
The Noise Suppression Transformers are available in
different levels of noise attenuation capabilities. The most
commonlyused are 100 & 120 dB. The coupling
capacitance between primary and secondary is direct 1: I
relationship with dB levels. Some of the graphs herewith
indicate different types of electrical noise observed in
typical Engineering industries & successfully eliminated.
Noise Suppression Transformers (Isolation Tranllformen)
are commonly used to:
Protect the Computers, CNC Machines and
Telecommunications equipment from damage due to
electrical noise. spikes etc.
In case the ground potential of system units are different
from each other and are exposed to the effect of instability
at high frequencies, Isolation Transfonner to the the right
choice in such cases.
Slightest leakage current (in the order of 100 micro ampere)
is permitted in operation of the equipments, especially for
computers and C.N.C. control equipments.
To shield large number of electronic equipment which
individually also are producing different types of electrical
noise at a common busbar typically CNC machines. Drives,
Hardening equipment etc. The use of NeT, being
preveats damage due to circulating noise
interference within them.
To protect against strong lightening, impulse noise, bus
short-circuit, accidental discharge of capacitors etc.
When multiple Noise Suppression transformers are used,
the suppression effects increase in proportion. Therefore,
the effects of cascading Noise Suppression TransfonnerS
is remarkable.
A) No impedance matching is required.
B) Characteristics are symmetrical to allow suppression effect
to work in both ways.
C) No secondary effects, such as noise generated by series
resonance at some frequency as with filters and other
similar devices.
By Function
Ultra isolation Transformers Typical Form or
Input Ou1put Rating Common
Single 1101 1101 0.4,1,2,3,5,8,
Phase 220 V 220 V 10,15,20,25
120 dB
.111; \I Ail; \1
100 dB . . . -, "' .. _,
Phase 25, 30, 40, SO, Standard
60,75,100, &
125, 1SO, 200, 120 dB
225, 250 & 300 on request
Regulation: 2-4% for 100010 change in current @ unity P.F.
Connections: Delta/Star for 3 phase
Operating power ractor : 0.75 lagging to 0.75 leading
Dielectric strength : 2500 V for 60 sees.
Resistance: Greater than 1000 Ohms
Coupling capacitance: 0.1 pffor 100 dB 0.05 pffor 120 dB.
Leakage current: less than 20 micro Amps.
Technically, any transformer that have no direct current path
between it's primary and secondary windings provides
'Isolation'. Other commonly used even if they
have a separate primary & secondary winding, are intended to
convert the input voltage to a more useful level and do very
little to attenuate the passage of noise or transients from primary
to secondary. Even though both are separately wound
Transformers, they are substantially different with respect to
construction, specification and perfonnance characteristics.
In Shielded Transformer, 4 low impedance paths exist for noises
to pass to secondary and vice versa, (as shown in drawing). A)
The high value of coupling capacitance, B) Maximum linkage
of magnetic field, C) The leakage current, D) Static transfer of
electrical noises.
Noise Suppression (UHra Isolation) Simple Shielded Transformer
High Noise Attenuating Poor Noise Attenuating Typical 20 dB - 40 dB
Characteristics 100/120 dB
No Electrical switching noises Amplifies electrical noises between 100Khz-
amplified and noises are not allowed 1 MHz an area where maximum switching
10 secondary of the transformer noises are generated
Electro static & Electro magnetic Noises are passed on through secondary to
noises due to welding also load.
MagneticaUy sensitive equipments MagneticaUy sensitive equipments cannot
can work near Ile installations. work near the instaHations due to high
electromagnetic field jtenerated.
r-B t Electrical Parameters
Noise Suppression (UHra Isolation)
Simple Shielded Transformer
1 Coupling Capacitance Better than 0.1 Pico farads (10"12) 1-10 Micro Farads (1fr6)
2 Inductance
., II high
3 Resistance Greater than 1000 meaa Ohms Greater than 50 mega Ohms
4 Leakage Current Less than 20 Micro Afl1)s (10"6) Less than 5 Milli Amps (10"3)
5 No!se Attenuation Better than 100 dB= 1/1 ,00,000 20 dB maximum = 1/10
6 Electro Static & Electro magnetic Yes, efiminates Cannot eliminate
By Construction
Sim Ie Shielded Transformer
Ultra Isolatation Trdnsfonner
_ T !3Qsf<!rmer
Core BIOllnsuialed
Faraday's Unique Box
Construction Of Noise
Supprerssion Transformer
(A) Coupling Capacitance

",' I '",
d .
"-'. / I ;,,,
I "'. (H) Magnaic
n' i': \......,.
!! ',.-
v; i,

(0) Static Trnn*
d (
, I:' (0
i: to
9: r
___ 1-'
L ....
llImiIatim leakage Cwmrt
Consumes extra quantity and better quaity of copper, lamination etc. Nonnal content of Copper, Iaminaoon & of insulating materials etc.
(50-70%) with content of special grade insulating materials
... ; ........ ; ....... ........ ........ ........ ; ............. : .. ; ..
H1:1r1zDn111110n.lIcondl/clvllion Y-** SOwall:lld\4llon
F..at dearing operation far from Ihe
'---+---+-t--1--' _ ....... +-.. -4-----....---.
. " ., ..... : .. " .. "
. . ... .. " .... ; ....... . .... ' ............ " .. > .. .
,." .
HotbIon ... 2fiO . \,.
loniJ-term voltage utf NauIdng from thundemorm'l 't
on UIIIty grid. .
!,\* t,'
FIGURE 23-7 Graphics of different power ,\)I
(From 1M Dranetz Field Handbook/or .
ysis, 1991.)
. ...... ;" ., ..... " .; ... "" ... ! ., ...... : ...... '. ' .. "
. "11"
. "TtllHlll\
"'-T'-"-'-'--"Tr;,J \I II 1/1 I
.... l r:ij ijVl.
. . .
v.rbIlO .........
Fault on utlty feeder cUing alarm.
............... ; .....
. .
PowIf.fllclor coneclon CIIpdDr. InIW Yd&.ge Gop i\-
cIcIatM do&in;. 0\IInh0at n ringing due to dR:Ut
.: ...... : .. ,.
. ................. ,; ..... .
Hut ....... "., .=1
1m.,... bI.nta C*IMd by ardng of poor c:cno""'L . The wt>
Ing 00CLn .,.., vaItIIIIge pMk when the 80R canducIa.
FlGUltE %3-7 (Contbllled)
Typical Transformer Network Of RLC
Primary Shunt
L e o k ~ 1m1Mpeda!1CICIN1C1---

N o ~ ~ ~ ~
Important Electrical Equations
U_sed For Isolation Transformer Design
Resistance (R) = --.L... RcacbulCe (XI-) :z 2'niL Reactance (Xc) - 1 flJz LC- 1
a 2't11C
Whcro Where Wbcrc
R = Resistance in Ohms X - Reactance X- Rcactmcc
p = Resistivity 2'rr = COIlstant 2'rr - COIlSfaot (2m1'LC
1- Mo.n long1h of windag f- FrcquODcy f= FrequCIIICy
a - Conductor Area L =- Inductance in Hcbcry ~ = Capacitanco n Faraday JLC
Noise Level dB = 20 log Atinueation
Noise Level dB _ 20 log 8u 1
Noise Level dB 20 log 1 mv
0.00001 mv
Noise level dB - 100
= 1
= -1
Simple Shielded Transfonner
Secondarv ---
-- Lamination
(A) Coupling Capacitance
(D) Static Transfer
Leakage Current
NomlaJ content of Copper, lamination & of insulating materials etc.
Faraday's Unique Box Shielding
Ultra Isolatation Transfonner
Various magazines. seminar papers and electrical
periodicals have been consulted while preparing this
I extend my sincere thanks to Shri Anil Kumar Agarwal,
Managing Director ofP. M. Electronics Limited, to permit
me to prepare the paper for presentation .
I also extend thanks to my colleague Mr. I. 1. Dasgupta for
reviewing this paper before publication.
F aradav's Unique Box
Inter Winding Shield
Construction Of Noise
Supprerssion Transfonner
Kiranmayee - Engineer - Business Development
Vijai Electricals Ltd., Hyderbad
Growth Scenario of World Energy
I :l
. /
_ 2111. 21115 2020 _ 2IIH
Soun:e : EIA. Wot1d Engery Projection. Plus(2008)
VIrId market of energy consul]'lPtlon Is
to Increase by 50% from 200
to 2030
Overan electricity c:onsumDtIon In India
Will grow from 609 billion 1cilowatt hours
In 2005 to 1,730 bilHon kilowatt hours
by 2030.
Furthermore, Africa, Brazil, Meldco and
South Korea will more than double their
consumption from 2005 to

...... " ... '".,,:,," .,
growtli in 'World InerBY Consumption
How to meet this demand?
Increase the Generation
Use Energy in a Efficient way
P.ffect of increase in eneTBJ C01ISUmptUm
16% '
2fo Gal
coal 1Inport_nr: ,..,eI tOf"" Gen.".clon - Pollut."t
High carbon emissions
Acconling to CIA. world carbon emissions will
Incruse steadily from 28.1 BMT( 2005) to
42.1BMT (2030)
Increased carbon Emissions
II - w.;.. ....... IMI ...
-lWiIn -.1owI ...
Rising Temperatures
(Global Warming)
Rising Sea level due
to thermal expansion
of seawater and melting
of glaciers and ice caps.
" -lIIiIiIWII-.1owI1itt
Da18: Intergoyemmental Panel on _ Change (IPCC) of
UnIIsd Nations El'Nironmenl (UNEP) .
'R!su{t of g{o6a{Wannino
Global WMming. The future is not far.
It is estimated that sea water will rise up by 13-14 meters and
many cities aaoss the globe wit! be wiped out completely.
A sea level rise of 1 meter would swallow about 90% of the sandy
beaches in Japan and 18% of Bangladesh.
The production of winter wheat will decrease by 55% in India and
15% in China bylhe year 2100.
9detliodS atfopted to red'uce tlie car60n emissions
to meet the i1lCTtase in energy consumption
Most of the coootries in the wor1d are adopting certain
Energy Efficient Standards for purchasing
- reduce the power consumption
- reduce carbon emissions.
Emphasis should be given on
distribution transformer, which
is the most important electrical
equipment showing
- vet)' large impact on the
networlc's overall cost
& energy consumption

_., 1 .. : .....: .. 4 ..

[W .... ......... " ... j .

Carbon credits In
Exchange for $ or (

What is a Carbon credit?
What is Kyoto Protocol ?
Carbon credits can be aequlred through:
1. Clean Development Mechanism (COM):
2, Joint Implementation (.II):
3. International Emission lrading (lET):
struetunt of the Carbon Market
Allowance based lransadlons

Carbon eredlts ean be acquired through:
Clean Development MecMnIam (COlI): project - based emission
reduction activities in deveIopilll countries.
Joint Implementatl= Two developed countries can invest in
emission mitigation . generating Emission reduction
Units (ERUs).
international Emission tnIdIng (lET): Credit deficient and credit
surplus countries (that have ratified the Kyoto Protoool) can
trade carbon credits at specialized exchanoes (European Climate
Exehange, Chicago Climate E.'tchange. UK's CO
(JlYM in lniluz
4:'J MIllIOn c.:I:Ks t:xpted !rOm c.:UM Keglslefe<l
PI'ojects Until 2012
India towards Kyoto Protocol
signed in December, 1997
and ratified in August, 2002.
National COM Authority
1 CER = 1ERU = 1 tonne of
CER.: certffted Ernisaon RaduaiOn
fRU: Emfssion Reductton Unit

. ,'.''''.' .. ''.'')'''
India's lPosition
Per capita Carbon dioltide emission of India is
amongst the lowest in the world. Contributes around
4% of world total.
Country wise ,er ca(tita CO
emission in 2004
(tonnes of
- India 1.2
- USA 20.6
- U.K 9.8
- China 3.8
- World 4.5
With continuous Effort of Government of India
average emission rate is showing declining trend
indicating larger commitment
Weighted average emissions rate (tC02/MWh)
2003-04 --- 0.85
2004-05 --- 0.84
2005-06 --- 0.82
2006-07 --- 0.80
status of Cfean CDewWpment 9hcIianism
q?rojects (Dmt) in India
1. National COM Authority (NCOMA) already given
Host country approvals to more than 700
2.296 Indian COM projects already registered with
CDM Executive Soard out of World Total of 868
projects (35%)
3. Main Projects; Waste heat recovery, Small Hydro,
Biomass based Power Generation, Wind Power
and Energy Efficiency improvement in Industries
like cement etc.
I nauzn Scenario-()lY.M
oExists high potential of Carbon credits.
oSaseline Carbon di-oxide emissions from power sector
already in place- First CDM country
oWide spectrum of projects with different sizes
oDynamic. Transparent & Speedy processing by Indian
National CDM Authority (NCDMA) for Host Country Approval
oVast Technical Human Resource
oStrong Industrial Base
Main Sectors of Green House Gases Emissions
- Energy Sector
- Transport Sector
- Agriculture Sector
- Industrial Sector
*Power sector is estimated to contribute around 50%
of total CO

Major Initiatives by the Government to
reduce GHG Emissions
Current OIY.M C17"ojects in Inauz
LANCO Kondapalli Power Pvt Ltd (LKPPL) is
an Independent Power Project (IPP) located at
Kondapalli Industrial Development Area near
Vijayawada in India. The Plant is a 368.144 MW
Combined Cycle Power Project operating on
Natural Gas as primary fuel. Based on a Power
Purchase Agreement (PPA) for a period of 15
years, the plant supplies power to the
Transmission Corporation of Andhra Pradesh

} -'
. .
Bachat Lamp Yolana (BlY)
Bharat lamp Yojana (BlY) Is a tripartite agreement
:. ... .. .,"' .. "LC:, uiSCOM & CFl Manufacturer.
Outline of BL Y:
1. Repladng Inefficient Incandescent bulbs with CFLs for
2. RedUCing the price of CFL and recovering the cost by
using COM.
3. 400 million light points presently estimated on
.incandescent bulbs could reduce energy consumption
upto 10,000 MW.
When CFL lamps can
contribute in reducing
carbon emissions,
why not the distribution
CJ(,'4:J{ql1vq POlltPSV'ECJ?{j'Y
/FPIClt:J{T7fR.M{SPO</1Jrf./RS INT.JfE
USA - DOE norms (Department of Energy)
Australia -MEPS (Minimum Efficiency Performance Standards)
has one in place for the Dry Type Transformers.
India - Bureau of Energy Efficiency (8EE) ..
In Asia, Vietnam and Philippines has ripously encoufilied the
Amorphous Metal Core Usage in Transformers.
Banclade$h has been prowri", these transformers from the year 1997.
First Cost is
a tiny %
ofTfr. TOC.


Use of AMDTs In the world
25% 8_IT-"&
801M ...
4% Initial cost
riPBrating Cost (Transformer losses)
are ignored. yet offer by far the largesl
. opportunity for savings.
96% Life cycle operating cost
Why initial cost dOminates purcfiasino process?
Operating Cost (Transformer losses) are ignored,
yet offer by far the largest opportunity for
End user, who pays for losses, is rarely involved
in purchasing process.
End user is not educated to the potential
Savings due to BEE star 3
_ (Low loss Energy Efficient Transformers)
RIlling REC BEE Star 3 Savings l_ .......
S.,.cIfIcIoUon Lon .. In 1nRs.
W.Us.,11O'l1o V ..... (RII.)
100%lNding loading
Nll II Nll II NLl LL AxNLL BlIlt
11kVC ....
25 100
685 30 1665 70 120 130115 1"36 1000 31251200
235 55 1
95 125 140 53778 12272 1000 56049100
100 260 11760 74
'726 186 134 60021 1'931 1000 81951760
ToCaI No loed Saved 10.3I11W
ToCaI SavInga In D1atribution Transformers on account of
I 16.I3Crorn
Savings due to BEE star 4

(Low loss Energy Efficient Transformers)
RatIng RIZC BEE Star 4 Savings Los_sval ... Qty. Total Net
Specifieation loa ... In In As.
w.u.. VaIue{Rs.)
100% loading 100% loading
25 100 1685 27 1608 73 177 31406 14373 1000 35778890
Ii3 160 11235 55 11085 125 1150 53718 18519 1000 62296000
100 260 1
60 75 11575 185 1185 19591 J 10506 14100 90096850
Total No load Saved 10,31_
Total Savinglln DiStribution Transformers on account of
/, . I2Crotn
l_erlos ...
of 'Energy 'Efficient 1:' ran.sfonners in ditimine tIie
car60n credits.

...... "' .......
Lei US know have a smalllUustralIoo based on the fads and figures ot
how can _ earn caIbon credit5 using Energy Efficient transtormers, say
Amorphous Metal Transformers in India.
Bask: assumptions .,...t;
let \he number of Amorphous distribution transformers installed in !he Indian
Power Systems (Average capacity 63 kVA) be: 5,00,000
Cost of -.gy at the lransfonner terminals : Rs.6.oo I kWH
Plant load factor (for generation) : 60 %

No-load losses in Transformer are important because ihey are throughout the
life of \he Transformer irrespective of the loading.
No-load loss of 63 kVA (eRGO core transformer): 1SO Watt
No-load loss of 63 kVA (Amorphous Metal Transformer): 45 Watt
Saving in No-load loss at distribution end of 5,00,000 transformers:
= X 5,00,000 = 61.5 MW
1000X 1000
Total saved at distribution end = 67.5 Wi
= 67,S 1000 8760 hoIn
'" 591300000 klNh I,....
!t s:;:: at !he distribution level, cuts down 0.91rg of Carbon emissions
Hence 1DtaI carbon emissions reduction = 591300000 x 0.9 = 532170000 Irgs of
Carbon emission = 532170 lOnSIyear.
One 100 01 C8tbon emission reduction when certified as a credit is called a CER which
is traded at 20 EurosICER.
So, 266OISO Tons IE 532110 CER x 20 Euros IE 10643400 ewoslyor IE Rs.

The abow Carbon Credits CIIn be earned till 10' years and Ictal earnings is Ra
The above calculation is applicable only if 5,00,000 Nos. of
Amorphous Transformers of 63kVA are used in the Power system
Network of India and this is a sample calculation to show how
many carbon credits can be claimetl using an example of single
rating.in India.
On the whole, laking all the distribution transformer ratings being
used in India, we can now imagine how much earnings we can
get by claiming the carbon credits using Amorphous Metal
Distribution Transformers. At the same time we can reduce the
carbon emissions in the distribution network.
I?(YI)lUlBD C)llP,ftCntY of IMIa
(M Q}( 17.01.20(8)
............. ,,--
11M a'.M( - 78,J77!M.'W
Thermal - 58,644 MW
( Coal - 52.105 MW, G .... MW, Lignite- 1,450 MW)
Addltion.1 capac{ty expected: New Re_.1IIes - 14,000 MW
: captive - 12.000 MW
Latest Feulble capacity : 78327MW
In order to meet the above energy
requirement, we need to install energy
efficient transformers in the distribution
network to reduce the carbon emissions.
. ...

l'J1ie ""Jona is ill our fiancfs a,uf saviTIfJ e1lergy is savi"o our p[allet
Vladimir M. Segal, Ph.D!
Over past few years there were dramatic changes in the market
for grain oriented silicon steels (GO) characterized by two most
pronounced trends:
1. Steady (and steep) increase in steel price and
2. Shortage of low-loss grades, especially HiB and HiBDR
(domain-refined) and thinnest RGO (regular grain
.oriented) grades (0.23 mm and less), which persists despite
the three-fold increase in price over four years.
It's outside the scope of this paper to discuss reasons for such
changes, while it's necessary to point out that, in author's view,
these changes reflect strategic shift towards higher costs of
energy and basic materials, which are here to stay in foreseeable
This paper results of the effort to address market needs
for the low-loss GO by developing grades with power loss lower
than that for Regular Grain Oriented steel (RGO), while keeping
Build Factor (BF = ratio of No-Load Loss (NLL) in a stacked
core to Power Loss in steel) equal or lowf;r than that for RGO.
Thanks to higher B8 (typical value = 1.87 T) and much larger
grains (average size = 35 mm) than conventional RGO the new
material should be susceptible to lowering power loss by domain
refinement [I]. Stacked cores built with domain-refined RGO
(no data was published earlier for such cores) showed no
increase of Build Factor for 0.27 mm cores and significant
decrease for 0.30 rom ones. Such a combination allows design
of transfonners with NLL values close to that with HiB-made
cores and fills the gap between RGO and HiB grades.
Six coils (three of 0.27 and three of 0.30 rom thick) were selected
to test the effect of laser scribing (LS) on power loss (Single
Sheet values). Continuous permeability diagrams were used to
ensure consistent 88 value within 1.87 - 1.88 T along the coil
length selected for the trials.
A COz laser was used in an experimental set-up. where
sheets of steel 600 rom long x 250 rom wide were fixed
on a rotating drum 400 nun in diameter. Drum's speed
was controlled to model the actual speed of strip passing
the laser beam in a future production line. Laser beam
was focused by a special optical system to produce scribes
(of pre -determined intensity) with 7 nun distance between
the scribes. Prelimi9nary experiments were done to select
power parameters of the laser impulse and find the
optimum distance between the lines comprised of
microdots produced by individual impulses of the laser
beam. The wave length and power of laser impulse were
selected to prevent damage of insulation coating. As. a
result, the lines produced by LS treatment were not visible
on the steel surface.
Results of experimental LS treatment of single sheet samples
presented in Fig. 1

:i 12
!! 10
: 8
i 4
Eft'eet of LaMr8crlblnO on Power Loss Decrease (S8 Ylllues. %)

r 0.27

. ----
0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
Flux Density. Tem
Fig. 1 Power Loss (Single Sheet values) decrease after Laser
Scribing, %.
Decrease of power loss on 0.27 mm steel is about twice that on
0.30 mm. While this agrees with the well known trend of higher
effect of LS on power loss for thinner grades of HiB (2], in this
case the effect is more pronounced. Aoalysis of this result has
shown that 0.27 mm steel has about the same grain size as 0.30
mm, but lower angular deviation of {II O} [00 1] grains from
perfect in-plane orientation (/i-angle). Earlier studies [3] show
that when grain orientation becomes "too perfect" {especially,
for P < 2} power loss increase due to rise in the magneto-elastic
energy required to create sub-domains, which are necessary for
alternating magnetization to occur. Localized plastic
deformation (due to thermal shock) produced by laser impulses
facilitates sub-domain microstructure and thus lowers the energy
loss during alternating magIll!tization [4].
Based on single sheet experiments, energy parameters of laser
impulse were adjusted (energy of an individual impulse was
slightly decreased and area of the focus point - increased). This
allowed for treatment of more than eight hundred of core
laminates (instead of sixty SS samples)'with inevitable variations
of grain size and orientation, magnetic characteristics and
. surface coating thickness. Six three-phase stacked cores (three
in each thickness) were built with identical in-plane design:
9OOx900 mm
. Cores with 0.30 mm GO had square cross
section: 280x280 mm
, while for 0.27 mm cores it was 140x 140
Step-lap joint with 6 laminates per group was used. Single-
phase cores were built with laminates from the limbs of three-
phase cores (as a result, single-phase cores had stack height
half that of three-phase ones). To ensure reliability oftrial results,
the No-Load Loss (NLL) ,,"d noise values were first measured
on cores built from as-cut (no LS treatment) laminates, then
cores were disassembled and all laminates were numbered in
order of their position in each stack. After that, each laminate
was subjected to LS treatment, re-stacked into identically sized
and shaped cores, and NLL and noise values were measured
3.1 Effect of Laser Scribing on NLL and BF
Change (in %) of average values for NLL and BF measured
before and after LS treatment for single-and-three-phase cores,
with laminates cut from 0.27 and 0.30 mm steel, presented in
Fig. 2 and 3.
Decrease ofNLL in both I-phase and 3-phase.cores made with
LS-treated 0.27 mrn steel at flux density < 1.5 T is higher than
anticipated from data for Single Sheet experiments (especially,
for I-phase cores). For cores made with 0.30 mm laminates
NLL in cores decreases more than power loss in SS samples
only for flux density> 1.5 T.
At B =/> 1.7 T the effect of LS on NLL of all four types of
cores becomes similar (still, slightly higher for 0.27 mm steel):
NLL decreases by 12 toI5%. That's despite the Power Loss
for 0.27 mm SS samples decreases (after LS) 8 to IOOIe more
than for SS samples 0.30 mm thick.

I j::
i 14
i 12
1 6
t 4
! 2
NLL decrees. after LS treatment. %




.0.27 mm I-pNse
.0.27mm 3phase
.. 0.:10 mm I-pNse
FluxDenstIy. T.st.
0.11 1.0 1.1 1.2 1.:1 1.4 1.5 1.6 1.7 1.1 l.t ZOO I
L ............... _ ........ _ ................. . _ .. _ .... _. __ .. _____ ..______ ... _._ ....... _. ________ -.l
Fig. 2. Effect of LS on NLL of 1 and 3 - phase cores (0.27
and 0.30 mm) . .1NLL = = lOO*(NLL tIefote - NLL .,.)/
NLL befON)' %
r--- 12
Chance of Build Factor aft.r lS, "
,' ....

__ ..J

- --
/ .0_27.1"Phase ,
0.3.1 ........
I- .. O.;17,i,,-

0.1 1.0 1.1 1.2 L3 1.4 1.5 1.6 1.7 1.8 1.1
f ..... Densley, T ....
Fig. 3. Effect ofLS on Build Factor of 1 - and 3 - phase cores:
(L\BF=IOO*(BF ...... - BF .... )/BF WON)' %
Specific reasons for such a significant difference of LS effect
on NLLand BF. especially, the crossover (at B = -1.5 T) of the
curves representing M3F=f{B) are being investigated. Overall,
effect of domain refinement on NLL of cores stacked with the
newly developed GO is similar to that for HiS grades.
For 0.27 mm cores decrease ofNLL by more than 10% resulted
from even higher decrease of Power Loss in steel itself, while
for 0.30 mm (3-pbase) similar of NLL (in percents) is
caused by decrease of BF (ai B> L5 T).- 10 pnnclpJe, higher
effect of LS on decrease of NLL for cores stacked with 0.27
mm steel follows the trend found for HiB steel: thinner steel is
more susceptible to LS treatment[2J, but it much more
pronounced in our experiments. Earlier studies of power loss
in GO GO with very large grains (> 30 mm) have shown increase
of loss for narrower strips [S). It was found that deviation of
[00 I J axes from rolling direction of the grains at strip edges
results in higher flux density in the middle of the strip. Domain
refmement facilitates rotation of the domains in mis-oriented
grains. which results in lower power loss.
3.2 Effect of Laser Scribing on Noise Level
Results of measuring level of acoustic pressure or core noise
(produced by vibration of laminates in a core due to
magnetostrlction) presented in Fig. 4.
ACOUSlic noise decrease after LS. 3-phase cores. dB

I :27 cores l-I
] O.30cores I
III f1l" r:

: I--- I---

r- I-- e-

;' f----

f-- '--- I-- -
- I--

- I-- I-- I--
1.3 1.4 1.S 1.6 1.7 1.8 1.9
TestlndUcdon. Tesla
FIG 4. Change of acoustic noise level (dB before LS - dB
after LS), 3-phase cores
Laser Scribing typically increases volume of closure domains
(sub-domain structure), which increases the noise level (2 to 3
dB)f'J. Instead, in our trials a decrease in sound level was
measured. It's very plausible that in large grained steel LS leads
to lower volume of closure domains:
Most relevant (from a practical point of view) results of this
paper presented in Table I.
There's significant difference in effect of LS on 0.27 and 0.30
mm steels: power loss (single sheet samples) in 0.27 material
decreases twice as 'much as in 0.300 mm. On the other hand,
only slight change of BF is for cores stacked with
0.27 mm, while for 0.30 mm cores there's a pronounced decrease
. .14
ofBF .. As a result, the.overall decrease ofNLL is very similar
for both steels.
Noise reduction of more than 2 dB, while relevant in absolute
value, is relatively small and needs to be verified in more
extensive production trials.
B test,
Decrease of Losses after LS for samples
and 1 and 3-phase stacked cores
Loss improvement, %
samples I-phase 3-phase
13.4 12.2 11.3
15.4 14.9 13.9
14.9 14.2 15.0
Change of Build Factor, 1 and 3-phase cores
B test, ABF=100* (BF before - BF after)/BF after
Tesla Single-phase Three-phase
1.6 1.0 - 2.6
1.7 1.0 -1.8
1.8 1.0 0.0
0.30 mm
B test, Loss improvement, %
Tesla samples I-phase 3-phase
1.6 6.1 13.1 8.8
1.7 6.8 13.3 11.8
1.8 5.6 14.7 12.2
Change of Build Factor, 1 and 3-phase cores
B test, ABF=100*. (BF before - BF after)lBF after
Tesla Single-phase Three-phase
1.6 8 3
1.7 8 6
1.8 10 7
Table 1. Effect of Laser Scribing on Power Loss of Single Sheet
samples and NLL and BF of cores stacked with 0.27
and 0.30 mm laminates.
1. Effect of laser scribing on power losses of conventional
grain-oriented electrical steels, by Vladislav Wiglasz and
Petr Pacl, of Magnetism and Magnetic Materials.
1992, Volume 112. Issues 1-3.1 Pages 186-188
The latest advance in very low core loss grain oriented
silicon steels, by K. Veno, N. Takahashi, and
T. Nozawa, Journal of Materials Engineering, 1990,
Vol.l2. Number I, March. pp 11-20
Relationship Between Total Losses Under Tensile Stree
in 3 Percent Si-Fe Single and Their Orientation
Near {110)[001]. by T. Nozawa, T. Yamamoto at aI., IEEE
Transactions 011 Magnetics, 1978, vol. MAG-14, No 4,
4. Changes in the magnetic domain structure of silicon steel
caused by heating with a fret(-running ruby laser, by M
Domagala at aI., Quantum Electronics, 1997, vol. 27, pp.
5. Some Peculiarities of Interrelation of magnetic Parameters
in thin silicon iron sheets, by Yu.N. Starodubtsev and V.
M Segal, Physics of Metals and Metallography, 1984. vol.
58, No.4, pp. 45-49
6. Magnetic domain refinement of silicon-steel laminations
by laser scribing, by S. Patri at al . Journal of Materials
Science, 1996, Volume 31, Number 7, January. pp. 1693-
Ester Transformer Fluids for Improved. Fire Safety, Operational
Reliability and Environmental Protection
Sabine Bowers,
M&I Materials Ltd
Manchester, UK
The need for reliable and secure power
transmission and distribution systems, in a world
that faces the two challenges of an ageing
transformer population and a rapid rise in demand
for new transformers, is putting huge pressure on
security of supply. These pressures have caused
an increase in interest in fire safe and operationally
improved dielectric fluids as viable alternatives to
mineral oil. In order for these fluids to have an
appeal for both existing and new transformer
installations, they must demonstrate that they offer
a high standard of electrical performance over a
long working life, providing a lifetime benefit and
that they fit in with today's requirement of ground
water and environmental protection. This paper
illustrates the advantages ester fluids offer in terms
of fire safety, moisture tolerance and operational
reliability and as a replacement for mineral oil in
critical locations. Environmental benefits and fluid
testing are also discussed.
Dr Russell Martin
M&I Materials Ltd
Manchester, UK
Synthetic ester dielectrics have been successfully
used for almost 40 years and continue to grow in
popularity. Originally developed as a non harmful
replacement of Askarels (PCB). Esters are also
increasingly used in major replacement. projects for
Mineral oils in fire critical locations.
In the last 10 years there has also been
resurgence in interest in the use of natural ester
dielectrics because of their obvious 'renewable'
Esters- are not as widely known as more traditional
transformer insulating methods such as mineral oil,
silicon fluids or dry type transformers, therefore
there is still widespread confusion about the nature
of ester dielectric fluids. This paper attempts to
explain how they differ from other types of liquid
insulation and looks in detail at a few of the key
advantages tllat ester fluids have to offer.
What is an 'ester'?
, Keywords: The chemical structure of Ester molecules and
their origin is very different form mineral oils. The
term 'ester' comes from the chemical linkage which
is formed from the reaction of an alcohol and a
fatty acid.
Esters, fire safety, moisture tolerance, retrofilling,
fluid testing, biodegradability, environment
Esters have been used as dielectric liquids since
the invention of the oil-filled transformer in the late
1880s. The earliest simple natural ester dielectrics
were subsequently found to be incompatible with
free breathing equipment, because of their
chemistry, and were gradually replaced by mineral
oils. Later, silicon oils made an appearance.
offering a high fire factor alternative. However.
these were found to be environmentally
questionable and more difficult in high voltage
applications. C] Synthetic esters, developed in the
1970s, were designed to overcome these
drawbacks and can operate in free breathing
transformers as well as in high voltage
The ester linkage occurs in both natural and
synthetic esters, but does not occur in mineral or
silicone oils. Table 1 shows the difference in the
chemical structure of esters, mineral oil and
silicone oil.
Synthetic esters
Synthetic esters are derived from chemicals. They
are usually the product of a polyol (a molecule with
more than one alcohol functional group) with
synthetic or natural carboxylic acids to give
structures where several acid groups (usually 2, 3
or 4) are bonded to a central polyol structure.
Importantly, the acids used are usually saturated
(no e-c double bonds) in the chain, giving the
synthetic esters a very stable chemical structure.
Natural esters
Natural ester dielectrics are produced from
vegetable oils, manufactured from renewable plant
crops. Plants produce these esters as part of their
natural growth cycle.
. The structure of the natural esters is based on a
glycerol backbone which is bonded to 3 naturally
occurring fatty acid groups. Again, these fatty acids
may be the same or different. There is one key
altterence to synthetic esters: many of the chains
present in natural esters are unsaturated (contain
C=C double bonds) and this means that they are
less oxygen stable. They are therefore considered
mostly suitable for hermetically sealed applications
and not recommended for retrofilling free breathing
Key differences between esters and other types
of transformer oils
Because esters are more expensive than
traditional mineral oil insulation, the operational
and environmental benefits of using esters have to
be well understood and of key importance to the
The structural and key properties of the common
currently used transformer fluids are summarised
in Table 1.
These key differences, making esters the preferred
choice in many applications today, will be
discussed in detail below .
Improved Fire safety
Fire safety is a key concern for today's users of
dielectrics. This is especially so when conSidering
their use in areas where a fire would be highly
damaging and difficult to control for example in
public buildings, apartment blocks, industrial
applications such a steel works, in subway tunnels,
or in financial centres.
Both natural and synthetic esters can offer a high
degree of fire safety, due to their high fire points.
Smoke density, the actual damage from the
transformer fire but also damage from pool fires
needs to be taken into consideration.
To reduce this fire risk, existing mineral oil filled
transformers are increasingly being considered for
retrofifling with esters. Importantly for
Name Mineral oil Silicone oil MIDEl 7131 MIDEl eN -----1
Refined oil
Type based distillate Synthetic oil Synthetic ester Natural ester
mixture of
CHI 01, CHl

di-alkyl silicone polymer
, r 0 \ \ /'
;yi-r In --yi
tetra ester
R 0
-.- t-
R 0
Plant based natural
r ,
l o.
I ?

Source Purified from oil Made from chemicals chemicals Extracted from crops

Very slow to Effectively does not Readily Readily
b;odegrade biodegrade biodegradable ___ _
---- - --- . ---
Performance P rf . . Excellent Good mo,'sture
e ormance senSitIVe
Sensitive to . moisture tolerance
moisture to mOisture tolerance
--- .. -.-------.---
Water saturation 55 220 2600 I 1100
----->2-50----f- __
'C 170 >350 + ___ >-.3_0--,-0 ___ . ____ . ___ >.-::3:..=5.0..:._______
Fire 0 KKK
____ , _________ ________ .. _______ L _ . _________ -'----___________ .
Table 1 - Common FlUId Properties
retrofilling. Esters and Mineral oils are comptatsly
miscible and can be used interchangeably in most
designs of distribution transformers.
Fluid type Flash point Fire point Class
OC ac
Mineral oil 160 170 0
Silicone oil >300 >350 K3
>300 >350 K2
>250 >300 K3
Table 2 - Fire properties of fluids
Fire Safety Tests in Situ
To demonstrate behaviour in a worst case
scenario, a synthetic ester (Midel 7131) filled
transformer was sacrificed to demonstrate that in
the event of a serious electrical fault, the fluid does
not contribute to any resulting-firs'fl.
The transformer, installed in a test rig, was
subjected to a massive power overload, designed
to simulate a lightening strike or serious fault. It
was intended that the transformer be destroyed to
simulate worst-case conditions.
The transformer is set up in the test rig:
Time =0
The overload is introduceo causing an initial flash.
The tank ruptures due to the initial pressure surgt:::
. Tirne = 1 second
The tank splits and the fluid pours onto the floor.
The resulting liquid pool has not ignited.
Time = 3-4 seconds
Only very little, low density smoke can be
observed. After 3-4 seconds, the initial flash
caused by igniting fault gases and vapourised oil
starts to rapidly die away.
Time;; 7 seconds
After 7 seconds, the initial flash has self-
extinguished. The pool of Midel 7131 on the floor
rapidly cools below its flash point and hence
cannot contribute to a fire. In critical areas, esters
are often the most appropriate fluid because of
their high fire safety credibility.
Fire Safety Test - Victim of fire
Another scenario commonly inllolving a
transfor! Iler is vvllc:re the transformer falls victim to
an external fire.
A test was conducted at the German Insurance
Company Allianz in their fire testing laboratory [3]
to demonstrate that in this scenario the transformer
fluid will not contribute to a fire. The test set-up
was as follows:
630KVA Transformer (365kg Mide17131)
Floor Area 10m X 6m
Fire Load 180kg Pre-Dried Cribs
12 thermocouples were attached to the
transfornler both inside anJ outside at different
points to measure the temperature rise during this
W(.'llt,,.. ""-.
Fig 1: location of the 12 thermocouples
The cubes were ignited and the temperatures
measured. The fire was observed during the test
duration of over 70 minutes.
! _... -----------

v --- ------ .. ...... , ..... , , . ....,---,- --1
'Ii ':: ::' U
: 1;","""",,,,,
I fll.;nI:oelcnclil"
112 1 3.l
.. 1)1
:; It 41
- .. j .. ....,.--r-..-,
40 ,4 02
Fig 2: Temperatures measured outside: temp/min

'4 i
Fig 3: Temperatures measured inside: temp/min
Inside the transformer, the temperatures of the
Midel7131 rose to a maximum of 180'C (bottom)
and 204'C (top) in the 53rd minute.
"No dangerous operating state arose in or on the
transformer filled with Midel 7131. Throughout the
duration of the fire, the gases escaping from the
pressure relief valve from the dome region could
not be ignited nor did the transformer leak at any
pOint. ,,[4]
The transformer filled with Midel 7131 did not
contribute to the fire incident. Despite the harsh
conditions of the test, the transformer was found to
be still electrically operational afterwards.
Operational Reliability - the effect of moisture
The ester linkages present in both natural and
synthetic esters make these fluids 'polar'. like a
magnet, these molecules contain regions (the
ester linkages), which are able to attract other
polar molecules.
Water, a polar molecule, is made up of two very
different types of atom, hydrogen and oxygen
Consequently, esters have a particular affinity for
water molecules in a way that mineral and silicone
oils cannot.
, .. ( 'H
R/ 'OR'
Fig 4 - How esters attract water molecules

,_. __ cO- ----- ---
t ---I
r..=:s,;.;;;,-&;l :
. i I
1-- -- - -_01
. -_01 I :
L _______ I
! .
----"1.1O...!.-------------.------------.-- ."-- --
Fig 5 - Log plot of water content vs temperature
Natural esters have 3 ester linkages per molecule
whilst synthetic esters may have 2-4 linkages per'
molecule. These differences become evident when
we consider the amount of water that will dissolve
in these fluids.
Approx water
saturation at 23C
Mineral oil 0 55
Silicone oil 0 220
Natural ester 3 1100
Synthetic ester 4 2600
Table 3 - Solubility of water in fluids
The solubility of water in all these fluids increases
with temperature. A logarithmic plot of the water
solubility against temperature, Figure 5, shows an
approximately linear relationship, with different
materials having different gradients [5 6]. Clearly
the more polar esters are able to absorb a lot more
water across the temperature range.
The absolute amount of weater that a dielectric fluid
contains can have a dramatic effect on its electrical
properties. Polar fluids have significantly more
water tolerance, and this can be clearly seen in the
effect of water content on breakdown voltage ( 8].
Non-polar fluids such as the mineral oil are
particularly sensitive to the absolute moisture
1Ct r- - - - - -
100 300 sw eoo ... IlOO
......... c:or-t(ppmt
Fig 6 - Plot of breakdown voltage vs absolute water
Operational reliability and extending
Transformer life
Paper (cellulose) is still the most common form of
solid insulation used in transformers. Cellulose is a
polymer structure made up of many glucose units
arranged in chains:
H060H 0
OH, \

The average number of glucose units in the
polymeric chains, called the 'degree of
polymerisation' DP, dictates how mechanically
strong the paper IS: the longer the chains, the
stronger the paper.
A key indicator of the potential working lifetime of a
transformer is the condition of the paper insulation
In a new transformer, the DP of the cellulose
insulation is typically between 1000 and 1200.
As transformer ages, the paper (cellulose)
begms to degrade and de-polymerise, forming
chains of shorter length (lower DP). The overall
effect of this is that the paper becomes
progressively weaker [9 1.0 11 12 13]. When the
average DP reaches 200, the paper insulation is
said to have reached the end of its working life.
This means that it should be completely replaced,
or the transformer scrapped.
A by-product of this de-polymerisation process is
the formation of water, which accumulates in the
paper structure. Worse still, the presence of water
causes an acceleration of the rate at which the
paper degrades. This leads to a situation where
or those in poor condition may
rapidly towards the end of their working
hfe, ultimately leading to failure. There is now a
wealth of scientific papers which show that a
. reduction in the amount of water held in the
cellulose insulation significantly extends the useful
life of the paper.
In this regard, the affinit.r. of esters for water has
been put to good use C ].
Retrofilling mineral oil transformers
!n leil 11.,;:: ui eXIsting
mineral oil filled transformers becomes a frequent
necessity due to theUAacceptabie moisture
content affecting the dielectric quality of the oil and
the ageing of the cellulose. For this reason
retrofilling the transformer with an ester fluid is
often considered, not only to increase fire safety
but also to reduce the frequency of maintenance
cycles and improve reliability of the equipment.
In a field study, a previously redundant BOOkVA, 30
year old wet mineral oil transformer was retrofilled
with a synthetic ester, MIDEL and the
results can be observed in Table 4.
(Mineral Oil)
7131 )
22 kV >65kV
51.4 ppm 400 ppm
Core Insulation
Table 4: Transformer performance before and after
Follow up studies C
J show that the ester absorbed
the water from the cellulose, and that after a short
period of time the transformer was able to be
returned to active service.
100, ,-
E eo r-

! 40: ...

00 o.s 1D 1.5 2-0 2.S 3.0 35 40 4.5 so
T-.(Y ... )
Fig 7: long term retrofilling results breakdown voltage vs
Long term retrofilling experience has demonstrated
that ester transformers continue to operate safely
even if ther have absorbed more than 400ppm
moisture. [ 7]
Both natural and synthetic esters are miscible with
mineral oil in all proportions. Synthetic ester can be
used in aU equipment, including free breathing
transformers and has been used in this type of
application for almost 30 years.
Ester fluid manufacturers will provide printed
guidelines for exactly how this can be done easily
and safely.
Environmental behaviour
The environmental properties of esters make them
increasingly attractive alternatives to mineral oil.
Both natural and synthetic esters are classified as
being 'readily biodegradable'. This means they
pass strictly controlled degradation tests carried
out according to DECO methods These test
methods are internationally established and
recognized. The behaviour contrasts markedly with
mineral oil (very slowly biodegradable) and silicone
oil (practical1y non-biodegradable).
:-----:-. .--:;.--.-.:..----S;IIcone-OM
Fig 8 - Comparative biodegradation rates
In addition, esters have demonstrated that they are
also non-toxic to land and water based life. These
key factors have prompted the German Federal
Department of the Environment
(Umweltbundesamt, USA) to classify these esters
as "non water hazardous"C
20]. Esters therefore
are less harmful to the groundwater.
Rapid biodegradability and non-toxicity are two key
factors that allow esters to be used where a leak or
a spillage into a watercourse for example, would
be an issue.
Maintenance of ester fluid equipment
The standard set of tests to determine fluid quality
for mineral oil can also be used with ester fluids.
Typical examples are: breakdown voltage,
neutralisation value and moisture content, among
many others. However, test results will be different
to those expected from mineral oil, and the
appropriate standards for ester fluids should be
consulted [21 22].
The range of DGA gases produced by esters are
the same as for mineral oil, namely hydrogen,
methane, ethane ethylene, acetylene, etc. When
diagnosing transformer faults, the diagnostic charts
of ratios (e.g. Rogers ratios or Duval triangle) can
be used with the mineral oil values to give a
reasonable guide [2324]. For example a strong
arcing fault will produce acetylene from both
mineral oil and esters. However it is true to say
that the amount and proportion of the fault gases
produced are slightly different from mineral oil, and
some care needs to be taken, especially in the
diagnosis of borderline cases.
The CIGRE group TF15 headed by Michel Duval
[25] is currently investigating DGA in alternative
fluids and is due to issue a formal report in due
course. The University of Manchester in the UK
has also carried out excellent work in this area [26
27]. .
1 V. Wasserberg: "Retrofilling of Transformers - Facts
and Arguments" Independent Condition Assessment and
Concepts for the refurbishment of Transformers 2004.
2 M&I Materials Internal video ~ o r d of tests carried out
by a European transformer manufacturer
3 "Fire Test on a Transformer Filled with MIDEL 7131"-
Allianz Fire Protection Service
4 "Fire Test on a Transformer Filled with MIDEL 7131" -
Allianz Fire Protection Service
5 Fofana, I.; Wasserberg, V.; Borsi, H.; Gockenbach, E.
"Challenge of mixed insulating liquids for use in high-
voltagetransformers.1. Investigation of mixed liquids" -
Electrical Insulation Magazine, IEEE
18, 3, May/Jun 2002 18 -
6 A.Burgess "Saturation levels of MIDEL eN at ambient
and 80C", 2007 - M&I Materials internal technical
7 Lance Lewand, Doble Engineering "Understanding
water in transformer systems" - Neta World report, 2002.
8 R.Martin - Internal Technical Report, M&I Materials.
9 L. Lundgaard, W. Hansen, D. Linhjell.T, Painter
"Ageing of oil impregnated paper in power transformers"
- IEEE Transactions on Power Delivery 2004.
10 D.J.HiII, T.T.Lee, M. Darveniza, T. Saha. "A Study of
degradation of cellulosic insulation materials in a power
transformer Part 1. Molecular weight study of cellulose
insulation paper" - Polymer Degradation and Stability 48
(1995) 79-87
11 D.J.Hill, T.T.Lee, M. Darveniza, T. Saha "A Study of
degradation of cellulosic insulation materials in a power
transformer Part 2. Tensile strength of cellulose
insulation paper - Polymer Degradation and Stability 49
(1995) 429 - 435
12 D.J.Hill, T.T.Lee, M. Darveniza, T. Saha "A Study of
degradation of cellulosic insulation materials in a power
transformer Part 3. Degradation products of cellulose
insulation paper" - Polmer Degradation and Stability M
13 T.Prevost "Degradation of cellulose Insulation in
Liquid filled Power Transformers" - EHV-Weidmann
Industries Inc. W-ACT12005 4th Annual technical
14 V Wasserberg, H Borsi, E. Gockenbach, I Fofana,
University of Hannover, Germany "Drying of Liquid
immersed solid insulations using a hygroscopic
insulating liquid".
15 A.Smith "Drying cellulose Insulation with MJOEL
7131" 2003. Internal Technical Report, M&I Materials.
16 A.Smith "Drying cellulose Insulation with MIOEl
7131" 2003. Intemal Technical Report, M&I Materials.
17 M. lashbrook, "Retrofilling Wet Transformers with
MIDEl 7131" 2007 - Internal Report, M&I
19 "Ready Biodegradation Of MIUl:l 7131 and MIDEl
eN". Cantest 2007 - M&I archive document.
19 MIDEl 7131 NWG status confirmed by letter by
20 MIDEL eN NWG status confirmed in writing from
21 8S EN 61099 (IEC 61099) "Unused synthetic organic
esters for electrical purposes", 1992.
22 8S EN 61203 (IEC 61203) "Synthetic organic esters
for electrical purposes - Guide for the maintenance of
transformer esters in equipment" 1992.
23 R.R.Rogers , "IEEE and IEC Codes to Interpret
Incipient Faults in Transformers, Using Gas in Oil
Analysis" IEEE Trans, Vol El-13, N05, Oct 1978, P
24 lance lewand, Doble Engineering "Using dissolved
gas analysis to detect active faults in oil insulated
electrical equipment" - Practical Oil Analysis.
25 Michel Duval "Dissolved gas analysis: It can save your
transformer" - Electrical Insulation magazine, IEEE Vol
5, Issue 6, Nov-Dec 1989, pages 22-27.
26 I Khan, Z.o.Wang, I. Cotton "Dissolved Gas Analysis
(DGA) of Alternative Fluids for Power transformers" -
University of Manchester, UK. Submitted to IEEE
Electrical Insulation magazine in 2007.
27 J Oai I Khan, Z.o.Wang, I.Cotton "Comparison of
HYDRAN and laboratory DGA results for electric faults in
ester transformer fluids" - University 01 Manchester, UK.
Submitted for the IEEE CEIDP conference in October

3D-magnetic measurement in joint part of Silicon steel model core
H. Yamaguchi, M. Ishida, H. Pfiitzner*
Steel Research Lab., JFE Steel Corporation, Kawasakidori 1-chome, Mimshima, Kurashiki, 712-8511, Japan
Institute of Fwuiamental and Theory of Electrotechnics, Bioelectricity & Magnetism Lab., Vienna University of Technology, Gusshausstrasse 27/351, A -1 040,
Vienna, Austria
Grain oriented Silicon steels show a high degree of anisotropy which favors magnetization in the rolling direction.
Other direction of magnetization occurs in joint regions which are characterized by interlaminar flux. For the purpose
of direct measurement of those flux behavior, flat sensor elements were developed which do not create additional air
gaps. The present study was carried out using a model three phase transformer core. To pick up the local flux around
the joint part thin copper film elements were vaporized on the surfaces of both sides of individual laminations.
Normal flux components were detected by frame coil arrangements, in-plane components by films with tip contacts
through the coating. The results illustrates that in joints, the incoming in-plane flux is transferred to flux
in the overlap region, until a high degree of global saturation is reached, a state when flux through air gaps between
laminations yields a tendency of more homogeneous magnetization.
1. Introduction
Loss values of transformer core have been reduced by
improvements of both material and core design.
However, local disturbances of the magnetic flux
distribution in joint regions cause excess losses which
may be influenced in a rather unpredictable way. The
reason is that, apart from flux in rolling' direction (RO),
there arise components in transverse direction (TO) and
in normal direction (NO).
Many studies of flux distributions in overlap regions
of joints in three phase transformer core have been
reported [1-3]. It is complicated due to requirement that
sensor elements should not yield artifacts as resulting
from introduced interlaminar air gaps. In order to avoid
this, sputtered Au film elements of less than 1
thickness have been applied on the coated material in
earlier work [4]. Thus it was possible to establish surface
search coils that detect interlaminar flux in ND.
A novelty of the present study is given by further
surface film elements with end contacts to the
conductive material for the additional detection of
surface voltages resulting from eddy currents. This
yields "virtual" cross-sectional search coils for the
determination of flux in RD or TD as earlier proposed
for the so-called needle method [5,6]. Here we studied
much smaller model cores where it was possible to
arrange eu films by means of vacuum deposition on
both surfaces of individual sheets of core material [7].
O-joint C-joint
Fig.1. Example of an investigated model core exhibiting
a single joint part.
RO .
Fig.2. Example of the end region of a lamination with 1
thick copper film elements representing a frame coil
for the detection of interlaminar flux as well as spot
contacts for that of in-plane flux.
2. Interlaminar flux measurement by vaporized
search cons for joint part
Aiming at compact components, the experiments
were performed by small model transformer cores with
50 mm lamination width including a single modified
joint region (Fig.l). They were built up from commercial
high-grade grain oriented silicon steel (HGO) materials.
For the joints were also studied in linear
arrangement with overlap in 4Y or 90, respectively.
Stacking with two sheets per layer was done with
alternate lap and step lap. The overlap length was 5 mm
in all cases, the gap length within 1 mm.
Fig.2 shows an example of a lamination with film
elements of about 800 J.tm width at the miter end region.
In the depicted case, for the determination of local BND it
comprises a 15 mm x 4 mm surface coil - up to three
coils being used for studies of local variations. Signal
pick-up was favored through slight side shift of the
manipulated lamination out of the pack.
For the determination of flux BGD in gap direction (or
RD and TD, respectively), film lines were used that end
in a spot contact of about 0.2 mm diameter where the
coating had been removed locally. Pairs of spot contacts
of about 20 mm distance were arranged in distances up
to 20 miD from the gap region. Equivalent arrangements
were made at the lower surface of lamination in order to
separate eddy current components. In-plane flux
components were determined from U
= Uab + Ucd with Uab
the upper and Ucd the lower voltage (Fig.3), following the
above mentioned virtual search coil concept.
b a

c d
Fig.3. Concept of in-plane flux detection by means of
spot contacts (see text).
Region 1
Region 3
Coil B
lap length 5mm
Fig.4. Schematic outline of flux components in an
overlap region of alternate step type ..
The general distribution of flux take-over in
alternate-lap joint is outlined in FigA. With two sheets
per layer, a symmetry area is given which lacks flux in
ND. Considering the central sheet, the flux which comes
in to region 1 tends to flow in ND to the bridging region
BR and leaves in RD (or TD) to region 3.
Only at moments of high B, flux components through
the air gaps are likely to arise. This is reflected by Fig.5a
(for cHGO) which shows the flux through a frame coil A
positioned closely before the overlap: For 0.9 T, all flux
goes in ND, while for 1.3 T and 1.7 T, a time period of
flux reduction is given which is compensated by flux
through the gap in the butt. Fig.5b shows the flux in ND
to the bridging region BR measured by Coil B which is
maximum at that point of time.
-0. 12L--________ --'
0.000 0.005 0.010 0.015 0.020
lime /sec.
(b) 0.12 Coil B
1.3T 0.9T
0.06 1.7T"
-0.12L-. ________ --'
0.000 0.005 0.010 0.015 0.020
Fig.5. Normal flux as detected by flat frame coils. (a) In
coil A closely before the overlap. (b) In coil B in the
bridging region BR.
0.09 .
:::... . ..
-0.09 ... .._
-0. 12L-___ ---' __ -'--___ ---'
0.000 0.005 0.010 0.015 0.020
0.12 CoilS
0.06 .


0.000 0.005 0.010 0.015 0.020
Coil B
Fig.6. Normal flux according to FigA, except for
detection with frame coils of 50 j..tm thick wire (see text).
The effectiveness of the film method is illustrated in
Fig.6 which for comparison shows results for the
application of thin copper wires of 50 J-tm diameter. Coil
B yielded widely unchanged time responses, however,
with reduced intensities (Fig.6b). On the other hand, coil
A (Fig.6a) yielded almost no flux which can be
explained by the high magnetic resistance of the
interlaminar air gap leading to error in measurement.
Next, influence of lap method to distribution of
magnetic flux was considered by double-side ND flux
probe. In the case of alternate lap with two sheets per
layer, smooth take-over flux was observed in upper side,
but almost no normal flux BND in lower side in Fig.7b.
Sensor position is shown in Fig.7a. It suggests that
incoming flux from the opposite side suppresses take-
over flux due to symmetric lamination. However
lamination which flown into would be saturated in
magnetic flux and it causes increase of iron loss.
Magnetic flux
upper side
lower side
Fig. 7a Schematic outline of flux components on
alternate-lap joint
Lower side
- - 0.7
-- -0.9
- -1.1
- - 1.5
-0.2 L-___ .....-___
0.000 0.005 0.010 0.015 0.020
Fig. 7b time variation of BND on alternate-lap joint with
B = 0.5-1.7T
Step-lap joint with two sheets per sh?wed
moderate offset-flux in lower side as shown In Flg.8b.
Avoidance of flux concentration in joint region was
observed according to asymmetric lamination (Fig.8a). It
would prevent stray flux by over saturation and increase
of iron loss.
Relaxation by
asymmetrical stacking
upper side
lower side
Fig. 8a Schematic outline of flux components on step-
lap joint
Upper side
-0.2 Lp-..... ===-_=-=6..::-=;-'-------'
Lower side
: 0.5
- - 0.7
- -1.1
- -1.5
0.000 0.005 0.Q1 0 0.015 0.020
Fig. 8b time variation of BND on step-lap joint with B =
3. Needle probe measurement on three-phase model
Fig. 9 shows the configuration of the probe for 2-
dimentional localized magnetic measurements. The
probe consists of two pairs of needles whose tips are
made of tungsten placed Smm from each other. Two
Hall probes are mounted between these two needles,
making it possible to measure the surface magnetic field
over practically the same area as the needle probe span.
The 2D-needle probe (Bx, By) and 2D-Hall probe (Hx,
Hy) are arranged on 3-phase stacked model transformer
core and scanned the whole core area automatically. The
flux density and magnetic field are measured as vector
components. The model core is composed of 100mm
wide sheets and SOOmm x SOOmm in outer diameter.
20-needle probe (Bx, By) ,
Hall probe
(Hx. Hy) 20 (X-V)
__ area
3-phase slacked core
Fig. 9 2D local measurement on 3-phase model
transformer core
Normal flux components on lap regions between
limbs and yoke are detected by frame coil of deposited
. eu film as shown in Fig.lO.
Fig. 10 interlaminar deposited coil on lap region
Fig. 11 Schematic outline of 3D-flux components
measurement in overlap region of alternate-lap type core
Fig. 12 2D magnetic flux distribution around T-joint
(a) oot = 30 degree, (b) oot = 150 degree
2D magnetic flux distribution around T-joint is
expressed in vectors as Fig.12. Shading means
magnitude of 2D magnetic flux. Fig.12a and 12b are
instantaneous states corresponding to zero excitation in
W-limb and V-limb respectively. Magnetic flux that
flowed into zero-flux limb was observed in both cases.
In the former, magnetic rotation was observed on around
the center of the yoke between T -joint and W -limb.
in the order of W, V and U'
in this experimental condition (W -V -U excitation),
rotation of magnetic flux was clockwise. Inverse
excitation condition (U-V -W excitation) brought about
anti-clockwise tum of magnetic flux.
Flow-in flux in V-limb reaches the opposite T-joint as
shown in Fig.12b, which is known as the circulating flux.
The flux in the two half width of V-limb was opposite 1 y-
oriented; the actual phase difference was approximately
50 degree.
The phase angle difference tloot is defined as the
difference between the phase angle of flux averaged over
one segment out of ten comprising the total limb width
and that in the total width.
In W-V-U excitation condition, magnetization in a
limb width progress from left side to right side of the
limb according to 3-phase excitation as shown in Fig.l3.
Similarly, phase shift in the transverse direction was
observed in yokes on the both side of T-joint due to
partial circulation of magnetic flux toward the corner
The lower four figures show time dependence of NO
flux (offset-flux) density across the lap region of each
joint part and t!Ie upper three figures show those of RO
flux density in each limb at 0.3-1.9T, 50Hz in Fig.I3.
Distinct ND flux was observed at every lap region by
deposited eu search coil and showed saturation. Its
magnitude is supposed to be dependent on insulator
thickness, interlaminar space (ND), joint gap (RO-TO),
sheet shape, and anisotropy of material. After saturation,
magnetic flux went straight through joint gap in the butt
at high B region.
Attention should be taken to phase shift of NO flux,
which was synchronizing with RO flux density of each
limb in the comer joint as shown in No.1 and 4 in Fig.
14. Joint No.2, 3 played an intermediate role between
different phase limbs. This phenomenon is observed for
example, at a peak of downward flux in U-limbat 270
degree and peak of upward flux in V-limb at 330 degree.
In this case, phase shift between them is 60 degree. NO
flux in No.3 lap was observed widely at around 300
degree. However, we obtained average NO flux at whole
lap region. It would occur from end to end as previously
mentioned. Practically, in-plane phase shift of magnetic
flux at lap was detected corresponding to that of yoke
and limb.
.0 20 40 60 80 100
x positiori /mm


11Ot-- .-:-- -,.
200 220 240 280 280 300
x position ,mn

2101-............... .1....0. .......................
lO 420 440 460 480 500
x position Jmm
20 40 60 80 100 20 40 60 80 100
Y position /mm y position /mm
Fig.13 phase shift in the limb and the yoke of three phase transformer core
2.0,..V-r--Imb,-;-' O,..3. .,.' .,.9T.,.' 5OHz..,...,.--""
-01 80 120 tao 240 300 -0.1 ..0.1 60 120 180 240 300
0.000 0.005 0.010 0.015 0.020 0.000 0.005 0.010 0.015 0.020 0.000 0.005 0.010 0.015 0.020 0.000 O.fI05 0.010 0015 0.020
time /SeC. time /SeC. time /SeC. !me Is;
Fig. 14 time dependence of RD flux in the limbs and ND flux at lap regions of three phase transformer core
Iron loss 'of transformer core is considered to be
separated to in-plane loss and off-plane loss. The former
is of RD-TD magnetization origin, which would be
deteriorated by rotation of magnetization due to
deviation from RD around joint part and distorted
excitation wave form in the limb. The latter is of ND
magnetization origin, which cause offset flux on the
surface of lamination and off-plane loss.
To predict the building factor of a transformer core,
accurate evaluation of these contributions is necessary.
The in-situ 3D magnetic measurement technique in the
present study must be useful tools on this point of view.
4. Conclusions
The present study reports the application of thin film
elements for 3D analyses of magnetic flux distributions
in joint regions of model cores. The main conclusions
are the following:
(a) Comparison of 50 11m wire elements with 1 11m film
elements demonstrated that the latter technique is a
prerequisite for effective analyses.
(b) Thin films arranged on the coating of lamination
worked as frame coils for the determination of
interlaminar flux components in normal direction.
(c) Films on the coating with end contacts to the
conducting material at spots of removed coating
yielded "virtual" search coils for the detection of in-
plane flux components.
(d) The application of the method in joint part of model
core illustrated that in the overlap region, in-plane
flux is transferred to interlaminar flux. However, in
instants of high induction, the latter is restricted due
to approach to saturation. This can be interpreted as
a high degree of homogeneity of the joint part is
(e) The series of measurements revealed that almost all
parts of three-phase transformer core undergo
distortion and phase shift in magnetization even in
the limb center part, which is totally excited to a
sinusoidal flux waveform.
[1] M.Jones, A.Moses and J.Thompson, "Flux
distribution and power loss in the mitered overlap joint
in power transformer cores," IEEE Trans. Magn. MAG-
9, pp114, 1973
[2] C.Bengtsson, H.Pfiitzner and Schonhuber, "On the
optimization of Il).itered overlaps in transformer cores,"
Physica Scripta, 40, pp629, 1989
[3] T.Nakata, N.Takahashi and Y.Kawase, "Magnetic
performance of step-lap joints in distribution transformer
cores," IEEE Trans. Magn. MAG-18, pp1055, 1982
[4] A.lIo, H.Pfiitzner, H.Bangert and Ch.Eisenmenger-
Sittner,"Sputtered search coils for flux distribution
analyses in laminated magnetic cores," 1.Phys. IV
France 8, pp733, 1998
[5] K.Senda, M.Kurosawa, M.lshida, M.Komatsubara
and T.Yamaguchi, "Local magnetic properties in grain-
oriented electrical steel measured by the modified needle
probe method," 1. Magnet Magn. Mater., 215, pp136,
[6] G.Luisus ami A.j''tlu1)c::', C l i t i ~ d l J::,'tdludlluH iiUU
Limitations of Localized Flix Density Measurements in
Electrical Steel," IEEE Tras. Magn. MAG-37, pp2755,
[7] H. Yamaguchi, H.Pfiitzner and E.Mulasalihovic, "3D-
measurements of magnetic flux on joint regions of SiFe
steel model core," J. Magnet Magn. Mater., 320, pp935,
Hiroi Yamaguchi, Masayoshi Ishida, Helmut P/iltzner,
D.V. Jagannathan, T.e.S.M. Gupta
Increasing power generation , transmission and distribution
demand more reliable performance of the Power Transformers
requiring mineral insulating oils of superior quality and
enhanced performance levels. Over the years due to cost factors
and improvements in transformer design technology , the
transformer sizes have shrunk but at the same time there is
appreciable increase in the severity ofthe transformer operating
conditions due to the extra high voltage requirements. These
transformers of more compact material efficient designs need
better circulating oils for more effective cooling with higher
electrical and ageing properties, lesser gas evolution and enough
gas absorption for safe operation and effective diagnosis of the
operating transformers. Also the new requirements of non
corrosive oils to ensure prevention of copper sulphide corrosion
on the conductor - paper insulation faced recently in many parts
ofthe world being attributed to the levels and types of corrosi ve
sulphur compounds in the oil combined with the extra severity
ofthe operating conditions. Working groups like Cigre , IEEE,
Oil Refiners, major OEMs, Utilities world over and Independent
laboratories such as Doble are studying this problem. Research
at elevated pace is called for to understand this phenomena and
to come out with proper recommendations to overcome the issue
of copper sulfide corrosion meeting the oxidation stability
Emerging trends in today's Power Transformer operations
have thrown the following new challenges that are required to
be met for trouble free performance :
1. Increase in Severity of Transformer Operating Conditions
2. Extra High Voltage Requirements I HVDC Systems
3. More Compact Designs require better circulation oils for
effective cooling
4. Higher oxidation stability requirements for longer trouble
free service life
5. Combating new problems like the Copper Sulphide
6. Lesser Gas Evolution requirements in Transformers
To meet these increased severe demands Transformer oil
requirements and specifications have undergone considerable
changes with increase in severity of some ofthe test parameters
and inclusion of more stringent tests and limits especially in
terms of Oxidation Stability and Corrosive sulphur .
Severity of Oxidation Stability tests has increased (
Table - I ) from 100 0 C with oxygen flow in IS 335 &
IEC 296 to 120
C with Air flow in IEC 60296/ BS 148

to ensure higher performance representative of actual
transformer working conditions for extended oil life in a
transformer. IEC 60296 Standard also stioulates more
severe oxidation stability requirements with stringent
limits for Special Application. requirements . Additional
tests like Power Factor Valued Oxidation (PFVO) &
Sludge Free Life (SFL) included in the Doble TOPS are
becoming more significant in deciding life expectancy of
More stringent Corrosive Sulphur tests ( Table - 2 ) like
DIN 51353 ( Silver Strip, 100 0 C , 18 Hrs) , Modified
ASTM D 1275 B (Copper Strip, 150
C ,48 Hrs) and
Cigre TF.A2.32.0 1 CCD- Closed Conductor Deposition
test ( Copper Strip + Kraft Paper, 150 0 C,72 Hrs) are
introduced in place of the less stringent earlier tests like
IS 335 Annexure B, ISO 56621 BS 5680 (Copper Strip
, 100 C , 19 Hrs) & ASTM D 1275 A (Copper Strip,
C, 19 Hrs) for better corrosive sulphur controls.
lEC 60296 has included 0.15 % max total sulphur content
for special application requirements. New test method for
detection of potentially corrosive sulphur involving
Copper strip and paper and examining copper by ASTM
D 1275 B method and determining total copper I sulphur
content of paper to evaluate the presence of copper
sulphide deposits on paper by SEM-EDX method is also
under consideration by lEe. SpecifYing Mercaptan SUlphur
content by ASTM D 3227 - 04 A with stringent
acceptance limit is also being explored.
Lower Gassing tendency limit ofless than + 5 is specified
in BS 148 for ensuring low gassing oils for effective
diagnostics by DGA , safer operations and also avoid risk
of insipient faults taking place in transformers getting
Break Down Voltage under impulse conditions is specified
as 145 minimum in ASTM D 3487 & Doble TOPS for
ensuring better insulation life of Power Transformers .and
HVDC Systems.
PCA, PCB, 2 - FAL Content are specified with stringent
limits .in IEC 60296 1 BS 148 for meeting the
environmental regulations
New properties of Lowest Cold Start Energizing
Temperature ( LCSET ) and Electro Static Charging
Tendency ( ECT) have been included.
Meeting these ever demanding specifications of High Oxidation
stability requirements, stable electrical properties, nm1 corrosive
requirements under the more severe operating conditions along
with the other properties is often contradictory and poses
enormous challenge to oil refiner to optimize the transformer
IEC 74 specified in IEC 296 -1982 100 0 C , 164 Hrs, Oxygen
IS 335 - 1993 & IS 12463 -1988 100 0 C , 164 Hrs , Oxygen
( For Inhibited Oil ) Plus RBOT- Rotating Bomb
Oxidation Test only for
Inhibited Oil
ASTM D 2440 specified in 1I00C, 72& I64Hrs.
ASTM D 3487 - 00 Oxygen
( Reap proved 2006 ) & Double
TOPS.- 2006
IEC 61125 C specified in IEC 1200 C , 164 Hrs ( U ) 1332
60296 - 2003 General Specification Hrs ( T ) 1 500 Hrs (I ) Air
IEC 61125 C specified in IEC 60296 1200 C , 164 Hrs ( U ) / 332
- 2003 for Special Applications with Hrs ( T ) / 500 Hrs ( I ) , Air
stringent limits for Total Acidity,
Sludge & DDf at 90 0 C
ASTM D 2112 - Additional test RBOT- Rotating Bomb
specified in ASTM D 3487 - 2000 Oxidation Test for Inhibited
(2006) & Doble TOPS ( 2006 ) Oil
Additional Tests specified in Doble - Periodic measure of Power
TOPS. PFVO ( Power Factor Valued factor till 140 Hrs when oil
Oxidation) is aged at 95 0 C.
SFL ( Sludge free Life) - Measures the time limit in
8 Hrs periods until oil forms
The IS: 335 Standard with its recent revision in 1993 laid down
in the country meant for the paraffinic type transformer oil is
attributing to only the basic requirements of insulation and
oxidation stability at lower severity levels where as it is still not
addressing to the more important performance requirements like
better cooling characteristics, lower gassing tendency, better
corrosive sulphur controls to meet the more severe new tests,
higher oxidation stability under more severe test conditions,
higher Break Down Voltage under negati ve impulse conditions,
lower PCA ( Polycyclic aromatics) levels and ensuring free
from the carcinogens like PCBs ( Poly chloro bi-phenyls ),
low 2-Fal and Furan content limits etc. Because of these
limitations the oils meeting the IS:335 specification can be at
the most considered as adequate only for the lower rating
distribution transfonners and for the extra high voltage power
transformers ( especially 220 kV and above ) it is becoming
essential to use only oils of upgraded specification for ensuring
enhanced perfonnance levels and longer service life.
International Electrotechnical Commission has taken efforts
in upgrading the oil standard in the year 2003 by replacing the
ISO 5662 specified in IEC 296-1982 Copper Strip, 100 oC , 19 Hrs
BS 5680 specified in BS 148 - 1984 Copper Strip, 100 oC , 19 Hrs.
IS 335 - 1993 & IS 12463 - 1988 Copper Strip, 100 0 C , 19 Hrs
( For Inhibited Oil)
ASTM D 1275 A specified in Copper Strip, 1400 C , 19 Hrs
ASTM D 3487 - 88
DIN 51353 specified in IEC Silver Strip, 100 0 C , 18 Hrs
60296 - 2003
ASTM D 1275 B specified in ASTM Copper Strip, 1500 C ,48 Hrs
D 3487 - 00 ( 2006 ) & Doble TOPS
(2006 )
Cigre IF. A2 .32.01 - Closed Copper Strip + Paper, 150 0
Conductor Deposition Test specified C,72 Hrs
by OEMs / Utilities.
lEC 62535 Ed.1 - under Copper Strip + Paper, 150 0
consideration by IEC- Detection of C , 72 Hrs Examining. copper
potentially corrosive sulphur by ASTM D 1275 B Finding
copper / sulphur content of
paper to evaluate copper
sulphide deposits by SEM -
EDX method
Specifying Mercaptan Sulphur
Content by ASTM D 3227- 04 A
is also being explored
earlier existing IEC Class I. II & III for uninhibited oils and
Class IA,lIA & lIlA for inhibited oils with one standard for the
three grades namely uninhibited, trace inhibited and inhibited
oils, differentiating them by the limits of antioxidant content
allowed and also increasing the severity of oxidation stability
test in terms of temperature along with different durations for
each of these grades. New concept of the lowest cold start
energizing temperature ( LeSET ) has been included and new
properties like electrostatic charging tendency ( ECT ) has been
added and values for the other properties have been revised.
ASTM D 3487, BS 148 ,AS 1767.1, DIN 57370/0370, Doble's
TOPS are some of the other upgraded transfonner oil standards
followed widely world over. These international standards ,
especially the IEC 60296 , are able to address the demands of
the present generation transformer oils more adequately than
the earlier generation IS: 335 Standard oil used in India. For
ensuring the oil quality to meet the more severe needs of EHV
Power Transformers. OEMs, Electricity Companies and Utilities
world over have developed their own more stringent transformer
oil specifications. Major such companies and utilities in our
country have also started following these international trends

and are upgrading their transfonncr oil specifications
Increased oxidation stability and making non corrosive oils by
removal of corrosive sulphur, both important for the
performance of insulating fluid are contradictory. Sulphur
containing molecules in lhe oils can give both negative and
positive characteristics to the oil as some type of sulphur
compounds can cause corrosion of copper and silver but can
act as peroxide destroying inhibitors in the oxidation process.
Some sulphur compounds (good stable sulphurs - Thiophenes ,
Carbazoles and sulphides) and aromatics (mono aromatics)
which are natural inhibitors help in better control of oxidation
stability and gassing tendency and hence are preferred in the
oil. On the contrary the sulphur compounds such as - Disulfides
( RSSR) breakdown at high temperature conditions and form
mercaptan sulphur (RSH) and hydrogen sulfide (H2S) which
are very corrosive to metals in the transformer. To separate
these compounds selectively within the similar molecular
distribution range becomes an extremely difficult task for an
hydrotreating operation. Hence corrosive sulphur removal and
also maintaining higher oxidation stability is a contradictory.
chaIlenge for the refiner requiring the right balancing act.
Good oxidation stability is essential in oils to maintain insulation
life for ensuring trouble free operation of Transformers for long
periods. Oxidation of oil gives rise to acids and sludge formation
and is influenced by the main parameters like oxygen,
temperature and also metals acting as catalyst for the reaction.
All oils contain certain amount of air even after degassing say
0.05 to 0.25 % oxygen by volume even in a dried and sealed
transformer which will participate in the oxidation. Heat
accelerates this deterioration, for example a temperature rise
of8 to 10C doubles the rate of deterioration. Oxidation process
comprises of four main stages namely initiation, propagation,
branching and termination.. It is the creation of free radicals
from hydrocarbons which is favored by heat , UV light and
high electric, stress fields. The hydrocarbon radical then reacts
with oxygen to form peroxide radical. The reaction will continue
by peroxide radical reacting with hydrocarbon to form a new
radical and peroxide. The peroxide fonned which is not stable
under the conditions of high temperature and electrical stress
will decompose to give two new radicals which will then react
with new hydrocarbons to form alcohol, water and new radicals.
The radicals will again react with oxygen and form more
peroxide and thus the chain reaction continues. The alcohol
will further oxidize to form' first aldehyde then carboxylic acid
/ ketone. The carboxylic acid may react with alcohol to form an
ester. In some of these reactions water is also formed which
will affect the dielectric properties of the oil and also cause
cellulose degradation.
Increase in severity of Transformer operating conditions with
Extra High Voltage Requirements and HVDe Systems need
higher oxidation stability requirements .due to higher operating
temperatures. Power Transformers often required to work at
the optimum or at times even on over load conditions resulting
in higher operating temperatures which demand non corrosive
oils with very low sulphur content and very high oxidation
stability. In HVDC systems where the electrical energy from
AC network of lower voltage is converted into higher voltage
DC ( rectification ) and at the other end of the transmission
line, the high voltage DC energy is again converted into low
voltage AC energy and fed into the AC network (inversion )
The transformer oil to work effectively in such severe
interchanging voltage and frequent switching operations has to
satisfy extra stringent levels of oxidation stability , corrosion
stability, negative gassing, lower viscosity and water content
and higher negative impulse breakdown voltage.
, "Oxidation of oil can be reduced as a result of optimum refining
yielding high oxidation stability by minimizing sludge
deposition, electrical losses, metal corrosion, electrical faults
and maximizing insulation life. Oxidation stability increases
with the degree of refining and hence with decrease ill aromatic
content but at the same time very severely refined or over
refined oils possess lessp.r oxidation stability due to the fact
that the natural inhibitors in the oil like the mono aromatics and
the good non corrosive sulphur compounds are also being
removed leading to the formation of oxidation products such as
acids and sludge Controlled selective refining is required for
ensuring the retention of these good hydrocarbons in the oil.
Certain types of sulphur compounds which are reactive if present
in the oil tum corrosive at the higher temperatures of more severe
transformer operating conditions. The copper corrosion by
formation of conductive copper sulfides on conductors and
insulating materials are likely to reduce the partial discharge
inception voltage and more importantly reduce the dielectric
breakdown voltage of solid insulation, which can result in a
dielectric puncture through the paper insulation. With the
increased severity of working condition of higher temperatures
and electrical stress, this copper corrosion problem due to the
higher levels of corrosive and reactive mercaptan sulphur in
the oil can cause catastrophic failures in a high voltage power
Different organo-sulphur compounds are present in transformer
oils, depending on the crude oil origin and the degree and type
of refining. The two types of reactive sulphur compounds namely
the hetero cyclic disulphides and mercaptan sulphur are very
corrosive to metal surfaces like steel, copper and silver present
in transformers and switchgears .Certain types of sulphur
compounds which are reactive ifpresent in the oil tum corrosive
at the higher temperatures of more severe tmnsformer operating
conditions. Studies on Transformer failures due to Copper
Sulphide Corrosion also came up with detection of Dibenzyl
Disulphides ( DBDS ) in oils which becomes reactive at the
more severe operating conditions. As some sulphur compounds
have affinity to metals , they may act as copper passivators
initially and subsequently may promote corrosion in the
transformers due to migration. Power Transformers often
required to work at the optimum or at times even on over load
conditions resulting in higher operating temperatures demand
non corrosive oils with very low sulphur content and optimum
oxidation stability produced by deep hydro treatment of low
sulphur crudes and also not containing any ODDS.
To meet these ever increasing challenges the oil refiners have
the continuous task of understanding the chemistry of various
hetero atoms such as sulfur in oil in different operating
conditions and strike the right balance of these hydrocarbons
by controlled optimum refining techniques as well as fine tuning
the refining process with very high degree of reliability to
achieve the best of the properties in oils.
Refining methods like deep hydro treatment and hydro de
sulphurization reduce aromatic and sulphur contents of the oil.
It is essential to select the right source of low sulphur crude and
subject it to selective refining for ensuring very low levels of
corrosive and mercaptan sulphur content in transformer oils to
avoid further problems during usage. Starting with naphthenic
crude oils of very low sulphur content and subjecting it to
more severe deep hydrotreating it is pvssible to bring down
these sulphur compounds to low levels in tansformer oils.
Although copper corrosion can not be reversed, it is possible
to avoid further significant corrosion by possible methods of
oil replacement and partial oil replacement with non corrosive
oils or passivation using metal deactivators. Addition of
Passivators to oil is suggested as a temporary solution but the
technology is still not fully clear in terms of any risks on the
long term stability of oils and possibility of side effects. Also
there are additional problems of dosing of Passivator while the
oil is in service for maintaining its levels. It is preferable to
select and use a non corrosive oil right from the beginning that
does not require any passivation.
Oil inhibition with DBPC ( Di - tertiary Butyl Para Cresol) , a
well established technology, helps to achieve enhanced
Oxidation stability levels in oils. To addres!i to the consequences
of overloading of transformers, IEC 60296 -2003 has introduced
the more stringent oxidation stability test procedure of lEC
61125 C which is able to discriminate oils with higher oxidation
stability from oils with lower oxidation stability. To increase
the precision of the test, gasses are passed through water with
the lower air flow of 0.15 litre I hour on oil sample in contact
with 90 mm copper coil at 120C for the stipulated longer
duration of 164 hours for uninhibited oils ( inhibitor not detected
) , 332 hours for trace inhibited oils ( up to 0.08 % max inhibitor
) and 500 hours for inhibited oils ( 0.08 to 0.40 % inhibitor.
There is an option in the lEe 60296 for stricter limits for special
applications and in some countries more stringent limits and
additional requirements and tests may be requested. These
spc;cial requirements are generally met with Inhibited Oils. The
ASTM D 3487 Standard also includes only Type 1.( Trace
inhibited oil with 0.08 % max inhibitor) and Type II ( fully
inhibited oil with 0.3 % max inhibitor) which are also adopted
in the Doble TOPS along with the Uninhibited oil specification.
The hydrotreated uninhibited oils even though are capable of
meeting the current general oxidation stability as well as
corrosive sulphur requirements are found to have some
limitations for meeting the special application oxidation stability
limits. The hydrocracked and Iso dewaxed paraffinic oils of
very low sulphur content are also considered as inhibited oils
for increased oxidation stability and control of corrosive sulphur
as tested and accepted recently by major OEMs. These paraffinic
oils also have better response to oxidation inhibitor as
compared to the naphthenic oils. Thus a new option opens up
for inhibited oils with existing naphthenic oils and Iso dewaxed
paraffinic oils for achieving enhanced performance levels. In
case of the inhibited oils, as the inhibitor will be utilized in
arresting the oxidation reactions while in service, the inhibitor
content should be monitored closely to replenish it to its original
level for maintaining the performance.
Impulse break down voltage is another important property
which is not commonly included in the specificatiuns except
ASTM D 3487 I Doble TOPS. Breakdown behavior with DC
impulses and heterogeneous gap is very different from the AC
breakdown strength. It is designed to simulate lightning pulse
striking a transformer during thunderstorm. This property is
independent of the contaminants influencing the normal break
down test. Dielectric impulse strength can be thought of as the
capacity of the oil for reducing short electric shocks such as
arcing of wires or the load on a transformer while being
connected and disconnected. The impulse strength of the oil is
critical in electrical equipments as it indicates the ability ofthe
oil to resist electrical breakdown under transient voltage stresses
( lightning and switching surges ). This performance property
is sensitive to both polarity and electrode geometry. The methods
used for testing this property are IEC 60897 and ASTM D 3300.
The ASTM D 3487 specification requires a minimum value of
145 kV negative impulse. Impulse breakdown voltage decreases
with increase in PCA Content. . It is possible to get negative
impulse strength values of above 200 kV in optimum refined
oils even with the PCA Content of up to 3 % maximum, however
above 3 % PCA level is not desirable in oils as it will result in
drastic reduction of the Impulse strength.
Some of the important oil properties of transformer oil like
Oxidation Stability , Corrosive Sulphur, Gassing Tendency,
PCA Content, Aromatic Content, Impulse Breakdown Voltage
etc are rather interdependent and also depend on the type of
crude, i.e Naphthenic or Paraffinic and the type and degree
of refining. For example, even though low PCA I Aromatic
Content is desirable in the oil for better electrical properties
enhanced performance levels , it is essential to take total
perspective of all oil parameters than working only in isolation.
and corrosion stability, it should not be at the cost of other
important performance parameters like oxidation stability and
gassing tendency. The refiner has to strike the right balance
of all the oil properties by carrying out controlled selective
refining to optimize its performance. While selecting
Transformer oils for specific high v o l t g ~ requirements of
power Transformers of say 120 kV and above requiring
Typical values of Uninhibited and Inhibited Hydrotreated
Naphthenic Oils are given in Table - 3 showing how the
oxidation stability ( IEC 60296 ) as well as the new corrosive
sulphur test requirements are met.
Specification VALUES VALUES
(U) ( I )
Aromatic Content, CA % ISO 60590 No general requirement 11 11
Total Sulphur Content, ppm ISO 14596 No general requirement 500 497
Corrosive Sulphur DIN 51353 Non Corrosi ve Non Corrosive Non Corrosive
( Additional Test) ASTM 01275 B Non Corrosive Non Corrosive Non Corrosive
(Additional Test) CigreTFA2.32.01 Negative Negative Negative
Antioxidant Content,Wt t% IEC 60666 (U) Not detectable Not detectable 0.38
( I ) 0.08 to 0.4
Oxidation Stability IEC 61125 C 1.2 max ( U) - 164 Hrs ( I ) - 500 Hrs
Total Acidity, mg KOH I gm 0.42 0.55
Sludge, Wt % 0.8 max 0.3 0.32
Typical values of Trace Inhibited and Inhibited Hydrocracketi Iso dewaxed Paraffinic oils are given in Table- 4 showing how the
special application oxidation stability ( IEC 60296 ) and the new corrosive sulphur test requirements are met..
Specification VALUES VALUES
Aromatic Content, CA % ISO 60590 No general requirement Nil 1.2 ) Nil ( < 1.2 )
Total Sulphur ISO 14596 0.15 ( 1500) < 0.005 ( 50 ) < 0.005 ( 50 )
Content, % ( ppm )
Corrosive Sulphur DIN 51353 Non Corrosive Non Corrosive Non Corrosive
(Additional Test) ASTM 01275 B Non Corrosive Non Corrosive Non Corrosive
( Additional Test) CigreTFA2 .32.01 Negative Negative Negative
Antioxidant Content,Wt t% lEC 60666 (T) 0.08 max. 0.075 0.38
( I ) 0.08 to 0.4
Oxidation StabilityTotal lEC 61125 C 0.3 max ( T ) - 332 Hrs 0.1 ( I ) - 500 Hrs
Acidity, mg KOH / gm 0.03
Sludge, Wt % 0.05 max NIL
Emerging changes and demands of Transfonner operating
conditions in the complex power systems have thrown new
challenges to the oil refiners to produce high quality transformer
oils meeting the latest standards and test requirements .. The
refiner is often required to take a tight rope walk to get the
balancing act right tor meeting the contradictory requirements
of high oxidation stability as well as totally non corrosive oils
. Particular type of oil may not be the criteria rather than a product
has to be tailor made to meet the particular requirements of the
Power Tmnsfonners.
With optimized refining of low sulphur naphthenic crude by
deep hydrotreatment , it is possible to produce high quality non
corrosive uninhibited transfonner oils meeting the general
requirements of high oxidation stability. However, for meeting
the further enhanced special application requirements of
oxidation stability and also non corrosivity , the new options of
considering inhibited oils with the existing hydrotreated
naphthenic and also hydrocr", ked Iso dewaxed paraffinic oils
are emerging ..
It is important to understand the influence of different oil
parameters when selecting insulating oils for the electrical
equipments especially Power Transformers. Characterization
of a good Transfonner Oil is based on its electro insulating
properties, cooling characteristics, chemical ( oxidation and
corrosion) stability, behavior under severe electrical stresses,
compatibility with Transformer materials, compliance to
environmental factors and satisfactory service.
Oil constitutes only 5 to 7 % of cost of transformer. It is easy
to compare the high cost ofTransfonner failures to the relatively
negligible differential cost of High Grade Oils. In the complex
power generation , transmission and distribution system ,
transfonner is like the heart in the human body that requires
the right quality blood to perfonn continuously and effectively
for a long period of time for a healthy life. Like the human
blood, transfonner oil of the right quality standard plays a major
role in ensuring longer and reliable performance of
Transformers thereby enhancing their service life.
I. " Copper Sulphide in Transformer Insulation"- Cigre
Report TF A2 -31 on behalf of SC A2.
2. "Copper Sulphide in Transfonner Insulation"- WG A2.32
Report- Electra, February 2007 .
3. Ricardo Maina, Fabio Scatiggio, Shuhhen KapiJa, Vander
Tumiatti, Michela Tumiatti and Massimo Pompilli.,"
Dibenzyl disulphide ( DBDS ) as corrosive sulphur
contaminant in used and unused mineral Insulating oils"
4. A.C.M. Wilson ., " Insulating liquids." Institution Of
Electrical Engineers, lEE electrical and electronic
materials and devices series; I by International
Electrotechnical Commission ( IEC )
5. " Specification for Fluids for Electrotechnical application
- Unused Mineral Insulating Oils for Transfonners and
Switchgears. " ,lEe 60296: 2003, Third Edition
6. .. British Standards Specification for Unused and
Reclaimed Mineral Insulating Oils for Transfonners and
Switchgears ... BS 148 : 1998
7. "Australian Standards Specification for Unused Mineral
Insulating Oils for Transfonners and Switchgears." AS
1767.1 :1999
8. "American Society for Testing and Materials Standards
Specification for Mineral Insulating Oils used in Electrical
Apparatus." ASTM 03487 : 2000 (Reap proved 2006.)
9. " Doble Transfonner Oil Purchase Specification." Revised
January 1 , 2006.
10. "VDE Specification for New Insulating Oils for
Transfonners and Switchgear. ".
DIN 57370 IVDE 0370-1,1978.
Saif Qureishi
(Email questionsorcommentsonthispapertosaif.qureishi@kryfs.com)
A proper value of the Building factor (OF) is important
while calculating the No-Load Loss (NLL) of any power or
distribution transformer. It is an empirically derived factor
which is based on the experience of the Transformer
Manufacturer (TM) and ranges from 1.08 to 1.35 for three
phase, three limb cores. Step Lap (SL) construction of
transformer core instead of conventional butt type
construction (or Non Step Lap (NSL) type) which is still
widely used by TMs in India, has been successfully used by
various TMs world over,.to reduce the Building factor in
Transformer cores by 5 to 8%, reduce the No Load current
and the noise level relative to conventionally stacked NSL
cores. This paper explains how SL laminations reduce tbe
No Load Loss in a transformer by considering the specifics
of magnetic Flux transfer in joints areas of a SL core versus
a NSL transformer core.
The magnetic circuit is one of the most important active parts
of any transformer. While the basic principle of transformation
of energy has remained the same for over a century, since the
first transformer was built, transformers have become more
efficient due to improvements in materials and more
sophisticated production processes (better manufacturing
technology). '
This paper examines one such production process which
reduces the NLL, No Load current and noise level of a
transformer and hence improves it's efficiency.
BF is a non dimensional parameter defined as the Ratio of (No
Load Loss of a transformer I core weight) in walts per kg to the
Epstein or Single Sheet watt loss (in watts per kg).
Therefore BF = (NLLmeasuredon transfamerin watts per kg)
(EpsteinorSingleSheetTest loss in watts per kg)
BF ensures that the NLL of a transformer does not equal to the
core weight of the transformer multiplied by the Single sheet
loss as defined by the producing Mill at the particular operating
flux density. For example, if the core weight is 200 kgs for a
100 KVA transformer, manufactured from M4 grade material
and operating at 1.7 Tesla, if the material used is C 12027 (i.e
conventional M4 grade GO having a core loss of 1.2 Watts per
kg at 1.7 Tesla) then the Maximum NLL should simply be 200
kgs x 1.2 w/kg = 240 watts. However, in practice it is never so
for a three phase, three limb core due the BF.
1. Core Geometry: BF is directly proportional to the Area
of Proportion of Comer Joints to the total core area, kinds
of joints (SL or NSL ) Air gaps, overlap area at joints,
2. Quality of workmanship of Core: BUrrs in Laminations,
Accuracy of dimensions especially at the corner joints
(precision of angles) , flatness of the laminations, dust
particles in between laminations, skills in core building
(squareness of the core), clamping pressure on the core
3. Grade of Material Being used: Whether material being
used is Hi- B Grain Oriented (HGO) or Conventional
Grain Oriented (CGO) which affects the stacking factor
of material, Insulation resistivity values of material (lR
values of coating on GO). Permeability of the material,
thickness of the lamination (which effects eddy current
Howeverthe major reason contributing to BF is Core Geomtery,
i.e the Area of Proportion of Corner joints to the total'area of
the core. Therefore in a smaller rating core (like 25 kva to 100
kva- varies from 1.25 to 1.35) where the BF is much hi5her
compared to a Power transfonner core like 25 MVA and above
(varies from 1.08 to 1.15) where it is comparatively lower.
As it is not possible to examine all these factors in a single
paper, we shall only examine the most important one, which
can directly reduce the NLL significantly- using SL laminations
for making transformer core and understanding the reasons for
the same, mainly the advantages in flux transfer in SL over
Both Hysterisis and Eddy Current losses when added make up
the NLL of a transformer. Due to the complexity of determining
each individually, generally designers use either of the following
two equations while calculating the NLL in a transformer:
(1) NLL= Wtx Kt x w
(2) NLL= (Wt- Wc) x w+Wc x w x Kc
Where w is core loss in Watts per kg of the eRGO material
used in core.
Wt is the total weight of the core
W c is the weight of the core at Corner joints.
Kt is the Building factor of the total core and
BF depends on various factors, some of the major ones being
the following:
Kc is the Building factor of the core at the corner joints.
A Cross Sectional view of behavior of Flux in a Conventional
NSL Core stack. is shown in Figure 1 below. It has mitred joints
and is usually built two laminations at a time with an off-centre
overlap of around 10 mm in the centre leg. As can be seen from
the diagram, the joints are staggered one on either side 5 mm
around the centre point.

Figure t- A cross sectional view of the Conventional Non
Step Lap (NSL) type joints and behaviour of flux in it, which
is mostly used in India.
A Cross sectional view of SL Laminations (6 steps) and the
behavior of flux is shown in Figure 2. As can be seen the joints
are staggered 3 on either side, around the centre point.

Figure 2 Behavior of Flux in the 6 SL joint.
Similar to an electrical circuit where, Electric current follows
the path ofleast resistance. in a magnetic circuit, Magnetic flux
follows the path of least reluctance (which corresponds to
Highest Magnetic Permeability). In fact, the situation in an
electric circuit and a magnetic circuit are analogous. Like
Resistance (R= VII) in any electric circuit,
Reluctance, R= FI cl
Where R is the Reluctance of the Magnetic Circuit in ampere
F is the Magnetomotive Force in Ampere - turns and
clis the Magnetic flux in webers.
As we see above in Figures I and 2 above, when the Flux in the
transformer core approaches the air gap at the comer joint in
the core, it has two options - Either to cross the Air gap at the
joint (where the Magnetic Permeability is much lower (=1), as
the Magnetic Reluctance of Air is much higher.
The second option for the flux is to cut across the insulation
layers of the laminations and move to the overlapping
laminations in the vertical direction (the direction perpendicular
to the rolling direction of the laminations) above or below where
the Magnetic Permeability is of the order of 101\4 and as a
result the Magnetic Reluctance is much lower.
Obviously the flux will chose option 2 i.e to cut across insulation
layers and transfer to the laminations overlapping above or
below it. However as CRGO saturates at a flux density of
approximately 2 Tesla this is the constraining factor for the
"ready to transfer flux" approaching the air gap at the comer
Let us consider a core which is operating at a flux density of
1.7 Tesla. As the flux approaches the "air gap" at the comer
joint, it has to decide between option 1 and 2 above. If all the
flux transfers to the overlapping laminations above or below,
in a NSL type joint where there are just two overlapping
laminations, then the flux in the overlapping laminations will
be 5.1 Teslal 2 = 2.65 Tesla per lamination which creates
"overcrowding of flux" and also much over the saturation limit
of the CRGO (which approx 2 Tesla).'Thus in a NSL type
joint, some of the flux will get transferred to the adjoining
overlapping laminations, however some part of the flux will
also have to jump across the air gap (which is option 1). This is
what is shown in Figure I above. Even the flux which gets
. transferred to the adjoining laminations in a NSL, increases the
flux density in the lamination above the salUration level which
automatically contributes to the saturation of the material at
joints and therefore higher NLL.
The flux which crosses the air gap contributes to "wasting of
flux" and therefore requires a higher no load current to achieve
the required calculated flux density. Further the over-saturation
of flux at the comer joints also leads to higher magnetostriction
of the core which is the main cause of noise level in a
transfonner. This transfer offlux in a Conventional type (NSL
type) core is more explicitly shown below once again for better
understanding in Figure 3: .
Figure 3: Diagrammatic representation of flux transfer to
overlapped sheets in a Non Step Lap joint
However the situation is different for SLcore. The "ready to
transfer flux" approaching the air gap has many more options
as can be seen in figure 2 above , simply because there are
more layers of laminations "available" for distributing this
"ready to transfer flux". As can be seen in the diagrammatic
representation of the 6 Step Lap core, the flux has six options
to jump instead of just two and therefore there is a more balanced
distribution offlux over the adjoining laminations which results
in very little flux jumping the air gap thereby also contributing
to lower losses at the comer JOints, as the flux density of the
comer joints remains around the saturation flux density of 2
Tesla. This is explained diagrammatically below in Figure 4 :
Step #

Figure 4: Diagrammatic representation of flux transfer to
overlapped sheets in a Step Lap joint.
It has been determined by Mechler and Girgis in (1) that the
Flux density in NSL joints at the air gap is high as compared to
in a SL joint and the flux density in the adjoining laminations
in a NSLjoint is much higher, as compared to SLjoint which is
discussed more in detail below with the help of diagrams and
The distribution of this flux in NSL is represented below in
Figures 5,6,7 and 8
- .-::::1
Figure 6: Corresponding Sketch of Figure 5 to identify lines
for line plots (not to scale).
x 1m]
0.000 +---+----;----1--;---+---+---+----1
0.070 0.080 0.(9) 0.100 0.110 0.120 O.lJJ 0.140
Figure 7: Magnetic flux density in the :amination along
line 1 of Fig. 6
___ 1______ __
,-- -- -:------J
:: :::::: -- -- i:: --
0.500 ------t------:-------:-------:------ I z[mJ
0.00:0 D.rms 0.0010 0.0015 0.0020 0.0025
Figure 8 : Magnetic flux density along line 3 of Fig. 6
As can be clearly seen from the above, the flux density in the
steel laminations along line 1 after the Air Gap reaches 2.7 Tesla
and this is approximately the same amount of flux found along
line 3 (which is normal or perpendicular to the direction of
rolling) in the steel laminations. It should also be noticed that
the Flux density in the air gap is as high as 0.7 Tesla.
Mechler and Girgis then repeated the same experimel!t with a
Figure 5: flux lines in a NSL core joint at 1.7 T . SL core and the results are shown below in Figures 9,10, 11
induction. and 12.

. ~
Figure 9 : Flux lines in step-lap core joint at B (overall) =
1.7 T.
Figure 10 : Sketch fur identification of lines in a Step Lap
2.040 ~ - - ~ - ....... --r--r--;:::=:I=::;)
2.000 t---+-.&-+---a;------ro--"-t-'"---'''-t---' Z ImJ
0.00 O.(Q)5 0.0010 0.0015 0.1m) o.m om
Figure 11 : Flux density distribution along line 9.
I &[TJ
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
.COO X !mj
o . ~ 0.070 0.080 O . ~ 0.100 0.110 0.120 O . I ~ 0.140
0.080 0.1m 0.100 0.110 0.120 0,13.) 0.140
0.070 O . ~ 0.(1) 0.10)" 0.110 0.120 0.1l)
0.(1) 0.070 o.ceo 0.(1) 0.100 0.110 0.120 0.13) 0.140
Figure 12: Magnetic flux density in all 6 different
laminations in a step-lap core joint along lines 1-6 of Fig. 1 0
As can be seen from the above, the flux density across the air
gap in a Step Lap joint is IS low as 0.04 Tesla (as compared to
0.7 Tesla for a NSL joint) and the flux density across the six
steps is at the saturation level of approximately 2.028 to 2.035
Tesla (as compared to over saturated level of 2.7 Tesla for a
Therefore it is clear that the SL construction facilitates a much
more efficient way of core building and empirically has been
found to !'!: !:::::: :!":. .. :: by 5 to 8
% over NSL cores which is a significant reduction in NLL.
Of course, there are various parameters in a SL core which have
to be optimized four of the important ones being the Number
of steps in a SL joint, the distance of overlap between the steps
, number of laminations per lay and the air gap. Best results
with a SL lamination can only be achieved when all these
parameters are at an optimum level. However it is not within
the scope of this paper to discuss optimization of these
parameters here.
There are two types ofSL laminations that can be manufactured.
Horizontal step lap and vertical step lap. These are shown in
Figures 13 (Vertical Step Lap) and 14 (Horizontal Step Lap)
below for better understanding. It is the designer's choice which
one to use as there is no significant difference in performance
of the core as long as the four optimization factors
in the previous paragraph are taken care of.
Figure 13: Vertical Step Lap (Yoke and side limbs shift in
the vertical direction)
Figure 14: Horizontal STEP-LAP (Yoke and side limbs shift
in the horizontal direction)
Figure 15 : View of the Variable length yoke plates
The variables while designing step lap laminations are the
1. No. of steps per book
2. No of laminations per step
3. Amount of over lap (shift) in the book between different
4. Type of step lap to be used - Vertical or Horizontal step
Ideally, if a core designer is not comfortable converting the
existing NSL core drawing to SL core drawing, the Lamination
manufacturer or processor should have the capability to do so
given the parameters of core diameter, window height and cross
sectional area of the core.
SL laminations are nothing new. They have been used in
Europe, USA and many other countries for more than a decade
now. Some Multinational transformer manufacturers in India
too use these kind of cores as they significantly reduce the BF
by 5 to 8 % ,thereby significantly reducing the NLL of the
transformer, reduce the No Load current and also reduce the
Noise level of a transformer. However it has not been
popularized in India because SME transfonner manufacturers
using their existing designs were comfortable with the NSL joint
design. Also if some of them did try to migrate to the SL design,
the manufacturing facilities were not available in India till some
years ago. Manufacture of SL laminations is a precision job
that can only be done on high precision Automatic cut off lines
(like Heinrich Georg) using only Prime CRGO coils, which
involves installing expensive capital equipment. As most of the
core processing in India till a few years ago was done on manual
treadle shears, it was not possible to manufacture SL laminations
even if the designers wanted these type of laminations. So the
designers did not change over to the more efficient SL design
as the manufacturing capability was not available within the
country specially with the lamination manufacturers. Also the
lamination manufacturers never bothered to invest in expensi ve
capital equipments as few Indian TMs used SL design
laminations which would be the major reason to invest in an
expensive cut to length line. This was a classical chicken and
However now many lamination manufacturers and transformer
manufacturers have installed CNC lines and it is possible for
step lap cores to be now manufactured in India. which can only
be done from Prime CRGO coils as CNC machines don't accept
second grade material. Therefore use of step lap core not only
ensures more efficient transfonners, it also ensures use of Prime
material, which has been the demand ofITMA for a long time ..
Thus it is strongly recommended that TMs should take
advantage of this ad\tancement in manufacturing technology to
reduce the BF and as a result the NLL and make their
transformers more efficient.
1. The NLL in a transformer is dependant on the BF which
is dependant on various factors like Core Geometery.
Quality of workmanship of flT ('ort" anrl the
grade of material being used.
2. The Area of the joints to the area of the entire core has a
major role to play in the BF increasing.
3. The BF can be reduced by using SL Laminations instead
ofNSL laminations, by a factor ofS to 8 %.
4. SL Laminations distribute the flux around the corner joints
more effectively than non SL Laminations as the flux has
more options to choose from.
5. SL Laminations thereby reduce NLL, No Load current,
Noise level (due to lower magentostriction of core) and
help productivity as they make core building faster and
more efficient.
6. Step Lap Laminations are of two types, Horizontal Step
Lap and Vertical Step Lap.
7. SL laminations can only be manufactured with Prime
material on Automatic cut off lines having the capability
to do so.
References :
1. Gunther F. Mechler and Ramsis S. Girgis, "Magnetic Flux
UlSlfloulions in Transformer Core Joints", IEEE
transactions on Power Delivery.
2. Takayoshi Nakata N. Takahashi, Y. Kawase. "Magnetic
performance of step-lap joints in distribution transformer
cores" IEEE Trans Mag.
3. N. Hihat, E. Napieralska-Juszczak, J. Ph. Lecointe, J. K.
Sykulski, "Computational and experimental verification
of the equivalent permeability of the step lap joints of
transformer cores"
4. Vladimir Segal, "Effects of Joint design and some other
parameters on Build factor of a 3 phase stacked core",
Proprietary report.
S. S.Y.Kulkarni, S.A.Khaparde, "Transformer Engineering
Design and Practice" CRe Press.
Venugopal s. Adya
Never before the challenge posed by Global warming, climate
change and green house gas emission was as alarming as exists
world wide have come forward fmd viable and sustainable
solutions to overcome this challenge.
In the past, tmnsmissionand distribution companies worried
little about global warming. Electric power utilities today are
looking for more efficient distribution transformation
accompanied by Energy savings and environment friendly
systems. Adoption of well planned T &0 system can help reduce
line and equipment losses and defer the need for power
generation thus impacting reduction of green house gas
emissions. Availability of adequate power is one of the key
factors for the economic development of a nation. However it
is a major challenge in India. Our country is facing high AT &C
losses of over 50% out of which T & 0 losses account for
approx. 19% resulting in the peak demand shortage of about
15%. This adversely affected the availability of power to the
users and hence the country needed to genemte more power to
meet up the short supply. Major cause of concern for India is its
heavy dependence on the thennal power which causes emission
of harmful gases. The situation could improve dramatically if
the T & 0 losses are brought down to a reasonable level of
about 10%.
Govt. of India by enacting Energy Conservation Act 2003 has
made its intention clear to bring about the much needed refonns
in the power sector and to implement energy conservation
policies. Consequently Bureau of Energy Efficiency (BEE) a
Govt of India organization has come out with the Standards &
Labeling (S & L) Program covering a wide range of Consumer
Electrical appliances & Industrial Electrical products to cut
energy waste and promote Energy Efficiency. Complementing
its efforts are Central Electricity Authority, Rural Electrification
Corporation Ltd. and Bureau of Indian Standard by framing
suitable standards. This initiative would go long way to help
conserve energy and promote Energy Efficiency.
Besides various issues suggested by the Indian Prime Minister in
recently announced National Action Plan on Climate Change,
following is actual extract regarding the Power Distribution
"India's current technical losses during transmission and
distribution are as high as 16% - 19%. By adapting HVAC
and HVDC transmission, the figure can be brought down to
6%-8% by using amorphous core transformers and up-
grading the distribution system (avoiding congestion etc.).
Distribution losses can also be reduced by adopting energy-
efficient transformers, which can use high-grade steel in
transformer core"
. 107
In accordance with the global trend, the Department of Energy
(DOE), US has issued a Final Rule in October 2007 for Energy
Conservation Program for Commercial Equipment: Distribution
Energy Conservation Standards which mandates
procurement of high efficiency Distribution Transformers.
Similar actins are awaited from other countries as well.
Electric utilities and equipment manufacturers world wide
developing and adopting innovative methods to cut energy waste
and reduce emissions. The most important strategy is to improve
the efficiency of the electricity distribution. Distribution
transformers are the vital links between the utility and the
consumer in the electric power distribution network. They are
also the second largest loss making component in the network.
However they are relatively easy to replace certainly in
comparison to lines / cables. Their efficiency could be easily
measured, improved, classified, labeled and standardized.
Distribution transformers are energized every hour ofthe every
year. Because of this, even small efficiency improvements can
yield big energy savings. Moreover, the decision to purchase
an energy-efficient transformer locks in long-term savings over
the life of the equipment (typically 25 years).
Amorphous Metal Distribution Transformers, made from the
new generation ultra efficient low loss electrical steel, otTer
options not only to save the scarce energy and improve
distribution system efficiency but also help to reduce green
house gas emissions and contribute to make our environment
cleaner for the next generations.
In the operation of distribution transfonner there are two types
of energy losses:
a) No load or Core losses - that occur in the magnetic core,
once the transfonner at all times over the life of transformer
regardless of load. No load loss represent significant
portion of energy lost during power distribution. It depends
upon the properties of. the core material used and the
operating induction.
b) Load loss or Copper losses - That vary on the transfonner
loading. It depends upon the properties of the winding
material i.e. copper / aluminium and its current density.
Thus any efforts towards reduction of core loss would have
maximum impact on improving the efficiency of the transfonner.
In the last few decades. Several improved grades of low loss
grain oriented electrical steels were developed to manufacture
high efficiency transformers. Around the same time a unique
electrical steel known as Amorphous Metal was developed
which provided significant reduction in the core loss that helped
to make high efficiency transfonners .
Amorphous metal distribution transformers (or AMDTs) are .
thought to be new but have been around for 25 years or so.
They use a metallic glass alloy in their cores. Amorphous metals
can be defined as bulk, structural, metallic materials whose
microstructure in the solid state is, unlike that of conventional
metals, non-crystalline, amorphous or 'glassy'. As a result of
this novel microstructure, amorphous alloys exhibit unique
combinations of properties such as magnetic performance,
hardness, strength, damage tolerance and corrosion resistance.
Conventional Metals Regular and Periodic atomic
Amorphous Metals Random atomic arrangement
Metglas Amorphous metal core aHoy 2605 SA I is produced
by a unique rapid solidification process where in the molten
alloy is cooled at the rate of approx. one million degree Celsi us
per second. Random molecular structure of amorphous metal
causes less friction than CRGO when a magnetic field is applied.
This unique property which allows ease of magnetization &
demagnetization significantly lowers hysteresis losses in
amorphous metals
Amorphous metal alloy is cast in the form of thin ribbon /
laminations. Thin laminations result in lower eddy current losses
as compared to CRGO. Thus the core loss of amorphous metal
core is significantly reduced by about 70% due to the combined
effect of lower hysteresis loss and lower eddy current loss The
amorphous core has a wound core construction with gaps
distributed at lap joints. The core cross section is rectangular
built from packets using several plies of amorphous ribbon for
ease of handling. Annealing the core under magnetic field results
in minimum core loss capability. Gluing the edges of the core
and curing at the specified temperature to impart desired rigidity
which facilitates installation of coils at a later stage. Inner and
outer core have layers of CRGO sheet make the core robust
and protect it against physical damage. Core's performance
parameters are verified in the final testing of the finished core
at specified magnetic induction using a sophisticated digital
power analyzer for accurate measurement.
Ready to use Amorphous Metal Core
Comparison of typical No load losses of CRGO and AMDT
kVA Number of No-Load Loss (WATTS)
Phases AMTs eRGO core
10 1 10 40
16 1 15 60
25 1 21 75
25 3 28 100
63 3 50 180
100 3 66 260
Table 1
Amorphous metal core transformers improve electrical power
distribution efficiency by reducing transformer core losses by
about 70% compared to silicon steel-based transformers under
sinusoidal load conditions.

Under the present situation the gap between traditional designs
and amorphous technology is shrinking. The difference in
quantity of core material (magnetic steel, amorphous metal) as
well as in unit price of both is now within 10-20010 span only.
Comparison of key features - Amorphous Metal and eRGO Steel Transformers
properties unit Amorphous Metal CRGO Steel
Density (glcml) 7.19 7.65
Specific resistance 130.00 45.00
Saturation flux density ( T e ~ i a )
2.03 J. .J7
Typical core loss at 50 Hz Watt/kg 0.22 @ 104 Tesla 0.89 @ 1.5 Tesla
Nominal thickness nnn 0.025 0.27
Space factor 0.86 0.97
Available form Ribbon/Foil Sheet! Roll
Annealing temperature C - 360 - 800
Annealing atmosphere Inert gas Inert gas
Other process Magnetic field annealing
Core construction & assembly Wound core construction with Assembly of Sheared & cut
distributed lap joints, ready to use core laminations, time consuming
Productivity Relatively high with less wastage Low
Process equipments cost Reasonable considering its high Cost built in to the laminations
productivity & low pay back period cost
Core design Flexibility to design various DTs Flexibility to design various DTs
Core steel grades Multiple grades Single grade of assured quality
Coil winding Concentric rectangular shaped Concentric round shaped coils
coils for better space factor utilizatiOJi
Reparability Possible Possible
No load loss Very low, up to 70% to 75% compared to High
conventional grade eRGO transformer
Transformer first cost Cost difference between CRGO & AMDT
is around 20%. However with loss
capitalization, AMDT becomes competitive
Table 2
The dramatic increase in the use of power electronics has
resulted in a considerable amount of higher harmonics distortion
in electricity systems. High frequency harmonics lead to
increased transformer core losses, especially in distribution
transformers that use conventional steel core materials.
Amorphous metal core distribution transformers are well suited
to providing low loss performance under higher frequencies.
This is due both to unique processing techniques that result in a
greatly reduced standard thickness for amorphous metal and a
unique chemistry and atomic structure that give a higher
electrical resistivity and low energy magnetic flux reversal.
Harmonic currents and voltages caused by non-linear load are
known to increase losses and over-heating of transformers. The
major impacts are dramatically higher energy losses than
expected and premature equipment failure.

Save energy and therefore reduce greenhouse gases and
other pollutants
Excellent option to reduce distribution losses and improve
Key to improving utility economics and enhancing energy
conservation efforts worldwide
Less expensive to operate
Superior electrical performance under harmonic conditions.
Possible to improve power quality and mitigate harmonics
. Lower temperature rise, Slower deterioration of insulation,
Longer life
Replacing older inefficient transformers improves
Some time back ERDA had carried out a study on the
performance of 3-Phase 25 KVA and 100 KVA AMDT and
CRGO Transformers under harmonic loading coDClitions
encountered at iRigalion pump sites. Findings of their report
established that the increase in the (core losses on account of
harmonic contents in the CRGO steel transformer was
significantly higher than that of AMDT.
The following table summarizes the study results of 3-Pbase
25 KVA traosfonoer where the increase in the core loss of
CRGO steel transfonoer was about 6.5 times higher than that
Transformer Losses without harmonic distortion (Assuming
Linear Load)
Loss' (watts) 25KVA 25KVA
Hysteresis 19 32
Eddy current 6. 65
Total Core Loss 25 97
Coil Loss 117 115
Loading level -43% -43%
Total Transformer Loss 142 212
AMDT Core Loss reduction
Linear Load Expected 70
Transformer Losses with harmonic distortion (Actual - Non-
Linear Load)
Loss (watts) 25KVA 25KVA
THD% 9.7% 8.7%
Hysteresis 19 32
Eddy current 49 509
Total Core Loss 68 541
Coil Loss 239 237
Total Transformer Loss 142 212
- Linear Expected
Total Transformer Loss 307 778
- Measured Actual
Total Loss Increase 165 566
AMDT Core Loss reduction 401
Non-Linear Load Actual
It is seen from the above study results that AMDT is much less
affected by harmonic conditions than the CRGO transformer.
If the harmonic associated with the pump site equipment have a
significant impact on the core losses, the cumulative effect on
increased losses throughout India could be substantial and needs
prompt attention and rugent remedial measures.
Due to the energy efficiency policies and sustained initiatives
of Ministry of Power, Govt. of India that in the recent years
there bas been a growing awareness about the benefits of energy
saving among electric utilities. Many progressive utilities / SEBs
in India have procured over 350,000 AMDT units which are
already in service with satisfactory performance and expected
Following is a typical example to demonstrate the enormity of
the Projected Economic Benefits of using AMDT in the Power
Distribution System
Situation today
Total installed power generation Capacity
(as on 31.08.2008)
145,627 MVA
Estimated number of DTs operating in the : 30 Lakhs
Indian Power system
Average annual Production of Distribution : 34000 MVA
Transformer (3 years)
Estimated number of DTs added to the
system per annum
Average DT capacity
No load loss of 63 KVA eRGO DT
No load loss of63 KVAAMDT
Average energy cost per kwh
Savings in No load loss per transformer
: 5.0 Lakhs
: 63 KVA
: 180 watts
: 50 watts
: Rs. 3.00
: 130 watts
Savings in No load loss of 3.0 million DTs : 390 MW
Savings in energy of3.0 mi1lion DTslannum : 3.28 billion kwh
(Operating at 8400 hours) .
System losses up to DT : 10%
Plant load factor for power generation : 70%
Saving of effective additional generation : 620 MW
capacity at DT
Cost of generation per MW (appro x) : Rs. 5 Crores
Saving in inveStment for : Rs.31 00 Crores
generation capacity
Cost of energy saved (approx) : Rs. 980 Crores
The cost savings that could be achieved is mind boggling, though
it is not practically possible to do so.
However it is possible that new additions can use Energy
Efficient Transformers like AMDTs.
The country has already lost enormous energy and valuable
resources for not utilizing the energy savings opportunities.
However it is high time the opportunity must be used at least in
the immediate future. In that case the energy saving opportunity
looks as given below:
Situation tomorrow (immediate near future)
Estimated addition of power genetation in : 78600 MW
the 11
Estimated average annual addition of
Distribution Transfonner
Estimated annual average addition of
Distribution Transfonner
Average DT capacity
No load loss of63 KVA CRGO DT
No load loss of63 KVAAMDT
850,000 nos
: 63 KVA
: 180 watts
: 50 watts
Average energy cost per kwh : Rs. 3.00
Savings in No load loss per transfonner : 130 watts
Savings in No load loss of 850000 DTs : 110 MW
Savings in energy of3.0 million DTs/annum : 0.92 billion kwh
(Operating at 8400 hours)
System losses up to DT : 10%
Plant load factor for power generation : 70%
Saving of effective additional generation : 175 MW
annual capacity at DT
Cost of generation per MW (approx) Rs. 5 Crores
Saving in investment for additional annual Rs. 875 Crores
generation capacity
Cost of energy saved per annum (approx) : Rs. 275 Crores
The benefit of energy saving accrues year after year during the
entire life of transfonner, which is about 25 years. Saving in
additional (avoidable) power generation not only conserves our
. scarce national resources which could otherwise be put to better
use and also our environment from degeneration.
The advantage of AMDT is not limited to reduction of losses
alone. Since these losses are converted into heat energy, cooling
oil inside the transfonner tank will be heated up and it will lead
to emission of green house gases. Less power generation means
reduced emissions and significant fuel savings. AMDTs help
utilities reduce hannful emissions such as sulphur dioxide,
Nitrogen oxides and Carbon dioxide, the pollutants that cause
acid rain and global wanning.
Several Electric utilities in the Asia-Pacific region have already
adopted or in the process of migrating to high efficiency
transfonner standards to reduce distribution losses, improve
their system efficiency and reduce green house gas emissions .
. Amorphous Metal Core Transformers have been their preferred
option to attain higher efficiency, accrue huge energy savings
besides achieving environmental benefits. Several thousand
AMDT units are serving successfully in the utilities of India,
ChIna, Japan, K..orea, ! alwan, t'nlllppmes, vietnam, Nepal and
Bangladesh. Further AMDT trials are underway in some of the
remaining countries.
Amorphous Metal Core Transformer Technology is mature,
proven and used extensively in several countries 'worldwide to
meet up with the requirement of high efficiency transformer
standards. It demands worthwhile attention and consideration
by both the transformer manufacturers and users as an option
to reduce losses and improve system efficiency. AMDTs offer
huge economic benefits not only in terms of energy savings
and improvement in system efficiency but also help to reduce
green house gas emissions, thus contributing to make our
environment greener and cleaner for the future generations.
The author is thankful to the management of Hitachi Metals
for their permission to present this paper.
ERDA Report - Transformer (Core Loss) under Harmonic
Conditions for 3-Phase 25 KVA & 100 KVA
Metglas Inc. DT Core Product Guide
Hitae:hi Metals Presentation
Venugopal S. Adya graduated with B.E. Honors in Electrical
Engineering from Karnataka University. His employment
experience covers Distribution Transformer Design,
Engineering, Manufacturing & Management, Technology
Development, and Technology Transfer & Business
Development in reputed companies like MPVY Ltd. (NGEF)
and AlliedSignal / Honeywell. Rich experience in AMDT
Technology and Extensive International exposures to best
practices. Currently Consultant exclusive to Metglas (Hitachi
Transformer with No-Leak Guarantees
Swapan Kumar Datta
General Manager (Technology)
Nu-Cork Products (P) Ltd. India
Bad gaskets and oil leaks are important topics to every utility. A
lot of Oil- filled electrical equipment tends to leak. Nearly every
such oil leak is past a gasket or seal. If no one else focuses on
these areas, oil leaks will forever be a part of our lives.
Incidentally the chronic leakage has been with larger, low
pressure, oil gaskets.
During service, moisture can gradually evolve into oil, increasing
the water content in the thin insulating structure. If the
transfonner is open for inspection, the insulation can absorb
moisture from the atmosphere. Atmosphere water is the main
source of the transfonner contamination. Three mechanism are
acting: absorption of water while direct exposure of insulation
to air (installation and repair works)- ingress of moisture into
tank in the fonnofmolecular(Knudsen) flow due to the difference
in water concentration in atmosphere and the oil in the tank, and
viscous flow of wet air into the transfonner under the action of
difference in pressure ( atmospheric and inside the tank).
If there is a leak, moisture can enter in the fonn of water or
humidity in air. Moisture is also fonned by the degradation of
insulation as transformer ages. Most water penetration is flow of
wet air or rain water through poor gaskets seals due to pressure
difference cause by transformer cooling. The most common
moisture ingress points are gaskets between bushing bottoms and
transformer'top and the pressure relief device gasket. It is
important to repair small leaks, the small amount of vi sible oil is
not important in itself, but it also indicates a point where moisture
will enter. It is critical for life extension to keep transformer as
dry as free of oxygen as possible. Moisture and oxygen cause;
the paper insulation to decay much faster than normal and formed
acids, sludge, and more moisture. Sludge settles on windings
and inside the structure causing, transformer cooling to be less
efficient, and slowly overtime the temperature rises.
Table 1: Upper estimating of the Rate of Water
Direct exuosure insulation to air
a) RH=75%, t=16h, 20C
b) RH=40%, 1= 16h, 20C
Water vauour molecular now

Via sealing capillaries

Via loosed gaskets
Viscous now of air; shipping condition: ,
Core and coil covered with oil
- Adequate sealing
- Insufficient sealing

Operation with open-breathing

Insufficient sealing drowned with rain water
Acids cause an increase in the rate of decay, which forms more
acids, sludge, and moisture at a faster rate. This is a vicious cycle
of increasing speed fonning more acid and causing more decay.
The answer is to keep the transfonl1er ory Tree
of oxygen as possible. Water can exist in the transformer in three
Free water, at the bottom of the tank.
Water can be in the form of water loil emulsion.
Dissolved water (is given in ppm in DGA).
Chemically bound water.
A gasket must fill the space between flanges. Since the pressing
force that compresses the gasket when new was essential to the
seal, it must continue to react by pressing tightly against the
flanges throughout its many years of life. While this might seem
obvious, and easy for any material to do, this is where many
materials fail. If the gasket material shrinks, or looses its elasticity,
it can no longer press against the flanges.
Gaskets get a lot of attention, when they do what we do not want.
Among those things, that we do not want them to do are:
1. Leak oil around the gasket.
2. Leak oil through the gasket.
3. Damage flanges by being too hard to compress.
4. Sticks so badly that the cover can not be removed without
5. Removing a old gasket puts shreds of it into clean spaces.
6, Contaminates, by partly dissolving or fragmenting into the
From the above list, most failures are of the first type, leakage
around the gasket.
Rate of contamination
Sorption with surface 1000 m
up O.5mm depth
13.5 kg
8.1 kg
. Less then 1-5 g per year
Less then 30-40 g per year
600 g per year
15 g in a day
6,000 g per year
200 g in an hour
Gaskets have several important jobs in sealing systems. A gasket
must create a seal and hold it over a long period of time. It must
be impervious and not contaminate the insulating fluid or gas
above the fluid. It should be easily removed and replaced. It must
be elastic enough to flow into imperfections on the sealing
surfaces. It must withstand high and low temperatures and remain
resilient enough to hold the seal even with joint movement from
expansion, contraction, and vibration. It must be resilient enough
to not take a "set" even though exposed for a long time to pressure
applied with bolt torque and temperature changes. It must have
sufficient strength to resist crushing under applied load and resist
blowout under system pressure or vacuum. It must maintain its
integrity while being handled or installed. If a gasket fails to
meet any of these criteria, a leak will result. Gasket leaks result
from improper torque, choosing the wrong type gasket material,
or the wrong size gasket. Improper sealing surface preparation
or the gasket taking a "set" (becoming hard and losing its
resilience and elasticity) will also cause a leak. Usually, gaskets
take a set as a result of temperature extremes and age.
Transformer is fitted with A) Thermo siphon b) Air cell in the
radiator C) Silica gel breather, in order to minimize ageing due
to oxygen and moisture. Due to leak through the gasket
transformer is under accelerated ageing mode. Sealing (mating)
surface preparation: Clean the metal surface thoroughly. Remove
all moisture, oil and grease, rust etc. A wire brush andlor solvent
may be required.
Caution: Take extra care that rust and dirt particles never fall
into the transformer. The results could be catastrophic, when the
transformer is energized.
is under 50 OF, wait about % to 1 hour after applying the material
to surfaces before bolting. If you are using cork-nitrile or cork-
neoprene, you can also seal gasket surfaces (including the edge
of the gasket) with this same material. Loctite makes other sealers
that can be used to seal gaskets such as "Hi-tack".
GE glyptol No. 1201 B-red can also be used to paint gasket and
metal surfaces, but it takes more time and you must be more
cautious about temperature.lfpossible, this work should be done
in temperatures above 70 OF to speed p:oint "urin::: All"", tl,A
paint to completely dry befo.re applying glue or the new gasket.
It is not necessary to remove old glyptol or other primer or old
glue if the surface is fairly smooth and uniform.
Caution: Most synthetic rubber compounds, including nitrile
(Buna N), contain some carbon, which makes it semi-conductive.
Take extra care and never drop a gasket or pieces of gasket into
a transfonner tank. The results could be catastrophic when the
transformer is energized.
Choose the correct replacement gasket. The main influences on
gasket material selection are design ofthe gasket joint, maximum
and minimum operating temperature, type of fluid contained,
and internal pressure of the transformer.
Cork-nitrile should be usee! if the joint does not have grooves or
limits. This material performs better than cork-neoprene because
it does not take a set as easily and conforms better to mating
surfaces. It also performs better at higher temperatures. Be extra
careful when you store this material because it looks like cork-
neoprene (described below), and they easily are mistaken for
each other. Compression is the same as for cork-neoprene, about
45 %. Cork-nitrile should recover 80 % of its thickness with
compression of 400 psi in accordance with ASTM F36. Hardness
should be 60 to 75 durometer in accordance with ASTM D2240.
(See published specifications for E-98 by manufacturer Dodge-
Regupol Inc., Lancaster, PA.)
Caution: Cork-nitrile has a shelf life of only about 2 years if
kept in normal storage condition at 27 ambient, so do not order
and stock more than can be used during this time.
Cork-Neoprene mixture (called chloropene) can a)so be used;
however, it does not perform as well as cork-nitrile. This material
takes a set when it is compressed i,\nd should only be used when
there is no expansion limiting grooves. Using cork-neoprene in
grooves can result in leaks from expansion and contraction of
mating surfaces. The material is very porous and should be sealed
on both sides and edges with a thin coat ofGlyptol No. 1201 B
red or similar sealer before installing. Glyptol No. l20lB is a
slow drying paint used to seal metal flanges and gaskets, and the
After rust and scale have been removed, metal surfaces should paint should be allowed to dry totally before installation. Once
be coated with Loctite Master gasket No. 518. This material will compressed, this gasket should never be reused. These gaskets
cure after you bolt up the gasket, so additional glue is not should be kept above 35 OF before installation to prevent them
necessary. If the temperature is 50 OF or more, you can bolt up from becoming hard. Gaskets should be cut and sealed (painted)
the gasket immediately. This material comes in a kit (part No. indoors at temperatures above 70 OF for ease of handling and to
22424) with primer, a tube of material, and instructions. If these reduce paint curing tinie.lnstalling neoprene-cork gaskets when
instructions are followed, the seal will last many years, and the temperatures are at or near freezing should be avoided because
gasket will be easy to remove later ifnecessary.lfthe temperature the gasket could be damaged and leak. Cork-neoprene gaskets
must be evenly compressed about 43 to 35 %. For example, if
the gasket is inch thick, 0.43 x 0.25 = 0.1 O. When the gasket is
torqued down, it should be compressed about 0.10 inch. Or you
may subtract 0.1 from inch to calculate the thickness of the
gasket after it is compressed. In this case, % = 0.25 so 0.25 minus
0.10= 0.15 inch would be the final distance between the mating
surfaces after the gasket is compressed. In an emergency, if
compression limits are required on this gasket, split lock washers
may be used. Bend the washers until they are flat and install
enough ofthem (minimum of three), evenly spaced, in the center
of the gasket cross section to prevent excessive compression.
The thickness of the washers should be such that the gasket
compression is limited to approximately 43 %, as explained
Nitrile "NBR" (Buna N) with 50 to 60 Duro (hardness) is
generally the material that should be chosen for most transfonner
Caution: Do not confuse this material with Butyl Rubber. Butyl
is not a satisfactory material for transfonner gaskets. The tenns
Butyl and Buna are easily confused, and care must be taken to
make sure Nitrile (Buna N) is always used and never Butyl.
Replace all cork gaskets with Nitrile if the joint has
recesses or expansion limiting grooves. Be careful to protect
Nitrile from sunlight; it is not sunlight re3istant and wi!)
deteriorate, even if only the edges are exposed. It should not be
greased when it is used in a nonmovable (static) seal. Whenjoints
have to slide during installation or are used as a moveable seal
. (such as bushing caps, oil cooler isolation valves, and tap changer
drive shafts), the gasket or O-ring should be lubricated with a
thin coating of DOW No. 111 or No. 714 or equivalent grease.
These are very thin and provide a good seal. Nitrile perfonns
better than cork-neoprene; when exposed to higher temperatures;
it will perfonn well up to 85C (195 OF).
Viton should be used only for gaskets and O-rings in temperatures
higher than 110C or for applications requiring motion (shaft
seals, etc.). Viton is very tough and wear resistant; however, it is
very expensive ($ 1,000+ per sheet) and should not be used unless
it is needed for high wear or high temperature applications. Viton
should only be used with compression limiter grooves and
recesses. Care should be taken to store Nitrile and Viton
separately, or order them in different colors; the materials look
alike and can be easily confused, and a much more expensive
gasket can be installed unnecessarily. Comparison and fill
requirements for Viton are the as same as those for nitrile, outlined
above and shown in table 2.
Gasket sizing for standard groove depths. Nitrile is chosen as
the example because it is the most commonly used material for
transfonner gasketing. As shown in table 2, nitrile compression
should be 25 to 50 %. Nitrile sheets are available in 1116 - inch-
thick increments.
the amount available to be compressed. (See table 2.). Gasket
sheets come in standard thicknesses in 1116 -inch increments.
Choose one that allows one-third of the gasket to stick out above
the groove if you can, but never choose a thickness that allows
less than one-fourth or as much as one-half to protrude above
the groove. Do not try to remove old primer from the groove.
Horizontal groove fill is determined by how wide the groove is.
The groove width is equal to the outer diameter (OD) minus the
inner diameter (IlJ) divided by two: Y2( OlJ-tl). Or just measure
the groove width with an accurate caliper.
The width of the groove minus the width of the gasket is the
room left for the gasket to expand while being compressed. For
nitrile, the amount of horizontal room needed is about 15 to 25
%. Therefore, you need to cut the gasket cross section so that it
fills about 75 to 85 % of the width of the groove.
For example, an 8-inch 00 groove with a 6-inch ID,
is = I inch. Therefore, the width groove is I inch.
Because we have to leave 25 % expansion space, the width of
the gasket is 75 % of I inch, or:y. inch. So that the gasket can
expand equally toward the center and toward the outside, you
should leave one-half the expansion space at the inner diameter
of the groove and one-half at the outer. In this example, there
should be a total space of 25 % of linch or (1/4 inch) for
expansion after the gasket is inserted, so you should leave 118-
inch space at the OD and J/8-inch space at the ID. See figure 14.
Always cut the outer diameter first. In this example, the outer
diameter would be 8 inches minus Y4 inch or inches.
Rectangular Nitrile Gaskets larger than sheet stock on hand can
be fabricated by cutting strips and comers with a table saw or a
utility knife with razor blade. Cutting is easier if a little transformer
oil or WD-40 oil is applied. Nitrile is also available in spools in
standard ribbon sizes. The ends. may be joined using a
cyanoacrylate adhesive (super glue). There are many types of
this glue; only a few of them work well with nitrile, and they all
have a very limited shelf life. Remember to always keep them
refrigerated to extend self life. The one proven to standup best to
temperature changes and compression is Lawson Rubber Bonder
No. 92081. The Lawson part number is 90286, and it is available
from Lawson Products Co. in Reno, Nevada, (702-856-1381).
Loctite 404 is commonly available at NAPA auto parts stores
and works also but does not survive temperature variations as
well. Shelflife is critical. A new supply should always be obtained
when a gasketingjob is started; never use an old bottle that has
been on the shelf since the last job.
When bonding the ends of ribbon together, ends should be cut at
an angle (scarfed) at about 15 degrees. The best bond occurs
when the length ofthe angle cut is about four times the thickness
of the gasket. With practice, a craftsperson can cut 15 - degree
scarfs with a utility knife. Ajig can also be made from wood to
hold the gasket at a 15 degree angle for cutting and sanding. The
Gasket thickness is detennined by groove depth and standard ends may be further fme-sanded or ground on a fine bench grinder
gasket thickness. Choose the sheet thickness so that one - fourth wheel to match perfectly before applying glue. Aj ig ground on a
to one-third of the gasket will protrude above the groove; this is fine bench grinder wheel to match perfectly before applying glue.
Table 2.- Vertical Compression for Rectangular Nitrile Gaskets
Standard groove Standard ribbon Recommended gasket Available to Available
depth width
(in inches) (in inches)
3/32 1/4
118 5/16
3/16 3/8
1/4 3/4
3/8 3/4
A jig can be fabricated to hold the gasket at 15 degrees while
cutting, sanding or grinding.
Note: Maximum horizontal fill ofthe groove should be 75 to 85
% as explained above in the circular gasket section. However, it
is not necessary to fill the groove fully to 75 % to obtain a good
seal. Choose the width of ribbon that comes close to, but does
not exceed, 75 to 80 %. If one standard ribbon width fills only
70 % of the groove and the next size standard width fills 90 %,
choose the size that fills 70 %. As in the circular groove explained
above, place the gasket so that expansion space is equal on both
sides. The key point is that the cross sectional area ofthe gasket
remains the same as the cover is tightened; the thickness
decreases, but the width increases.
Caution: Nitrile (Buna N) is a synthetic rubber compound and
as cover bolts are tightened, the gasket is compressed. Thickness
of the gasket is decreased and the width is increased. If a gasket
is too large, rubber will be pressed into the void between the
cover and the sealing surface. This will prevent a metal to metal
seal, and a leak will result. It is best if the cross sectional area of
the gasket is a little smaller than the groove cross sectional area.
As cover bolts are tightened, the thickness ofthe gasket decreases
but the width increases so that cross sectional area (thickness
times the width) remains the same. Care must be taken to ensure
that the gasket cross sectional area is equal to or slightly smaller
(never larger) than the groove cross sectional area. This will
provide space for the rubber to expand in the groove so thilt it
will not be forced out into the metal-to-metal contact area. lfit is
forced out into the "metal-ta-metal" seal area, a leak generally
will be the result. When this happens, our first response is to
tighten the bolts, which bends the cover around the gasket
material in the metal-to-metal contact area. The leak may stop
(or more often not); but the next time the cover is removed, getting
a proper seal is almost impossible because the cover is bent.
Take extra care sizing the gasket, and these problems would not
Caution: On some older bushings used on voltages 15 KVand
above, it is necessary to install a semiconductive gasket. This
type bushing (such as GE type L) has no ground connection
between the bottom porcelain skirt flange and the ground ring.
The bottom of the skirt is normally painted with a conductive
paint, and then a ground. The gaskets are typically a
semiconductive neoprene material. Sometimes, the gasket will
have conductive metal staples near the center to bleed ofT these
charges. When replacing this type gasket, always replace with
like material. If like gasket material is not available, use cork-
thickness compress compression
(in inches) (in inches) (in inches)
118 1132 25
3/16 1116 33
114 1116 25
3/8 1/8 33
112 118 25
Thin metal conductive shim stock may be folded over the outer
perimeter around approximately one-half the circumference.
These pieces of shim stock should be evenly spaced around the
circumference and stick far enough in toward the center so that
they will be held when the bolts are tightened. As an example, if
the gasket is 8 inches in diameter, the circumference would be
pD (i.e. 25.13 inches) in circumference. 50 % of25.13 is about
12.5 inches. Cut 12 strips I-inch wide and long enough to be
clamped by the flange top and bottom when tightened. Fold them
over the outside edge of the gasket leaving a little more than I -
inch space between, so that the shim stock pieces will be more
or less evenly spaced around the circumference.
Note: Failure to provide a path for static electric charges to get
to ground will result in corona discharges between the ground
sleeve and the bushing flange. The gasket will be rapidly
destroyed, and a leak will be the
Bolting sequences to avoid sealing problems: If proper bolt
tightening sequences are not followed or improper torque applied
to the bolts, sealing problems will result. The resulting problem
is illustrated in figure 1.
Figure -1, Bowing at Flanges
A slight bow in the flange or lid top (exaggerated for illustration)
occurs, which applies uneven pressure to the gasket. This bow
compromises the seal, and the gasket will eventually leak.
Proper bolting sequences are illustrated for various type flanges/
covers\in figure 2 bolt numbers show the correct tightening
The numbers do not have to be followed exactly; however, the
diagonal tightening patterns should be followed. By using proper
torque and the illustrated sequence patterns, sealing problems
from improper tightening and uneven pressure on the gasket can
be avoided. Use a torque wrench and torque bolts according to
oe 10
. OJ 10
4 2
10 o . 0
-0 C1F1CUlM FOIJft.aot. T
to' PO


1 3
b b
the head stamp on the bolt. Check manufacturer's instruction
book for pancake gasket torque values.
Figure 2 - Bolt Tightening Sequences
Recommended Method:
Utilize DTl's
To reduce cross-talk, tighten bolts to 50 % of torque value
during first pass.
Use a cross pattern when tightening (12 -6, 3-9).
Second pass to 75 % of load.
Third pass until you see orange or 100 % of torque value.
Oil preservation system:
Conservator system
Complete oil system in transformer
Minimal routine maintenance
Gas accumulating relay
Identical transformers with different oil quality results due to
long term leaks
Case Study
TI- Transformer oil preservation system well maintained
with no-leak.
T2- Transformer oil preservation system not properly
maintained and with very heavy leak due to poor gasket.
Change in Moisture Content:
Water Content
04/16193 05128104
2.5% Upper limit
ReIIitIYe s.turation


Change in 1FT and Neutralization #
Interfacial Tension

04/16193 OS/28/04

Neutalization Number OJ Upper limit
0.3 ........... rw--r""II!:I .... ___ ......
9 0.2+--------------___
0.1 t----;.:=::::===-.;----1

04116193 05128104

All of a sudden ingress of free water may kill the transformer
immediately. E.g. 400MVA, 220Kv transformer fell due to
breakdown of the oil space between the bushing and the tank
after sharp cooling with rain-fall and ingress of about SOOg of
water through the broken sealing of the draw-lead bushing.
Ifa large cover is sealed with the cheapest material, between the
cheapest flanges, and with seriously flawed joint designs, oil will
always find its way out. It may be after the warranty period, but
it will always seem to be before the equipment would otherwise
have needed service. If we stop buying new equipment with
untested materials, or with joint designs we know it to be bad,
the problem might go away. If equipment manufacturers and
material supply and service companies were to offer really long
no-leak guarantees, the future might begin to look drier.
For transformers with air cell (diaphragm) oil preservation
system, careful analysis ofDGA, increase in oxygen is precursor
of leakage in diaphragm.
One common failure is the shrinkage of the compressed part of
the gasket. It seems to be a result of compression and heat. With
cork and cork-rubber materials, this can occur whether or not
the sample is immersed in oil. Most solid rubber materials are
still elastomeric after testing; the white PTFE "rope" material
never shrinks, but it cannot tolerate much transformer flange
movement in service. It is intimidating to realize that most samples
fail. This should not be too surprising, since most of our gaskets
leak after 5 to 10 years of seryice at much milder temperatures.
The oil preservation system must be. designed to prevent the
entrance of moisture into the oil over the life of the transformer.
This requires rubber in the expansion tank bags that will prevent
air leakage through the bag. It is surprising that almost no utilities
<;!,prity m ~ x m u m air leakage rate through the rubber. The gasket
must be designed to prevent leakage. Any added cost for good
gasket systems and rubber bags results in less maintenance over
the life of a transformer and extended life.
Bushing gasket leak may permit free water to enter the
Jransformer, dropping onto the core & clamping structure. Some
of the core insulation may bum, and sludge & carbon provide
over heating of the oil, probably for the current path across the
core lamination and to the unintentional ground.
If moisture in transformer is doubled, life is reduced by half.
1. Moisture in Transformer "A cup full of trouble" by A.K.
2. "The Answer to your Transformer's future lie in the oil "by
A.K. Yadav
3. "Dangerous effects of moisture on the dielectric strength
and ageing of oil paper insulation" by A.K. Yadav in Doble
India Colloquialam 2006.
4. FIST 3-30 Transformer Maintenance, United States
department of the Interior, Bureau of Reclamation
5. "Water in Transformers - So What!" P.J. Griffin, Doble
Engineering Company, National Grid Condition Monitoring
Conference, May, 1996.
Mark Ridgway
In 1997 a new transformer core construction method was created
at AEM Cores Pty Ltd This core is createu oy a m(1l,;iliul;; W iii"i1
precisely folds laminations which are then easily and quickly
assembled into a new type of core, the Folded or UNICORE.
The UNICORE is a modern development in wound core
technology. Throughout this paper we will use the terms:
Folded Core to refer to laminations that are machine folded.
This technology has been adopted by most of the major
international transformer manufacturers around the world and
is now being adopted by the leading manufacturers in India.
Wound core to refer to the traditional method of winding a core
on a mandrel and then deforming it under pressure to arrive at
a rectangular form factor. Some Indian manufacturers use this
Stacked Laminations to refer to the traditional manually cut
and notched, stacked lamination. This is by far the most common
method of core manufacture used in India today.
In focusing on the special characteristics and needs of the Indian
market, this paper demonstrates the superiority of the Folded
core over wound and stacked cores,
Due to limited market acceptance and penetration, and the lack
of availability of the material, amorphous metal core
transformers have not been considered in this paper. The focus
of this paper is to study transformer core production from CRGO
which is widely available in large quantities sufficient to satisfy
the considerable and immediate requirement for transformers
in India.
Stacked core transformers have been familiar to Indian
transformer manufacturers since Independence. Commonly,
engineering books, periodicals, magazines etc. cover the
construction, features and performance of stacked core
transformers. Little written material is available for Folded and
wound core transformers. However the working principles of
all of these types of transformers are the same and are based on
Faraday's Law of Electromagnetic Induction. The same formulas .
are applicable for calculation of core area, flux density, current
density, losses, impedance, efficiency regulation etc. It is simply
the orientation of the laminations that is changed.
For stacked cores, following the lengthy process of assembling
the individual laminations which form the core, the coils are
slipped onto the assembled core limbs. In Folded or wound
core transformers, the core laminations (in packets) are slid into
the stationary coils. This difference in construction technique
greatly decreases the transformer assembly time.
Rural village electrification under APDRP and RGGVY schemes
iijam agenda items for Indian power utilities.
Available statistics indicate that some 1,25,000 villages are yet
to be electrified. To complete electrification, Millions of
distribution transformers will be needed, very good % of which
are single phase types.
The main advantage of single phase distribution transformers
is that the shorter length of LT lines reduces line losses and
prevents people from using power without paying for it. With
the planned schedule of forthcoming electricity demand and
generation, there will be a huge demand for small rating
distribution transformers and the application of Folded core
transformers will lead to huge saving of electricity by way of
less current loss.
India's transformer industry has a number of challenges which
make a Folded Core transformer a superior choice over wound
cores aI!d stacked cores.
Preventing unscrupulous manufacturers using second grade
materials or used lamination material for making transformers
has'been discussed in many conferences and periodicals.
Unfortunately, these poor practices by some manufacturers still
continue in the manufacture of small to medium sized stacked
core transformers.
To prevent transformer manufacturers from using inferior
materials, manufacturers can be encouraged to make distribution
transformers with Folded core technology instead of stacked
core technology. The continuous and precise nature of Folded
core manufacture prevents the use of inferior materials.
Three Phase Folded Core (450Kg)
Machinery for making Folded cores (for example the
UNICORE) requires continuous strip to be fed from a roll of
CRGO on a dereeler or decoiler. This prevents the use of used
laminations. The UNICORE machinery also requires accurate
strip thickness and good quality rolls of CRGO to ensure the
high feed accuracy necessary to make the precision Folded core.
This makes it difficult to use inferior materials whilst at the
same time providing substantial economic benefits. A Folded
core results in an energy efficient transformer. and guamntees
the buyer will benefit from quality materials and construction
techniques. This directly benefits the Indian Government's goal
of electrification and therefore the Indian people.
In India, small to medium distribution transformers have
typically been made from stacked laminations. A stacked core
is typically hand cut and assembled leading to potentially large
variations in quality and performance.
The following issues can lead to sub-standard transformer
Variations in material thickness leading to mismatched
laminations at the joins
Variations in material grades leading to non-uniform heat
distribution, and therefore to high failure rates
Potential for cut burr leading to higher losses
Potential for inaccuracy of mitre cuts and joins, again
leading to higher core losses and failure rates.
These production variables combined with human error mean
that the core loss of stacked core transformers can differ from
person to person and batch to batch even if the designs and
nomii.1ated materials are the same. Wound cores, also involve
large operator input due to their less highly developed
technology. This results in similar variability issues as stacked'
cores. This variability can lead to core losses significantly higher
than the mill test certificates.
A Folded core is the answer to the above problems. Produced
by a computer controlled machine, the Folded core has precise
geometry and is free of variability due to operator input. The
core is of rectangular type with distributed cuts along the limbs,
which results in rapid assembly and minimal core loss. Various
dimensions like window length and width, strip width, build up
and strip thickness' are fed into a computer attached to the
UNICORE Control Software - Core Designer Screen
Folded cores benefit from negligible variability due to minimal
operator input and consistent material thickness and quality.
Engineering design factors can therefore be reduced, resulting
in more efficient use of material and cheaper, more reliable
The Folded core is also superior to traditional wound core as
the manufacturing process only stresses a small part of the core.
In wound cores the entire core is stressed, particularly by the
r r o c ~ s th:d invnlv_ "'!"PP7;n:: lh,:, ~ ~ : ~ : : ~ ~ ~ to change the
wound shape. After annealing, a Folded core can be produced
with almost zero destruction factor.
The uniform magnetic properties of Folded cores enables
manufacturers to consistently offer very low no-load loss and
current transformers
Stacked cores need multiple cutting/shearing and notching
operations to form a fully mitred core assembly resulting in
core material scrap of around 3 to 4% for small rating
Folded cores need only I or 2 horizontal machine made cuts
with a result of zero wastage of core material. At the end of
production of a batch of stacked core transformers, it is common
to see unused laminations remaining in various sizes. This is
also counted as wastage.
This potential total wastage of 5 to 6% for stacked core increases
the stacked core transformer cost compared to a Folded core
In stacked core transformer production, the core is made from
loose laminations of multiple cut sizes and strip widths. The
assembly operation starts with the smallest strip width first and
involves building the magnetic circuit by laying down the
matching limb and yoke laminations. The operation continues
until all the steps or strip widths are completed. The stacked
core assembly is done on an inter-leaved basis. An overlap of 5
to 10 mm is kept at the core joints to make the assembly possible,
This overlap is also necessary to offer some mechanical strength
to hold the assembly rigid. The.other steps, as per the design,
are added one above the other until the assembly is made
complete. Depending upon the size of the transformer, it takes
approx. 2 to 3 hours for 2 workmen to complete a stacked core
assem.bly. Working space and storage space requirements are
A further advantage of Folded cores is that the core and coil
assemblies may be constructed in parallel and then bought
together for final assembly. This also serves as an aid to repair.
The coil may be wound onto a bobbin or former, which can
then be put through processes such as drying ovens without
waiting for the core. This also yields energy and space savings
in the drying process leading to further cost savings.
One of the significant benefits of Folded cores is their very
short assembly time. Typically when manual assembly time for
a stacked core is 2 hours, a similar dimensioned Folded core
can be assembled into the coils in approximately half an hour.
Single phase DUO UNICORE ready for assembly
Folded cores "self locate" during assembly ensuring more
accurate core building. Folded cores also enable assembly with
less stress and therefore lower losses than wound cores. The
working and storage space are also greatly reduced.
Folded corl.! transformers save a considerable amount of
manpower and production time, leading to significant cost
savings and increased output capacity. This allows existing
factories to greatly increase production and therefore meet
India's electrification goals.
The limb section of a stacked core is usually made round to
accommodate cylindrical coils. By forming a round section,
approximately 6 to 8% of core area is lost. Numerically this
can be written.
Gross Core Area = 0.94 1td
Where d = Core diameter.
The factor 0.94 varies with core diameter and the number of
core steps. With a core diameter of approximately 125mm
having 10 to 11 steps, the space factor may be taken as 0.945.
For smaller rating transfonners (say 25 KVA and below), the
space factor may become as low as 0.92.
The limb section of a Folded (and wound core) is rectangular
in shape and because of this, the space factor is close to 'unity'.
The fonnula now becomes:
Gross Core Area = Strip Width x Build Up
Where Build Up is the stack height in a Folded core.
Since the space factor is improved by 6 to 8%, a Folded core
can yield lower core weight with respect to stacked core for the
same geometrical configuration.
The manual fabrication of stacked core laminations is a complex
process. The process starts with slitting of the mother coil into
narrow widths. The number of widths depends upon the design
criteria. For example, for a 100 KVA stacked core transfonner,
there may be as many as 10 steps.
After slitting, the next operation is the mitre shearing of
laminations to the required lengths. It is again a long process
and continues until all the steps or widths have been processed.
Then the top and bottom yoke laminations must all be notched.
All these manual processes make the stacked core transfonner
very slow to produce vastly reducing the output oftransfonners.
The number of manual processes adds to the cost of the final
Folded cores remove the need for the process of multiple slitting,
shearing & notching. These cores need only one slitting
operation to produce the strip width specified by the design.
The AEM Cores UNICORE machine permits the simultaneous
manufacture of2 cores when the strip width is below a certain
size, thus doubling the core production rate.
Lamination emerges from UNICORE machine
The UNICORE machine eliminates the manual shearing and
notching operations. Slit CRGO is fed to the UNICORE machine
which progressively folds and cuts the laminations automatically.
The entire core cutting and building operation is managed by
only one or two persons. This is in comparison to a stacked
core plant which will require significantly more people and area
to operate. With this large number of people and multiple manual
operations, control of errors and quality variation is far more
difficult. As a result, stacked core transfonners are loosing
popularity to Folded core transfonners. The main driving force
is that the Folded core transfonners can be made to a consistent
high standard, very quickly and at reduced costs.
Magnetic and electrical characteristics of a Folded core
transformer are far superior than a manual stacked core
transfonner because:
It is a machine fonned core comprising precision computer
controlled lamination fonning.
The number of joints in a lamination are reduced.
Handling factor (Destruction factor) is very low, and is
further reduced by. annealing
Space factor is unity (The folded core is rectangular in
Human manufacturing variations are almost eliminated.
The above advantages mean that:
No-load loss is reduced
No-load current is reduced.
Other performance figures are similar to a stacked core
transformer (when it is precisely manufactured).
Property Stacked Core Transformer Folded Core Transformer
No. of cuts per layer or lamination 6 or more (3 phase) . Not more than 2
Magnetic Properties Variability due to manual cut and assembly Consistent due to automated production
Production Rate Slow Fast
Working Space High Low
Noise level Problems due to joint quality Low due to precise geometry
Core framing brackets Heavier
Insulating Materials
Internal clearances
Core Mass
Overall Transformer Cost
Stacked core transformers have been familiar to Indian
transformer manufacturers since Independence. Folded core
transformers are a modem development which has entered the
market relatively recently and is being used by the major
international manufacturers around the world.
This paper has examined the merits of Folded core transformers
over stacked core transformers in the Indian transformer
manufacturing context. The following significant advantages
have been demonstrated:
Less Needed
Accurate geometry allows tighter design
Greatly reduced production time leading to higher output
of transformers
Consistent, high quality cores resulting from computer
controlled manufacturing process, lead to reduced energy
consumption and increased efficiency
Reduced manufacturing cost while maintaining high .quality
Increased production of transformers through the use of industry
leading Folded core technology will assist the Indian
Government achieve it's electrification goals.
The utility and usefulness of Grain Oriented Stack Core
transformers have been discussed at length in many international
seminars, periodicals, magazines and even in books also. These
days Amorphous Metal Wound Core Transformers also have
gained a little space for discussion in the seminars. However,
these two topics are always discussed separately. No one has
taken any pain to address a comparison between the two as yet
in any common forum. Our aim is to identify the merits and
demerits of each category of transformers and finally conclude
the topic with few comparison tables, which will provide the
power utilities a fair idea while selecting the specific grade of
transformers for use in the distribution systems.
1. Introduction
Power sector has grown -immensely from the time of
independence. Against an installed capacity of 1713MW in
1950 the installed capacity today is more than I, 35,000MW. It
has been targeted to install additional 75,OOOMW by the end of
11 fifth year plan, thereby the total installed capacity would be
2, 1O,000MW by 2012.
sharply increasing costs of electrical energy are forcing
electrical supply authorities to recognize the critical importance
of the cost of electrical losses. Electrical utilities are increasingly
required to operate their net works more efficiently and to reduce
the total real running costs of the equipments. Transformers
(after transmission lines) are the largest loss making components
in the electrical net work. There are about 25,26,517 Nos.
Transformers with a total installed capacity of 759,239 MW
(as on 31-3-2004) in the power system. It is estimated that
about 35 billion KWH of energy is lost in transformer every
year. The total losses due to transformers in the electrical net
work in India exceeds 6 per cent of the total electricity generated,
equivalent to about 8-10 per cent of the total loss from the
The enormous stock of transformers is often overlooked as a
source of cost savings. In fact, avoidable losses from the
distribution transformers currently in service would mean release
of about 650 MW of generating capacity, which could instead
be used to serve customer loads (from CEA report).
2. Grain Oriented Stack Core Transformer
Cold roller grain oriented silicon steel is used as core material
for transformer. In recent years, more importance is being laid
down on energy savings. Transformer manufacturers are
encouraged on minimizing losses. Earlier transformer
manufacturers had no option but ro use a particular grade of
CRGO Steel, as the availability was restricted only to grade
M4-0.27 mm. But with the passage of time and with the advance
technology, we are now with finp nfrunn
Steel, commonly known as HI-B and Lazer Grade.
3. Comparison of various grades of CRGO Steel
Most commonly in India, distribution transformers are
manufactured by using M4-0.27 mm grade materials. But in
order to minimize the losses, the following changes in grades
may be considered:
a) CRGO M4-0.27mm : Replaced by - HI-B-MOH-
O.27mm, affecting a reduction
of losses by 16 per cent.
b) HI-B-MOH-0.27mm Replaced by - HI-B-MOH-
0.23mm, affecting a reduction
of losses by 8 per cent.
c) HI-B-MOH-0.23mm : Replaced by - Lazer grade
0.23 ZDMH, affecting a
reduction of losses by 8 per
It is understood from the above that by using better grade of
materials, the losses can be reduced to (16+8+8)=32 per cent
than that the conventional core being design previously with
M4-0.27mm. materials.
With the availability of new grade ofCRGO Steel, the no load
loss of standard rating REC transformers may be reduced to as
RATING Previously Re\'ised
3 Phase) guaranteed guaranteed
No Load Loss No Load Loss
16KVA 65 W SOW
25KVA 100 W 68 W
63KVA 180 W 123 W
lOOKVA 260W 177 W
Reduction of losses not only minimize the running cost, but
also helps to reduce the core frame size (expect amorphous metal
transformer), thereby helping to reduce the cost of winding and
oil. We can now design a cost effective low loss better efficient
transformer by incorporating the above grade of materials.
4. Processing of CRGO Steel.
Processing of CRGO Steel includes the following operation:
(i) Core Slitting
(ii) Core Shearing (Mitring and notching)
(iii) Core Annealing
Let us add few paragraphs on each of the; above operations
highlighting the basic needs to get quality output.
Core Slitting:-
Slitting operation itself is to slit the mother coil into narrow
hoops. In the case of electrical steel, slitting operation causes
deterioration of magnetic properties of the mother coil.
In order to achieve good slitted coils without much deterioration
of magnetic properties, the cutting tools ofthe slitter lines need
to be adjusted properly. Satisfactory results cannot be obtained
without proper control on this part. In case the sharpening of
cutting blades are not satisfactorily attended or the blades are
not properly aligned, considerable amount of burrs may results
in the cutting edges. Moreover, sometimes the coating insulation
along the cutting edges may also get scratched because of the
mismatching of the cutting blades. The blades are needed to be
re-aligned on the basis of lamination thickness.
In case the cutting edges burrs are more than 40 microns, the
cause of such high burrs beyond limit may be investigated and
eliminated before further processing. In all cases, the process
should pass through a de-burring machine before taking them
for core assembly.
Core Processing:
After slitting, the next operation is to process the slitted coils,
which includes mitring and V-notching. The two side limbs are
cut at an angle 45 Degree. The top and bottom laminations are
punched as 'V' at the centre and thereby creating a notch to
recei ve the centre leg laminations.
The processing of CRGO Steel is a continuous operation an.d
the processors therefore need to be vigilant on the sharpenness
of the cutting tools. Otherwise excessive burrs at the cutting
edges may cause unwanted air gaps in the joints during core
assembly, resulting high no load losslcurrent.
Core Annealing:
During slitting, mitring and notching operations, the laminations
loose some of its magnetic properties. The grains get distorted
due heavy mechanical stresses produced during cutting
operations. To regain the grain orientation, the laminations need
further stress releaved annealing. Annealing is done to achieve
the following:
(i) To reduce the mechanical stresses in the laminations and
to yield optimum magnetic property.
(ii) To prevent contamination of steel with oxygen and/or.
(iii) To retain or enhance the insulation quality of lamination
These qualities are generally achieved by at a
temperature in the range of 1400 to 1500 Deg, F (760 to 845
Deg.C), preferably with a protective atmosphere. The
continuous roller furnace is generally employed for annealing
eRGO laminations of reasonable uniform sizes.
s. Core Building.
Various types of core stackings for three-phase transformers
are adopted during assembly. Each scheme has its Own
advantage as well as limitations. The manufacturers should
look into the applicability of each scheme before going for final
design of core building.
Apart from the conventional type of core building, there are
some more schemes, commonly known as 'Step-T
construction. The construction of 'Step-Lap' may be of 'cross-
step' or 'longitudinal-step'. These types of core building
schemes have specific advantages over other conventional
schemes, especially towards magnetic properties i.e. no-load
loss and current. But in India. 'step-lap' construction is not
very popular as the cost of processing equipments is huge and
beyond the reach of small manufacturing units.
6. Merits and demerits of Stack Core Transformers:
Let' us categories the merits of Stack Core Transformers as
(a) Though the CRGO electrical Steel is an imported
material, it is widely available in India through various
reputed vendors throughout the country (like KRYFS,
(b) The of CRGO Steel is easy which does not
need very sophisticated processing Equipments (except
automatic cut to length, CNC control machine).
(c) Cost of processing equipments is cheap and is affordable
to even small manufacturers also.
(d) The stack core constructions offer wide flexibility to the
designer in terms of shape, size, dimensions, grade and
(e) Core assembly does not require any special expertise.
Workmen having normal skill can haQdle the assembly
operation very successfully.
(f) With the use of improved grades of CRGO laminations,
the no-load loss can be.reduced to 32% than that specified
by REC. In numerical value, no-load loss specified by
REC for 25 KVA transformer as 100 Watts may be brought
down to 68 Watts with the use of superior grade ofCRGO
(g) With a bigger frame size of core assembly, the load loss
can be reduced considerably to offer a better cost effective,
low loss, energy efficient transformer.
(h) Concentric coils both for HV and LV offer a human friendly
construction. The coils can be made single or sectional as
per the requirements of design of criteria.
(i) Stack core transformers are made on the basis oflarge scale
production in India, because of its simplicity and easy
(j) Stack core transformer are easily repairable, even at site
also. No. Special infrastructure is required to undertake
such repair.
(k) Because of the wide availability of this raw materials in
India, stack core transformer can cater to the requirements
of the early delivery if desired by the buyers.
(I) Last but not the least -the overall cost of stack core
transformers is affordable to the buyers.
With above, let us conclude the discussion on 'Merits' of the
CRGO steel stack core transformers. We would like to add few
lines on the demerits of the stack core transformers.
(a) Based on the design criteria the cut to size laminations are
stacked one by one and are clamped at top and bottom to
form an integral part of the core assembly. Such stack
core assembly also needs space, as well as it is a time
consuming process.
(b) In case the operation of the slitting, cutting and punching
etc., are not made as per available guide lines, or the core
assembly leaves excessive air gap at the joint, we may loose
some of the valuable magnetic properties.
Baring the above two demerits (i.e. weaknesses), all other
properties come under merits (i.e. Strength). However, the
weaknesses can be converted to strength by proper planning
and improved workmanship. Stack core transformers is called
as user's friendly electrical equipments and can serve much
longer life iJ properly maintained.
7. Amorphous Metal Core Wound Type Transformers
There has been constant search for transformer core materials
which may have the least loss. Iron-Boron-Silicon amorphous
alloy has evolved as the low loss material for distribution
transformers. Mother metal when cooled to solid state at a very
high rapid rate retain a random atomic structure which is not
crystalline. This metal is called 'Amorphous'. This resembles
with glass and also referred as 'glass metal'. Need to achieve
the required cooling rate restricts the thickness of the metal to
about 0.025mm. i.e. almost 1/1O
of the thickness of the
conventional CRGO Steel.
Due to less thickness and low saturation factor, larger core and
consequently larger coils and tank size are required as compared
eRGO core transformer. The problem of less thickness has
been overcome to some extent with the development of
amorphous metal strips. This is achieved by compacting number .
of thin ribbons. The strip is commonly known as "POWER
CORE" and up to a thickness of 0.25mm has been developed.
Amorphous strips are four times harder than CRGO Steel.
Hardness along with reduced thickness makes slitting and
shearing difficult. The brittleness property of amorphous metal
has also made it un-friendly to the transformer manufacturers.
Due to these and other limitations, the amorphous core
technology has been limited at present to very few customers in
India and abroad.
Amorphous metal core has some merits. The non-crystalline
structure and random arrangement of atoms gives low field
manetitation and high resistivity. Due to low field magnetization,
hysteresis loss is low and due to high electrical resistivity eddy
currents are suppressed.
IH of Amorphous Metal aHoy vis-A-vis CRGO
8. Propert . '
Steel .
u ...... i(S of amorphous magnetic alloy are given
The important I'" '1
',t, of the conventional CRGO Stlcon Steel.
below along"''' ,
Amorphous CRGO Silicon
Alloy Steel of
M4-0.27 Grade
Saturation 11\1' ""I\siIY
1.55 Tesla 2.03 Tesia
at: 25C 1 ()(n ' . ..-_
1.49 Tesla 2.09 Tesia
Core Loss:( W"II
0.2 IW/Kg at 1.02W/Kg.at
1.40 Tesia l.60 Tesla
Exciting Pmn"
0.37VA/Kg. 1.4VNKg.
at 1.40 Tesla at1.60 Tesla
130 45
DPH-IO.3 RB-76
Thickness ill f\
25-30 275
- .'-'-
. . \"
Space factol ,II I '_
78-82 95-97.5
Sensitivity hI
Appreciable Negligible
Higher Lower
Higher Lower
Available Fell "'
Ribbons /Foils Sheets/Rolls
Market avuilnllllll\
In the hands Easy Available
of Monopoly
Flexible and not
Locked in the
hands of the in the hand of
supplier anyone
I" ... haractenstic of amorphous alloy IS
The most ntH lit ' ..
'\ 11\'\1"" l"l\V core loss and magnetlsmg current.
obViously Its I' \ I'l\'perties, practically all other properties
Except thcse " \ .' , 11' fi '
I ., of amorphous a oy are 10 enor
. 1 d th<'IIIIII"'"
me u 109 f ('Ih i\ \ which offer a great challenge to the
to 0 ., ',m: them so tl)at the low olss property is
engmeers to 0\ { 1\ \ ,
. d' II' ,,1'\'" ,,,lst effecttve way ..
exp oite lJJ II
.tUlUS Gore lransformers
9. Cost of n,,1 -
\" .. 1;\[ core saturates almost at 1.55 Tesla
The amorp loll'"
r as CIH Ie \ s,..,..: saturates as hIgh as 2.03 Tesla. Thus
whe e I I "'1" transformer result in increase of core
orphous 1111' ,I ,
'. 1;\\\" and insulating oil. Overall cost of
size condm 1111,
, . II:tl\$ti.)rmer approxlmate.20 to 30 per cent
amorphous nil' .
. h '1\ "\\(l".aal core transformers.
costher t an lill
, """,rphous Core lransformers
10. Derner II ' .'
"I' ,,,'n:' transformers has excellent magnetic
Though amOI!' \1 '
. I " l,'\\ no-load loss and current, yet It could
roperties (Own' ' f' h' h' .. 1
I' u1 bJian Market because 0 Its Ig mltIa
not get por
1111 , f' 'b '
, \l {O the high rate offallure 0 dlstn utIOn
cost. Morcov('I, " "0,::
transformers in service, REC bas recommended the use of
suitable protective devices (MCCB with matching characteristic
HV on secondary and primary sides respectively to protect
it from external fault and short-circuits which occurs frequently
on LT lines. With such recommendations. the amoIphous core
transformer become more costlier and remain beyond reach to
most of the SEBs.
Further upon failure of amorphous transformers beyond
: ...... most of the SEB's become helpless in
undertaking repair of such failed transformers departmentally,
as they do not have sufficient known how well as infrastructure
to handle such repair works.
Looking to such laminations, it is suggested that the buyers
should not get mixed up the issues of purchase ofCRGO steel
transformers with amoIphous metal core transformers. They
should be dealt separately in two different tenders, may be
called separately in two separate NITs. A separate tender may
be invited for Amorphous metal transformers, ifbuyer so desire
to purchase such transformers on accadomic point of view. It is
further stressed that SEBs may go in for buying a limited
quantity of amorphous metal transformers for monitoring their
performances in the present days distribution network, as well
as make themselves friendly in maintenance and repair of such
11. Comparison:-
We would like to conclude the discussion with a brief
comparison table, highlighting the following facts, which will
give the buyer a fair idea while purchasing transformers in their
forthcoming tenders.
SI CRGO Steel Stack Amorphous metal
No Core Transformer Core Transformer
Widely available through Availability is in the hands
various reputed vendors of one or two monopoly
(like KRYFS, Mumbai) suppliers.
The processing of CRGO It is available in market in
Steel is easy and does not built in shape of Pre-
need very sofisticated determined dimensions. No
processing equipments scope of further processing
Costs of processing Costs of processing
equipments are cheap and equipments are very high
are affordable to even small and beyond reach to mmost
manufacturers of the Indian Manufacturers.
4 Stack Core TIansformers Amorphous Core
offer wide fle&ility to the Transformers do not offer
design engineers in terms of any flexibility to the design
shapes, size, dimensions, engineer as the shapes,
grades, and thicknesses size,dimension, grades and
thicknesses are pre-
determined and fixed.
5 Core assembly does not Core is available in Wound
require any special form already built in
expertise. Workmen having rectangular shape, except
normal expertise can handle filling of top or bottom
the assembly operation very which need of expertise.
6 Various grades of CRGO Only one grade is available
Steels are available in the
market. The designer has to
chose a particular grade of
material based on the loss
characteristics specified in
7 The design is flexible. With The desin is locked in the
a bigger frame size of core hand offew Monopoly core
assembly, the load loss can supplier.
be reduced considerably to
offer a better cost effective,
low loss energy efficient
8 Concentric cylindrical coils Concentric but rectangular
both for HV and LV offer a shaped coils make the
human friendly process of making coils a bit
construction. The coils can tough. Sofisticated winding
be made single or sectional machine with coil pressing
as per the requirement of rollers need to be employed.
design criteris. No special
winding machine is
9 Large scale production is Rate of production is slow
possible specially for unless a huge infrastructure
distribution transformers is built, like automatic
with minimum conveyer.
10 Easily repairable, even at
site also
It can be repaired by the
amorphous transformer
manufacturers only, as lot of
special infrastructures are
required to handle such
repair work.
11 Electrical: Performance No-load loss exceeds may
parametert are fairly Stable fold after repqir.
even after repair and are
within (+) 10%
12 CRGO S.teel in the form of Amorphous metal core is a
either sheets or rolls can be tailer made product and is
kept in stock to cater to generally procured after
satisfy early delivery. getting a confinned order
13 Cost of CRGO stack Core Cost of amorphous
Transformer is reasonable transformer is Reasonably
high by around 20-30
14 Stack core assembly needs Amorphous core
additional space in the work transformers do not occupy
place. It is a time any work place, as it is
consuming process. available in assembled
15 In case the operation of As it is available in built-in
slitting, cutting, punching, form, loosing of magnetic
deburring etc. are not made properties due to wrong
as per quality guidelines, or processing does not arise.
the core assembly leaves However, lacing the top or
excessive air gaps at the bottom yoke back to their
joints, we may loose some position to close the
of the valuable magnetic magnetic circuit need
properties. utmost care. Any deviation
may cause excessing no-
load loss and current
16 The no-load loss and current The no-load and current are
are reasonably high. This reasonably
can be counter by offering
a reasonably low load loss.
17 Saturate at 2.03 Tesla Saturate at 1.55 Tesla
18 Space factor of CRGO Space factor of amorphous
laminatiolls are as high ad core is as low as 78 to 82%
95 to 97.5%
12. Conclusion:
Each category of transformer has some merits as well as demerits
also. It is my honest appraisal to highlight them to the knowledge
of the buyers. CRGO stack core transformers have definitely
some edge over amorphous metal transformers and that could
the reason why it is dominating the entire world market including
India. Because of the workmen friendly construction, easy
availability of raw-material, easy repairable and last but not the
least low cost, the CRGO stack core transformers have proved
to be users friendly in India and also globally. It is known to all
of us that the amorphous metal core transformers have some
attractive magnetic preperties, i.e. low-no-Ioad loss and current.
By accepting these two properties, all other properties including
market availability, to those of CRGO steel
transformers. This could be one of the great challenges to the
working engineers to overcome them so that the low loss
property of amorphous metal is exploited in the most cost
effective way.
The amorphous metal core transformer is not a new technology
in India. The then lEMA (now IEEMA) had organized an
International Seminar in New Delhi exclusively on amorphous
Metal Core Transformer way }lack 1976. 32 good years have
passed and amorphous metal core transformer still could not
share even 2% of our national requirements. Had there been
good advantage of amorphQus metal core transformer, it could
have been occupied a lion share by this time .. But that has never
In this connection, I would like to repeat a paragraph which I
have already stated earlier. Amorphous Transformers look to
be economical only when we put in on theoretical calculation
during total owing cost based on a life cycle of 25 years. This
approach of calculation is not justifiable, as the distribution
transformers in India do not have a.life span more that 6 to 8
years. Even on an argument point of view, we accept a life
span of even 10 years, the amorpholls tr::msformers oroved to
be costlier than conventional CRGO stack core transformers.
Looking to all such limitations, it is suggested the buyers should
not get mixed up tQe issue of purchase of CRGO steel
transformers with that of amorphous metal transformers. Each
category of transformer should be dealt with separately. There
should be two different tenders with two different specifications.
A separate NIT may be invited for amorphous metal
transformers, if the buyers so desire to purchase such
transformers on accadomic point of view. It is further stressed
that the SEBs may go in for buying a limited quantity of
amorphous metal transformers for monitoring their
performances in the present days distribution network, as well
as make themselves friendly in maintenance and repair of sllch
Peter Keller
1. Theory
During the drying process, moisture molecules in the insulating
material must be transported from inner layers to the surface of
the material. It appears logical, that for such task a pressure
tiiffprentiH 1 ;c;: T!!e ,.,.':!!e!" partial pressure which
is build-up within the insulation depends on the temperature
and on the moisture content. In various publications, the main
importance of temperature is described in detail. This is a
deceive factor for a drying process, as it:
determines the drying speed (moisture extraction)
influences the drying quality (residual moisture)
influences the paper quality (Depolymerisation)
2. Drying time
The moisture transport from the inner insulation layers to, the
surface, depends mainly on the water vapour partial pressure
on the inside the diffusion coefficient of the transformer board
and to a much lesser degree on the final vacuum archived on
the outside of the insulation.
According to the law of Fick, the speed of drying V (f), i.e. the
extraction of moisture in function of time from the insulation
v (f) = D '" dcldx (% * em/d)
V (0 = Amount of moisture in gr, kg, or %,
which diffuses in-function of time
through a specific surface
dc / dx =Moisture profile
D = Coefficient of diffusions (cm2/d)
d = Time (days)
Diffusion Coefficient of Transformer Board
low-density, oil-free,'under vacuum ofO,13 - 13 mbar



120 110 lOOt
of diffusion
.--' -'
./ ./

./"" -/

i coefficient
o 2.5 5 7.5
I from 120
! to at
Recent comparative measurements to the established source of
Lampe showed dearly. that the coefficient of diffusion 0 rises
very strongly with increasing temperature and will decrease to
lower values with lower moisture content in the insulation. (e.g.
the moisture extraction from 5% to 4% takes far less time than
the moisture reduction from 2% to 1 %).
3. Depolymer;sation
The mechanical strength of insulation material (cellulose) is
basically determined by the so-called degree of polymerization.
The properties ofthe insulating material also have an influence
on the speed of depolymerisation
Limits are set for the drying temperature through material
characteristics of the paper. If the process is carried out at higher
temperatures, an increased depolymerisation will take place.
Besides temperature, the depoiymerisation rate also depends
on the following factors:
time during which the paper is exposed to the temperature
moisture content of the paper
presence of oxygen
4. Heating uplheat transfer to the drying object
Based on the previous statements, it is important, that the heating
up of the drying object is executed in the most efficient way.
The basis of the vapour phase technology is, that the heat transfer
to the drying object is assured by mainly condensation of a light
solvent and to a lesser extent by convection. The efficiency of
condensation and convection depends mainly of the turbulence
and velocity of the solvent vapour in the Autoclave respectively
on the surface of the drying object. Optimal heating up time are
achieved, ifplant components (evaporator) ensure the highest
possible turbulence/velocity on the surface of the drying object
during the whole heating phase.
5. Summary and consequences
The drying temperature is to be held at a high level, yet holding
in consideration the existing moisture content, in order to
achieve an optimum drying speed and quality. When raising
the drying temperature, the existing moisture level must always
be taken into consideration. To extract as much moisture as
possible at lower temperature of the insulation, helps to decrease
the depolymerisation
1. Applied technologies
Nowadays, the following criteria are applied in order to
determine the end of the drying process:
Measurement of the final vacuum with a total pressure-
measuring instrument (e.g. Pirani measuring device). In
addition a compression vacuum meter (McLeod) is being
used to measure the non-condensable gasses and vapours,
which under these conditions mainly results in the "air
partial pressure".
The resulting vacuum values achieved with the Pirani
vacuum meter are dependent on the gas and vapour
composition of the environment. 1t therefore ,is
recommendable to employ a gas-independent total
pressure-measuring d e : ~ : ; : ; .
Measuring of the Specific Water Extraction Rate (SWER),
which is expressed in gr(H
0)Ih.t , i.e. grams of water
extracted per hour and per ton of insulation. This method
used in combination with temperature pressure
measurement, is widely used these days by any large power
transformer man ufacturer. Yet, it must also be considered
that the SWER rate also depends to a certain extent of the
build-up of the insulation (voltage range, diffusion
characteristics, etc.)
The SWER rate is determined through measurement of
the water vapour partial pressure (P H
0) in the gas/vapour
mixture, which has been pumped offby the vacuum plant.
All above mentioned parameters are part of an indirect moisture
measurement of the transformer insulation and require in depth
understanding of the drying technology for correct interpretation
2. Adsorhtion moisture measurillg iltstrument VZ 4061475
The VZ406 instrument has been successfully used for years. It
measures, the water vapour partial pressure prior to the
preliminary vacuum pump through adsorption of the water
vapour with phosphor pentoxyde. The hereby-resulting pressure
difference corresponds with the water vapour partial pressure.
Through automatic determination of the suction capacity of the
preliminary vacuum pump(s), corresponding to the intake-
pressure, the SWER rate can be accurately and reliably
3. Compression moisture measuring instrllment HVPM
This unit has been recently developed for the determination of
the SWER rate without the use of chemicals. In addition direct
measurement of water vapour pressures inside of blocks of
insulation allows indirect determine the moisture con ten t in
thick insulation plates (as per piper curves).
The SWER rate can be determined by compressing the gas
sampled in the autoclave and measuring the water-, solvent
vapour and air partial pressure Simultaneously the SSER rate
(Specific Solvent Extraction Rate).
As this new measuring system operates on the basis of the
compression! condensation principles, it is virtually maintenance
4. Moisture measuring instrument with dew point sensor
This unit has been recently developed for the determination of
the SWER rate without the use of chemicals especially in less
critical application like the drying of distribution transfonners
in HA V plants. In order to overcome the typical disadvantages
of dew point sensor working at the limit of their measuring range
at vacuum level in the range of 1 mbar and below, the sensor
can operated at higher pressure levels than in the autoclave This
instrument allows the determination of above mentioned S WER
rate in conjunction with the measurement of the dew point,
pressure and integration of the suction capacity
5. Dew point measlIrement as stand alone measurement
This method will not provide sufficient information for
determination of the drying end. This due to the fact, that
dewpoint measurements varies in relation of the insulation
weight versus the suction velocity of the vacuumplant
6. Summary and consequences
The combination of the SWER rate together with the
measurement ofthe water vapour partial pressure inside of thick
insulation blocks, is an excellent basis for the exact termination
of the drying process with much less room for interpretation,
particularly when drying power transformers.
I. Vapour Phase Drying plant with JET evaporator
Vapour phase is a process for drying windings and active
parts of power transformers which were treated comp letel y
under vacuum, i.e. under absence of oxygen allowing
higher process temperatures compared to the classical Hot
Air Vacuum (HAV) plants. This will lead to significantly
shorter process times and better drying results
Nowadays with increasing worldwide and local
competition and energy cost, a VPD plant must operate in
a most efficient, secure and reliable way to guarantee low
operating cost and best drying results.
Therefore MICA V AC developed a new patented JET
evaporator system taking all experiences and in depth
knowledge about physics of drying into account.
JET evaporator
A suitable white spirit oil (solvent) is used as a heat transfer
medium. It is heated up in a small heat exchanger, injected
in the autoclave were it evaporates in order to heat-up the
drying object under vacuum, whereby this solvent vapour
condenses on the object and heats it up. The water in the
insulation evaporates and diffuses to the surface. The water
vapour and the solvent vapour (as dragging gas) are fed
into the condensation system, where they are separated
from one another in liquid phase. The solvent will be fed
back into the autoclave and re-evaporated until the max.
object temperature is reached.
This. so called heating up phase will be interrupted by
Intermediate r e s s u r ~ lowering (IPL) phases, allowing
water extraction at an early process stage which will result
in a shorter overall process time and reduced
vapour temp.
object temp.
Heating Press. Low. Fine vacuum
Finally the autoclave will be evacuated to the lowest
possible pressure in order to evaporate all remaining
solvent and water out of the insulation in order reach the
lowest achievable residual moisture
Key features of the JET evaporator system
Significantly increased turbulence in the autoclave and on
the drying object compared to any other evaporator system,
as vapour is discharged all along the evaporator,
corresponding to virtually the whole autoclave length and
the JET effect keeps all vapour particles in the autoclave
circulating at high speed. This fact leads to a better heat
transfer which is of most importance at the end of the'
heating up phase were the temperature difference between
the solvent vapour and the object is limited due to the max.
acceptable solvent vapour inlet temperature
Only the required minimum amount of solvent is filled into
the system, thus no unnecessary heating-up of solvent
(economising energy)
This also results in minimised danger in unlikely case of
air break-in during drying process due to minimal amount
of heated solvent in the system.
The evaporator built-in into the autoclave avoids pressure
and temperature loss in pipes resulting in both, a more
effective heating-up and.1ower energy consumption.
The energy consumed by the evaporation is used for heating
of the parts to be processed, even during distillation.
Distillation will be carried out during intermediate pressure
lowering phases and pressure reduction phase.
Reduced space requirement, as no separate evaporator
room is needed.
T. Vijayan
Impedance of Axially-split multi winding Transformer
A simplified approach
Use of axially-split multi transfonners is
recently. Axial split transformers comprises of transformers
having 2 or more axially separated winding placed on a single
magnetic limb. These are used to inter connect different circuits.
Axially-split transfonners help to introduce certain impedance
between windings, to limit short circuit currents. 'Such inter
connectivity is possible with use of more than one transfonner
also. Use of more number of transfonners increases the cost,
losses and space requirement.
3 winding transformen
In a conventional 3 winding transformer the impedance
characteristics can be represented simply and completely by
assuming each circuit has individual leakage impedance. This
can be visualized, applied and justified by the equivalent circuit.
Terminal 2
Terminal 1
Terminal 3
The above diagram represents equivalent net work of 3 winding
transformer.Z 1 represents impedance of winding I ,Z2
represents impedance of winding 2 and Z3 impedance winding
It is evident from above circuit that, when loading a pair of
winding there will be no current flow in other winding. 1be
arrow in fig 2 shows the flow of current during short circuit
test, for impedance verification.
3 winding Transformer - Axially-split
. But in the case of axially separated transformer with 2 HV
windings connected in parallel and 2 LV windings, axially-split
and connected to separate circuits, impedance between 2 LVs
is affected by third winding.

.. 1T'-- __ a
Fig 3
Z 1. Z2 & Z3 are impedances between each pair of windings
with out considering the effect of third winding. During short
circuit test of LVI-LV2 pair, current flows through HV also.
Even though impedance calculated between LV I and LV2 is
high, effective impedance is less than impedances of HV-LVI
and HV-LV2 added together.
Fig 4
Calculation effective impedance
Zl is the impedance between HV-A and LVI
Z2 is the impedance between HV-B and LV2
Z3 is the impedance between LVI and LV2
All these impedances are on common MVA and with out
considering the effect of other windings.
Effective impedance is calculated as below
HV-LV1+LV2 - Zl parallel Z2
= (Z1+Z2) parallel Z3
4 winding Transformers -Axially-split
In a three winding transfonner the impedance between HV-LV I
and HV-LV2 is not affected by other winding. But in the case of
4 winding axially-split winding (see fig) impedance between
any pair of windings is affected by other windings.
Fig 5
Paths of current flow during short circuit test is shown in fig6
Fig 6
Effective impedance is calculated as below
HV-LVI = ZI parellelZ6 +Z2
= ZI*(Z6+Z2)
HV-LV2 = Z3+Z4+Z5
LVI-TRY =Zl parallel Z2
LV2-TRY =Z3 parallel Z4
LVI-LV2 = Z7 parallel (Zl +Z4) parallel (Z2+Z3)
Because ofthe effect of other windings in effective impedance,
designing a 4 winding axially-split winding is real challenge to
designers. Following methods are useful to achieve reasonable
I. Keeping both Tertiary winding with out connecting in
2. Increasing tertiary inductance by special winding
3. Use of reactors to increase impedance of tertiary.
1. Transfonner Engineering - Mr S.V. Kulkarni, Mr. S. A.
2. Transfonner Engineering
(Impedance Characteristics of Multi circuit transforrners)- Mr.
Abstract There is switching of DTs in India causing frequent
inrush current and cold load pick up problem. In this paper a
method is presented to determine the probability of occurrence
of inrush currents and loss of life of DTs due to inrush current
and cold load pick up. Higher loss of life during inrush for
aluminum wound DTs is expected due to poor creep behaviour
of aluminum. This is evidentfrom the experience li'ith aluminium
and copper wound DTs. Results for 25, 63 and 100 KVA
aluminium wound DTs are discussed.
Keywords Distribution Transfonner, Inrush current, Cold load
pick up (CLPU), Hot spot temperature (HST), Loss of life,
Mechanical stresses, Creep.
1. Introduction
Restoration of power distribution network after a prolong
interruption, produce a load demand significantly higher in
magnitUde and different behavior than pre-interruption
condition. It can be categorized in four phases based on
magnitude and nature of the current. These phases are inrush,
motor starting, motor running, and enduring phase. In first inrush
phase the current magnitude rise up to 10 to 15 times the pre-
outage value for few cycles. It is mainly because of
magnetization ofDTs; however, the starting current of cold lamp
filament also has minor effect. In the second phase, the
magnitude of the current increases up to 6 times the nomlal
value due staring current of motors and this phase sustain for
about a second. The third phase causes because of the increased
current required for acceleration of motors and it lasts for about
15 seconds. These three phases may cause excessive mechanical
stress to degrade the winding insulation. The fourth phase is
due to the loss of diversity among the control led
loads and it continues until the nonnal diversity amongst the
loads is restored. The duration varies from few minutes to hours
depending upon various factors such as weather, nature ofload,
duration of interruption etc. The load current magnitude in this
enduring phase may be nearly 2 times of the nonnal load[7].
This phase causes high HST in the windings of the transfonners
In India, yearly failure rate of DTs are in the range of 15% to
25% as against 3 to 5% in Western countries. The major reasons
for the failure of DTs are thermal, electrical and mechanical
stresses on transformer winding insulation. The life of DTs is
nonnally gauged from the life of its insulation in proximity of
Hot Spot Temperature (H ST) . The insulation deterioration due'
to thennal stresses is slow but continuous [1-3] and may be
high during CLPU(7], due to electrical stresses the deterioration
occurs where field intensity becomes more than a particular
value and it is very fast [1,2]. Mechanical stresses due to short
circuit and switching inrush current playa major role in the
deterioration of transformer winding insulation.
. and thereby accelerated loss of life of the winding insulation.
In India most of the utilities have frequent tripping due to
demand exceeding generation, and poorly maintained
distribution lines. Thus, repeated switching of the transformers
take place causing CLPU and the transient magnetic inrush
current. This inrush current may be as high as 10 times or more
of rated current and persists only for about one half ofa second.
Inrush current peak of equal or more than 60% of the rated
short circuit current cause local forces of the order of magnitude
as those during short circuit and hence similar damage to the
winding that may be due to short circuit [4]. This current
produces excessive mechanical stress and intense localized
heating which may be sufficient to cause elongation of aluminum
conductor may be due to creep. Inrush currents which produce
forces in order of the short circuit forces cause large damage on
the winding as it occurs more frequently than that of the
short circuit current (as switching is more in India) and with a
significant longer duration [5]. Thus at every occurrence of
the inrush, there is some degradation of aiuminum conductor
and its insulation. After many such inrushes local heating may
take place in the winding causing ultimate failure of the
CLPU problem is being observed by Indian utilities in some
The 1iterature survey revealed that, not much attention.has been
paid to estimate the probability of failure of DTs subjected to
very frequent switching which leads to the occurrence oflnrush
currents and CLPU problem. It is envisaged that excessive
failure rate of aluminum wound DTs may be due to this
phenomenon. Also CLPU, thennally degrade the insulation of
the windings. It is therefore, in this paper attempts were made
to evaluate the effect of inrush current and CLPU on the failure
tendency of DTs. Other reasons of failures ofDTs are discussed
in detail in the earlier publications of Gupta [8].
2. Computation of Loss of Life
The loss oflife estimation depends on the parametirs affecting
the service life. Transfonner operation at high temperature
causes reduced mechanical strength of both conductor and
structural insulation. These effects are of major concern during
periods of transient over current like inrush conditions or short
circuit conditions (through-fault) when mechanical forces reach
their highest levels and occurrence CLPU problem. An algorithm
has been presented in the following steps for computing the
loss oflife ofDTs of the given rating due to Inrush current and
2.1 Computation or loss or lire due to Inrusb Current
I. Inrusb Current
Poor generation and load shedding cause repeated tripping of
poorly maintained distribution lines. Thus repeated switching
of transformers takes place. Every time when a transformer is
energized a transient current much larger than the rated
transformer current can flow for several cycles. This is caused
because the transformer will always have some residual flux
density and when the transformer is re-energized the incoming
flux will add to the already existing flux which will cause the
transformer to move into saturation. This transient current may
be as high as ten times the full- load current and persists for a
short duration (about one-half second). This transient current
is referred to as Inrush Current. Though they persist for a short
duration but repeated currents causes intense localized heating
which will cause increase in HST of the windings and will lead
to damage of insulation, production of excessive mechanical
stresses, and voltage drop at the consumer's end.
The peak. value of the inrush current [6] can be estimated under
saturated conditions by the equation (J)
i = J2v (2Bmcosa + Bns -B.wl )
rIIllLt.J 1 1 B (I)
( (oL
) + Rck m
v = Applied Voltage (Volts]
Lair = Air core inductance of the DT [Henry]
Rdc: = DC resistance of the transformer windings [Ohms]
Bm = peak value of induction in iron at the moment of
switching [Tesla=Vs/m2]
8,... =:: saturation flux density of core material [Tesla]
(B,... =0.83* 8.)
B ... = saturation flux density of core material [Tesla]
(B ... =1.125* B.)
a is the switching angle [degrees]
II. Algoritbm
Step l:j:omputation orsbort eireuiteurrents and Tbresbold
Compute the short circuit current isc for DTs of the given rating
using (2) and threshold value i'h using (3)
i'h = 0.6 x i.<<" (3)
kli m Asymmetric factor whose value is decided based
XlRratio with reference to IS:2026, Part- J -1977 Clause 16.11.2
i",.... = Phase current of the transformer [Amp]
Zpu = per unit Impedance
tban tbresbold value i'h
a) For B, (residual flux) values varying between 0 to nB_
(n is the maximum percentage value) in steps ofm repeat
the following steps:
i) Compute peak value of inrush current, for each switching
angle (0 to 360 deg) using
ii) Compute the current ratio (peaks of inrush current to the
short circuit current) for each switching angle.
current ratio( a) = irlllaJ a)
Where a =:: Switching angle [O:s; a :s; 360 ]
iii) Find the number of switching angles, for which the inrush
current is greater than the threshold value i
Let r = Total
no. of angles at which irmu.r ;;:: ith
iv) Compute the probability of switch on causing deterioration
of windings (Pd) the residual flux Br
Pd(B )=-
r 360
b) Calculate the probability of switch on causing irma. e-i
> . _ L pdanger
P{i_ - I ... ) - =----
Step 3: Com.,utation of Mechanical stresses
When the inrush current becomes equal or greater than the
threshold value, the mechanical stresses developed due to inrush
current are same as that during the short circuit current. The
mechanical stresses for two winding, three phase transformer
having a core type construction and concentric winding with
tapings placed within the body of the outer winding is calculated
a) Calculate Hoop stress ( (jm('Ul1 )
Hoop stress is a mechanical stress acting circumferentially
(perpendicular both to the axis and to the radius).
Hoop stress (j meul/ ) = K [kg/cm
11' I'll
K(Al) = O.02(k.fi)2
h. = winding height [mm]
The value of kFl is obtained from step (I). The
value of hoop stress shall be less thait 700 kglcm
for aluminum.
b) Calculate Radial Bursting Force (F)
These forces tend to increase the distance between the windings
and causes damage to the windings.
2trO"m,m,I phN
Radial bursting force ( ) = [kg]
t5 = Current density [Amps/mm2]
c) Calculate internal axial force (F.):
Axial forces bend the winding turns in a vertical direction. These
forces also increase the pressure acting on spacers between coils.
F = (_) 348
(" Z h [Kg]
1'1/ IV
Sn = Rating of the transformer [KVA]
The negative sign indicates that force is acting towards the
d) Calculate maximum compressive pressure in the radial
spacers (P)
The one-third ofthe internal axial force (F ) exerts on the outer
windings of DTs. Therefore the force on the
windings is
= -- [kg/cm
A=Total supported area of the radial spacer [cml]
e) Calculate the Resultant force on the conductor just belowl
above the spacers
The maximum value of the hoop stress is two times ofthe mean
value and it is on the inner layer of the outer windings. Therefore,
the resultant stress on the conductor just below/above the radial
spacer of the inner layer of most stressed coil is
Resultant Stress in kg/cm2 = 2 x O"III(OUII.' + P + 40
Where 40 Kg/cm2 is the tightening force taken into account.
Step 4: Computation or Loss or life (LOL) per switching
Compute loss oflife per switching taking into account the failure
statistics of the DTs.
" S' h' Yearly Failure Rate on accounl o/Switching
L per Wile mg
No. 0/ Switch on pq Year
Step 5: Computation of loss of life due to inrush greater
. .." ..
UlillI un: LlUt:MIUIU value
Calculate the loss of life due to inrush current using step (2)
and step (4)
LOL due to Inrush Current = LOL per Switching
P( i
HI. Results
For analysis putpose, data from four aluminium wound DTs of
Uttar Pradesh and Punjab with rating 1I000/433V, 100kVA,
63kVA, 25kVA is collected and is given in Appendix. The
maximum value of the residual flux density for CRGO grade
M4 used in the considered DTs is 83% of the maximum flux
density. Consider the residual flux density values varying from
o to 0.83B
(n=0.83) in steps of .01 (m==84) and various
switching ranging from 0 to 360 degrees, the peak value of
inrush current is calculated using step 2 for the considered DTs.
The calculated maximum inrush current (peaks) at zero
switching angle with different residual flux density is shown in
Fig.l. for all considered DTs.
.. 25 kVA (UP) .,'
., t I .- - 25 kVA (PUN)
; I
60 -I" lOOkVA(UP) i
! 30
: :l- ........ __ ........... ...
o U I 1.5
Residual Aux Density. Brs (lesla)
Figure 1. Peak Inrush Current (irmax) vs. Residual Flux Density
Figure I shows that for low rating DTs, the probability of
occurrence of inrush current greater than the threshold value is
at every residual flux density compared to the high rating DTs.
Therefore, for low rating DTs, if the number of switching per
day is more at zero switching angles, the occurrence of inrush
current peaks is not safe for the DT for all values of residual
flux density.
The probability of switch on causing deterioration of the
winding, for different current ratios (irma. lise) as explained in
step 4 is calculated. The calculated probability with residual
flux density is shown in Figure 2, 3, 4 & 5 for the considered
DTs. It can be clearly seen from the figures that for low rating
DTs, the danger of the deterioration of the windings is more
than for high' rating DTs. As the rating of the DT is increasing,
the effect of the inrush current is decreasing. This may be the
reason for the failure oflow rating DTs is more than high rating
.r: T . ---- ----:--------r-----rl
"c I' i, c.-t 11110>0.6 i f2-S-KVA (Upj-: I
'-. c.-tralIo>O.7! ---- - - ........... --:
. -_ .. c .... rallo>O.8 \
. .. "."
l ...... .
I -
tn 0.1;-
>. "

D- 0-------..1.------'------'- _______ --' _______ _______ L-
o 0.2 04 06 08 1 1.2 1.4
Residual Fu Density. Btw (Testa)
Figure 2. P(ifN&)C>i'h) vs. B, for 25kva. DT (UP)
.c:; o'3i------:----------.----- -:-------- ---, -
! ralKl>O.6

.. 0.25" . - - curra1l81io>O.7 .
,g I C\lI'IIR
co '
. 02L


.0 '
.. '
" ..... .
0----- -------.:.--- ----
___________ " .+_ L ____ " ____ , __
o 0_2 0.4 0.6 0,8
1.2 1.4
ResiciJal FIuJr. Density. Bres(Tesla)
Figure 3 .. P(irmax>i
) vs. Br for 25kva DT (Punjab)
0.35----- ---------,-- --- --- ------
: 0_3

Q 0.1:
t 0.00
0... 0
curren ratio>O.6
, curren ratio>O.7,
63 KVA(UP)
0.2 0.4 0.6 0_8
Residual FiuK Dersity. 8'es(T asia)
... 4 (
t t
.' .. ,
1.2 1.4
Figure 4 .. P{irmax>i
) vs. B, for 63kva DT (UP)
ClliTefl ralio>O.6
, cUtYert ratio>O_7:
---- .- _._- ---- -- ------- ,-
.. '
" .
a.. 0 __________ .J __________________ __ j ______ ______ l.. ______ __ __ __'____;
o 0.2 0.4 0.6 0.8 I 1.2 14
Residual Fu Bres(Tasla)
Figure 5 .. P(irmax>i.
) vs. Br for IOOkva DT (UP)
Now assuming one switching per day and failure rate of
aluminum wound DTs as 15% on account of inrush, the failure
probability of DT per switching is found to be .000411. The
probability of dangerous switch on causing i
i'/I is
calculated using step 2 for DTs considered and is shown in
Table I .
The loss oflife ofDTs due to the peak inrush current is calculated
developed due to short circuit current have been calculated.
These stresses are the same when the inrush current becomes
greater than the 60010 of the short circuit current. The calculated
mechanical stresses developed on the primary windings for the
considered DTs is given in Table 3 and Table 4 respectively.
Table 1: Probability of switch on causing irmax>=i'h
Particulars Probability of switch on
(KVA) causing irmax>=i
i Ii
rmax 51:
>=0.6 i Ii >=0.7
nnax 5e
lise >=0.8
UP 100 0.0768 0.0296 0.0012
63 0.1007 0.0507 0.014
25 0.1253 0.0739 0.0329
PUN 25 0.0984 0.0486 0.0125
Table 2: Loss oflife of DTs due to inrush current
Particulars Loss of life of DTs
lise >=0.6 i Ii >=0.7
rmaJl $C
i Ii >=0.8
nnu 5e
UP 100 0.0054 0.0139 0.3425
63 0.0041 0.0081 0.0293
25 0.0027 0.0027 0.0039
PUN 25 0.0023 0.0033 0.0051
Table 3: Calculated mechanical stresses during short
circuit on HV windings ot DTs
Particular UP PUN
KVA 100 63 25 25
Bmall (Tesla) 1.685 1.69 1.647 1.585
amen (kglcm
) 1l0.8 77.7 40.67 43.84
Fr(kg) 53466.6 33214.4 10464.2 11417.98
Fc;(kg) 1757.28 1175 517.3 556.3
P (kg/cm2) 9.039 6.089 3.454 3.419
Table 4: Calculated stress on the most stressed conductor of
HV windings of DTs
Particula UP PUN
rsStress (Kglcm2) 100 63 25 25
Hoop Stress 110.8 77.7 40.67 43.84
Compression Stress 9.039 6.089 3.454 3.419
Resultant Stress 270.64 201.49 124.79 131.1
2.2 Computation of loss of life due to Cold load pick up
CLPU is the most severe non fault condition faced by the
utilities. This problem occurs where the power outages are very
frequent. The thermostatically controlled loads such as electric
space cooling/heating devices and industrial illumination are
major factors for CLPU problem. Fig.6 shows typical load
during CLPU. During the time J1T, the ioad increases to many
times the normal load because of the loss of diversity during
this time .. The later phase shows the enduring behaviour and
takes hours to settle and the steady state phase.
'41 .,
o Time (minutes)
Fig. 6 Load during CLPU
The load during period J1T is about 1.5 to 2.5 times of the
normal load in the developed countries due to large
thermostatically controlled load and the load are switched on
in a controlled manner to reduced excessive loading on the
transformer during CLPU. For Indian environment; this may
be taken as 1.5 to 2 times of the normal load but there is no
provision to switch on the loads in a controlled manner to
overcome CLPU problem. The CLPU problem occur very
frequent-ly in Indian utilities causing excessive degradation of
the winding insulation near winding HST.
II. Algorithm
With typical values of the loadings on the DTs , HST and loss
of life during CLPU is calculated by the procedure presented
as below:
i) Compute Factor for load loss (k) =
Load on the Transformer
( Rated Load on the Transformer )2
ii) Total loss at rated load = no load loss + load loss
iii) Total loss during time J1T = (kx load loss) + no load loss
iv) Compute L=
( Total loss during time l:!.T )
Rated total loss of the Tramformer
v) The wmding temperature rise at end of l:!.T minutes after
switch on, Or = (1- e -;,:-) x x L
Where Til' = Thermal time constant of windings
0lt. = Winding temperature rise of rated load
vi) Winding Temperature 0wr = 8
+ 8 amb
Where 8
= Ambient temperature
vii) Calculated HST at time fi
::::; 1.2 x 8
+ 8
viii) HST at rated load, 0HOl' ::::; 1.2 x 8
+ 8
Where a = 0.0865, to = 24 x 60 minutes
III. Results
Considering the load during CLPU fourth phase time of 30
minutes (l:!.T =30 minutes) as 2 times (i.e n=2), the average
winding temperature rise ( 0" ) as 50C and ambient temperature
as 35C, HST is calculated at the end of time l:!.T and Rate of
loss oflife and is given in Table 5. The thermal time constant of
the windings is taken as 15 minute. The loss oflife during th is
period l:!.T is equivalent 4.6 to 5.76 days and HST may touch
as high as 160Uc. Due this phenomenon high failure rate DTs
has been reported in some areas of India and likely to be more
severe in due course of time.
Table 5: Calculated loss of life due to CLPU problem
Total loss Total loss
loss of at
at rated during Time life ri
Load (W) tiT (W) in days
100 1940 4685.6 160 5.76
63 1415 3341.6 157.4 4.6
25 785 1853.6 157.4
4.6 .
25 770 1838.6 158.8 5.2
3. Conclusion
An algorithm has been presented for computing the probability
of occurrence of inrush current greater than threshold value and
loss oflife ofDTs due to inrush current and CLPU. The life of
the low rating DTs are found to be more affected by the inrush
current compared to the high rating DTs. With the increase in
the KVA rating ofDTs, the effect of the inrush current reduces
because. of the nearly same residual flux for all ratings DTs.
The creep behavior of aluminum is very poor in comparison to
copper. The use aluminum conductor for DTs subjected to
frequent switching is not recommended; however for low loss
transformers aluminum conductor may be used with lower
current density ( say <1.1 Almm
The CLPU is another severe problem for Indian environment
with increasing thermostatically controlled load. It causes
accelerated degradation of winding insulation. This is more
severe for small DTs having lower thermal time constant. To
overcome CLPU problem a proper restoration scheme [7] along
with higher capacity DTs are to be used. Further investigation
from field data will be of great help to device optimal strategies
that may reduce the life of the transformers.
[I] Rengarajan, S., et al., "Behaviour of High Voltage Machine
Insulation System in the Presence of Thermal & Electrical
Stresses," IEEE Trans. on Electrical Insulation, vol. EI-
20, no. I Feb. 1985.
[2] Ramu, T.S., "On the Estimation of Life of Power Apparatus
Insulation under Combined Electrical & Thermal Stresses",
IEEE Trans. on Electrical Insulation, vol. EI-20, no. I,
Feb. 1985.
[3] Desai, B.T., Nair, N.C., and Gupta, H.O., "Inconsistency
In Standards For Exact On Line Over Load Time-Limit
Computation Based On Hot Spot Temperature Insulation
Deterioration", IEEMA Journal, pp. 27-32, July 1993.
[4] Michael Steurer and Klaus Frohlich, "The Impact oflnrush
Currents on the Mechanical Stress of High Voltage Power
Transformer Coils", IEEE Transactions on Power Delivery,
vol. 17, no. I, Jan 2002.
[5] Laszlo Prikler, Gyorgy Banfai, Gabor Ban, Peter Becker,
" Reducing the magnetizing inrush current by means of
controlled energization and de-energization oflarge power
transformers", Electric Power Systems Research, no.76,
pp 642-649,2006.
[6] Ramsis Girgis, Ed teNyenhuis, "Characteristics of Inrush
. Current of Present Designs of Power Transformers", IEEE
Power Engineering SOCiety General Meeting, pp. 1-6, June
[7] Kumar, Vishal, "Power Distribution System Operation
under Cold Load Pickup", Ph.D thesis, Department of
Electric Engineering, liidiiiii vi
Roorkee, India, Dec. 2006.
[8] Gupta, H.O, "Transformer Reliability", Eleclricallndia,
vol. 47, pp. 118-123, Aug. 2007.
[9] Gupta H.O., 'H.O. Gupta, Design and Manufacturing of
Reliable Distribution Transformers CBIP
Seminar on "Distribution Transformers - Failures and New
Developments, Lucknow, Feb. 6-7,1997.
Transformer Data Core design parameters
Rated Power:l00kVA flux Densilty Bm(T): 1.685
No Load voltage: 11000/433 B, =O to .83 ofB
Mean dia (mm): 213.25 B =1.125*B
sat m
No ofturns:3344 Core Dia(mm): 110.909
DC load loss (W):920 Window height (mm):470
Rated Power:63kVA
No Load voltage: 11000/433 flux Densilty Bm(T): 1.69
Mean dia (mm): 195 Core Dia(mm): 97.1288
No ofturns:4356 Window height (mm):445
DC load loss (W):640
Rated Power:25kVA flux Densilty Bm(T): 1.647
No Load voltage: I 1000/433 Core Dia(mm):76.8505
Mean dia (mm): 160 Window height (mm):405
No of turns: 72 16
DC load loss (W):3 77
Rated Power:25kVA flux Densilty Bm(T): 1.585
No Load voltage: 11000/433 Core Dia(mm):77.7817
Mean dia (mm): 159.25 Window height (mm):380
No ofturns:7304
DC load loss (W):380

Indigenous development of Cast Resin Multi-secondary
1600kVA Transformer for Regulated High Voltage Power
Supply (RHVPS).
Regulated High Voltage Power Supplies (RHVPS) are
commonly used in high-energy particle accelerators. RHVPS
is a modular power supply in which kV level modules (40 or
80*) are cascaded to generate desired level of voltage/power
[I]. One of the most challenging tasks involved is to provide
input power to number of rectifier modules with required
isolation (inter-winding and winding to ground). This is
accomplished by deploying multi-secondary (large numbers, say
40 secondaries) transformers. This RHVPS concept was
realized for the first time in the country with development of
oil filled multi-secondary transformer. A pair of3.3MVA, 11 kV/
(940Vx40) has been successfully demonstrated, isolation of
6kVDC (inter-winding) and 160kVDC (all secondery tl) ground)
The next generation power supplies are unitized with indoor
installations. This has created the demand for dry type multi-
secondary transformers in compliance with safety regulations.
This paper presents FEM analysis, manufacturing issues and
testing of the prototype resin cast coil. On the manufactured
prototype, inter-winding isolation is tested up to 6kVDC and
125kVDC with respect to ground. Following this activity full-
scale manufacturing of the 1600kVA, llkV/ (1.lkVx 40)
transformers are initiated.
KEYWORDS: RHVPS, multi-secondary, inter-winding, CRT,
High-energy particle accelerators are modulated by high
voltages/power. Certain uses like Neutral beam heating of
Plasma in TOKAMAK require high power. A typical Neutral
beam has few tens of megawatt power. Typical example is a 5
MW neutral beam injector at the Institute for Plasma Research
(IPR). Ion source provides H+ ions, which are accelerated at
55kV. The new age particle accelerators may require millions
of volts for atomic particle acceleration. To meet the demand,
modular power supplies are best suited based upon the facts
that they are easy to m,aintain, well-controlled, regulated, can
have sharp rise and fall time and have better serviceability.
The chosen topology for such applications has introduced
multisecondary transformers (figure 1 a & 1 b). The design of
these transformers accommodates the need to feed modules
individually in isolation mode. Earlier, the capability has been
demonstrated in oil-cooled transformers successfully and an
attempt has been made in Resin Cast [2], [3]. Dry type multi-
secondary transformer is proposed under an MoU between IPR
and BARC for development ofa lOOkV, 25A RHVPS. Unlike
oil filled transformers, dry type transformers are manufactured
up to 33kV voltage class only. This has posed the limitation on
the achievable kVDC isolation and resin system to support the
same. Also to assure the manufacturing feasibility, a prototype
resin cast coil, section of LV cast with four coils having radial
dimensions same as with actual transformers (1600kVA, llkV/
1.1x40), was proposed. By use of FEM analysis, size &
clearances for the winding conductor in epoxy dielectric is
optimized to limit maximum electric stress - 6k V /mm.
Figure la Topology of RHVPS.
Figure Ib Pair of 3.3MVA, llkV/40X 940 multi secondary
transformers commissioned at IPR
Design of a prototype was motivated due to the fact that this is
a first ever attempt by industry to go for such non-standard cast
resin transformer (CRT) with multisecondary design and high
voltage requirements. Based on available data and FEM
analysis, dimensions. size and clearances were chosen. A one-
tenth section (4 LV coils) of LV of actual transformer (l600kVA,
1 IkV/1.1kVx40) design was preferred to cast as prototype. The
HV coil (llkV class) is standard practice for industry and has
not been considered as part of the prototyping.
A model has been developed to analyze the electrical stresses
in and out of the resin medium. A cylindrical electrostatic shield
to reduce corona in intercoil region is provided between LV
and HV coils. Simulation has been done for different thicknesses
of resin at radial/top/bottom regions. Each winding of the coil
is subjected to DC level by virtue of power supply topology.
The last coil in the series is subjected to the highest voltage. In
order to reduce the thickness of resin at top and bottom, dummy
coils of larger diameter are used to shield top and bottom coils
electro-statically. The schematic is shown in the figure 2.
Figure 2 Schematic shows the internal arrangement of coils.
ANSYS mUltiphysics has been used for FEM analysis. Due to
the symmetry of coil distribution, only the end conductor of the
winding, shielding conductor aed electrostatic shield/screen
were chosen for modeling. The FEM results show that stress
levels have decreased from 6.2kV/mm to 4.SkV/mm with
dummy coils at lOOkV as shown in figure 3 &4. Simulation
shows a reduction from 12.4kV/mm to 9kV/mm at higher
voltage of200kVDC. Based upon these results, the dimensions
are chosen as inputs for the design.
Figure 3: Simulation results without dummy coil (max. stress
Figure 4: Simulation results with dummy coil (max stress 4.5kVi
Industry procedure was followed for the manufacturing of
prototype. The coils were assembled in moulds and put into the
vacuum chamber. Huntsman make resin (CY 205/HY 905),
McLube make releaser (MA2021), 300-mesh size of silica and
Binani n:aake copper strands were used in the cast. The
manufacturer, based upon his experience, chose the hardener
(DY061), plasticiser (DY040) and their mixing ratio. The resin
mixture was poured into the mould in fluid form. Curing &
post curing followed as per the recommended procedures. The
finished product was examined mechanically as well as
This cast was first of its kind to be manufactured. The cast did
not show a good insulation resistance (IR) value on megger
test. Skewing of screen was found in the cast. Also, development
of cracks along the periphery was observed during the post-
curing phase (figure Sa). These issues were not expected, as the
manufacturer never observed them before. One of the reasons
envisaged was the mass involved in the cast. The OD nOmm
and ID 460 mm of the coil cast gave resin thickness of 155 mm,
which is bulky. As stated earlier in FEM analysis that 60 mm
thickness of resin is required at the top and bottom and inside
periphery of the coil. In order to deal with such issues, following
arrangements were made during next assembly.
I. Introduction ofFRP isolators at ID side of copper coils
and at the ID of the coil cast.
2. Placing of screen over FRP isolator.
3. Use of fibreglass matt in bulk resin region.
After these incorporations in assembly, the second cast was
produced. The megger value increased to the desired level and
no skewing was observed. The whole cast was found intact. A
few lines of cracks were observed (figure 5c). The cast was
subjected to electrical testing and excellent results were
Figure 5: (a) Picture after casting offiest coil. Skewing of screen
and crack lines are seen along the inner periphery. (b) Assembly
before casting (c) The final cast with dimensions.
High voltage withstand test
HV injection all terminals shorted
Testing of coil with HV
Figure 6: Testing arrangement for the cast.
Testing terminals
All coil terminations are shorted and HV is applied between
screen and shorted terminations.
Before HV application, all the adjacent winding terminations
withstood megger test for Imin.
IR value> 2000MOhms @5kVDC
HVAC test at site
S.No Applied Duration of Remarks
voltage (kV) application
I. 45 60 sec No Corona
2. 64 60 sec No Corona
3. 75 60 sec Initiation of
audible Corona
4. 86 60 sec -do-
S. 93 60 sec -do-
6. 97 60 sec -do-
7. 99 60 sec -do-
8. 100 60 sec -do-
9. 102 60 sec -do-
lO. 105 60 sec -do-
ll. llO 60sec -do-
HVDC lu-bouse test
S.No Applied Leakage current Remarks
voltage( kV) (micro Amp)
1. 60 not readable No corona
2. 80 Not readable Initiation of
audible corona
3. 90 100 -do-
4. 100 <200 -do-
5. 110 <200 -do-
6. 120 <200 -do-
7. 125 <200 -do-
Duration of application at each voltage step was 60sec
*The measurement scale does not show below 100 microamps
and the scale is in steps of 100 microamps.
All the readings were between 100 and 200 microamps.
After HV application, all the adjacent winding tenninations
withstood megger test for lmin.
IR value> 2000MOhms @5kVDC
CRT for high voltage application -lOOkV or beyond with
muItisecondary features call for investigation in the area of bulk
resin castings. thermal management and advance approach in
the dielectrics to make the overall equipment compact and cost
effective. The prototype development has demonstrated the
capability of manufacturing but with limitations in database and
experience. The lack of database in areas like FRP interaction
with resilly Cross linking pbenomena in bulk resin, curing.
mixiDg ratio & composition of resin for bulky/compact design,
in view of HV applications, open up investigations for future
Authors are thankful to M/s. Ames Impex Pvt. Ltd. for
mauufacturing the prototype resin cast coil.
1. "High-voltage, high-power, pulse-step modulators for the
accurate supply of gyrotron and other beating devices"
W. Forster, J. Alex, 25th International Power Modulator
Symposium, 30 June - 3 July 2002.
2. "Studies on the behavior of multisecondary transformers
used for regulated HV power supplies"
N P Singh. UK. Banwh. PJ. Patel. S. K. Mattoo and NBI
Team. Institute for Plasma Research. Bhat.
Flision Engineering and Design 75-79(2005) 127-13
3. "Development of Multisecondary Transformers for
regulated HV Power Supplies"
N P Singh. U.K. Baruah. P.J. Patel. L N Gupta and S. K.
Mattoo, Institute for Plasma Research. Bhat, Gandhinagar
J U Mamtora, Transformers & Rectifiers (India) Ltd.,
IEEMA JOllma174-75 August 2006
12th November 2008 (2
Timing: 11.30 AM to 1.00 PM
A Low Intervention Strategy for the Removal or Mr. Ian Bowden
Maintenance of Low Levels of Moisture in Managing Director
Power Transformers. Transec (Uk) Ltd ., 26 Great Queen St.
London WC2B 5BB
Overhauling of Transformer With Mr. N. S. Murthy
Manufacturing Know-How Technical Director
Pan-Electro Technic Enterprises Pvt. Ltd
Hyderabad - 500 018.
Condition Assessment of Solid Insulation in Mr. V. V. Pattanshetti,
Power Transformers By Transformer Oil Analysis. Mrs. S. Vijaya Kumari, Mr. K. Dwarakanath,
Central Power Research Institute,
Bangalore- 560080.
Regeneration of Contaminated Oils Mr. Prashantha S. K,
Mr. Narasimhan S. R.
Manager - Marketing
Fowler Westrup (India) Pvt. Limited
Bangalore 560 099
Life Extention of Power Transformers Mr. J. S. Batra,
Director IP & C,
Bhakra Beas Management Board,
Chandigarh.- 160019
Failure of LCTS In Northern Grid - A Case Study Mr. K. K. Arya,
Mr. S. P. Singh Gaharwar,
Mr. R. P.Aggarwal,
Mr. Vikram Singh
Central Electricity Authority,
New Delhi- 110 066
Failures of on Load Tap Changers (OLTC) Mr. R. G.Yadav, ED (OS),
in Power Transformers in Powergrid Mr. G. S. Sarkar, AGM (OS),
Mr. V. K. Bhaskar, Chief Manager (OS),
Mr. Gunjan Agrawal, Dy. Mgr (OS)
Power Grid Corporation of-India Ltd.
Gurgaon - 12200 I
CHAIRED BY : Mr. Uttam Khobragade. General Manager, BEST, Mumbai
REPORTARf BY : Mr. Sa if Qureishi, Mis KRYFS Power Components Ltd., Mumbai
~ ~
Ian Bowden-
Power Transformers constitute a high capital invesbnent within
the electricity generation, transmission, and distribution sectors,
as well as power intensive industries. With the high, and sharply
increasing cost of replacement, it is essential to consider methods
of life extension.
This paper will explore the use of molecular sieve technology
to process transformer oil online, continuously removing
moisture, keeping the oil in a dry condition, and extending the
life of the transformer. .
The design life of a power transformer is around 30 to 35 years.
In fact the typical 'Time to failure' of a large generator
transformer working at constant full load is 18 to 24 years, and
a transmission or distribution transformer working at half load
can be 40 to 60 years. { I }
Fault statistics over a range of transformers indicate the classic
'bath-tub' curve, {2) where, following initial commissioning,
the majority go through to experience failure in the predicted
elld of life rise of the curve. However, what is also noticed is
the period of time between the initial fall in the bath-tub curve
before the 'towards end of life' rise in the modern transformer
appears to be shorter than with older generations of transformers.
Whether this is due to cost pressures leading to the abandonment
of what might be considered 'overengineering' now, but was
standard practice in the last quarter of the 20th century, is
difficult to say. Another explanation could well be a completely
different way of using these expensive assets, for example 'Two-
shifting', [G S U Transformers]; the adoption of single rather
than duplex transformers in Distribution Transformer
configurations; ere-rating' by, for example adding additional
cooling capacity or fans; monitoring devices which will predict
hot -spot temperatures, and allow transformers to run for limited
periods above their plated rating. {3}
All these and, no doubt, other pressures, can be justified possibly,
by a financial analysis in terms of an immediate return. To take
the last two examples, a relatively inexpensive addition may
avoid pUrchasing a new transformer with the associated capital
The Ageing process.
It must be understood that paper insulation acts like a sponge
and will adsorb large volumes of water. It is not uncommon to
find 99% of water in a power transfornler will be contained
within the paper cellulose, in windings, connections, pressboard
spacers etc., but less that 1 % will be dissolved into the insulating
oil. {4} When a sample of oil is taken for analysis, because of
this ratio, the level of moisture in ppm can only be indicative of
the total water content of the transfornler. A number of methods
can be used to establish the relationship between the dissolved
water in oil, and the volume of water contained in the solid
insulation, presented as a percentage of moisture by weight of
insulation ..
The paper insulation and oil within most transformers are
manufactured from organic materials, which will deteriorate
during their lifetime. Insulating oil is a renewable commodity,
and can be changed, although only with caution. New oil is less
hydroscopic than aged oil, and will not absorb the same quantity
of water; this may lead to free unabsorbed water in the tank
with the inherent danger of flashover. Paper insulation is not
renewable. Once degradation of the cellulose insulation has
advanced to a point where mechanical strength is lost, effectively
the transformer has reached end oflife, and will most likely fail
during the next exposure to high electrical stress, for instance,
as a result of a through fault. The catalyst for this deterioration
The Ageing Catalysts.
HEAT - Temperature Cycles - Unless redundancy has been built
into the transformer rating at the specification stage, a
transformer is always likely to operate at or near to the rated
value. As the transformer load cycles, temperature of the
insulating oil rises and falls. During this process the hydroscopic
properties change. As the temperature rises, the oil becomes
more hydroscopic, and will adsorb water made available from
. the paper insulation. The ppm value of moisture rises in a direct
relationship with the temperature, ~ l t h o u h with some time
lagging constant. Conversely, when the temperature drops, the
oil becomes less hydroscopic, and the water is absorbed back
into the paper insulation. This is clearly demonstrated in the
graph FIG I showing an extended period of on-line monitoring
temperature vs moisture using a Vaisala HMP228 probe in
flowing (not static) oil. It underlines the importance of giving a
top of tank temperature reference to any oil sample taken.
- Hot Spot - The hotter the top temperature the greater the effect
it has on insulation. For every six and a half degree rise in
operating temperature the paper insulation life will half, based
on a moisture content in the paper of 0.3% by weight. However,
the rate of ageing will be up to 50 times higher at moisture
content 9f 5% in paper. {I }
OXYGEN - This movement of water back and forth across the
oil to paper interface accelerates the degradation process within
the cellulose, which is primarily a long chain glucose hydro
carbon molecule. During the degradation a monomer can break
its bond with the hydro carbon reducing its length and strength.
The bond consists of Hydrogen and Carbon molecules which,
if Oxygen is available in the oil, will combine to give CO, C02
and H20. The by-product of the ageing process is therefore
MOISTURE - The relatjonship between the rate at which the
paper insulation ages and the water content of that insulation is
an exponential one. Typically for every percentage rise in the
water content of the insulation a reduction of between 2 and 4
times the life of that insulation can be expected. FIG 4.
To achieve longevity of the paper insulation system, one or all
of the above catalysts must be addressed.
HEAT - The temperature is most often dictated by operational
constraints, but care in planning, and predictive hot spot
monitoring can reduce the temperature profile, and therefore
aid the ageing process.
OXYGEN - If the oil is to be de-gassed to remove oxygen,
other gasses will also be removed, and may effect the vital tool
of Dissolved Gas Analysis. An/operator must weigh up the
balance of losing the of an accurate DGA
or gaining the benefit of the reduction of internal production of
moisture through the paper ageing process.
MOISTURE - Can be managed from the start of life of a
transformer by ensuring that water is not made available to the
paper insulation from the time of commissioning. There are a
number of mechanisms that will allow water to enter and
accumulate in the transformer. Fitting and maintain:ng (every
month) a silica gel breather will help keep the air space in the
conservator dry, and mitigate a potential source of water which
would otherwise be dissolved into the oil. Attention to careful
maintenance to ensure that gaskets are sealed. Any indication
of oil on the outside of a transformer means that moist air can
be drawn into the transformer tank, and add to the accumulation
of dissolved water. However, the single largest source of water
is as a result of organic deterioration, or ageing. .
A number of systems claim to remove water from transformers,
but it is worth considering the optimum approach for individual
transformers. For instance, a large transformer with oil volumes
exceeding 40,000 litres of oil, which shows a percentage
moisture in paper of 4 to 5% will require a different strategy to
a much smaller transformer with only 10,000 litres of water,
showing the same degree of wetness. Again, a different strategy
might be adopted for a new or mid-life transformer.
Extracting water from the oil is relatively simple, can be
done ysing heat and vacuum or oil filtration processing.
However, as noted previously, this is only addressing perhaps
as little as I % of the problem, unless the extraction plant can be
left on the transformer for many months. Very quickly, the water
that has been extracted from the oil is replaced by the water
already present in the paper insulation, in order to re-establish
the natural aquatic equilibrium.
A large transformer with high oil volumes that is wet, could
best be served by being taken out of service, and with oil
removed, Low Frequency Heating (LHF) applied to the
windings. Generating an even heat throughout the winding to
110 deg C. water can be successfully removed over a relatively
short period of time. {5}
This is of course a 'high intervention' strategy which involves
downtime and lost production.
A 'low intervention' strategy could be more appropriate for new
ttansformers, strategic spare transformers, wet transformers with
smaller oil volumes, or larger transformers that need to be
maintained in a dry state. A continuous on-line process offers a
'iow imervemiuu' Lij<1L UVl .lUlt:jlere or upset the
natural balances within a transformer, does not involve
downtime, but gradually extracts water on a continuous basis.
Background: Continuous processing of transformer oil on-line
to remove moisture is not new. In the 1960's oil filters were
developed, but largely the absorbing materials used were not
able to bond the water molecules, and once saturated, started
giving back the water to the oiI, leading to the risk of free water
within the transformer.
Only in more recent years, with the arrival of technology to
control new dehydrating materials very closely, that the process
has become viable and offers an effective solution to combat
the accumulation of water in ternls of:
o Cost, both to purchase, install, operate, and maintain.
o Performance, in terms of the efficiency of the adsorbtion
o Ease of retrofit or installation on any target transfonner.
o Ease of maintenance to exchange cylinders charged with
new adsorbate.
Process: Transformer oil can be pumped continuously over
zeolite adsorbate beads that are in such a way as to
give correct turbulent flow, without channelling, ensuring free
flow through adjacent beads, without excessive pressure drop,
to maximise the adsorbtion potential of the zeolite. The
relationship between the bead size, molecular pore size, flow
rate, and back pressure are critical factors in maximising the
efficiency of any molecular drying system.
The pore size is selected to minimise the adbortion of gases,
which generally have a greater molecular size that H2O, and
therefore cannot be adsorbed by the zeolite. Also water has a
greater molecular polarity, so it will be adsorbed preferentially
over materials with a lower polarity, which is the case for most
dissolved gases.
The use of this technology therefore has a minimum effect on
adsorbing gases, and hence interfering with the balance ofDGA.
FIG 3. gives an example ofa trace ofDGA taken at intervals on
a transformer which has a molecular sieve fitted, but experienced
two through faults in May 2001 and in January 2004. The faults
can clearly be identified in the sharp rise in Ethylene, Methane
and Ethane. At the points of regenerating the molecular sieve
cylinders there is no signiflC8Dt change in the key gases from
the point of removing saturated zeolite, to replacing directly
with new zeolite.
The Oil flow can be pre-filtered via a Pall cartridge to extract
sludge or fibres in a mid to end of life transformer, that would
otherwise block the pores of the zeolite. The oil is 'post' filtered
in any case through a 10 micron particulate filter to ensure no
contamination from the unit, or entrained within the oil, passes
back into the transformer. An advantage of this is t.'.c
of ionised particles in the oil that could cause a flashover under
an over- voltage fault, e.g, lightning strike, or switching surge.
Air traps, bleed, and non-return valves will ensure air is not
introduced into the transformer in the oil flow, and allows
effective and easy commissioning.
Monitoring: Electronic monitoring can be introduced to give
continuous reading of temperature and moisture by introducing
probes into the oil flow at the input and output points of the
molecular sieve. A 4. - 20mA or 10V analogue signal can be
brought out to SCADA, or read locally. In this way a good
indication is given of the overall wetness of the transformer,
but also the degree of saturation of the zeolite beads. Monitoring
can be extended to send an alarm on loss of oil flow, or on the
loss of power supply.
Re-genention: When the output reading of moisture matches
at the input, saturation is complete, and the cylinders can
be exchanged for replacement cylinders. The saturated cylinders
are weighed, and the contents removed. The cylinders are
cleaned, dried, and under controlled conditions are re-filled with
new zeolite beads and new transformer oil to IEC 60296. The
cylinders are re-weighed, and the difference between the before
and after weights give the weight, and therefore volume of water
removed from the transformer.
A typical standard Unit, designed to adsorb up to 13 litres of
moisture, can be manufactured and installed ( in the U K) for
less than Sk. Its running cost is whatever a continuous 230/
240 v single phase load of 1, I amps costs and there is no
maintenance cost, apart from the cost of replacing cylinder sets
from tir.ne to time.
The actual time taken to adsorb its full design capacity is almost
impossible to predict, because there are so many variables: oil
temperature, viscosity - highly dependent on temperature and
very influential on actual flow rate through the On - line Unit,
age of oil and extent of entrained fibres, etc. Records are
available of moisture adsorption values in excess of 10 litres in
some cases, in approximately 12 months' operation.
1.0. Background.
The client operates, among several other 33 k V primary voltage
Oil- filled transformers. two Rectifiers in particular, support a
commercially critical production process. These two units are
virtually identical in all respects. being to the following
Voltage Ratio: 331 0,13 kV DC output rectifier
Capacity: 6,98 MVA OFWF Continuous rating. [i.e.,
pumped oil through an oil/ water heat
exchanger, with pumped secondary.]
Load: Half -wave rectifier, in 24 bour operation.
In general, both perform similarly, except that for unknown
reasons, one of the two has tended to run slightly hotter than
the- other. Typical Top Oil temperatures are up to 53 C and
Winding temperatures indicated of up to 75 C on the warmer
of the two: the Plant to which the Molecular Sieve Unit was
installed. The secondary, i.e., water, pipe runs of the heat
exchangers are marginally different between the two, which may
affect the thermal behaviour. The moisture content of the oil
has been a concern, especially as the transformer rectifiers were
only some 14 years old at the start of this project.
2.0. Molecular Sieve Unit Installation and Monitoring
It was decided by the client to install a Molecular Sieve Unit to
one of the two Rectifiers.
'Transformer No: I', the warmer of the two was selected. The
installation and commissioning was completed on November
07 2002. A Vaisala Moisture Monitor was also installed on a
temporary basis on this Transformer and on the neighbouring
transformer, which had no Molecular Sieve Unit fitted to it.
The intention was to leave the monitors on line for approximately
2 weeks, with each of their outputs for ppm and oil temperature
connected to a 2 - channel logger, set to record measured DC
voltage at 15 minute intervals.
At the end of each fortnight ( or near to this) the monitor/logger
sets were removed from site and the stored data from the loggers
downloaded, each to produce a two - trace Chart: moisture ppm
and temperature. To calibrate the chart traces, samples of oil
were taken on each visit, as close I:j.S possible in time to the time
at which the monitors were either first switched on or, at the
end of any period, to when they were switched off. These
samples were taken at each visit, except the first when the
Molecular Sieve Unit was actually installed. In general, the
laboratory analysis results, from the Karl Fischer titration
method, were used from the first visits of any period, except
when weather conditions were judged to have contaminated the
readings, in which case the second set of sample results were
used and 'back-loaded' into the chart.
3.0. Results.
Over the 12 months, a noticeable reduction in ppm was achieved
on the transformer to which the Molecular Sieve Unit was
connected. diverging from the No: 2 transformer, for which the
moisture content remained the same. In fact, there may have
been a slight increase, from calculations using the Piper Chart,
but only of a lower order of magnitude than the experimental
tolerance to be expected with the sampling and analysis.
The results for Transfonner No: 1 have been plotted in the two
charts attached:
'Transfonner No: 1 Chart 1 Nov. '02' FIG 1 and
'Transfonner No: I Chart 4 Nov.' 03' FIG 2
By calculation, it is estimated that a volume of moisture of the
order of 8,5 litres has been adsorbed by the Molecular Sieve
Unit from the insulating oil oftransfonner No: I, with the result
that the moisture - in - paper percentage by weight has reduced
from 2,95 % to 2,15 %.
5.5. Conclusions.
Although the Dielectric strength of the oil in this Transfonner
has consistently tested at over 60 kV and there is virtually no
fibre content, a reduction in moisture such as this can only help
increase the reliability and longevity of the transfonner.
The results achieved here have been on a relatively low oil
volume - 13 000 litres - which is only some 15% of the oil
volume on a very large Generator Transfonner. However, the
oil temperature is consistently at least 50 0 C, and the higher
the temperature the more pronounced the Water activity', i.e.,
the movement of moisture between paper insulation into the oil
on temperature rise, as can be seen on the charts. This is one of
the values of monitoring, as opposed to a 'spot' reading. For a
given temperature, it is possible to obtain significantly different
sample moisture ppm results, depending on whether the sample
happened to have been drawn on a rising or falling temperature
The Molecular Sieve offers an economic 'low intervention'
strategy to the removal of water not just from the oil, but
gradually from that absorbed into the paper insulation by using
a continuous on-line filtering process.
This strategy is ideal for protecting new transformers from
accumulating water from the date of commissioning. It is also
appropriate for the maintaining of low levels of water in large
transformers, or for the lowerine of ' ~ ~ ~ ~ : : : : ~ : : : : ~ ~ ~ ... u ... : LV
medium sized transfonners.
In keeping the transfonner dry, the rate of ageing will be reduced,
the dielectric strength of the insulants maintained at a high level,
giving increased protection to life threatening events, such as
through faults. The result offers life extension to the transfonner,
with a significant impact on capital budgets.
1. Professor D.J. Allan - Removal of moisture from
transfonners in service April 2005.
2. V. Sokolov: " Understanding Failure Mode of
Transformers" Proceedings of EuroTechcon 2005 ( as
3. J.B. Shimwell The Case for On-Line Drying of Oil-filled
Transformers. A Risk/Benefit Analysis - Euro Techcon
4. Y.Du,A.V.Mamishev, B.C.Lesieture, M.Zahn & S.H.Kang:
"Moisture Solubility for Differently Conditioned
Transformer Oils" IEEE Transactions on Dielectrics and
Electrical Insualtion Vo1.8 No: 5: Oct 2001
5 P. Koestinger:" Drying of Power Transfonners in the Field,
Applying the L F H Technology to Oil Reclamation "
Proceedings of EuroTechcon 2005 (as above)
Reactor 1 - Trace I




u .......
Graph showing the history ofDGA on a 132KV Grid Transformer with Molecular Sieve fitted

Water content X
FIG 4. Graph showing the
relationship between the rate of
ageing and the water content as
a percentage by weight of the
Limited. Page 8 of 8
The transfonner manufacturing techniques have undergone lot
many improvements due to increase in technical skill ofIndian
Engineers & technical collaborations with multinational
Some of these techniques are not familiar to all customers I
manufacturers I erection contractions.
It has been observed that several contractors are undertaking
overhaul contracts without the knowledge of manufacturing
features and spoiling the equipment in stead of overhauling a
healthy transformer.
Few examples of such cases are discussed below:
1. Lifting of tank cover with low rating slings without
knowing that the transformer is cover mounted i.e. core
& coil assembly is fixed on cover.
This will result in bending the cover ends due to inadequate
handling capacity & breaking of on load tap changer if it
is cover mounted.
2. Lifting of tank cover without disconnecting earthing leads
if core earthing & core clamp earthings are taken out and
fitted on cover.
3. Un-Tanking the Core-coil assembly.
3.1 Without removal of earthings connections from tank wall.
3.2 Without removal of bottom feet locking with shoes if
3.3 Without withdrawing tap switch handle.
3.4 Without disconnecting directed oil flow pipes fitted on
bottom yoke clamps & tank body bottom if possible.
4. OLTC disengagement, withdrawal of OLTC diverter,
OLTC engagement & assembly ofOLTC diverter which
is a very skilled job.
5. Disconnecting high voltage condenser bushing leads if
there is a middle joint at bushing seating. Dismantling
and .... reassembly of condenser bushings at a very larger
These aspects are only a few and several other critical
features are existing.
It is recommended that manufacturer of origin should be
only entrusted with this activity.
As a general guidance normal sequence for overhauling
& site repairs are detailed below.
Periodic overhaul of power transfolmers is desirable during
planned shut down periods which will help in taking preventive
measures to avoid any major mishappenings and also to increase
life of transfonners.
This period can be judged in follov";r;g : ; : ~ ~ : ~ ; ~ ; ; : ;
A. Any faulty gases are generating (dissolved gas analysis
will indicate)
B. Any increase in humming sound.
C. Any leakages (weld I gasket joints) are observed.
D. Any dis-colouring or smoke taking place on terminal
E. Oil BOV & PPM values are deteriating.
Even if no abnormalities as indicated above are taking place
once in 5 years it is desirable to take up overhauling works.
Following sequence is suggested.
I. Disconnect bushing connections, clean and reconnect.
2. Man hole (inspection) covers to be opened.
3. Drain complete oil form main tank isolating cooler bank
and conservators.
4. Inspection of active part (core coil assembly) through man
5. Tightening of all bolted joints and retaping of all leads
wherever found loose.
6. Tightening coil pressure screws.
7. Tightening yoke bolts I core colts
8. Removal of bottom oil slag.
9. Cleaning the core - coil assembly through hot oil jet.
10. Replacement IRecirculation of oil depending upon values.
II. Boxing up the transformer.
12. Vacuum puJIing.
13. Inspection ofOLTC diverter switches by taking out
14. Replacement of o LTC oil.
15. Opening cooler bank and conservator connected valves
which have been closed.
16. Topping up and air release.
Following sequence is suggested.
1. Disconnect bushing connections.
2. Isolate cooler circuit and conservator.
3. Drain complete oil from main tank as well as conservator.
4. Dismantle all mountings like bushings, motor drive unit,
cable boxes etc.,
S. Drag the transformer to service / repair bay,
6. Open the main cover
7. Disengage the tap changer / tap switch
8. Take out the cover
9. Un tank the core-coil assembly
10. Inspect thoroughly
11. Tighten all bolted joints, pressure screws, core and yoke
12. Remove the bottom oil slag in tank.
13. Replacement of all gaskets
14. Re-tanking the transfonner
15. Verification of core - clamps & clamps tank earthing
16. Inspection of OLTC diverter switches by taking out.
17. Tank cover mounting
18. OLTC matching
19. Boxing up
20. Vacuum putting
21. Filtered oil filling
22. Dragging back to plinth
23. Mounting all accessories.
24. Equalizing with cooler bank and conservator
25. Toping up and air release
26. Additional filtration due to exposure of core - oil assembly
if oil with stand values are not adequate.
During the inspection oftransfonner active part, manufacturers'
responsible engineer and customer to record all the observations
and to ensure that no foreign particles are left out inside the
transformer tank.
For desired values of oil and for filtration activities original
methods as followed
S ite e p i ~ save time & money (transportation to works & backO
Following aspects are to be considered before deciding for site
1. Detailed assessment of damage and decision
replacement / repair.
2. Time available for outage of equipment (Availability
of spare transfonner).
3. Available of spares like spare winding, bushing or
other damaged components and site for replacement
including fresh gaskets I oil seals.
4. Availability of facilities at site such as
4.1 Repair bay (covered bay having adequate lifting
capacity i.e. EOT Crane of sufficient capacity to
handle un-tanking height and weight of active part)
should be free from fumes, steam, dust & carbop
etc. should have well lighting, fire protection
equipment and source of power & compressor air.
4.2 Lifting tackles like slings, D Shackles standard tools,
hydraulic jacks welding set, brazing equipment
should be available.
4.3 Platforms for working at core & winding at Core
Clamp height & storing unlaced top yoke
4.4 Cradles for dismantling & reassembly of windings.
4.S Sheet steel trays for removed oil soaked winding
4.6 Electrical heater I halogen lamps / hot air blowers.
4.7 Drying of oil in oil storage tanks.
4.8 Low voltage test equipments.
4.9 Oil test equipment.
4 . .0 Consumables like cotton tape, terelene tape, twine
thread, clean cloth, cotton, waste tarpaulin &
Polythene sheets.
S. Availability of required skilled / unskilled man
power to assist manufacturer.
6. Technical data like manufacturer's drawings fault
finding chart, agreed repair I replacement procedure
& check list.
Following are the drawbacks in site repairs.
I. High voltage test is not possible (transformer can be
recharged based on L.V. test and healthiness of repair can
be confirmed only after recharging).
2. Additional burden in arranging facilities at site.
3. Manufacturer normally guaranty for only repaired portion.
4. Delays in repair in case of any further damages are noticed
during repair activity which was not anticipated early and
replacement of material will take time.
S. Drying and filtration process will take a longer time due
to exposing of core and windings during repair
6. Disputes in repair price, scope time and cost of material.
The following recommendations are suggested.
1. Replacement of one full winding in case of failure of
winding in a particular limb.
2. Replacement of all windings in case of failure of
transformer beyond IS years of service.
V. V. Pattanshetti , S. Vijaya Kumari & K. Dwarakanath
power transformers are important components of power system
network. Information on the condition is normally obtained by
testing the transformers for tests like Insulation resistance, Tan
delta, Recovery voltage testing and others. All these are being
done by taking the transformer offline. New test that has been
started in the year 2000 in the country called furafural content
test / fi!ran test has attracted lot of interest as it is done on oil
sample which can be drawn on the transformers on line. The
information obtained is highly reliable and can accurately give
the condition assessment on regular basis. The paper covers
the basics of the test, the instrumentation and the CPRI studies
on the power transformers in the country.
Transformers are the largest, heaviest and often costliest single
piece of equipment in a power system. It plays vital role in
transmission and distribution system. Failures can cause losses
in terms of equipment cost (1.5 - 5 crores) and losses due to
non availability of equipment are very high (I 00-200 crores).
These hard facts indicate the importance of careful design,
manufacture and condition monitoring during service. There is
a scope to study the vulnerable materials in the power
transformer, degradation pattern, indicative tools of this
degradation, the factors influence these degradations. In this
paper, materials used in transformer construction, with a special
focus to the solid insulation bas been covered.
Materials of Power Transformer construction
Transformer can be considered as made of three major type of
materials. First can be metallic consisting of core steel, winding
conductor and tank. Second can be liquid insulation. Third can
be solid insulation made of cellulosic materials like paper press
board and others. Others like paints, varnishes gums and other
extraneous materials form a small part. It will be interesting to
look into these materials based on degradation aspects.
Metallic materials: Core steel
The purpose of a transfonner core is to provide low - reluctance
path for the magnetic flux linking primary and secondary
windings. During this function core experiences energy losses
due to hysterisis and eddy currents flowing within it. These
heat up the core and alternating fluxes can cause. Specialised
steels conSisting of cold rolled grain oriented steels are used
for the purpose. BS 60 J -1973 gives specification of such steels.
Metallic materials: Winding conductors
High conductivity alloy copper is recommended for the winding
conductors. Silver is normally used up to 0.01-0.02% to give
the conductors the required mechanical strength and
conductivity BIS ] 432-1987 gives specification of this material.
Degradation of metallic materials during the course of service
is considered minimal. The joints, bolts, nuts and leads normally
get over heated and melt leading to failures. Some times due to
over heating of winding anns shorting takes place leading to
inter tum faults and failures. Rectification of such faults at site
is difficult and it has to be done at the manufacturers
Liquid Insulation: Power transformers normally use petroleum
based oils having carbon number between C] 5-C25 having good
oxidation stability and good dielectric properties. Petroleum
oils are very sensitive to contamination like water, particles,
polar degradation products and others. These contaminants can
render them unsuitable for continued use in transformers. BIS
1866 gives the limits of the properties at which oi I can be
partially reconditioned by filtration under vacuum. If the
properties are very bad specially with respect to tan delta and
specific resistance oil can be replaced with vcrgin oils. BIS 335
gives the specification of new oils.
Solid insulation: This is an important material in power
transformers. These are made up of cellulosic materials like
- paper press boards. Insulating paper is wound round the copper
conductors to have proper insulation to the total assembly. In
a typical power transformer, if about 40 KL of oil is llsed then
about ]2 tonnes of paper and pressboard materials will be used.
Good insulation is always a must for any transformer to run
Any damage to solid insulation due to degradation
IS almost permanent and is almost difficult to reverse the
damages. It can also be considered that if the solid insulation
has lost its life means transformer has reached end of its service.
Hence, it is very interesting to know the chemistry of its
degradation. Assessment of its condition and Estimation of its
remaining life
Failures of Power Transformers.
It has been generally accepted that transformer mainly due to
the following causes
1. Design aspects (30%)
2. Manufacturing defects (30%)
3. Material defects {I 5%)
4. Poor Maintenance and operational aspects (25%)
Transfonner failures can be broadly classified as
I. Winding failures
2. Failures in the magnetic circuits
3. Failures in the insulation
4. Structural failures
Failures due to internal short circuit in HV coil are mainly due
to deterioration ofthe insulation and in6ress of moisture in the
oil. Under over load conditions or the fault currents resulting
in over load, cause over heating of winding by which insulation
becomes brittle resulting in failure. This cauSes severe damage .
to the windings and transformer core. Before this failure occurs
the insulation will be deteriorating and if we are able to assess
this before hand remedial measures can be taken. Remedial
measures can be operating the transformer at low moisture
levels. Once the deterioration has reached a level, improvement
in condition of insulation is almost not possible. Under these
conditions, condition assessment can only be used for remaining
life estimation and subsequent replacement of transformer ..
Solid Insulation Condition assessment techniques.
Various methods are conventionally used for the condition
assessment of transformer as whole. These techniques give
information about the health of insulation including that of solid
insulation. Trend in variation of parameters like insulation
resistance, dielectric disscipation factor and partial discharge
levels will give useful insight about the condition of insulation
as a whole. Methods like recovery voltage test can give useful
information about the moisture retained in the coil and core
assembly. Tests like Surge voltage and frequency response
techniques measure the response to the applied surge voltage
or frequency. The pattern can be necessarily be compared
with another winding of similar design or the pattern of the
same winding taken when it was healthier. Little more details
on each of the above techniques would be useful in
understanding the insulation are covered in the later sections.
Insulation Resistance Measurements
In a power transformer typically three phases each having HV
and LV windings between them the core and the insulation matrix
is available. It is possible to measure the resistance between
them. The resistance measured is the sum of the resistance of
. the solid and liquid insulation. Typically four schemes of
measurements are generally made.
I. HV Vs LV with LV grounded to the tank
2. LV Vs HV with HV grounded to the tank
3. HV vs LV without grounding
4. Bushing insulation with respect to test tap
Measuring the resistance after a definite time and finding their
ratios will give us the polarization index of the transformer
insulation. Initially in a transformer IR will be very high
(- 2000 MOhms). During the course of service due to the
degradation of liquid and solid insulation the values will be
going lower. Moisture and other polar degradation products in
presence of higher thermal and electrical stress will accelerate
th is process.
Recovery Voltage Method
In this method, charging and discharging of the transformer
winding with the help of DC voltage. The polarization spectrum
obtained can be used to assess the condition of the solid
insulation. An estimation of moisture present in the winding
can also be made using this information. The method can be
useful in providing infonnation about the condition of solid
insu lation and over alJ health assessment of insulation between
Surge vultage comparison metbod
This method uses the response of transformer conductor in the
windings in presence of insulation on application of surge
voltage. The rising pulse of surge voltage applied spreads along
the winding coil and generates a ringing pattern which is typical
in nature. If the insulation is faulty the pattern gets deviated
original pattern or a healthy winding present on the other side.
Major cause of transfonner failure is due to the fault in the
winding turns. This method will be useful in such cases.
Frequency Response Analysis
This is the latest and powerful diagnostic test on power
transformers. Wide range of frequencies the step voltages are
applied. Impeda!1ce of transformer winding will be measured
for range of frequency. Comparison ofthe results to a reference
test will indicate the damage to the transformer windings. Use
offast fourier transform techniques very clear and unambiguous
information about the movement of the windings can be
There is a relation between the configuration of the winding,
core, resistance, inductance and capacitance. There exists a
pattern of response over the range of frequency for the given
circuit. The changes in the geometric configuration alter the
impedance network and in turn alter t ~ transfer function.
Charges in the response pattern will reveal a wide range offailure
All the above tests are either done during the maintainance
schedules or a special shut down has to be taken for undertaking
these tests .
Direct assessment of condition of solid insulation
Insulating paper drawn from the transformer can be tested for
tensile strength and degree of pol mer is at ion. This necessitates
removal of representative paper and then getting it tested for
the properties. The test can be very infrequent and causes
problems to the system in long run.
S.D.Mayers Inc. developed a new method of testing the solid
insulation which involved the testing transformer oil its self for
the degradation products of solid insul8tion and thereby
assessing the degradation status of paper can be arrived at.
Presently large data is available on this analysis and has been
practiced world wide indicates the .use and popularity of this
testing. Insulating oil with which the solid insulation is in
equilibrium acquire the degradation products (furan derivatives)
by dissolution. Contents of these derivates can be quantitatively
evaluated up to the level of ppb. Range of concentrations of
these derivatives can be categorized for degradation levels of
Central Power Research Institute, Bangalore established this
facility in the year 2000 and large data for the power transformers
in the country has been developed and is continuously being
updated. Details ofthe cellulose chemistry, the instrumentation
and the technique are as follows.
Molecular chemistry of paper
Paper is a natural cellulosic polymer consisting of chain of
glucose units coupled together witb a glycosidic ether linkage.
1)rpically a cellulose molecule of paper consists of 1200 units
(degree of polymerization) of glucose coupled together.
Physical strength of paper depends up on the lengths of chain.
If the chain length is reduced its strength will be reduced. The
degradation of paper, ultimate,ly means the chain breakdown in
to smaller chains. In this process the physical properties of the
paper get affected. Loss in physical properties of paper
insulation brittle. This causes blisters I micro cracks in the
insulation leading to partial discharges in the insulation and
ultimately lead to the failure of the transformer.
Figure: 1 Structure of Celllulose
Cellulosic paper in presence of heat and reducing environment
rearranges to form furan deri\lates. This degradation process is
essentially assisted by moisture and thermal stresses in the
transformer environment.
Figure 2 : Degradation of Cellulose to free glucC'se and their
Heat & Moisture

The glycosidic linkages gets broken and glucose structure gets
rearranged to furan ring structure which is much more stable as
compared to glucose structures. Five important derivates of
furan are generally considered as important and the
concentration of these is directly proportional to the degradation
levels of paper insulation.
o Q
o CHO tK)CH 2 0 CHO
MUNW.t2fALt s-HYDfm'Mt11YlfURFURAi. (5HMF)
These are five furan derivatives which are important these are
I). 5-Hydroxy Methyl furfural (5HMf), 2) 2- Furfuryl Alcohol
(2FOL), 3) 2- furfural, 4) 2- Acetyl Furan (2 ACF) and 5) 5-
Methyl -2 Furfural (5 MEf).
High Performance Liquid Chromatography
Chromatography is the branch of chemistry which is ever
important and is used in various applications which range in
enrichment of uranium 238 to purification of water. The method
is used to separate the ions, gases, and liquids of similar
molecular structure and can be quantitatively estimated. Levels
of accuracy can be much below ppb levels. Normally the mixture
of the material to be separated is passed through a column
consisting an adsorbent material which has varying affinity of
retaining these components and get them eluted with the help
of carrier I e)ueant medium. Hence each of the component
comes out the column at different retention time and gets
detected with the help of specific detectors.
High Performance Liquid Chromatography (HPLC) essentially
consists of a pump or a system of pumps which can pump
continuously the eluent (which can be one or two component
systems). The eluent is continuously flowing to the column.
Through an injector the sample is pushed in to the stream of
elueant. The stream passess through the column (nonnally
reverse phase C 18 for furan analysis). The column has different
retention times for the 5 furan Hence they get
separated and will be coming out the column at different
retention times. These can be detected by their absorbance of
light due to their molecular structure. All the five deri vates are
not detected with good sensitivity witb UV-Vis detector. Hence
the evaluation is done at two different HPLC Chromatograms
of standard calibration mixture and the typical forms are shown
in figures given below.
at 22hm
Figure 3 Chromatogram at 220 nm
at 276nm
t (min)
Figure 4 Chromatogram at 276 nm
wavelengths to get appropriate' set}sitivity for all the five
components: Peak intensity and area under curve are directly
correlating their concentration in the sample.
Calibration of instrumentation is possible by preparing the
standard mixture of these five derivatives at different
concentration and their peak intensities I area under the curve
can be used to develop the calibration curve and then
quantitative estimation of the concentration of these derivates
can be made.
CPRI has carried out studies in which the samples of paper
were aged at different temperatures and the degree of
polymerization and furan levels were studied.




800 600
.'20C I
i 140C i
! 0160C I
L..: __
400 200 0
Degree of Polymerisation
Figure 5 Correlation of Degree of Polymerisation with
Statistical analysis was also carried out on the furan
concentrations and the age of commissioning for power
transformers in the country.
FFA Concentration va Transformer Age
" ... __ ...... -"--'- .. --r ....... _ ..... _ ..J ...

t 8
t . ... .. ....
o 10 20 30 40 50 60
v .... s-. Commlnlonlng
Figure 6. Correlation of furan concentration with the age of
power transformers
However,lndian conditions are much more severe as compared
to the European conditions as ambient temperatures are almost
double of an average the (36-45 C) European country (0-25
C) and hence the life estimations oflndian power transformers
are generally lower. Hence the concentrations of furan will be
comparatively higher for Indian conditions.
Condition of transformers can be assessed by fixing limiting
values which are being followed internationally based on
S.D.Mayers study during the formative years of development
of this test method.
Total furan. ppb
o - 100
101 - 250
251 - 1000
1001 - 2500
Low reliability
Rewind I Replace Solid insulation
It is imperative that even in Indian conditions such limiting
values are to be followed for having the condition assessements.
Limiting values can be reworked to some extent for suiting
conditions. It has been recommended that the testing on regular
basis can be carried out on annual basis along with the
recommended tests mentioned in BIS 1866 - 2000 would be
very useful in condition assessement of transformers.
Furtural coment test can be very useful tool to assess the
condition of solid insulation, infonnation about which is very
difficult to' get.
The information is useful in Remaining Life Assessemnt and
remedial action plan for the transformers in service.
1. Large scale survey of furanic compounds in operating
transformers and their implications for estimating service
life. John R Sans, K.Muge Biligin and Joseph J. Kelly
(Sen ion member IEEE) at International Symposium on
electrical insulation, Arlington, Virginia, USA, June 7-10,
2. Furanic compounds in electrical insulating liquids by High
Performance Liquid
Chromatography Standard Test Method. ASTM D 5837
3. Mineral insulating oils - Methods for the determination of
2-furfural and related compounds. IEC 1198-1993 Geneva,
4. of average viscometric degree of
polymerization of new and aged electrical papers and
boards. Standard test test method ASTM D 4243-1993
5. Measuring and understanding the ageing of kraft insulating
paper in power transformers. Technical report TR(T) 217.
Issue 1 March 1996 by Mushtaq Ali, Alan Emsley, Christian
Eley, Richard Heywood and Xinhua Xaio Addressee.
6. Degradation of Electrical Insulating paper monitored with
high performance liquid chromatography by J Unsworth
and F. Mitchell, IEEE Y transactions on Electrical
Insulations Vol 25, No.4 (1990).
7. Field studies of furan formation in transformer fluids as
an indicator of damage to paper insulation. By D.P.Myers
and J.R.Sans. Minutes of fifty ninth annual international
conference of Doble Clients Section 10.6.1, Boston, MA
USA (1992).
8. Transformer life extension through proper re-inhibiting and
preservation of the oil insulation. By J.J.Kelly and
D.P.Myers at IEEE Transactions on industry applications.
Vol 31 No.1 January / February 1995.
9. Condition diagnosis and monitoring of transformers -
Possibility of estimating transformers life time. By
Boisdon, M. Carballeira et ai., .CIORE, 1992.
10. Diagnostic tests of high voltage oil-paper insulating
systems (in particular transformer insulation) using DC
dielectrometrics. CIGRE Session 1990.
II. Remaining life announcement of power tninsformers,
S.Vijayakumari et ai. Published in CIGRE session 2008.
Prashantha S. K, Narasimhan S. R.
All high capacity transformers in use today have to cope with
voltages in excess of 400k V. Therefore it becomes imperative
(0 maimaill oplll.Uui pwperties of the transformer oils
by rigid control over moisture, dissolved gases, particulate
contamipation and acidity.
Transformer oils deteriorate with time and the decay begins the
moment the transformer oil is shipped from Manufacturer's
factory. High temperatures, the presence of oxygen and water,
the catalytic action of the materials within the transformer, all
combine to result in oxidation and cracking of the transformer
oil. The by-products of this oxidative process are acidic in
nature, which results in an ever-increasing rate of deterioration
of the transformer and its oil. The resuiting sludge build-up
reduces the effects of the oil driving the whole decay
mechanism at an increasingly accelerated rate. This sludge
eventually precipitates out for solution and forms a heavy tarry
substance which adheres to the insulation, the side walls of the
tank, lodges in ventilation ducts, cooling fins and so on ...
When this type of contaminated oil, completely unfit for further
use in the transformer is examined, it may be found that at least
80% of the hydrocarbons present in this oil are unchanged and
can be reused if the oxidized products are completely eliminated.
Hence this type of worst contaminated oil cannot be filtered
through normal thermal-vacuum filtration process and hence
has to be discarded or can be re-used using the Regeneration
technique developed by us.
We, Fowler Westrup (INDIA)PVT LTD, (FWL) through our
. continuous R&D efforts & collaboration through our technology
provider , Redragon - Canada have developed a new
Regeneration system, which consists of both Regeneration &
Reactivation columns wherein the specially treated media can
be used for more than 300 times. We have observed that the
regeneration process is less strainful on the transformer
insulation than the alternate process of draining, flushing and
refilling the transformer with new oil. Hence based on our
in the filtration field for over 40 years, we are
confident that if the contaminated oil is regenerated through
the technology developed by us, the damage to the transformer
insulation can be prevented and the life of the transformer will
be extended for another 20 years. Hence in this paper, we have
highlighted the importance of Regeneration process which will
enable us!o save the most precious asset in today's world called
" OIL".
drive to tighter tolerances of modem transformers and electrical
apparatus results in greater electrical stress in insulating material
and fluids. To handle these stresses, oils are required to have
better dielectric strength, and lower residual water content must
be maintained to reduce the speed of oil aging. The periodic
mu:l proper treatment of these insulating fluids will result in the
improvement of the properties of the entire insulating system
of power transformers and will extend the effective lifetime of
the asset.
Many thousands of oil filled transformers containing huge
quantities of insulating oil are in service today. As you are aware,
the mineral oil in these transformers is intended to,
Provide dielectric strength of the transformer insulation
Provide efficient cooling
Protect the transformer core and coil assembly from
chemica! attack
Prevent the build-up of sludge in transformer
When the properties of the oil have changed enough that oil
can no longer satisfactorily perform anyone of the above
mentioned functions, the oil is said to be bad. Continuing to
operate a transformer with bad oil significantly reduces the
transformer life expectancy.
The principal functions of the insulating fluid are to serve as a
dielectric material and an effective coolant. To perform these
functions, the insulating fluid must possess qualities at
the time of initial impregnation of the transformer core and filling
of the tank at the manufacturer's factory and which must be
maintained at the same level in field operation if optimum
performance is to be guaranteed. Typically these qualities are
well defined by international and national standards.
Based on these functions of good oil, the goals to be reached
Operating the transformer in the" sludge free range"
Preventing the premature transformer failure due to
transformer oil oxidation products
Conserving two very valuable resources - Oils &
We will now go through the chemistry of oil degradation,
2. INTRODUCTION: "New" oft does not exist even in a newly installed transformer.
Transformer oil is in an environment that leads to degradation Oil deterioration begins as soon as the oil is placed in the
in its desired properties and even in a proactive maintenance equipment at the factory prior to shipment. Traditional
t programme, transformer oil will lose its insulating & cooling distinction has been made between deterioration and
L .. properties over a penod of time. The rating requirements and 159 contamination,
1. Contamination: In general, an oil will be considered to be
contaminated when it contains moisture or other foreign
substances that are not products of oxidation of the oil
2. Deterioration - This tenn will be used to denote only the
effect of oxidation.
As stated in the fonn of chemical reaction, we will now see as
how good oil goes bad, where oxygen is the primary cause,
Unsuloie iiyaf'UaerOODS lin loe 011) + Oxygen + Cataly5ts+
accelerators = Oxidation by products.
Transfonner oil is adversely affected by oxygen, moisture (as
catalysts) and heat (as an accelerator).
The American Society for testing and materials (ASTM) has
established that oxidation of an oil results from a process that
begins when oxygen combines with unstable hydrocarbon
impurities under the catalytic effect of the other materials in the
transfonner. Certain other factors accelerate heavily the sludge
reaction. As you are aware, transfonner oil has a special affmity
for oxygen. Degrading cellulose especially via heat also supplied
a source of oxygen. Finally the natural oxygen inhibitors in new
insulating oil are gradually depleted with time, increasing the
rate of oxidation throughout the in-service life of oil.
The catalysts are primary substances that increase the rate of
oil degradation without being consumed in the process which
1. Moisture: Moisture can enter the insulating oil externally
via a leak, as condensation or internally through the
chemical process of oxidation. Water is a major catalyst in
oil oxidation. All forms of water-contain additional oxygen.
2. Copper & Oxygen: All of the preceding shows the
beginning stages of what is inevitable process of
deterioration. Fortunately this is a long-tenn process,
catalyzed also by copper (windings) and iron (core) and
finally accelerated by several other factors.
3. Accelerators: Several secondary factors increases oil
oxidation. These are called accelerators and consists of,
Shock Loading
Surge Voltages
High electrical stress
Unfortunately, the terminal stage of the deterioration process is
sludge, the visible sign that the oxidation process has long been
at work. Sludge is a resinous partially conductive substance,
which is moderately soluble in oil. Sludge comes about when
the acids attacking the iron, copper, varnish, paints etc and these
materials comes in to solutions and combine together. This
sludge eventually precipitates out for solution and forms a heavy
tarry substance which adheres to the insulation, the side walls
of the tank. lodges in ventilation ducts, cooling fins and so on '"
Even though the oil gets deteriorated badly due to the above
factors, relatively few of the estimated 2870 hydrocarbons
present in the oil have reacted with oxygen. Further and more
important, the same oil can be used again for its original purpose
after the removal of the oxidation products through regeneration
As you are aware; if the acidity content present in the oil is too
high ( i.e above 0.25 mg KOH/g oil) , then the oil ceases to be
a good one and filtering the same. through normal thermo-
vacuum filtration process will not bring the oil properties back
to its virgin condition. Hence this highly contaminated oil has
to be regenerated instead of mere filtration.
Regeneration is the procedure followed if oil filtration is
insufficient to bring back the oil to an acceptable condition. In
other words, Regeneration is the restoration of the oil to, or
better than, virgin oil specification. The difference between
regeneration and purification is that purification cannot remove
substances such as acids, aldehidtj. ketones etc in solution and
hence cannot improve the quality of contaminated oil. Under
these conditions, the oil has to be regenaated.
Regenerations systems typically remove the products of aging
from used transformer oil. By products that are known a"
secondary contaminants are created as a result of natural aging
of a transformer in service coupled with inadequate maintenance
regimens .. The aim of regeneration is to remove the decay
products from the oil before they cause damage to the
transformer insulation system. A well-maintained transformer
will prevent a wet core condition and ensure that the transfonner
always operates in the sludge free zone.
Hence the process of Regeneration is absolutely necessary which
rejenuvates the oil thereby preventing damage to the transfonner
insulation which helps in increasing the life of the transfonner
by another 20 years ... Experience gained by us in this field has
also shown that the regeneration process is less strainful on the
transfonner insulation than the alternate process of draining,
flushing and refilling the transfonner with new oil
Regeneration can be carried out by two processes,
In earlier days, Reclamation of oil were carried out in crude
unconventional methods which conswned a lot of space and
reclamation was carried out through unhealthy safety standards
apart from polluting the atmosphere.
Historically, Fullers Earth has been used to reduce the acid
number on a one-time batch basis. There are obvious
disadvantages to this, not least the environmental impact of
dumping oily waste into a landfill am! the increasingly regulated
waste management sector prohibit this dumping at any cost.

I t I. t LAC"l

t ,-
W ...=.l

Limitations of conventional method:
Media gets saturated after single use, i.e USE -
Specialized Disposal Methods are required.
Frequent Replacement of Media has to be carried out.
Labour intensive & time consuming process
Less productivity resulting in increased running cost
Oil is a key component in the Electrical and automotive industry.
Oil which is used as a lubricant or as an insulator loses its
properties over a period of time due to accumulation of
contaminants like moisture, metal fragments, chemicals
reactions etc., These contaminants increase the acidity value of
the oil, degrades the viscosity and hence the oil loses its original
lubricant & insulating qualities and tends to become abrasive
and conductive. The very same oil which once facilitated the
functioning of the equipment, now starts to hinder the
performance of the equipment and if unattended to will result
in breakdown of the equipment.
The simple solution is to maintain the oil in good condition.
Replacing the oil periodically is a very expensive exercise and
regeneration through conventionid method will result in
unhealthy production standards apart from polluting the
atmosphere. That's where the regeneration technology adopted
by FWL comes in very handy. The regeneration system brings
back the contaminated oil to its original virgin oil condition at
a fraction of the cost and the system can be incorporated in the
same premises, or kept mobile for multiple installations. The
effective regeneration of oil is effective rejuvenation of the
whole equipment.
Advantages of FWL developed Regeneration system:
1. The system will consists of Regeneration & Reactivation
2. Specially Treated Media will be used
3. Media can be reused for more than 300 times, i.e. USE-
4. Productivity increases as more ,oil can be processed with
same media
5. No specialized disposal methods are required
Used, dirty oil is drawn into the Regeneration system by means
of an inlet pump. The dirty oil passes through a coarse mesh
strainer that removes solid contamination. The oil is then heated
to 60 Deg C, typically by low watt-density oil heaters. This
reduces the oil viscosity and allows for first stage filtering at
reduced differential pressures. At this stage the oil is typically
filtered to 50 micron. The oil is then fed to the regeneration
columns filled with specially developed media, which rely on
an adsorbent principle and restores acid, Interfacial Tension,
color, power factor and other parameters equivalent to new oil
Exiting the columns, the oil is then fed on to a final polishing
filter that typically reduces particle contamination to less than
I micron and it also passes through a vacuum vessel where under
reduced atmospheric pressure, residual water and entrained
gases and air are extracted by a vacuum pump system and
exhausted. The newly reclaimed oil is pumped out of the vacuum
vessel through the outlet pump. Based on the input oil and final
required parameters, the oil may need to pass through this
procedure again. When coupled with fine particulate filters plus
a high vacuum degasser system, virtually all oxidation by-
products are removed and the oil returns to its original, virgin
oil status. The regeneration process also removes corrosive
sulphur and metals from the oil.
The adsorptive capacity of the media will produce a yield of
around 5 liters per kilogram and the media. should then be
reactivated. The reactivation process relies on a thenno-vacuum
or thermo-pressure technique that typically will take ~ 4
hours. Upon completion, the media is now reactivated and can
be re-used. There is a marginal decay in the yield with each
reactivation that detennines the useful lifetime of the media.
Operational oil parameters
Parameter Unit Contaminated Treated Oil
oil before after
Regeneration Regeneration
Appearance Cloudy, Sparking, clear
brown & transparent
Colour 3.5 LO.S
Breakdown Kv 15 70
Gas content % 10 0.1
IFT@2llc N/m 0.02 0.04
Neutralisation mgKOH/g
Value oil o.s 0.03
Oxidation Stability
Neu,traliBation mg KOHlg - 0.4
Value oil
Resistivity @270c ohm-em l.OxlO
Resistivity @900c ohm-cm O.lXIO
Tan delta @90oc
1.0 0.002
Water content ppm 200 5
The regeneration process incorporates the thermo vacuum and
fine filtration process. Through regeneration, the following
results are achieved.
1YPica1ly the reactivation process can be performed around 300
times before the yield is low enough to warrant replacement of
the media. Optimized regeneration systems utilize dual banks
of columns that allow for continuous operation of the system
by reclaiming oil in one bank and reactivating the same in
another column.
The moisture of the oil is reduced to less than 5 ppm
The acidity is reduced to less than 0.03 mg KOH/g of oil
The dielectric strength is improved to better than 70 kV
The 1FT is improVed up to 0.04 N/m
The tan delta of the oil is improyed to less than or equal to
Degradation of the solid insulation has been halted
To arrest the process of oil deterioration, it is necessary to
maintain the oil in good condition. This can be done by
Monitoring the oil condition on regular basis
M a i n ~ i n silica gel breathers in an active condition.
Repairing the oil leaks as soon as they occur
Starting the transformer with a good quality oil
Purification of oil as soon as the moisture rises above 20
ppm or the dielectric strength drops below 50 kV
Monitor the acid levels of the oil and regenerate the oil
before it reaches the critical level of 0.5 mg KOHIg of oil.
Benefits of regeneration oil:
Life Extension of a limited resource
Economically advantageous
Extended life of expensive capital equipment
Minimal Running cost
Huge Savings due to media re-activation
Fully customizable
Regeneration is the greenest technology as oftoday and is future
proof. Infact, it is a revolution in the making for the industry
where the resultant savings for the customer will be enormous.
AU this at a lower cost than the price of new oil. The suitability
and stability of this oil has been well oroven in field use for
well over 50 years and has been purchased by a number of
utilities and transformer manufacturers as an alternative to
purchase new oil.
Fowler Westrup had pioneered a revolution in the 80's with the
uNIRMAL" transformer oil filtration machines through its high
vacuum technology processing. Today we are pioneering another
revolution in the industry with the introduction of our Oil
Regeneration technology in India. Basically regeneration is a
safe, highly efficient process where the media used in filtration
is reactivated and reused multiple times before charging a new
media. This amounts to lengthening ofHfe of media from 1 or 2
uses to nearly 300 times. FWL is one of the few in the world to
master this niche technology and hereby proudly say that we
are the only one in India who can fully design, build, install,
commission and support such advanced systems ranging from
600 - 10000 litres per hour capacity with both mobile and Skid
configuration. AU this translates to huge benefits to the users of
our systems. Spectacular savings in raw material costs, drastic
reduction in storing and dumping area & material, reduction in
dependence on oil supplying companies, fully automated
handling, simultaneous cycles of regeneration of oil &
reactiva.tion of medi.a. Added benefits are that it can be fully
customIzed and programmed to handle different oils like
transformer oil, capacitor oil, automobile lube oil and cable
oil. We have developed this technology along with our
collaborators; Redragon of Canada and the same have been
developed for Indian conditions and customized for every client
of ours. All models are customisable to suit the specific needs
of the application. We take responsibility from studying the
client requirement, recommendation of solutions, design.
implementing the solution and even train personnel to maintain
the systems, and most important of all supporting them fully in
effectively incorporating our system into the client's
infrastructure. Hence we provide total solutions for oil
processing systems. We have been serving the world market
for the over 4 decades and have catered to various markets
varying from small electrical contractors to big corporate &.
government agencies.
To conclude, insulating oil when properly maintained can give
practically unlimited extension oflife. Only a limited percentage
of the crude available are used for conversion to transfonner
oil and hence it is our responsibility to preserve the life of the
most precious asset in today's world called "OIL".
The authors are pleased to acknowledge MIs Fowler Westrup
(INDIA)PVT LTD managel11ent for their kind permission to
present this paper at this prestigious conference.
. -- .--
J. S. Batra
Power transformer is the vital link between generation, transmission and distribution
constituting transformation capacity about four times the generation capacity. This is the most
expensive equipment in a transmission system but unfortunately, their failure rate is quite
substantial and alarming which leads to huge repair cost as well as revenue loss apart from
interruption of power supply. Only a few of the transformers die of old age but most of these are
killed due to improper up keep. In order to extend life of transformer it requires appropriate
preventive measures, condition monitoring based on periodic electric & oil tests and sensitive
protection & annunciation system apart from effective operation & maintenance.
Power transformer is the most expensive single equipment in the electrical transmission
system but failure rate of transformers in our country is quite substantial and alarming. These are
vital links between generation, transmission and distribution and constitute transformation
capacity more than four times the generation capacity .The failure of the transformer not only
results into huge loss by way of its repairs but also causes loss of revenue on account of
interruption of power supply. Only few transformers die of old age but most of the transformers
are killed. At present, main point of deliberation is how to stop killing of the transformers which is
not possible alone through effective operation & maintenance but it also requires care like human
body by adopting preventing condition monitoring based on periodic electric & oil tests
and providing sensitive protection & annunciation system. Although On line condition monitoring
is a future trend but very simple Hydrogen monitor can be introduced for transformers
requiring frequent DGA testing.The safety precautions for safety of. human beings and the
eqllipment are of paramount importance and hence list of DO's and DON'rs for safety is also
included in this paper.
Power transformer is the most expensive equipment in the electrical transmission system
but failure rate of transformers in our country is quite substantial and alarming. The main causes
of the failure of transformer are design defect, manufacturing defect, material defect, poor
operation & maintenance, surges & short circuits and last not the least loose contacts. In order
to have reliable & long life and to prevent frequent failure of the transformers following steps
may be adopted:-
a) Effective operation & maintenance
b) Preventive measures
c) Maintenance based on periodic condition monitoring through various tests
d) Sensitive protection & annunciation system.
The transformer needs to be protected and maintained through routine and
preventive maintenance to enhance its life and to obtain trouble free service. The
operation & maintenance schedule in broad and brief is given here under. However
for details relevant manufacturer instructions/maintenance manual and relevant
standards be referred to.
Work to be carried out
Periodicity I .. if
... ................. "" ... t. ....... 'VLU. y
Hourly Take remedial measures if temp. rise is
Check oil temp. rise
more than 3SoC .
Check MW, MVAR voltage & -do- Take remedial measures if values are
current. above normal.
Check winding temp. rise. -do- Investigate if temp. rise is persistently
more than 4SoC and take remedial
Check the oil level in Main & Ol TC Each shift If low, top up with good oil. Examine
Conservators and Bushings. transformer for leaks and necessary
steps to attend leakage be taken.
Check condition of silica gel & oil in -do- a) If silica gel is pink change by spare
breathers charge. The old charge may be
reactivated for use again.
Note: -As per recommendation, b) Ensure that there is no breathing of air
acrylic pipes for containing silica gel to the silica gel breather container from
should be used. any other source except from oil cup.
c) If oil level in oil cup is low replenish
immediately ..
Check indicating lamps for working -do- Investigate reason and replace lamps, if
of pumps & fans. fused.
tank pressure (if -do- It N2 pressure low check & attend leakage
applicable) and recoup N2 pressure.
Replacement of oil and cleaning of Quarterly Replace breather oil. Clean air passage,
vent holes of breathers if reQuired
Examine bushings for cracks & dirt Half yearly Clean to remove dirt. Replace i! cracks
deposits. are unwarranted.
Check aI/ the earthing connections -do- Tighten if found loose.
& bushing clamps for looseness.
Check for dielectric strength of -do- Take suitable action to restore quality of
oil(BDV) in transformer and tap oil.
Check cooler fan and pump -co- Replace burnt or worn out contacts or
bearings, motors, starters, other parts. Check current drawn by
relays and operating mechanisms & motors by clip on meters.
oil flow indicator and their Auto start
Check condition of jumpers & their - do- Replace the burnt out jumpers and
clamps tighten all the nuts & bolts of jumper
a) Check ALL alarms, their circuits, - do-. (a). Clean the components, contacts or
ACIDC fuse,links etc. replace contacts and fuses, if
b) Check actual tripping through necessary.
OTT,WTT, BRT, PRV & protection
relay_s and inter tripping -do- (bl Take suitable action, in case of non-
,,? ..........:;
. .
15 Check for vermin proof-ness of -do- Take remedial measure measures to
cubicles and terminal boxes and make these vermin proof.
their cable entry.
16 Check oil in pockets of OTI and Yearly Oil to be recouped if required.
17 Check and calibrate oil temp. and -do- Adjust if required.
winding temp. gauges and check
18 Cleaning of arcing horns and -do- Set the gap if required.
checking gap.
19 Measure DC resistance, tum ratio, -do- Compare with the previous values &
magnetizing & short ckt. Current, investigate reasons for change in value.
magnetic balance, IR, PI & IR should by taken preferably during
dielectric Absorption value of same month every year.
Mtnsformer windinQs.
20 a) Check control circuits of OL TC a). If faulty, take suitable action to set
driving Mechanism in independent -do- these right.
b) Driving Mechanism of off load tap -do- B). TTR should be taken at all taps by P
changers where existinQ. & Tcell
21 Inspect all moving parts contacts,' -do- Clean, adjust or replace as required.
brake shoes, motors etc. of on load
tap chanQer driving mechanism.
22 Check gasketed joints and oil -do- Only leakage/seepage joints should be
leakage tightened/attended
23 Check earth resistance of T/F -do- Take suitable action if earth resistance is
earthing & neutral earthing (Dry high.
24 Check Healthiness of -do- . Take remedial action if required.
balloon/Diaphragm where feasible
25 Physical inspection of radiator outer -do- Cleaning 'Of Radiators outer surface with
surface compressed air/water.
26 Physical check of oil level in S.O.S) Physical Check of oil level in conservator
conservator (not applicable for N2 yearly with dipstick method where glass gauge
filled and air cell type) not provided.
27 Testing of oil in transformer as per S.O.S} Analyze the test results for taking
ISS: 1866: 2000, DGA as per Two yearly appropriate action.
ISS: 10593-1992 / I EC-599
including furane
28 Check operation of Buchholz relay Two yearly Adjust float, switches etc. as required.
by forcing air in it.
29 Testing of TIF winding and -do- Change in v l u ~ s should be analyzed.
condenser type bushings for
a) TAN Delta
30 Painting of Transformer Three Use recommended paint for outdoor duty.
31 a) OL TC in main tank
i) Check contacts of diverter Five yearly Replace worn out parts and oil.
switch of OL TC having non-arcing or 10000
selector switch.
ii) Check non arcing selector switch
b) OL TC out side main tank
i) Check contacts of diverter switch
and non-arcing seJector switch of
Internal inspection of conservator
and T/F during capital mtc.

10 years at
the time of
Five yearly
or 10000
10 years**
Replace worn out parts
Replace worn out parts and oil.
vvash WHUjjflY alKl GOre wlin clean
good hot oil. Inspect winding, top
pressure plate & its bolts. Tighten all bolts
nuts, coil-clamping screws, check the
float mechanism of conservator. Replace
all Gaskets.
Keeping in view the basic mantra of "PREVENTION IS BElTER THAN CURE" it is very
essential to adopt preventive measures to minimize the repercussion of factors affecting life of
transformer . The primary factors that contribute to the deterioration of vital constituents and
lead to eventual failure can be categorized into two groups, i.e. external and internal factors:-
a) External Factors
i) Through faults specifically the sustained ones.
ii) Over loading.
iii) Over voltage conditions.
iv) System conditions of low frequency and high Voltage.
v) Breaker failure to operate.
vi) Moisture and oxygen through breather & joints
b) Internal Factors.
i) High temperature & hot spots.
ii) Mechanicallboseness/deformation.
iii) InsiJiation failure.
iv) Moisture content from cellulose
The former can be prevented by providing suitable protection such as high set
instantaneous protection on 33 and 11 KV lines, fast operating distance protection schemes on
outgoing 66 KV and above lines, definite time high set over current protection on transformer,
LBB protection at least for breaker controlling HV s;de of transformer, over current protection
for controlling over loading, and over fluxing relay to check over flUXing conditions. The
moisture and oxygen are the number one enemies of transformer and their ingression during
breathing can be prevented by providing Nitrogen sealing in the transformer or by providing air
cell breather. Nitrogen seaing due to its positive pressure even prevents entry of moisture
through joints and as such be preferred over air cell breather.
The later, apart from providing suitable protection schemes such as differential relay,
restricted earth fault relay, neutral displacement relay, buchholzl OSR relay, oil and winding
temperature devices, also require periodical carrying out of electrical tests and oil tests to
detect abnormality, jf any, in the transformer. Various off line methods based on heat & vacuum
PJocess such as vacuum oil filtration set and induction heating are being used for dehydration
of the transformer. In order to remove moisture from the cellulose. it requires continuous
dehydration process for longer period depending upon state of equilibrium corresponding to hot
spot temperature. In case offline heat & vacuum process is continued for longer period, it will
deteriorate characteristics of oil due to excessive heating apart from outage of transformer for
significant period. It is not possible to ensure complete dryness of transformer through off line
dehydration system as such on line dehydration system based on thermo siphon filter using
alumina and system using superdri cartridges are gaining ground.
The external factors responsible for damage of transformer can be checked by adopting
requisite preventing measures stated above but internal factors can be taken care of through
effective maintenance of transformer. Now maintenance based upon condition monitOring of
the transformer arrived at by carrying out periodical tests on the transformer and its oil is
gaining ground. Condition monitOring of power transformer in service has been a continuous
process and has seen many improvement over the years. Although several diagnostic tests
such as dielectric loss angle. dissolved gas analysis. tua-cill ';uld;y.,i., dlt1 dYCli;ClUit; UUl tJ1t;-
commissioning test form the basis for subsequent periodic tests to be conducted at site. The
pre-commissioninglperiodic tests to be carried out at site are as under:-
The values recorded during the pre-commissioning of various electrical tests given here-under
are compared with the tests results recorded at manufacturer's works. These values form basis
for comparison with the test results obtained during subsequent periodical testing of the
transformer. The periodicity of electrical tests is one year in 88MB ..
Test description Purpose Acceptance norms
I.R.Value and To check over all condition of insulation of T/F. P.1. value < 1.0 =dangerous
polarization index 1.0 -1.1 =poor, 1.1 -1.25
= questiona,ble, 1.25 to 2.0 =
Satisfactory , > 2.0 = Good
Vector group To check correctness of HV & LV winding connections As per commissioning results
Turn Ratio Test To check correctness of voltage ratio. Comparable with
commissioning results
Magnetizing current To check inter tum insulation. -do-
Short circuit current To check continuity of HV & LV windings. -do-
Magnetic balance To check core assembly condition and flux distribution -do-
Winding Resistance To check tightness of winding connections and jOints. -do-
Capacitance & Tan To check condition of capacitance bushings Tan delta should be Less then
Delta of bushings 0.007 at 20C.
Capacitance = + 10% to - 5% &
rate of rise less then 0.1 % per
Capacitance & Tan To check inSUlation I any displacement of windings Comparable with previous
Delta of windings. values
Through stability To verify CT ratio
secondary wiring Ditt. currents la = Ib = Ie = 0
test connection/vector group etc for diff.protection
Protection relays & To ensure healthiness Icorrectness of protection Permissible value is 5% of the
indicating system and indicating instruments set value
Bay eqpt. i.e. CT. To ensure healthiness Ifunctionality of CT. PT/CVT. Results should be close to
breaker. breaker. LAs .isolators rated values
LAs .isolators
To give idea of various values of electrical tests , the pre-
commissioning values recorded for 66111KV,10/12.5MVA ,ABS Make Transformer T-2 (Sr.No.
AAF-02) commissioned at PGI, CHANDIGARH are given here under:
PRE COMMISSIONING TESTS OF 101125MVA 66I11KV ABB Make Transformer T 2 .
J -
Top oil temperature; 35
C 20 See. 60 See. 120 Sec. 600Sec. P.I. (600S/60S)
Using Meggar make kit
type MEG 10-01 at 5KV HV-E 25.8 37.2 41.3
,.." ~
..... ,.
;JV.I 1.'-'\,,1
ValueslnGO LV-E 13.8 22.5 32.5 58.8 2.61
HV-LV 20.3 28.2 42.0 76.5 2.71
Using Biddle make kit S.NO. 4646
Tap No. Ratio TapNo. R-Ntr-n Y-N/y-n B-Ntb-n
1. 6'.300 1. 6.2840 6.2850 6.2840
2. 6.225 2. 6.2110 6.2120 6.2110
3. 6.150 3. 6.1390 6.1390 6.1390
4. 6.075 4. 6.0670 6.0680 6.0660
5. Normal 6.000 5. Normal 5.9930 5.9940 5.9930
6. 5.925 6. 5.9210 5.9210 5.9210
7. 5.850 7. 5.8490 5.8500 5.8480
8. 5.775 8. 5.7760 5.7770 5.7770
9. 5.700 . 9. 5.7030 5.7040 5.7040
10. 5.625 10. 5.6310 5.6320 5.6320
11. 5.550 i 1. 5.5590 5.5610 5.5590
12. 5.475 12. 5.4860 5.4870 5.4870
13. 5.400 13. 5.4130 5.4140 5.4130
14. 5.325 14. 5.3340 5.3350 5.3340
15. 5.250 15. 5.2630 5.2630 5.2620
16. 5.175 16. 5.1820 5.1830 5.1830
17. 5.100 17. 5.1100 5.1100 5.1100
I Fed single phase AC supply on R-N YoN B-N ron yon bon
, RIY 18 phase of HV side and
224.8 176.8 48.0 37.40 29.55 7.92
I measured induced voltge of all
109.0 224.7 114.7 18.32 37.40 19.02
I the phases of HVIL V side.
46.7 177.6 224.7 7.78 29.58 37.40
Fed 41 OV three phase AC R-N YN B-N r-n yon bon
supply on HV IL V side and kept
1.86mA 1.23mA i.80mA 38.8mA 25.3mA 37.3mA
the other winding open and
measured the current.
Fed 41 OV three phase AC supply on R-N YoN
bfV side and LV winding short
circuited and measured the current 6.68A 6.67A 6.75A
on HV side.
,"--, ~
Winding temperature = 26C
Tap No. HV Winding (mQ) LV winding ( mQ )
Scope make TRM-1 04 kit, R-N Y-N B-N r-n y-n b-n
S.NO. 117PS292 5 666.7 660.5 664.2 11.655 11.553 11.546
3 phase 418V supply ted to HV and B & b shorted ...
". Q Y-y = 348V A-r= 348V
A-y= 388V
Y-r = 388V . r
8-n = 40.3V N-n = 201.5 N
Vector group Y Nyno Is confirmed
Top oil temperature; 35
C, Eitel make kit, S.NO. 2006191, Range 0-12KV
Conti guration Measured Values
Measure Mode ~ Operation Test Voltage Capacitance (pf) Tan Delta %
HV to Tank GST 5 KVrms 9243 0.374
10 KVrms 9244 0.365
LV to Tank GST 5 KVrms 6177 0.334
10 KVrms 6177 0.324
HV-LV UST 5 KVrms 14380 0.365
10 KVrms 14380 0.354
Bushing On Terminals Measured Values
Test Voltage Capacitance (pf) Tan Delta %
1U Bushing 5 KVrms 219.6 0.620
(Sr. No. 6707210)
10 KVrms 219.6 0.607
1V Bushing 5 KVrms 215.9 0.583
(Sr. No. 6707222) 10 KVrms 215.9 0.573
1W Bushing 5 KVrms 213.6 0.600
(Sr. No. 6707206) 10 KVrms 213.5 0.595
THROUGH STABILITY TEST ABB make differential reay type RET- 541
IA = 6.2A 18 = 6.2A Ie = 6.3A la=38.7A Ib=38.8A I
Diff. currents la = Ib = Ie = 0
OL TC FUNCTIONAL TEST. i) Local operation of. OL TC checked along with operation of step by
step contactor, raise & lower limit switches and tap changer in
progress etc. ii) OL TC digital position indicator checked
TRIP & ANNUNCIATION All alarms and trip sionai checked and operating properly.
Tripping through protection schemes, WIT, OTT, PRY, BR and all the
alarms/annunciation at RTCC and C&R Panel checked and found in
working order. Functioning of oil & winding temperature repeaters and
tap change indicator checked and found in order.
COOLER CONTROL CKT. Auto / manual control of main Istand by pumps and annunciation
checked and found to b in working order.
BREAKER TRIP & Tripping of breaker through transformer protection checked
SUPERVISION independently for both trip coils . Trip ckt supervision in pre & post
closing mode also checked.
OIL TEST P =29.9x101Z0 cm t a ~ = .0019 ppm=11 ,BDV =68 I FP =156
2.3.2) OIL TESTS.
The oil is tested for electrical properties, DGA analysis and furan content analysis
periodically to access condition of insulation. The periodicity of oil test for electrical properties
and DGA is two years for transformers upto 10 years of life and annually thereafter. The furan
test is got conducted based on condition of transformer.
i) Common Oil Tests AnalysiS.
Condition assessment of liquid insulation can be made by periodiC testing of oil for tests
like Interfacial Tension, Flash Point, Neutralization Value, Electric Strength, Dielectric Dissipation
Factor, Specific Resistance. Water Content and Sludge Content. The permissible values of these
tests shall be as per 15-335: 1993 for New Oil and as per 15-1866: 2000 for unused oil filled in
T/F and oil in use in TIF
ii) Dissolved Gas Analysis
Electrical faults such as short circuits. loss of cooling. over loading and over
voltages lead to the de-composition of cellulose forming hydrogen(Hv. methane (CH
). acetylene (C
), carbon monoxide (CO) and carbon dioxide( CO
) as gaseous
products. These gases are ultimately absorbed in the insulating oil in transformer. The analysis of
dissolved gasses in insulation oil is useful not only to detect electrical faults but also to determine
hot spot temperatures and their origin. The relative amount of these gases are generally
determined by gas chromatography. The technique is well established and has been practiced for
years in most power supply stations for assessing the internal condition of transformers. The
i:lterpretation of results of dissolved gases based upon key gas method and various ratio
methods AS PER IS: 10593: 1992 I I EC-599 and Roger's Ratio help3 to pin point exact nature
of fault in the transformer.
iii) Furan Analysis.
The condition of paper insulation can be assessed in three ways a) by analyzing
dissolved carbon oxides. b) measurement of degree of polymerization, c) furan analysis. The
measurement of degree of polymerization suffers from draw backs regarding difficulty !n collecting
paper samples, location specific nature of test and complexity & accuracy factors involved in test.
The overheating of solid insulation leads to cellulose decomposition and evolution of CO and CO
gases. These two gases are also produced during thermal decomposition of oil. Therefore,
measurement of CO and CO
cannot be used"as an un-ambiguous indication of paper degradatiqn.
The furan test overcomes much of these difficulties as the test is simnar to the
conventional oGA test with which utilities are familiar. The ageing of paper produces several oil
soluble by products most predominant the furanoid compound (FFA). The furan analYSis by
periodical sampling of oil of the old transformers helps to determine condition of paper insulation of
the winding and leads. The analYSis of furan results can be done based on S.D. Myers Inc.
experience and CPR I experience.
Other tests like partial discharge, frequency response analysis and recovery voltage
measurement and infrared thermography are not most common tests. However, these may provide
very useful information in some of the cases. These tests and oiVelectrical tests may be
complementary to each other.
Despite effective operation & maintenance preventive measures and maintenance
based on condition monitoring through various electrical, oil & other tests suggested above, any
fault/abnormality developed during intervening period of two consecutive periodical testing schedule
remain undetected. To arrest such type of abnormality at an early stage the sensitl\le protection
scheme and annunciation system is required to be provided so that extent of damage may be
reduced. The detail of protection & annunciation system required to be provided for effective
protection of transformer and periodical testing thereof Is as under:-
The protection schemes, control system and indication/alarm schemes are tested during
pre-commissioning as well as periodically to ensure security and reliability of operation during any
fault/abnormality. Unreliable protection tystem and annunciation schemes will let the
abnormalitylfault go un-detected. The periodicity of testing for protection and annunciation system
is half yearly in BBMB. .
I) Protection Scheme.
Transformer faults can be broadly classified in the following categories:-
a) Winding & terminal faults
b) On load tap changer faults
c) Core faults
d) Tank & tank accessories faults
e) Abnormal operating conditions
f) Sustained or un cleared faults
Transformer protection will cater to the requirement of these faults.The practices in
application of protective relay schemes vary widely depending upon the importance, capacity,
operating voltage, types& number of windings and methods of neutral earthing etc. The following
relays and protective devices are provided for the protection of 100MVA capacity trensformer.
These are periodically tested for their operation just above the threshold value and restraint below
the threshold value. The safeguards provided by these are given below against each:
S.No. Description Safeguard provided against.
a) Percentage bias differential relay. All types of internal electric faults.
Circulating current high impedance
restricted earth fault.
c) IDMT over current relay
(HV & LV)
d} . Over fluxing relay.
e) Definite time High set
over cu rrent relay.
f) LBB relay.
g) Neutral displacement relay
(for 400/220/33 KV TIF only)
h) Buchholz relay/OSR
i) Oil temperature trip device.
j) Winding temperature trip
k) Pressure relief device.
I) Master trip relay.
Only internal earth fault with very high sensitivity
to detect earth fault even at incipient
Over load and back up protection for faults.
Damage/over heating of core due to over
Stressing of transformer due to prolonged un-
cleared through faults.
StreSSing due to un-cleared electric fault as a
result of HV breaker stuck up.
Earth fault in tertiary winding.
Gas formation due to internal fault and hot spot
Over heating due to various reasons.
Over heating due to various reasons
associated with load current
Excessive gas pressure due to severe internal
Low current rating and insufficient number of
contacts of aforesaid protection relays.
Magnetic oil gauge
lightening Arrestors
Nitrogen injection system
lightning / switching surges
Fire protection
With the introduction of numerical relays in the system we have added advantage in
respect of self diagnostic feature, disturbance/event recording, front & rear port communication,
time synchronization. accuracy & sensitivity and other in-built features i.e. LBB, High set over
current, over fluxing etc. These features in numerical relays help in precise analysis of fault and
reduce outage time. The instance of frequent tripping of 4OO1220KV leT Bank1 at 400KV GSS,
BBMB, Panipat in the absence of numerical relay is detailed under Case-6.
ii) Alarm and Indication Scheme.
Alarm and indication scheme is very essential for the operator to know any abnormality in
the system. The trip and non trip alarms and status indications to be checked are given below:-
a) Trip alarm. Relating to all the relays and protective devices.
b) Non Trip alarm. Relating to buchholzlOSR relay, OTI, WTI,
OLTC and
and DC supply failure etc.
c) Indications.
2.4.2 Diagnostics
PRO, MOLG, non flow of oil, abnormality in
fan/pump control system. AC
Status indication of OL TC tap positions, OL TC
operation in progress. fan/pump on, DC & AC
controlling breakers, isolators etc.
Relay alarm & facia indications appearing on C&R Panel should be thoroughly
investigated and remedial measures taken to eliminate the problems. Some of the important
events ha b d' h f 10 b -
ve een analyse
Int e 0'
wing ta Ie.
Symptoms Effect Possible causes Items to be checked
Differential relay. Tripping of the Intemal fault Collect gas from Buchholz's relay & get it
Buchholz's relay & transformer tested for DGA to know nature/cause of
PRD operates fault. Perform electrical tests on
simultaneously. transformer. Compare values & energize
transformer after eliminating the cause of
Only Differential -do- -In zone fault Maintain shut-down status and carry out
relay operates. transformer tests as described above to
check healthiness of transformer.
-Mal-operation In case of mal-operation, check relay
alongwith its control circuit and take action
Only Buchholz's -do- Check gas in the relay.
relay operates. -Internal fault Maintain shut down status, If gas is present,
then take action cis.per sr. No.1 above
-Mal operation If there is.' no gas investigate reasons of
ma!:Qperation & take action accordingly ..
Only PRO -do- Faulty operation Maintain shut down status.
(Pressure relief Investigate reasons to know cause of
device) operates. operation of relay & take action accordingly.
Protective device -do- Defect in OL TC Maintain shut down status. Check tripping
for OLTC mechanism. Remove & check diverter
switch & take action as required.
Problem in motor Motor drive Problem in electrical Check electrical circuit & take action
drive of OL TC inoperative circuit. accordingly.
Problem in motor Check & set right the problem.
Drive keeps on drive circuit.
Motor protective Problem in motor Check push button.
switch trips stop button.
during operation
of motor drive. Motor
current Investigate reasons of overload & take
action accordingly.
Buchholz's Alarm
Over current!earth Tripping
fault relay operates. transformer
-Trapped air.
-Built up of gas
of Extemal fault!
De-en6rgize the transforme,.
Perform gas analysis & eiectricai t 8 s i ~
Analyze the results & take action
Investigate reasons of relay operation &
take action accordinQIy.
Low indication on Low oil
Magnetic oil gauge alarm.
level -Drop in temperature
Top up oil & take measures to plug leakage.
WTI/OTI (Abnormal AlarmlTripping
rise in temperature)
-Incorrect setting!
De-energize transformer in case of alarm &
maintain shutdown status in case of
Check the instruments. Adjust & repair as
-Failure of cooling Chec.k cooling system & take action
system. accordingly.
-Sludge in oil.
-Short circuited core.
-Transformer over-
Filter oil to eliminate sludge.
Check exciting current and no load loss. If
high, inspect core and take remedial
measures as required.
Reduce the load within limit.
Although sensitive protection system is provided to detect fault developed in the
transformer during the intervening period of two consecutive periodical tests but this stage of fault
detection is much delayed vis-a-vis incipient stage. The only answer to fault detection at incipient
stage is on line monitoring of the transformer which will bring revolution in this direction and the
damage rate shall be reduced to minimum. On line monitoring of transformer is in fact a future
trend but critical transformers in the system which require frequent DGA testing may be provided
with very simple & basic on line Hydrogen, monitor.
Hydrogen is one of the fault gases produced by most types of faults. Hydrogen monitor
consists of a selected permeable membrane coupled with an electrochemical gas <;foetor. H2
dissolved in oil, permeate through membrane and react with O
from ambient air. This reaction
~ generates electrical current that is measured as a voltage drop across a load resistor. Signal
generated by detector is proportional to the rate at switch H2 permeates through membrane. This
rate varies according to concentration of dissolved combustible gases and temperature. The
electronic circuit, which amplifies detector signal, introduces an appropriate temperature
correction to produce a reading solely based on concentration of dissolved combustible gases.
H2 monitors are already in use. These are early warning devices that alert maintenance
personnel regarding developing fault conditions that could lead to failures. In time, as the need
for monitoring evolved, the manufacturer of H2 monitor developed the product to improve its
operational reliability, under varying thermal and environmental conditions and included CO as an
additional monitored fault gas. H2 and CO have been selected as they provide very early signs of
degradation in insulation system. H2 Is indicative of degradation of oil and CO that of solid
ceUulose insulation.
The safety precautions for safety of human beings the equipment are of
paramount importance and hence Jist of DO's and DON'T's for safety is given as under;
Disconnect the unit & ground it on both sides. Fulfill the conditions of PTW before
working on it.
notice board 'MAN AT WORK' in control room as well as at site of work.
3. Switch off auxiliary supplies to OL Te panels, Cooler control panel & Relay panel.
4. Attach the caution tags" DO NOT OPERATE THE SWITCHES" while working on
energized unit.
5. Make sure that fire fighting equipments are availabfe at work place & adjacent to the
6. Check and thoroughly investigate the transformer whenever any alarm or protection
7. Insulated oil and insulation of winding and connections are inflammable, watch for fire
8. While working inside the transformer, frame comprehensive guidelines and follow these
9. Don't overload the transformer beyond values specified in IS: 6600.
10. Don't parallel transformers which do not fulfill required conditions of paralleling.
11. Don't handle the off circuit tap switch (if any) when the transfonner is energized.
12. Don't energize the transfo'rmer unless the off circuit tap switch (if any) handle in locked
13. Don't leave off circuit tap switch (if any) handle unlocked.
14. Don't operate OL Te unless its chamber is filled with oil.
15. Don't meddle with the protection circuits.
16. Don't change the aetting of the relays or alarm unless authorized to do.
17. Do not allow oil temperature to exceed beyond the limits prescribed by the
manufacturer/quality of oil during drying out of transfofT]ler oil.
18. Don't s.moke on or near the transformer.
19. Don't allow unauthorized entry near the transformer.
20. Don't keep marshalling box doors open & they must be locked.
21. Don't use low capacity lifting lacks for jacking.
22. Don't use fibrous cleaning material as it can deteriorate oil when mixed with it.
23. Don't mix the oil, unless it contonns to specifications mentioned in the mtc. Instruction
24. Don't open the transformer during adverse weather conditions such as rainy, cloudy &
foggy weather.
The periodical carrying out of electrical tests & oil tests stated above and provision of
sensitive & reliable protection & annunciation system help to know the condition of transfonner
and nature of abnormality /fault, if any, present in the transfonner , at an early stage. Based on
periodic electrical, oil and other tests abnormalities were noticed in 5 NO. transformers i.e. 2 NO.
111132KV. 32.5MVA AJstom make unit transformers one each at Kotla and Ganguwal Power
Houses, 1 NO. 175MVA, 111220KV Transformer at Bhakra Right Bank. 1 No. 12.5116MVA
,i32133KV Transfonners at Ganguwal Power House and 1 NO.100MVA, 220/66KV Transformers
at Rohtak Road Delhi Substation . Remedial measures adopted to correct the abnormalities are
detailed here under:
On 23-9-2002. 11/132 KV, 32.5 MVA unit trdllsrormer No.2 at Kolla Power House
showed abnormality with the operation of buchholz relay. The gas was .noticed in the relay. All the
electrical and oil tests carried out at site were found to be normal except LT windings (delta
connection) resistance and oil DGA test with following values:
LT Resistance in ohms.
r-y = 0.0075 y-b= 0.00541 b-r= 0.0075
DGA Result.
Total Gas = 2.4 ml.
= 252 ppm, C
= 464 ppm, C
= 547 ppm.
= 3.85 ppm. H
= 2773 ppm ~ = 796 ppm.
a) L T resistance clearly indicated loose connection outside the delta in the portion
extending from delta formation point of r1-y2 to red phase bushing.
b) The dissolved gas composition based on Key Gas Method indicated hot spot in oil of
temperature range greater than 150 C and based on Rogers ratio table showed slight
overheating in the range of 200 C to 300' C .
Physical Inspection.
On physical inspection of transformer through l T side inspection window, the nuts of the
bolted joint connecting copper strip of R-phase delta formation bus with copper strip extension to
Red phase bushing were found loose and check nuts of these nuts were found missing . After
tightening the nuts and providing check nuts, the L T resistance of all the three phases was
identical and comparable with manufacturers works test results. Subsequent DGA results of oil
were found in permissible limits
DUring periodical testing of 11/132 KV, 32.5 MVA, unit transformer No.2, on dated
4.10.2002 at Ganguwal Power House, all the electrical tests and oil tests were found in order
except L T winding resistance which was recorded as under:-
L T Resistance in ohms.
r-y =0.00535 y-b=O.OOn b-r = 0.00537
The values of L T windings resistance indicated loose connection inside the delta in the
portion of Yellow phase winding extending both sides to meet at Yellow and Blue phase delta
formation buses.
Physical Inspection.
On physical inspection by entering into the transformer tank through inspection window,
apparently the nuts of bolted joints were not noticed to be loose, however on tightening the nuts
the L T resistance became within limits. In this case, the loose connection was detected before the
formation of hot spot and decomposition of oil into gases:
During trial run of 175MVA, 11/220KV Mis Transforrnator Co. JSC, Togliatti - Russia
transformer installed at Bhakra Right Bank in the year 1996, it was observed with infrared system
that on full load the tank body temperature at some locations such as flanges of bell cover,
bottom covers and between the plates of L.V. turrets went up to 120C. All other parameters
were well within the permissible limits. After detailed discussions, the following points were
considered and implemented.
Additional non-magnetic shunts were provided on the inner side of L.V. side tank
Replacement of magnetic steel plates with non magnetic stainless steel plates
between L.V. turrets.
Replacement of M.S. Nuts and Bolts with non-magnetic stainless steel Nuts and
bolts on L.V. turrets.
Increased the gap between L.V. side bus-bar and tank wall and by changing the
L.V. side leads with copper bus bar.
Atter implementing the above modifications, the temperature on full load came down to about 45-
50C .
. Case-4.
During periodical testing of transformer oil of 12.5116MVA, 132133KV, N.EJ. make
transformer (commissioned during 1976) the value of was found to be 0.0285 x 10
Ohm-cm at 90C as against required minimum value of 0.1 x 10 2 Ohm-cm. Further the IR values
of transformer were also very low, i.e., 25 mega Ohm at 45C and remained almost constant (i.e.,
for 15, 60 and 600 Sec.), even there was no improvement in IR value despite carrying-out
dehydration of transformer oil for many days. It was therefore decided to replace the transformer
oil after carrying out induction heating. There was appreciable improvement in the oil test results.
Tan delta, resistivity and IR values of the transformers were, 0.028 at 90C, 1.54x10
Ohm cm at
90C and about 300 mega ohm at 4yOC. The transformer is giving trouble free service thereafter.

During periodical oil testing of 220/66I33KV, 100MVA Crompton Make Transformer
commissioned during 1982 at 220KV GSS, BBMB, Rohtak Road Delhi, the test values in re$pect
of tan delta, specific resistivity, furan contents & IR value were found to be 0.29 at 90C, .06x10
Ohm cm at 90 C, 6764 ppb & 40 mega ohm at 25C respectively. The transformer was
overhauled during last quarter of 2003 and complete oil was replaced with new oil. The test
results recorded on 29.11.2004 & 9.5.2005 were found to be satisfactory. The values of tan
delta, resistivity, furan contents & IR values after one & half year of overhauling i.e. on 9.5.2005
were found to be .0032 at 90 C, 21x1012 Ohm cm at 90 C, untraceable and about 500 mega
ohm at 50C respectively.
Frequent tripping of 4001220KV 450MVA ICT Bank-1 at 400KV
400/220KV 450MVA ICT Bank-1 at 400KV GGG Panipat tripped number of times during
the period 22.8.03 to 27.8.03. The transformer has been provided with percentage
differential relay type DTH-32, high impedance current Circulating relay type CAG-34, neutral
displacement relay, VOG-14, back up over current protection, over fluxing relay and other
protective devices. In some of the trippings during the aforesaid period only CAG-34 operated
and in other trippings both CAG-34 & DTH-31 both operated. After every tripping the transformer
was thorough tested but no abnormality was noticed. Finally a numerical relay type KBCH was
installed in series with DTH-32 relay at about 7.30PM on 27.8.03. Transformer again tripped on
28.8.03 at 00.30 Hrs. (Le. first time after installation of KBCH) and the current recorded in the
KBCH relay were analyzed. From the current recorded it was inferred that 220KV Y phase CT
current vanished and resulted unbalance current for operation of DTH-32 & CAG-34 relay. CAG-
34 being more sensitive than DTH-32 relay operated alone those cases when unbalance current
was below thresh hold level of DTH-32. The 220KV Y phase CT was replaced and thereafter no
tripping has occurred on ICT Bank-1. Thus numerical relay helps in analysis of trippings and
reduces outage time of the transformer.
Based on the test results of oil, following 4 No. Power Transformers have been identified
by BBMB to chalk out contingency plan so that any emergent situation arising due to failure of
any of the given below transformers may be dealt without any repercussion thereof.
S.No. Life of Transformer Furan result (PPS) Other oil
1. 40 years
2. 39 years
3. 23 years
4...... 23 years
Overhauling done
8.8 (Tan
& Resistivity .066)
Dec. 2003
Due to effective maintenance and monitoring based on periodical testing of transformers,
several transformers in BBMB, as per details below, are giving satisfactory service even after
completion of their useful life of 35years.
uprating of
Ufe of Transformer
35-40 years
40-45 years
45-50 years
>:>v years
No. of Transformer
4 (replaced due to
the generators)
In nut shell, we can say that effective operation & maintenance,
appropriate preventive measures, condition monitoring based on periodic electric
& oil tests and sensitive protection & annunciation system enhance reliability of
transformer and detect abnormality, if any, at an early stage which limit the
extent of damage to the transformer & thus extends life of the transformer. Last
but not the least is safety of human beings and the equipment which is of
paramount importance.
There is a great need to be effective in detecting and locating
faults on power transformers in service. Any sudden failure may
not only cause loss to the transmission utility but can also cause
load loss to consumers apart from causing Grid disturbances
There have been some noticeable failures of Power Transformers
in the Northern Region in last two years. With these handful
cases we have tried to analysis the genesis to locate the root of
the problem of ICTs failure in Northern region.
1. Failure or ICT- I (315 MVA) At 400 KV Bamnauti on
On 09.02.2008 ICT-I at Bamnauli sis was operating at load of
85 MW. Buchhloz Alarm appeared at 9.30 AM and at 10.30
AM the transformer was taken on shut down and all the tests
such as capacitance and tan delta, insulation resistance value,
DC resistance,.. tum ratio, magnetic balance and magnetizing
current were conducted. Simultaneously oil sample was sent to
CPR I for DGA Analysis. After examining the report, and the
test results found to be normal, the transformer was put on no
load at 5.40 PM on 11.2.2008.0n 11.02.2008 at 8.41 PM with
sudden blast all the out going feeders of 400KV SiSto were
tripped along with tripping of transformer breakers and the
transformer got fire. The fire fighting system was operated
immediately, however, due to high intensity fire the same could
not be quenched immediately. Thereafter the fire tenders were
called from different localities including MES, I.P. Delhi fire
and fire was eventually controJled at 3.00 AM on 12.02.2008.
Photo of damaged transformer
This transformer manufactured in 1993. and commissioned in
2000 at At 400 K V Bamnauli sub station of OTL. On 14.05.2004
tan delta & capacitance of 220 kV and 400 kV bushing were
measured by MIs CPRI. The measurement of capacitance and
tan delta of bushings and windings was again carried out on
09.12.2006. One 52 kV bushing of B phase was replaced on
14-12-2006 No other major repair work of transforn:ter was
carried out since commissioning.
The failure of transformer is attributed to insulation failure (inter-
tum insulation failure) and I defonnation in windings. The
deformation in winding indicates flow of high short circuit
current for longer duration. There is no Disturbance Recorder
at 400kV substations to record such system faults and its
2. Failure ofICT-1 at 8allabhgarh sis on 1.5.2006 and Grid
disturbance thereafter:
A 400/220 kV, ICT -I at Ballabhgarh failed due to internal
arcing inside Y phase OLTC chamber. Fire broke out due to
splashing out of large quantity of oil from OLTC top cover.
Operation of over current protection on 220KV side indicated
that fault was almost below nOKV bushings and large portion
of winding was not involved.
The Grid disturbance was triggered at 18.l1hrs, by the failure
ofInter Connecting Transformer (lCT-I) at 400kV Ballabhgarh
sub station due to an internal Y-phase to ground fault. Restricted
'Earth Fault (REF) protection of the transformer operated and
the fault was cleared by opening of the breakers on both the
400 kV and 220 kV side. This ICT caught fire immediately and
the fire could not be controlled with the mulsifier system. Prior
to tripping of above ICT, 220 kV Samaypur-Dadri had tripped
from Samaypur end at 18.11 hrs (59ms before) on LZ 96.
An inquiry committee was constituted under the chairmanship
ofMS, NRPC and its reports states that at Hrs a
second fault on Blue phase (of RYB) to ground occurred, when
the fire damaged the Lightning Arrestor (LA) and the
transformer bushing on the 220 kV side of the ICT-I.The jumper
connecting the 220 kV side Circuit Breaker (CB) Blue phase
pole (top terminal) to the transformer bushing broke off from
the bushing terminal and fell down. This fallen conductor
connected to the top dead terminal of the vertical CB chamber
bridged the CB and created Blue phase -Earth fault from the
bottom live terminal of the CB to the operating mechanism
. chamber. The 220 kV side isolator had remained closed and
hence the broken terminal of CB had remained charged from
220 kV side.
The diagram is shown below: ==! /
-- II
= r ,1 /./
-"G:J" : ,,:.,
1-i r
The Blue phase to ground fault thus created falls within the
zone of protection (CT to CT) of the ICT#1 but outside the bus
bar protection zone of 220 kV Bus-2B at Samaypur (BBMB)
substation. Since the CB on either side of the transformer was
already open, the fault would have appeared as a case of Breaker
failure (Trip command & Current in the CB circuit) and would
have initiated back up tripping of all circuit breakers connected
to the 220 k V bus#2B at BBMB Samaypur Thus the fault would
have been cleared in 300 ms (max) as per protection scheme.
This protection however, could not be initiated since DC supply
was unavailable (operators switched off the DC supply to the
ICT#l bay at 18:15:11.710 Hrs). The .fault was now an
undetected bus fault (but beyond the zone of bus bar differential
protection of220 kV Bus 2B) at 220kV BBMB Samaypur and
had to be cleared by remote end ,back up protection in Zone-li.
This is a case of ineffectiveness of mulsifier system of the I CT.
The transformer got burnt due to inadequacy of its mulsifier
system. The committee recommended that there was a need to
review the design of the mulsifier system, including the
provision of Nitrogen injection, may be by a specialist
consultant. The nonns of separation and 1 or fire wall also may
be so reviewed, for suitable amendment if necessary in the safety
3. Failure of leT (No- IV) on 28.04.2006 and ICT -II on
9.5.2006 at Mandola sis:
These CGL make ICTs failed due to external faults. ICT-IV
failed due to failure ofY phase 220KV CVT in Bus-IV. Bus
Bar protection on 220K V side did not operate and hence could
not clear the fault. Fault current continued flowing for more
than 1050 ms through lCT and caused mechanical damaged!
failures of windings! core of the Transformer.
ICT-llalso failed due to winding insulation failure caused by
stressed due to external short circuit in the line. There was an
external fault prior to failure of the ICT. The line was being
restored, but since being a permanent fault , the line tripped
immediately on switch on to fault (SOT F) feature.
Electromagnetic forces due to fault current of the order 2500 A
might have resulted into stressing of winding insulator and
caused puncturing it. The inter fault resulted thereafter caused
operation of REF, differential and Buchhlolz relay. The
transfonner caught fire which could not be quenched by fire
fighting system.
4 Failure of ICT-I at Bawana sis on 29.5.08
ICT -I at Bawana sub station of DTL failed Due to tertiary
windings fault and bushing damage. The transformer caught
fir!' !lnrt thp fir!' rnl1lti not be quenched its fire fighting system.
On 29/0512008 315 MVA ICT No-l and 4 were running in
parallel connected with 220 kV bus and in feeding 240 MW
each. At 12:20 Hrs on 315 MVA, ICT-I Buchholtz alarm
appeared and same was reset, but as pre cautionary measure, it
was decided to check the ICT. Accordingly request was made
to system operation for shut down on ICT. After receipt of
NRLDC code at 13:05 hrs, the process of isolation started and
when 220 kV C.B of ICT-lswitched off, this resulted tripping
of ICT on differential relay with heavy sound.
On physical inspection it was observed that there was damage
00400 kV <OR" phase bushing including 220/33 kv bushings,
LAs ,CVTs, Marshaling box etc. The emulsifier system which
was on auto mode started functioning and in the mean time fire
tenders were called to extinguish the fire.
S. Failur-e of ICT at Greater Noida and Gorakhpllr in UP
ICT -II (325 MVA. 400/220 kV) at Greater Noida failed in
March 2008 due to internal fault and HV bushing failure. This
ICT at Greater Noida damaged completely on 400 kV side
bushing failure and in this case also the emulsifier could not
the fire and got melted in fire.
At Gorakhpur on 22nd June 2008, a 1996 manufactured ICT,
commissioned in 2006, failed on 33 kV tertiary bushing fault
at 0121 hrs on 2216/2008. The followings protection flag found
on relay panel: Differential-lII1 operated (Y Phase), PRY
kV side) operated, Buchholz trip & Alarm Operated, 86 trIp
relays both side.
It was observed that part of gasket between porcelain and
metallic part ofY- phase 33 kV terti;uy bushing had come out
from its position and oil was pouring heavily. Y Phase 52133KV
bushing found damaged. It was observed that lower porcelain
portion of bushing was broken and some part was lying on 33KV
side main tank junction box. Also flashing observed on lower
metallic flange of bushing.
6. Form.doa of excess gas in 4001220 kV ICT (315 MVA) at
It is observed that Main protection schemes for 220KV lines of
PSEB stations does not operate properly which results into
delayed fault clearing. It has been reported that during pa.st
couple of years there have been some incidents when fault In
220-kV transmission system of SEBs have caused tripping!
damage of 400/220 ICT's due to non-clearing of fault by their
protection system. In June '06 un-cleared fault in PSEB 220kV
Malerkotla SIS switchyard caused prolonged feeding of fault
pOWERGRID 315 MVA ICT, which subsequently resulted
in de .. elopment of DGA problem in the ICT. Later this
tranSfonner was shifted to other low loading location and a
new ICT was installed at Malerkotla.
The Electromagnetic forces due to fault currents flowing through
power transformers affect life of transfonners because of clamp
looseness, core overheating etc. This problem has also been
encountered at not only at Malerkotla, but also at Mandaula,
station. Matter has been ...
Protection Coordination Committee of NRPC and it is
recommended that Bus-Bar protection has to be provided in all
220kV Substations.
Analysis of failure of ICT in Northern Region
The leTs at BaUabgarh, Grater Noida, Mandola -IV, Gorakbpur
and Bamnauli sis failed and got completely damaged because
of non effectiveness ofmulsifier system. The proper functioning
of mulsifier system was also in question, particularly the
pressure of the hydrant and the choking of the nozzles. At Graeter
Noida , Ballaghgarh and Bamnauli, the nozzles, pipes and
associated equipment were also got melted in the fire, which
indicates that water supply was not adequate and was stopped
before fire quenching. At Bawana ,the impact of large that high
pressure pipe got sheared and all the pressure was
released from the damaged pipe instead of water getting sprayed
through high velocity jet on the transformer.
At Mandaula sis there have been more than 260 nos earth faults
on 220 kV system and similar is the case at Malerkotla where
faults at 220 kV system are passed on to 400 kV With each
earth fault causes stressing of transformer insulator and also
electro magnetic force at acting on different parts of the ICT
like core clamps arrangements, winding, bushing etc affects'
their mechanical integrity. Due to short circuit current there may
be damage of windings I core and may lead to fire. The thermal
and mechanical faults are causing damages to lCTs in Northern
Region. At present stage, utilities generally correlates with the
DGA results and trends, and this allows the detection of electrical
faults. For thermal and mechanical faults there is a requirement
for in-depth analysis and need for more effective tool to
understand significant improvement in the fault assessment
effort. Frequency Response Analysis (FRA) test should be
conducted to check mechanical integrity of ICTs in Northern
Regional on the ICTs which have also been subjected to short
circuit forces. FRA test should also to be conducted routinely
for ICTs wherever considerable short circuit current (more than
2KA) flows through transformer winding for through faults and
time duration more than 250 ms.
Suggestion for improvement
I. Review of Protection setting
Protection settings (Back up over current relays) for all
ICTs need to be reviewed by in a conservative manner in
view of the lCTs being subjected to a high frequency of
abnormal faults on the 220KV systems. Also the
replacement of conventional Electro-mechanical/static
relay by modem numerical relay could help in getting such
valuable information. The fault clearing time of protective
system needs to be reviewed. Adequate measures need to
be taken to protect the transformer from severe short circuit
II. Review of tbe adequacy and maintenance of fire
figbting and mulsifier system
Almost in all the case stated above the fire could not be
quenched by the mulsifier system of the ICTs .. There could
he maintenance lapse by some utilities and the nozzles
might have choked. Regular fire fighting Hydrant and high
pressure water (HPW )system operation checks are
mandatory to ensure reliable operation ofHPW pumps and'
DO operated HPW pump in case of auxiliary supply failure.
Actual operation of complete HPW system should be
checked annually during shut down. Manufacturers should
obtain the actual data from the site and review the adequacy
of the mulsifier system.
As per section 4 (0 ii (c ), electricity rules 1956, which are
being replaced by CEA ( measures related
to 'the transformers shall be protected by an
automatic high velocity water supply or by carbon dioxide
or BCF ( Bromochlorodifluromethane) or BTM
(Bromotrifluromethane) fixed installation system or
Nitrogen injection and drain method)."
In an event of bushing getting damaged etc., Nitrogen
injection would provide dampening effect on the oil, which
comes out from the Transformer.
III. Testing of short circuit withstand level of Power
The short circuit withstand level of transformer is being
verified by empirical calculations. No short circuit test has
been conducted on the transformer due to non-availability
of indigenous testing facility. Therefore, such failures raise
doubts about the short circuit withstand capability of
transformers being manufactured in the country.
Transformers being procured need to be type tested for
external short circuits as per IEC-60076-5 to ensure
mechanical integrity for electromagnetic forces.
Frequency Response Analysis (FRA) test should also to
be conducted routinely for lCTs wherever considerable
short circuit current (more than 2KA) flows through
transformer winding for through faults and time duration
more than 250 ms particularly ICTs connected to PSEB
and UP system.
IV. Proper storage of ICT at site:
The transformer after receiving at site is generally kept
Nitrogen filled under prescribed pressure for a maximum
period of 5-6 months and it is not advisable to store the
transformer beyond such a long period. In such case, it is
recommended that transformer should be filled with oil
and oil to be filtered at regular interval. During that period
regular testing like Capacitance and Tan Delta, Break Down
value of oil, DGA of oil, Furan test and degree of
polymerisation should be conducted as per manufacturer
V. On Line Tap Choger (OLTC) Inspection:
As OLTC operations are limited in Indian power system
due to highly conservative approach by Power Engineers.
It is suggested to carry out inspection and cleaning of
diverter switch contacts every 5 yrs to ensure healthiness
of tapping contacts or after 50,000 operations as per
recommendation of the manufacturer ..
VI. On line monitoring for Transformers:
Transformers Nursing Unit for monitoring oniine DGf\.,
Bushing Tan delta, partial discharge etc, may be installed
at stations where 3-4 transformers are installed or for
critical units. The possibility of carrying out on line
filtration, dry out and degassing ofTransfonners for critical
units which can not be taken out of service dud to system
constraints may be explored.
1. CEA report on ICT Failure at Bamnauli.
2. Grid incident report of dated 1
May 2006.
3. DTL report on Bawaoa ICT faiJure
A. ~ ~ ~ ~ ~ ~ ~ . : : : f : - ~ ~ T...'!l!lC'!.. ;;;: !C'Ts at Greater Noida and
5. Other inputs from POWERGRID, and state utilities
6. Different OCC proceedings of NRPC
R. G Yadav, ED (OS), G S. Sarkar, AGM (OS),
V. K. Bhaskar, Chief Manager (OS), Gunjan Agrawal, Dy. Mgr (OS)
1.1 Transformer is the single costliest equipment in a
substation which is quite reliable & efficient piece of
equipment. On Load Tap Changer (OLTC) is the only
dynamic critical component in an otherwise static piece
of equipment and hence reliability of this component
assumes a high degree of importance. Large power
transformers are expensive, critical elements in any utility
transmission systems. Any failure in these equipments
may lead to problem in evacuation of POWER affecting
the power supply to consumers lutilities as well as huge
financial losses. Any major Transformer failure may take
restoration time varying from six months to more than a
1.2 About 325 numbers of Power transformers (including
72 HVDC converter transformers) having OLTC are under
operation in POWERGRID at 116 substations located
all over India. In POWERGRID, Transformers have
suffered failures as shown in Exhibit-I. As seen from the
graph, the major reasons of failure are Bushing failure
(30 %), Winding failure (20 %) & On-Line Tap Changer
(OLTC) failure (20 %). Our transformers are fitted with
OLTC of different make like ABB, BHEL& Easun MR
(under technical collaboration with MR). All the OLTCs
(except original MR, Germany) have suffered failure in'
one component or other.
Component WISe Failure in Transformer

o 5 10 15 20 25 30 35
% ofT otal FaikJre
1.3 Diverter Switch in a separate tank, below which tap
selector, Motor drive unit along with drive shafts & bevel
gear and Protective (Oil surge) Relay together make up
the On-line Tap changer (OLTC) as shown in Exhibit II
& III. OLTC consists of about 1500 components and
about 19 are critical component, failure of which may
lead to
complete OLTC failure. OLTC failures are dominated
by faults of a mechanical nature (spring, bearing, shaft,
iU(Xi1C1uj:'Hl),lvlivwc;u uy fault of an electrical nature
such as coking of contact, burning of transition resistors
and dielectric insulation problems.
Exhibit III:
Diverter Switch
Exbibit II: OLTC
1.4 In POWERGRlD, most of the Transformers are having
On Load Tap Changers. HVDC Converter Transformers
. are ilaving large number of OLTC operations due to its
inherent requirement (operation controlled by reactive
power controllers) however in Auto Transformers, OLTC
operations are very few. The failure results however
indicate that failures of OLTC in Auto Transformers are
quite large. Most of these failures have been observed
in recently commissioned Transformers. In some cases,
failure ofOLTC has resulted into failure of winding also.
In this paper, POWERGRID experience of failure of
Power Transformers due to problem/failure in On load
Tap Changer has been enumerated.
2.0 Condition monitoring of OLTC
Due to recent problems faced by us, the reliability of
OLTC became an issue, few questions which arose in
our minds:
~ h o u l we draw oil for DGA on regular basis?
What gasses might be the key indicators of
deteriorating contacts?
Should we check all OLTCs just with thermography?
Should PO detection be used on OLTCs?
Another consideration is; What is detectable first and
which piece of test equipment should be used?
The first level of detection is sonic/ultrasonic noise
. analysis, (arcing/unusual sounds) which can be detected
by listening with a fault detector. Burning contacts, partial
discharges and signature patterns that are abnonnal are
the early stages of an OLTC degrading.
The second level of detection is through the analysis of
the oil. Although arcing is expected in the OLTC, and
arcing gasses are always present at some level, hot metal
gasses should not be present in large quantities. In addition
to evaluating OLTC oil for gasses, oil quality tests should
also be perfonned (dielectric strength (BDV), power
factor, etc.) ..
The third level of detection is through thennal analysis.
Our experience has been. that a profile of + 1 OC or more
on the OLTC tank is a critical situation, and a positive
differential between the main tank to the OLTC
compartment is a trigger for further iilVestigatio(l.
However all these detection techniques already in use in
many utilities have some limitations. We have taken oil
sample for DGA from diverter switch chambers from
few in-service transformers. Interpretation of OGA of
diverter switch chambers is a ticklish issue and needs
engineering judgment / experience as many fault gases
are produced during switching operations. We ensure
BOV more than 50 and moisture content less than 30
ppm in the OLTC chamber under operation. We carryout
infrared thennography of complete transformer including
OLTC chamber during annual maintenance.
3.0 Failure of transformers due to OLTC
3.1 Failures in Old transformers: The OLTC failures in
old transformers are few. First OLTC failure in
POWERGRID as observed physically occurred in tap
selector wherein Five out of seven bakelite supporting
structure for tap selector lead connection were found
brokeR as shown in Exhibit IV & V.
In another transfonner, there was blast inside the OLTC
chamber & diverter switch was thrown out by breaking
top plate on which diverter switch gear assembly is
mounted & oil caught fire causing damage of 220 kV
Bushing & nearby cables, OLTC conservator tank. OLTC
operating MS, OLTC & main Buchholz Relays. (Exhibit
VI & VII).
Failure of Aluminium casting below di.ierte; switch
through which tap selector is attached . Isor/?", been
Exhibit IV
Exhibit V
Exhibit VI
3.2 in newly commissioned transformer: Recently
we have few OLTC failures in new transfonners during
commissioning or with in a short span (warrantee period)
after commissioning which is a cause of concern.
In one of transformer, 0 LTC failed during tap-changing
operation under charged (NO-LOAD) condition. In this
case, Bakelite moving Contact B!"idge engaging with fixed
contact was found broken and lying <)Il the transformer
floor inside the tank as shown in Exhibit VIII & IX.
Exhibit VIII
In another transfonner, OLTC failed during tap operation
under load. In this case also bakelite Top row Contact
Bridge (connected to Odd numbered leads) found broken
into pieces. (Exhibit X & Xl).
Exhibit :x.
Exhibit XI
The components involved in failures in OLTC include
bakelite contact bridge (may be due to inadequate contact
pressure), dielectric failure of Tap selector FRP Bar &
supporting structure for tap leads (may be due to defective
material), dielectric failure of Diverter Switch FRP Bar
(due to moisture ingress inside paper based support bars),
oil surge relay (mal-operation due to faulty reed relay),
bevel gear assembly (may be due to misalignment) etc.
Each failure has been analyzed by a committee in
association with manufacturer of the OLTC to arrive at
a root cause cf failure.
4.0 Preventive & Corrective Measures:
In view of the failures in OLTCs of newly commissioned
transformers in recent past, detailed discussions with
OLTC manufacturer were held as well as factories of
various OLTC manufacturer were visited to review the
various processes in manufacturing & testing. It has also
been observed that OLTC being manufactured under
technical know-how ofthe principals which with passage
of time checking of critical technical requirement got
diluted. The OLTC manufacturers were advised to adhere
to the systems & procedures being adopted by their
principals and get the jigs & fixtures required for accuracy
in manufacturing and improving the reliability of the
Based on the discussions with the transformer
manufacturers & OLTC manufacturers, following actions
are being implemented by OLTC manufacturer and
respective Sites during manufacturing installation, pre-
commissioning as corrective measures to minimize the
probability of further failures:
1. The OLTC manufacturers has been asked to carryout the
following as routine tests on OLTC before getting installed
on transformer:
SOOO mechanical operations on OLTC
Continuity tests on all taps
Contact resistance on all taps may be measured and
compared before & after operations
Ensure the tightness of connecting lead terminals by
torque wrench
Oval tube on which contact bridge is supported may
be checked for its strength;sizes, straigh!!!e::, ' : " . ' : : : ~ >
and other dimensions by using special fixtures at the
2. OLTC manufacturers to arrange for training programme
for Site erection supervisors at manufacturer's works on
regular basis to familiarize the user with erection & testing
of o LTC.
3. For future commissioning of Transformers, following
checks have been recommended to identify the defecti ve
OLTC before commissioning:
a. With the transformer un-energised, one complete
cycles of operations (a cycle of operation goes from
one end of the tapping range to the other, and back
again). Check continuity of winding during this test.
Ensure that the voltmeter needle does not deflect to
b. With the transformer un-energised, and with the
auxiliary voltage reduced to 85% of its rated value,
one complete cycle of operation.
c. With the transformer energized at rated voltage and
frequency at no load, one complete cycle of operation.
5.0 Conclusion
Power Grid Corporation of India Ltd (POWERGRID),
a Central Transmission Utility (CTU), is engaged in bulk
power transmission from Central Sector Generating
Stations to Load Centers across the country with
Reliability, Security and Economy on sound commercial
principles. Recent failures of OLTC in newly
commissioned transformers have shaken the confidence
of our operations staff and in few cases they have stopped
operation of OLTC on load due to fear of failure which
is a major concern for us. OLTC failures have led to
winding failures due to which transformer have gone under
forced outage for months as it involved repair at
manufacturer's works or diversion from other construction
projects. Such outage is costing dearly to POWERGRID
as we are loosing financially for the equipment outage.
OLTC & Transformer manufacturers must take up
systematic failure analysis and corrective measures to
be implemented to improve the reliability in the operation
The authors are thankful to the POWERGRID
Management for publication of this paper. The views
expressed in this paper, are of authors only.
Oleh W. Iwanusiw, P. Eng.
Consultant, ELTEL Industries, Bengaluru
The classical voltmeter-am meter-wattmeter method, using
traditional instruments and instrument transformers, is
revit!tl't:u iUlU ":'Ul This is followed by
an outline of an electronic metering system capable of
accurately measuring the characteristics and power loss of
power transformers over a wide range of current and voltage.
Appended to the paper are descriptions and specifications
for a polyphase metering system operating over the range of
0-2,500 amperes and 0-138 kYo
The method of power transformer characteristics and loss
measurement to be described here is conventional and should
be well known and understood. It is the classical voltmeter-
ammeter-wattmeter method. The methodology for the test
equipment, however, is very up to date and capable of high
accuracy over a wide range of voltage, current, power factor
and frequency.
The requirements of the TLMS is to accurately measure the
characteristics of power and distribution transformers, primarily
its impedance and power loss. Such characterisitics are measured
during the final testing of the transformer and consist of short
circuit tests to determine the impedance and 'copper losses', and
open circuit tests to determine the 'iron losses'. The test sample
may be a single-phase transformer, or a three.:.phase transformer
connected 'star' or 'delta' requiring the appropriate connections.
It is interesting to note that international or nation specifications
dealing with transformers do not specify the measurement
accuracy for such tests, but some do specify the connections that
must be used. Thus for example, the three-wattmeter method of
power measurement is specified for a three-phase test sample,
regardless if the sample is 'star' o.r 'delta' connected. The
measurements system must be also capable of accurately
measuring power at low power factors, as some of the
measurements may involve power factors as low as 1 %, or even
As suchll system must be periodically calibrated, it should offer
features that will allow such calibrations to be carried out
conveniently and with ease. One such calibration method or
system is described in the NBS Technical Note 1204,
"Calibration of Test Systems for Measuring Power Losses
of Transformers". This system outlines recognised procedures
and equipment that can be effectively used to calibrate
transformer loss measuring systems.
The test methods available for transformer loss measuring
systems (TLMS) are limited. One of these is the voltmeter-
ammeter-wattmeter method, the other is the bridge method. For
typical power transformers that involve power factors in the range
of 0.3 ... 0.1 for core loss, and 0.1 ... 0.01 for copper loss, the
YAW would be the choice. The bridge method would be the
choice for measuring very low loss transformers or shunt reactors.
Such specimen may have power factors as low as 0.3% and their
measurements require special considerations.
This method is probably the oldest and most widely used method
used to measure the characteristics of electrical equipment
including transformers. Two possible connections for the method
are shown in Figure 1 and Figure 2.
PT CT===

Figure I. Voltmeter-Am meter-Wattmeter Connection.
=== CT PT

Figure 2. Voltmeter-Ammeter-Wattmeter Alternate
When used on equipment with linear characteristics, and when
testing with a sinusoidal source of low impedance, only the
voltage, current, and power need be measured. As this condition
can not be met under normal operating conditions, the test method
typically must consists of measuring at least four quantities.
These are the rms voltage, the flux voltage, the rms current and
power. Additional quantities of interest are peak voltage and
peak current. Both, rms and flux voltages are measured so that
the magnitude of harmonic distortion in the test voltage can be
appreciated and applicable corrections made. Knowledge of the
peak current and flux voltage during open circuit tests allows
the manufacturer to determine the exact operating point of the
core material.
Traditionally, the rms voltage and current was measured using
either a dynamometer or a moving iron instrument with their
characteristic squa&e-Iaw scales. The flux voltage is measured
with a moving-coJr-DCllflanent magnet movement associated with
a rectifying circflit. The wattmeter for power measurement
useo a uynamomeler movement. t be peak values were
traditionally not measured due to the lack of peak responding
The limitations or drawbacks of the voltmeter-ammeter-wattmeter
method using traditional instruments are many. These included
in the dynamic limitations of the ammeters and voltmeters, due
to their square-law scale. This necessitates the use of many ratios
of instrument voltage and instrument current transformers. It
would not be uncommon to have available more than ten different
ratios on the current transformers, and ten ditl'erent voltage
transformers to cover current up to 5000 amperes and voltage up
to 120 kV. Its operation was cumbersome - as it necessitated a
shut-down of the test circuit and only then changes of ratios could
be made on multi-ratio transformers. Often one instrument
transformer would need to be replaced by another.
The measurement of transformer core loss presented few
limitations. The power factor of this test circuit being in the
range 0[0.3 to 0.1 allowed the use of conventional, unity power
factor, or the low power factor (10 - 20%) wattmeter. The most
serious limitations of this test circuit was the availability of a
wattmeter with a multiplicity of current and voltage ranges that
could be changed under load.
The measurement of copper losses, I)n the other hand, presented
many problems. This would be especially true when testing large
power transformers where the test circuit power factor may have
been 0.0 I or even lower. This would result in an indication of
only a few even on most sensitive (low power factor)
To be accurate, the test results had to be corrected for the errors.
of the instruments and instrument transformers. The voltmeters
and ammeters were corrected for the scale errors of the
instruments as well as for the ratio errors (RE) of their associated
instrument transformer. The correction of the wattmeter was
much more complex and needed to be corrected for:
scale error of wattmeter,
phase error of wattmeter,
ratio error of current transfonner,
phase error of current transformer,
ratio error of voltage transformer,
phase error of voltage transformer.
The corrections due to the phase angle errors (PE) are more
important than those due to the ratio errors at the low power
mmutes. Tius could result in an error of up to45% of measured
power at 0.01 power factor. One can see that the scale errors of
a 0.2% accuracy class wattmeter are rather small when compared
to the possible phase angle errors of instrument transformers in
low power factor circuits.
To be able to make accurate power measurement at low power
factors one must have available accurate and clibrated
instruments and instrument transformers. A phase angle
uncertainty of only 1 minute on the CT, PT and wattmeter may
result m a total error of up to 9% when measuring power at, a
power factor of 0.01.
In addition to the errors of the instruments and instrument
transformers, one must examine and correct for the power losses
in the measuring circuit and measuring equipment. This can be
readily seen from Figure I, where the power reading should be
corrected for the 12R losses in the current transformer and the
current circuit of the instruments. The losses measured in
Figure 2 should be corrected for the V
R losses in the voltage
circuit of the instruments. These corrections, although appear
small and trivial, can amount to a percent or two in very low
power factor test circuits.
The circuit shown in Figure 2 is usually preferred because the
voltage measurement can be perfonned directly at the terminals
of the test specimen eliminating all lead-drop corrections from
the measurement. The lead-drop error') are usually larger than
the losses in the voltage or current transformers discussed above.
In summaI)', one must conclude that accwate measurement of
losses using conventional instruments and instrument
transformers require numerous corrections. Many of these
corrections are voltage, current, and burden sensitive and
therefore vary from set-up to set-up. Needless to say, the
instruments and the instrument transformers need to be calibrated
on all ranges and possible burdens so that extensive tables of
corrections can be prepared.
With the advent of solid state and digital technologies and a
judicious choice of current and voltage sensors, it is possible to
improve the voltmeter-ammeter-wattmeter method substantially.
By using digital electronics rather than electro-mechanical
meters, and coupling these to voltage and current sensors - rather
than transformers, the operating range. the accuracy and linearity
of the metering system can be improved. As modern electronics
allows virtual open-circuit operation of voltage and current
sensors, such sensors remain linear over a wide range.
Instead of electrodynamometer instruments, the signals from
factor typically encountered here. To complicate the situation t d Ita d t d d . a
curren an vo ge sensors are Igl Ize an processes tn
the ratio and phase errors of the instrument transformers are load computer to provide the required readings. There are no more
current or test voltage sensitive, and have to be accurately individual instruments for each measured quantity, all of the
determined from tables or graphs. The correction could be quantitIes are computed and displayed on an appropriate screen
relatively large if instrument transformers of commercial, instead for the operator to see. As the analog-to-digital technologies are
of precision, accuracy class were used in the measurement. Even well advanced and so are the programs that compute "rms" and
if instrument current transformers of 0.1 Class were used, the "average" ("flux voltage") values, what becomes important is
total phase angle correction could have been as much as 15 the proper sensing of the voltage and current parameters. For

. y
this reason, the remainder of this paper will discuss the sensing
of the voltage and current, rather than converting the signals into
a reading of voltage, current and power.
Practical voltage sensors include:
Magnetic voltage transfonner,
Capacitive voltage transformer .
Resistive divider,
.. Capacitive voltage divider,
Amplifier aided capacitive divider.
t t
The magnetic and capacitive voltage transformers were very
quickly eliminated as neither oftem is capable of providing the
linearity over the operating range. Similarly the resistive divider
was rejected because of the heating of the high-voltage, high
resistance elements. What remained, was the capacitive divider
- either direct, or in an amplifier aided inverting configuration.
The amplifier aided configuration was selected as it requires
considerably less shielding and guarding than does a plain divider
and is shown in Figure 3. The actual construction of the high-
voltage capacitor is most important in order to maintain linearity
and stability. A concentric "tube-in-a-tube" was selected as it is
similar to the construction of "Standard HV capacitors" that
have been found to exhibit the required characteristics of stability
with respect to temperature, linearity with respect to voltage and
suitability of high-voltage operation. It must be pointed out that
when the capacitive divider is used in Figure 2, it does not add
to the losses being measured by the wattmeter.
Practical current sensors include:
* Shunts,
* Current transformers,
* Compensated current transformers,
* Inductive current sensors.
As there are practical difficulties in using shunts at high voltage,
their use was eliminated. Current transformers are the most
sensors for current and can be designed to meet very
stringent accuracy and load requirements. Even the best of these,
are limited in their dynamic range to about 100 to 1,
WhICh is not sufficient of this application. The compensated
current transformers, however, do offer improved linearity and
accuracy. The type of compensation used, makes them more
for one or another applications. Some of the more
SUitable configurations include:
two stage design.
zero flux design,
amplifier compensated design.
The two stage CT design of Brooks and Holts 11.1 is the oldest,
but for some reason not used by the industry very often. This
two stage CT is capable of extremely high accuracy and very
wide dynamic range. In addition, the error curves of the two
stage CT are very flat and change very little with the load current. .
The CT is illustrated in Figure 4.
Figure 4. TWO-Stage t.:urrtJt Transformer According to
Brooks & Holts.
The operation of the two-stage current transformer can be
explained as follows. The transformer is wound in such a way
that the difference in ampere-turns between primary (HI - H2)
and the secondary (X 1 - X2) is applied to core B and this results
in a current in the tertiary winding (Tl - T2). When properly
burdened, and assuming and error of 0.1 %, the secondary current
will be 99.9% of ideal. If the tertiary winding also operates with
an error of 0.1 %, and supplies 99.9% of the remainder, the overall
accuracy of the transformer will be in the vicinity of 0.0001 %,
or one part per million. The above operation is maintained
provided that the windings are burdened individually - as shown
in Figure 4. Such individual burdening and current summation
is easy to accomplish in electronic circuits.
The loads that need be considered are:
* two-wire, single-phase,
* three-wire, three-phase,
* four-wire, three-pahse.
According to standards, two element metering is not allowed on
three-phase, three-wire, circuits. This means that only single-
phase metering is used, thus for three-phase circuits one must
use three single-phase circuits, regardless if the load is Delta or
Wye connected. The reason two-element metering is not allowed
is because the power reading in such a connection comprises of
the difference of two large readings. -Any small scaling errors
between the two elements leads to large errors in the overall
The example below indicates the problem of two-element
metering at low power factors:
Line current 1000 amperes.
Line-line voltage 14,400 volts.
Power factor 0.01
True power
Wattmeter # 1
Wattmeter #2
+7,325 kW.
-7,075 kW.
It can be readily seen that even an error of only 1 % in either
wattmeter will cause an error of approximately 30% in total
power determination.
The modem, or up-to-date metering system avoids all of the
problems associated with systems using CTs, VTs and
conventional instruments. The major improvements are as

linear voltage and current sensors re quiring no corrections
over the designed range of voltage and current.
automatic or manual current and voltage ranges switching.
accurate power measurement with full- scale values at
power factors as low as 0.01.
digital presentation of readings.
frequent updating of all readings.

prOVISion for obtaining all three-phase readings
provisions for viewing and saving the readings as well as
Typical specifications for such a system would be:
0-138 kV, operating seamlessly over the range of
200 volts to 138 kV.
0- 2,500 A, operating seamlessly over the range
of 0.5 to 2,560 amperes.
Accuracy 0.3% at 1.0 PF, 0.6% at 0.1 PF,
3% at 0.01 PF.
1. H.B. Brooks and F.e. Holtz "The Two-stage Current
Transformer". AlEE Transactions, Vol. 41, June 1922, p.
Brian Sparling, Jacques Aubin
Abstract - Moisture content of solid insulation is a
persistent concern for a power transformer as it causes
several detrimental effects on the insulation's inteeritv.
Moisture content assessment is often derived from a single
oil sample submitted to a Karl Fischer test in laboratory
although it is recognized that a single oil sampling cannot
reveal the moisture content in paper if the oil temperature
is unstable. On-line monitoring systems are available and
continuous recording allows integration of temperature
variations and the computation of a dependable value for
moisture content in paper. Collection of data over a long
period allows calculation of moisture content of the various
components of the solid insulation system even if they are
at different temperatures and characterized by different
diffusion rates. Field data is presented for an application
on a large power transformer along with the model used
to derive the water content of insulating paper from water
content in oildespite continuously varying temperatures.
Index Terms - liquid-immersed power transformers,
monitoring, insulation system, partition curves.
Moisture management in power transformers is a persistent
concern especially for aging units. Extensive drying procedures
are applied at the manufacturing stage and sustained efforts
are deployed in service to maintain high dryness. The effect
of moisture on insulation aging is well documented along
with the detrimental effect on insulation strength and partial
discharge inception level.
It has also been demonstrated that at high temperatures, the
residual moisture in winding insulation can trigger the release
of free gas bubbles, thus creating an immediate threat to the
dielectric integrity of the insulation structure. Assessment of
water content in solid insulation is rightfully an essential part
of any comprehensive condition assessment program. In the
former version of IEEE Std 62 - 1995 (1) the moisture content
in solid insulation was defined as follows:
Dry Insulation
Wet Insulation
Very Wet Insulation
In the more recent IEEE Std C57.1 06 - 2002(2) the permissible
moisture level in paper is inferred from values of water content
in oil, assuming thermal stability and moisture equilibrium
between paper and oil:
Maximum water content Equivalent
in oil water
Transformer ppm content in
rated voltage paper
50C 60C 70C
Up to 69kV 27 35 55 3%
69 to 230kV 12 20 30 2%
230kVand 10 12 15 1.25%
The traditional method of moisture monitoring calls for oil
sampling at regular intervllls. The oil sample is then processed
through a Karl Fischer titration method that provides the total
water content in oil in parts per million (ppm). Most of the
water is in the form of dissolved water and is available to
move from the oil to the solid insulation as the transformer
progresses toward equilibrium. However, some of the measured
water is chemically bound to chemical agents such as by-
products of oxidation. This bound water is only partially
available to migrate from the oil to the paper. As the oil ages,
the quantity of chemical agents due to oxidation increases
and these agents provide additional sites for the water to bind
to. Some of the water may also bind to particles in suspension
in oil, and this water would not be fully available to move to
the solid Insulation. In spite of these constraints, this method
remains the most commonly used to assess the moisture content
of solid insulation. Equilibrium curves have been developed
to relate absolute water content in oil to water content in paper
(Figure 1).
Application of these curves impJies that the transformer is
under a thermal equilibrium.
o 10 20
I / I
1 17 /
/ ./
/ /
I ./

'/ ./


o 50
rem.,.rature, cC
30 40
.... v V

- -
PPM uot.turt In 011

...... v

00 1
Figure t - Equilibrium Curves for Moisture Partition
Between Oil and Paper (ppm vs. WCP)(3)
Measuring directly the relative moisture saturation eRH %)
makes it unnecessary to consider the type or condition of the
oil. The relative saturation is the percentage of full saturation
and is the most representative figure of the water available
in oil for transfer to the paper. At equilibrium the relative
saturations of both components of the insulation system (the
oil and the paper) are equal. Commercial on-line moisture
monitoring sensors usually provide a relative saturation
measurement as well as the measurement of the oil temperature
at the location of the moisture sensor. From these two
measurements (and with knowledge of the oil saturation curve
for this type and this age of oil is available), the relative
saturation can be converted in absolute water content in oil
in ppm (WCO). At equilibrium, the relative saturation of oil
is the same as the relative saturation of paper in contact with
the oil. In this condition, the equilibrium curve relating water
in oil to water in paper (Figure 2) can be applied to determine
the water content of paper (WCP).
.. 12

=a 8

Figure 2 - Equilibrium Curve for Moisture Partition
Between Oil and Paper (RH 0/. vs. WCP)(4)
It is important to recognize that in normal operation, even
though a stable thermal condition can be created, there still
are temperature differences, and therefore different moisture
contents, in the various parts of the insulation structure.
For example, we may consider a naturally cooled transformer
with a temperature drop of 20C between top oil and bottom
oil at full load. Assuming an ambient temperature of 20 C
and thermal equilibrium, the temperatures shown in Figure 3
for oil and winding are typical. If a moisture sensor, installed
at the bottom oil temperature, shows a relative saturation of
22 %, we can deduce, from the curves in Figure
2, the moisture content indicated in Figure 3, for the different
parts of insulation. These different moisture levels need to
be considered in the condition assessment of transformer
12 1.7 2.2 3.3
MoIsture content (%)
Figure 3 - Moisture Content on Critical Insulation
In order to understand the diffusion process, it is of interest
to consider the total amount of water in each component of
the transformer. Table 1 shows the weight of each component
estimated for the transformer described in Section 5. The thick
insulation and the thin barrier can be assumed to be at the
same temperature, in line with the average oil (54C) while
the windinginsulation is at the average winding temperature
(66C). With these assumptions, the water distribution among
the various components has been calculated and the results
are shown in Table 1. It can be seen that in this thermally
stable condition, about half of the water is in the thick
insulation; as for the oil, the solid insulation can be regarded
as an infinite reserve of water.
Table 1 - Moisture Distribution Under Steady Thermal
Oil Thick Thin Winding Total
Insulation Barrier Insulation Water
Total Mass 45,000 2,250 kg 900 kg 1,350 kg
Full Load 26 ppm 2.4 % 2.4 % 1.6 %
Conditions 1.2 kg 54.0 kg 21.6 kg 22.0 kg 98.8 kg
Off Service 1.3 2.2 % 2.2 % 2.2 %
Conditions ppm 49.4 kg 19.8 kg 29.5 kg 98.8 kg
0.1 kg
In practice the thermal stability conditions are never achieved.
Beside load variations, the transformer is submitted to daily
and seasonal temperature variations. With temperature
conditions continuously varying, the equilibrium curves cannot
be applied directly; therefore it is not possible to assess
correctly the water content of insulating paper from a single
sampling of insulating oil and a Karl Fischer test in laboratory,
even if the oil temperature is noted at time of sampling.
Temperature changes t i l displacement of water from paper
to oil or vice versa. The diffusion rate for this process depends
on the temperature, but also on the thickness of solid insulation,
on the area of contact between circulating oil and paper, and
on the moisture content of paper. The diffusion time constants
reported in literature for thin insulation (1 mm) are summarized
in Table 2(1). Diffusion time constants are also indicated for
thicker materials, assuming the time constant increases with
the square of the material thickness. It is recognized that in
a transfonner the construction is complex and the equivalent
thickness of insulating material has to be estimated for each
specific structure.
Table 2 - Diffusion Time Constant for Oil- Impregnated
Pressboard Insulation (in Days)
Insulation Thickness
tmm 2mm 4mm
0.9 3.6 14
4.2 17 67
40C 20 9 317
93 373 1493
These slow diffusion rates make it necessary to average the
recorded data over a long period. Moreover, the averaging
period has to be adjusted according to the oil temperature
and the thickness of the insulation to be assessed. This is
best done with on-line monitoring providing data acquisition
of moisture in oil and relevant temperatures.
Several on-line monitoring systems are available to record
'the variations of moisture content in insulating oil. However,
the concern over moisture content in paper might not warrant
by itself the deployment of expensive monitoring systems over
a large population oftransfonners. An interesting solution is
to combine the real-time moisture in oil measurement with a
more essential monitoring function such as the real-time
detection of fault gases dissolved in oil.
for recording of SIgnificant parameters such as top oil
temperature, bottom oil temperature, and load, which are
required for an assessment of the solid insulation moisture
Figure 4 - HYDRAN M2 Sensor With Display of
Typical Recorded Data
With such a device, it is possible to record data over a long
period. This data can be processed with a suitable algorithm
that takes into account the variations of winding and oil
temperature on the dynamic process of water migration.
Relative saturation needs to be recorded along with the oil
temperature at the sensor. Top oil and bottom oil can also be
recorded along with load to derennine the temperature in the
insulation components that are of most interest. This data is
averaged over time to account for the slow diffusion rate
involved in the migration of water between paper and oil.
In August 2003, a 50 MVA, 230 kV transfonner was equipped
with HYDRAN M2 sensor and a continuous string of data
was made available from the data logging function for off-
line processing. It is an old unit, core type, 55C rise with a
nitrogen blanket. The sensor is mounted on a spare cooler
outlet at the bottom of the tank between the two sets of coolers.
This is the avenue selected for the HYDRAN* M2 that
combines two functions. The first one is the detection of
incipient fault in the transfonner insulation. On-lirie monitoring
Figure 5 shows one week of data where the daily variation
of hydrogen and CO has long been recognized as an excellent
method for the detection ~ f dielectric and thennal problems. of moisture content in oil is clearly visible along with the
Detection at an early stage is the best means of reducing the relative saturation and the top oil temperature. Since the thennal
risk of failures and forced outages. This functionality has been conditions are continuously changing, it is impossible from
provided by the Hydran technology for many years. For a a single measurement to assess the water content in insulating
minimal additional cost, a moisture sensor is now available paper. Depending on the sampling time, the computed value
with the HYDRAN M2. The moisture detector is a thin-film of water content in paper at top oil temperature could give
capacitive sensor that is sensitive to relative saturation. The any value from 1.3 % to 2.5 %, which is almost a two-to-one
use of capacitive sensor for the measurement of oil relative variation. However, if this data is recorded over a long period,
saturation was pioneered by TV Oommen(6) and these sensors it is possible to detennine the moisture content of the various
are now available from many sources. The RH % reading components of the insulating system taking into account
can then be converted to absolute moisture content in ppm. physical characteristics of solid insulation, diffusion rate and
As an option, configurable analog inputs are provided to allow temperature.
80 30
50 25

10 i
10 5
0 0
19-Oct 2Q.OcI 21.()ct 22.()ct 23-Od 24-<ld 25-Oct 26-Od
Figure 5 - Data Logging for One Week in October 2003
Figure 6 shows the main data recorded over a period of six
months from September 2003 to March 2004. It can be seen
that the loading is fairly stable between 0.5 p.u. in the fall,
reducing to 0.4 p.u. in winter. The transformer is running
quite cold with top oil temperature ranging from about 45C
in the fall to 25C in winter. Using the top oil temperature,
load and transformer characteristics, it is possible to calculate
the bottom oil temperature and hotspot temperature. Daily
variations have been filtered out from these data to better
show the long-term trend.
50 1.20
40 1.00

_ ..



SejMJ3 0c:HI3
DIo-03 Jm.04 Feb-04 Mw-04
Figure 7 - Derivation Of The Diffusion Time Constant
And Water Content In Winding Insulation
A second component of interest is the thin barriers providing
main insulation between windings and between winding and
tank. In the main insulation, the highest electric field is often
found on the barrier closest to the winding. The moisture
content in this area is therefore critical for dielectric strength
of the main insulation. It is known that an increase in moisture
content from 0.5 % to 3.0 % may reduce by 50 % the inception
voltage for partial discharges (7). The bottom part of the barriers
is more at risk because the cooler temperatures will lead to
higher moisture contents. The barriers' temperature is the same
as the bottom oil temperature.

8 30

i 20 ... 3.0

0:60 .
8.20 t
10 0.40

Sep-03 Oct.03 Nov.03 IJeo.03 JarHl4 Fe/).()4 Uar.()4
Figure 6 - Data Logging Over a Period of Three
The insulation stnlcture is made of several components that
need to be evaluated in term of moisture content.
The first component of interest is the winding insulation paper
the moisture content has a direct influence on insulation
aging and on the probability of releasing bubbles should an
overload occur. This component is fairly thin and experimental
evidence has shown that in regard to diffusion rate, disk winding
can be assimilated to a I-mm papers). As can be seen in Figure
7, the average hot-spot temperature, over the first month of
data logging, is about 50 C leading to a diffusion time constant
of about 9 days. But, as the hotspot temperature drops, the
diffusion time constant rises exponentially to reach above 60
days in some cold winter periods. This is taken into account
in the calculation and averaging of the water content in paper.

... 0 WP 'di
C in WIn ng _'L--"'---........ -" S
1.5 0
1.0 I
FIb-D4 War-04
Figure 8 - Long-Term Computation of Moisture
Contcnt in Winding Paper and Thin Barriers
Figure 8 shows the temperatures of interest and the moisture
contents calculated for winding insulation and barriers. I t can
be seen that the moisture content of the barrier can be
significantly higher than the moisture content of winding paper.
This difference will increase with load as it is mainly sensitive
to the temperature difference between winding hot-spot and
bottom oil.
In regard to thick insulation, data would need to be averaged
over a much longer period. The diffusion rate of these
components is difficult to assess since thickness varies over
a broad range, the surface in contact with oil is also very
variable, as is the actual oil flow on these exchange surfaces.
It is recognized that a large segment of the water in the
transformer is stored in the thick insulation. The moisture
content of these components does not lend itself to an easy
assessment. Fortunately, thermal aging for these components
is usually not a concern and the electric field sustained by
the thick insulation is usually much less' than the stresses of
the thin barriers.
The objective to determine the water content in the solid
insulation has been presented, together with the negative
consequences of high moisture levels inside the transformer
tank. The question raised most often is: now that we know
the situation, what do we do?
The obvious answer is to remove the moisture from not just
the oil, but where most of the water is hiding, in the solid
insulation. The objective is to achieve a level that will not
have the negative influences such as excessive aging of the
solid insulation and the oil. This is a task for which several
techniques have been developed and used in the past with
various degrees of success (8).
a) Traditional Hot Oil Circulation
The traditional technique of circulating hot oil through the
transformer has been widely used in many countries around
the world. This technique has some limitations, starting with
the transfomers' removal from service, to the point that once
performed, with many passes, the moisture level in the oil
rises after a few days or weeks once back in service.
The limiting factor in this process only removes the moisture
from the oil and very little from the solid insulation. Once
the transformer is returned to service, the heat generated in
the windings due to the load, 'push' the moisture from the
paper into the 'dry oil'.
b) Low Frequency Heating (LFH)
This technique requires the transformer to be removed from
service, with portable equipment brought to the site. The
transformer is disconnected from the grid, and a variable
frequency high current power supply is connected to the
transformer, together with the traditional oil treatment and
vacuum pumps.
The transformer is energized at a frequency ofless than I Hz,
and high current in order to heat the windings of the transformer
to at least 800C to 900C. Energized from the primary side,
with the secondary side is shorted out (as if a heat run were
being performed).
Depending on the size of the transformer, two passes or more
may be required. This technique is quite effective as it uses
heat and vacuum to remove the moisture from the solid
c) Stationary Molecular Sieve Technology
Permanently installed molecular sieves have proven to be very
effective in drying transformers with oil capacities of up to
30,000 litres. The benefits of this technique are that the
transformer is not removed from service, but depends on the
transformer in service to generate the heat required to move
the moisture from the paper into the oil. From there the oil is
circulated through the molecular sieve at a slow rate (typically
100 litres per hour) from which the moisture is removed with
the use of water absorbent beads.
This technique is obviously a long term one, and to monitor
its effectiveness, a meniter S ~ ~ ~ : ~ ~ ~ !-!Y!:,?_A_,!"f ~ ~ ~ . , . ; . ~
the ability to calculate moisture content in the solid insulation,
as described in this paper, provides a most effective method
to know when to stop the drying process.
The effect of moisture on thermal aging and the reduction of
dielectric strength are concerns for numerous aging
transformers. The moisture content of insulating oil can be
readily assessed from a Karl Fischer titration in laboratory.
However, if the transformer is submitted to daily temperature
variations, a single sampling can lead to large errors on the
assessment of water content in paper. On-line monitoring over
periods of several months allows for a more dependable
Moisture content in winding insulation should be treated
separately from moisture content in the main insulation,
Depending on the type of cooling, the temperature at the
winding hot spot can be very different from the temperature
at the bottom of the main insulation; thus leading to very
different water contents. Because of temperature and insulation
thickness, the moisture diffusion rates can also be quite different
for these two components.
Combining moisture sensor with more monitoring functionality,
such as dissolved gas, makes it more cost effective and allows
for dependable determination of moisture content in transfonner
solid insulation.
1. IEEE Std 62-1995, IEEE Guide for Diagnostic Field
Testing of Electric Power Apparatus - Part 1: Oil Filled
Power Transformers, Regulators and Reactors
2. IEEE Std C57.106, Guide for Acceptance and
Maintenance of Insulating Oil Equipment
3. T.V. Oommen, "Moisture Equilibrium Curves - Use and
Misuse" Doble Client Conference, Boston, April 2003
4. T.Y. Oommen, "Moisture Equilibrium in Paper-Oil
Insulation Systems," Proc. Electrical Insulation
Conference, Chicago, October 1983
5. T. Y. Oommen, "On-Line Moisture Sensing in
Transformers" Proceedings of the 20th Electrical!
Electronics Insulation Conference, Boston MA, October
6. B. Noirhomme, RD. Sparling, J. Aubin, and P. Gervais,
"A Practical Method for the Continuous Monitoring of
Water Content in Transformer Solid Insulation," EPRITN
Substation Equipment Diagnostic Conference VIII. New
Orleans, February 2000
7. K. Carrander, "The Variation of PO-Strength for Oil-
Paper Insulation With Temperature, Moisture and Loading
Time," ASEA Transformer Seminar, December 1987,
Ludvika, Sweden
8. Georg Mandl, "Life Time Extension of Power
Transformers with On-Site Drying of Active Part",
Electricity Forum, Annual Transformer Forum, Toronto
Canada, September 2004
Doble is a registered trademark by Doble Engineering
IEEE is a registered trademark of the Institute of Electrical
Electronics Engineers, Inc.
EPRI is a service mark of Electric Power Research
Institute, Inc.
Trademarks of General Electric Company.
All brand and product names mentioned in this document
are trademarks or registered trademarks of their respective
P.K. Kognolkar, B. V. Ragbavaiab, A. Awastbi,
Sarita Dongre, Himangsbu Roy
In order to control large high-voltage distribution networks and
to maintain correct system voltages on industrial and domestic
supplies it is a common practice to provide means of on-load
voltage variation on the majority of main transmission and base
load substation transformers.
The tap changer is designed for operation at the current of
appropriate transformer. The diverter switch and selector switch
which is capable of making and breaking load current in addition
to selecting tap, is housed in an oil filled container and separated
from the transformer oil. The driving mechanism provides access
to the operating mechanism and a suitably positioned window
allows easy reading of the tap position indicator.
There are three rotary switches used as three phases. Each phase
consists of a circle of fixed contacts equally spaced around a
drive shaft. Rotation of the drive shaft causes the spring loaded
moving contacts to pass from one fixed contact to the next,
connecting a transition resistor between the tappings whilst the
tap change is being made.
SwitchIng tests are performed in accordance with IS 8468: 1977
& lEC:602141 :2003 to check performance in the most onerous
conditions for which the On-Load Tap-Changer (OLTC) is rated.
Switching duty test mainly consists of service duty test and
breaking capacity test . Service duty test demonstrates the
making and breaking capabilities of the contacts, without
replacement, for the number of tap-change operations as per
standard at the switched current and at the relevant recovery
voltage. An important aspect is the large number of operations
i.e.20,000 as per IS and 50,000 as per lEe and it requires
continuous power for performing the tests . The Breaking
capacity test demonstrates the current corresponding to twice
the maximum rated through-current and its relevant rated step
While designing the tap changer, mostly two operating cycles
namely flag cycle & asymmetrical pennant cycle are being l;lSed.
.Details are given below:
a) Flag cycle: A method of performing a tap change operation
in which the through current is diverted from main switching
contacts before the circulating current starts to flow. Tap
changers employing flag cycle are normally used with power
flow in both directions as shown in Figure-I.
Figure 1
E-Relevant rated step voltage
R Transition resistor
I -Maximum rated through
W & Z -Main contacts
X & Y - Transition contacts
Duties on Main & Transition contacts for Diverter Switch with flag cycle
Maio Contacts Switched Current Recovery No. of
Voltage operations
Switched Recovery No. of
Contacts Current o ~ a g e operations
YztElR+I) (E+RI) N/4
Yz{EIRI) (E-RI) N/4
YiElR+1J (E+RI) Ni4
r Yz(EIRlt (ERI) Nf4
b) Asymmetrical pennant cycle: A method of performing a
tap-change operation in which, in one direction of movement
of the switch, the circulating current starts to flow before the
through current is diverted from main switching contacts, while
in the other direction of movement the through-current is
diverted before the circulating current starts to flow. This cycle
requires that the through current connection is at one end of the
transition impedance when this is carrying circulating current.
Tap changers employing asymmetrical pennant cycle are
normally used with power flow in one direction as shown in
Figure 2
E-Relevant rated step voltage
R -Transition resistorI -Maximum
rated through current
T-Main contact
S - Transition contact
Switching tests include service duty tests and breaking capacity
tests. These tests are applicable to the diverter switch and
selector switch. Selector switch is a device capable of making,
carrying and breaking the current The diverter switch is used in
cOnjunctIOn WIth a tap selector to carry, make and break the
current in circuits which have already been selected.
1. Service duty test
The contacts of the diverter switch and the selector switch are
subjected to a number of operations corresponding to 20,000
or 50,000 tap change operations as per IS/IEC when carrying a
current corresponding to not less than the maximum rated
through current and relevant rated step voltage.
2. Breaking capacity test
This test is performed at a current corresponding to twice the
maximum rated through current .and at its relevant rated step
voltage for 40 operations as per IS/lEe.
During operation of the tap changer from one tap to the next
tap position, the making and breaking of both the contacts takes
place. At the time of contact separation, the arc is generated
and this arc is extinguished at the first current zero or within a
half cycle. The recovery voltage across the contacts is measured
and the arcing time is checked which should be less than 12ms
as shown in figure-5. The contacts of the tap changer should be
able to break the maximum rated through current at the relevant
recovery voltage without any abnormality for the duties
mentioned in the standard.
These are checked during switching tests. This test helps the
manufacturer to decide the contact life and the performance of
the tap-changer
Methods of switching duty as per IS/IEC
There are two methods given in the standard, Transformer
method & Resistance method.
Circuits/Methods used by CPRI
The resistance method is being used for switching test for the
first time in India and the facility has been developed in-house
to carry out the test by CPR I Bhopal. This unique facility in the
country helps in fulfilling the demands of utilities and
manufacturers. This method is also adopted by KEMA,
We are also in process of developing switching test facility by
transformer method. The facility is expected to be ready within
18 months.
A oscillogram of switching test on diverter switch of
rating 300k V, 700A, transition resistor 1.21 Wand relevant rated
step voltage 1700V conducted at 700A switched current and
847V recovery voltage on main contact ow' and 352.5A
switched current and 853V recovery voltage on transition
contact 'X' is shown in Figure-5. Test circuit diagram using
resistance method to perform switching tests for Diverter Switch
with flag cycle type tap-changer is shown in Figure-3
Test circuit diagram using resistance method to perform
switching tests for selector switch with asymmetrical pennant
cycle type tap-changer is shown in Figure-4.
The OLTC under test and the condition of the contacts before
and after switching tests is shown in photograph nos. PI, P2 &
P3 respectively.
The test set up available at On Line Testing Station, CPRI,
Bhopal is used to perform the large number of operations for
switching tests on tap changer. The suitable method suggested
in the standard namely resistance method is used to perform
the tests. This method helps in fulfilling the demands of utilities
and manufacturers.
This method has been used to perform the switching tests on
33kV, 66kV,132kV and 300kV class tap changers With the help
of above method, CPR) is able to perform the tests at the
maximum rated through current of700A with the relevant step
voltage of I 700V.
Apart from the switching tests; the following tests have been
conducted at CPRI, Bhopal for which the facility is available
in-house. The details are given in Table-I.
Sr. Test details No. of tap
No. changer t$)Sted
1. Temperature rise of contacts 15
2. Short circuit current tests 25
3. Transition impedance test 5
4. Mechanical test 12
a) Endurance test
b) Sequence test 02
5. Dielectric test
a) Lighting impulse test 40
b) Separate source AC 44
withstand voltage
Number of failures on transformers have been reported due to
abnonnal perfonnance ofOLTC in the course of its life period.
The establishment of this facility in India will enable utilities
and manufacturers to get OLTC tested to satisfactory
performance for switching duties and breaking duties.
The switching test is highly essential to check the satisfactory
performance of the tap changer in the service conditions and to
avoid the frequent maintenance of the switching contacts as
well as to prevent the contamination of the oil used in the
switching chamber. This test also helps the manufacturer to
decide the contact life and its performance during the service,
since arcing phenomena leads to the wear and tear of the
Jleferenee:-IS:8486,1977 Amendment 1,1980 & lEe 60214-
It GImrit ....... UtIli'll ,. ...... nce IMIMd to perfomI awItI:hIIIg tuta on cllwrw swIIcII
with flaG qcIe
---., I
y-lrl r

.'.a2.IU""" 11-
PTI_T_fw __ v .... otW.xc-.
PT2_T_ftor __ V_olZ.VC-
C'f -c.-,.....,...... (<< C"""", at .., .. X Co.tadI
CT2cw.-T ...... -... ... --.c-., z. YC_
cc-..... ofr_ ...
W.Z ......

h.t ... it (11'9'.'" uJlng 1'Ki4;tanC41 method to.,.mmn ,wltchinv leat' ror Iector
$wllch with asymm'tlrlcal pennant eyele type tap-ehanger

t-..H.v.CIft ....
.. , .... .1')""' ....... ,1MiIIClI.I
"'."2 ........ .. rT .. sc.-.-.e ....
a,,. C'TJ-C ...... .. TASc......

T ... _
"1.T ..l.T . . _' .. ...
... ...... .,,. ...... c ........ eI,. .. a....-
I. +-________________________ -J
/4,. !----------------'

_ JS2lo\

Arc-li tl 11
/01> t-------------' L.J LJ L..J J
')--'--.-0 -- I
]0 _
P2-Diverter switch contact condition before the test P3 - Diverter switch contact condition after the test
Ashok Kumar K. (Asst. Manager (QA-I&T
The objective of this paper is to provide an overview of
harmonics, and the effects of hannonics on transfonners. In
particular an attempt has been made to emphasize the effects of
harmonics on core losses oftransfonners.
power systems have always had hannonics present. As early as
the1890s, hannonies were associated with distorted voltage
and current wave shapes observed on transmission systems.
They did not cause a lot of problems in industrial settings or
office buildings as equipment was less sophisticated. Over the
last 20 years. the proliferation of electronic devices has brought
the subject up-front and personal.
The significance of harmonics in power systems has increased
substantially due to the use of electronic loads such as
computers, laser printers. welding machines, UPS systems,
fluorescent lights, adjustable-speed-drive motors in light
industrial, commercial. and residential loads, solid state
controlled loads and other high frequency producing devices.
Due to their tremendous advantages in efficiency and
controllability, power electronic loads are expected to become
significant in the future, and can be found at all power levels,
from low-voltage appliances to high voltage converters. One
result of this is a significant increase in the level ofhannonics
and distortion in power system networks. This total harmonie
distortion (THO) may impact adversely on other consumers
connected at the same point.
An important consideration when evaluating the impact of
harmonics is their effect on power system components and loads.
Transformers are major components in power systems which
are affected severely by harmonics and hence the need for
investigating the harmonic problems is obvious.
The principal effect of non-sinusoidal voltages on the
transfonner's perfonnance is the generation of extra losses in
the core Non-sinusoidal currents generate extra losses and
. heating of the conductors, enclosures. clamps, bolts etc, thus
,reducing the efficiency of the transfonner & accelerating the
loss oflife of the insulation due to the additional heating of the
windings. These non-sinusoidal currents drawn by the load also
cause voltage distortion that results in increased core losses.
high temperatures for every increment of approximately 6C,
half the life of the transformer is reduced.
2.1. Short - term risks
The main risk, for short-term failures, is the reduction in
dielectric strength due to the possible presence of gas bubbles
in a region of high electrical stress. These bubbles may develop
in the paper insulation, when hot spot temperature rises suddenly
above a critical temperature.
2.2. bong-term risks
Cumulative thermal deterioration o(the mechanical properties
ofthe conductor insulation will accelerate at high temperatures.
If this deterioration proceeds far enough, it may reduce the
effective life of the transformer. Total owning cost (TOC) also
gets increased drastically.
Transfonner losses are segregated as no-load loss and load loss;
and the total loss can be expressed by the equation (I).
= Pc+ P
P T - Total loss, watt
Pc - Core or No load loss, watt
P u - Load loss, watt
Load loss is subdivided into J2R loss (P.2
) and "stray loss".
The stray losses can be further divided into winding eddy
losses(p EC) and structural part stray losses (P OSl)' Winding eddy
losses consist of eddy current losses and circulating current
losses, which are all considered to be winding eddy current
losses. Other stray losses are due to losses in structures other
than windings, such as clamps, tank or enclosure walls, etc.
The total load loss can then be stated by equation (2)
No Load loss is subdivided into Hysteresis (PH) and eddy current
loss (P CEC); Hysteresis loss is a magnetic loss, whereas eddy
current loss is an electrical loss.
The total no load loss can then be stated by equation (3)
This will lead to a reduction in expected life span of a distribution . 3.1. Harmonic effects on transformer losses
tl'ansfonner and the method of calculating the reduction in life
. span is clearly explained in IEC 60076-7: Loading guide for
. oil- immersed power transformers. It is clearly stated that at
The contribution made by harmonic currents to different loss
components of the transformer is described in this section. The
IV;';' V) UI" H<1IUIVIII\" IVaU1Ul:>
are the PR loss, winding/core eddy current loss, other stray losses
and hysteresis loss.
3.1.1 Harmonic current effect on (2R loss
Ifthe rms value of the load current is increased due to harmonic
component, the PR loss will be increased accordingly.
3.1.2 Harmonic current/voltage effect on P c
Eddy current loss (P r,..) in the power frequency spectrum tends
to' be proportional to the square of the load current and the square
of frequency. It is this characteristic that can cause excessive
. winding loss and hence abnormal temperature rise of
transformers supplying load currents. This excessive loss can
be calculated using the below equation.
P EC: Winding eddy current loss
h = Harmonic order, 1,2,3, etc.
h""J( = The greatest harmonic order to be considered
Ih = Current at harmonic order h, amperes
I R = Rated current, amperes
P EC,R = Eddy current loss at rated current and frequency,
Eddy current loss in core also can be calculated using the same
formula by substituting voltages instead of currents.
3.1.3. Harmonic current effect on P OSL
It is recognized that other stray loss(P OSL) in the core, clamps,
and structural parts will also increase at a rate proportional to
the square of the load current. However, these losses will not
increase at a rate proportional to the square of the frequency, as
in eddy current losses. Studies at manufacturers and other
researchers have shown that the other stray loss in bus bars,
connecting and structural parts increase by a harmonic exponent
factor of 0.8 or less.
This excessive loss can be calculated using the below equation.
3.1.4. DC Components of load current
Harmonic load currents are frequently accompanied by a DC
component in the load current. A DC component of load current
will increase the transformer core loss and will increase the
magnetizing current. Higher DC current components (greater
than the rms magnitude of the transfonner excitation current at
rated voltage) may adversely affect transformer capability and
should be avoided.
Both No load & Load losses are affected by the presence of
harmonics in load currel1ls and voltages.
t:-JJt:/":j;) un jUUU jU.).)t:.) un: 4:1t:urty UUI.:UIllt:nu:U III many
standards. Unfortunately very few studies are done on the
variation in no load loss. Infact. increment in no load losses in
a distribution transformer due to harmonics also has a
significant contribution to the capitalization cost when
operating in longer term.
Hence the time has come to analyze the effects of harmonics on
core loss and to find the solutions at manufacturer send
The frequency dependence is particularly strong in the core
losses and stray losses.
The design and operation of transformers assumes that they will
be operated at one single frequency of sinusoidal excitation.
However the presence of one single frequency is an ideal case
which does not occur in practice. There will always be some
harmonic frequencies present in addition to the fundamental
Both the magnetic core and the current carrying conductors in
the windings ofthe transformer generate heat during transformer
4.1. Core Loss
Transformer cores are designed for a specific frequency of
operation. The core size, the peak magnetic flux density in the
core and the winding voltage per turn around the core are all
interrelated with the frequency of excitation. Thus, any change
in this frequency of excitation for specific cores will affect the
transformer operation. In particular the core losses will be
affected by a frequency change.
For a specific core, ifthe frequency is increased, then the losses
will be greater than designed for, While the percentage content
of the hannonics may be low, although they are increasing all
the time, the frequency scaling oftheir associated losses means
that they have a disproportionately high component of the total
For example a 5 % magnitude of fifth harmonic will increase
the overall losses in a core by about 40%. Prevention of
harmonic generation by loads is thus required to limit the
increased transformer energy loss that accompanies their
4.1.1. Hysteresis loss
Hysteresis loss results from friction as the magnetic domains
attempt to follow the change in the alternating applied magnetic
field, the faster and more often these reversals of directions
occur, the more heat is produced. Thus hysteresis loss is
dependent on the frequency of the applied magnetic field. An
increase in the frequency of the exciting current will increase
the hysteresis losses by giving more reversals per second.
The generally accepted relationship is that hysteresis losses scale
up linearly with the frequency, but there is some evidence that
the scaling rate of the losses is a little faster, probably varying
with frequency to the power of about 1.3.

'.I.l. Eddy Current Loss
The other energy loss mechanism occuning in core material is
that due to eddy current loss. Any electrically conducting
rnaterial which is subjected to an alternating magnetic field will
bave eddy currents induced in it by the varying magnetic field.
These eddy currents then generate heat in the same way that
any electric current flowing in a conductofwiU, namely by ohmic
heating (PR) generation.
.-DC; :::.:!:.:=== scales linearly with
the frequency ofthe magnetic field (from Faradays law) and as
the eddy current heating scales with the square of the current,
the eddy current heating/loss has an exact square law increase
with the frequency of the exciting current of the transformer.
From the above it is very clear that both the components of
core losses are affected by harmonics and increase with the
4.2. Analysis
t is well known fact that no load loss in amorphous transformers
AMT) is 70% less when compared to CRGO transformers. In
act, AMT savings are larger than expected during non linear
oad conditions. Same was observed by Japan industrial users
Simple analysis is given to show the effects of harmonics on
ore loss and the advantages of amorphous core during non
inear load conditions for conventional 100 kVA, 11/0.433 kV
ransformer . In general, hysteresis and eddy current loss ratio
n CRGO is 50:50 where as in amorphous core is 70:30. Now
the calculations are as follows
Table - t
Linear load condition
R 0 Amorp ous
Core loss 250W 75.0W
Hysteresis loss 125 W 52.5 W
Eddy current loss 125 W 22.5 W
Harmonic loss factor (similar to K-factorin load loss calculation)
epends on harmonic spectrum of the load, but not on core d
material of the transformer and hence the same factor can be
pplicable for both the types of transformers a
ncrease in core losses, for Hysteresis loss factor 1.1 and Eddy
urrent loss factor 1.5,are as follows. (These factors depends
n harmonic spectrum of the load)
Table - 2
Linear load condition
CRGO Amorphous
Core loss 325.0 W 91.50 W
Hysteresis loss 137.5 W 57.75 W
Eddy current loss 187.5 W 33.75 W
nence II S CIC;d.UY C;\,IUC;IIL LUd.l lUI UJI;:>I; Pc11U\,.Ul,U tV;';'
the incremental core loss in eRGO transformer is 75 W where
as for amorphous transformer is only 16.5 W.
Graph-1: Loss comparison

350 .-_______________________________
Linea r N on Linea r Linear
lClHyst _EddY:
AM T-Non
U near
Same analysis can be applicable for all the ratings of
transformers and for different harmonic loss factors.
4.2.1. Cost of core loss for the above example can be evaluated
as follows
a) Using an "A Factor" as normally used in Total Owning Cost
(TOC) calculations. A Factor is typically 190.
TOC = Initial Cost of Transformer + Cost of the No-load
Losses (A No load loss) + Cost of the Load Losses (8*
load loss)
Where A = Cost per rated watt of no load loss
B = Cost per rated watt of load loss
b) Calculating the annual cost of losses as follows:Annual
Cost = Rs/K who x Core Loss x Factor
Rs/Kwh is the cost of power in Rupees per kilo-watt hour
(Say Rs 4)
Factor = Hours per yearll 000 = 8.76
Then. Annual Cost = 4 x Core Loss x 8.76
= Core loss x 35.04
Cost comparisons using the above methods are shown in
Table 3 and 4. .
Table - 3: Core loss Cost using TOC Method
(A Factor = 190)
Sinusoidal Non-Sinusoidal
load (Rs) load (Rs)
eRGO 47,500 61,750
Amorphous 14,250 17,385
labte - 4 : Lore toss cost usmg Annual Lost
Method (RslkWh = Rs 4)
Sinusoidal Non-Sinusoidal
load (Rs) load (Rs)
I Year to Years 1 Year 10 Years
CRG 8,760 87,600 11,388 1,13,880
Amorphous 2.628 26.280 3,206 32,060
The data in Tables 3 and 4 shows that there is asignificant
increase in core loss cost when the load is non-sinusoidal.
Many transfonners designed to operate at rated frequency have'
had their loads gradually replaced with non-linear loads that
inject harmonic currents and voltages. These harmonic currents
& voltages will increase load loss & considerable core loss and
hence cause abnonnal temperature rises which will decrease
the ex.pected lifetime. Such conditions require either transformer
de-rating, as much as 50% capacity when feeding loads with
extremely distorted current waveforms, to return to the normal
life expectancy or upgrading with a larger and more economical
unit. Application of this rating system to specify a transformer
for a particular environment requires knowledge of the
fundamental and harmonic load currents & voltages predicted.
In almost all the cases field measurements are required to
diagnose problems at a specific location, by analyzing
distortions in load currents and voltages.
Before doing the above exercise one can look into the
advantages of amorphous core transformers and save the energy
during non linear load conditions.
H1 au \:M\:U:)IUil Ul uus SlUU)' W lil IW1UI;; W Iln
enhanced loss separation data with various load conditions.
[1] IEEEStdC57 .11 0-1 998(R2004): Recommended practice for
establishing transformer capability when supplying non-
sinusoidal load currents"
[2] IEC 60076-7: "Loading guide for oil immersed power
[3] IEEE Std 519-1992 IEEE Recommendedpractices &
Requirements for harmonic control inelectrical power .
[4] IEC 60076: 2000 "Power Transformers", 2ndEdition
[5] Paul K. Goethe and William D. Goethe, Loss evaluation -
will non-utility transformers be next?, Coil winding show
- Rosemont, IL, 1995.
[6] Paul K. Goethe and William D. Goethe, payback times for
premiums paid for high efficiency dry-type transformers,
Coil Winding Show - Rosemont, IL, 1996.
[7] Nathasingh, David and Pruess, Chris, "Amorphous Metal
Transformers-Performance under hannonic conditions",
proceedings of Allied Signal Conference, New Delhi, India,
April 1999.
[8] "Impact of Amorphous metal based transformers on
efficiency and quality of Electrical power distribution " -
R.Hasegawa, Fellow, IEEE and D.C.Pruess
[9] ERDA report "Transformer performance under hamlOnic
conditions in industry", Electrical Research and
Development Association, Vadodara, India, July 2000
'J Ii .tA .., 1 Vii VI'! 1 U.K.J.Vl.1!.K:S ANU AU
Ashok Kumar K. (Asst. Manager (QA-I&T
The objective of this paper is to provide an overview of
hannonics, and the effects of harmonics on transformers. In
particular an attempt has been made to emphasize the effects of
hannonics on core losses of transformers.
power systems have always had harmonics present. As early as
the1890s, hannonics were associated with distorted voltage
and current wave shapes observed on transmission systems.
They did not cause a lot of problems in industrial settings or
office buildings as equipment was less sophisticated. Over the
last 20 years, the proliferation of electronic devices has brought
the subject up-front and personal.
The significance of harmonics in power systems has increased
substantially due to the use of electronic loads such as
computers, laser printers, welding machines, UPS systems,
fluorescent lights, adjustable-speed-drive motors in light
industrial, commercial, and residential loads, solid state
controlled loads and other high frequency producing devices.
Due to their tremendous advantages in efficiency and
controllability, power electronic loads are expected to become
significant in the future, and can be found at all power levels,
from low-voltage appliances to high voltage converters. One'
result of this is a significant increase in the level of harmonics
and distortion in power system networks. This total harmonic
distortion (THD) may impact adversely on other consumers
connected at the same point.
An important consideration when evaluating the impact of
harmonics is their effect on power system components and loads.
Transformers are major components in power systems which
are affected severely by harmonics and hence the need for
investigating the harmonic problems is obvious.
The principal effect of non-sinusoidal voltages on the
transformer's performance is the generation of extra losses in
the core Non-sinusoidal currents generate extra losses and
high temperatures for every increment of approximately 6C,
half the life of the transformer is reduced.
2. t. Short - term risks
The main risk, for short-term failures, is the reduction in
dielectric strength due to the possible presence of gas bubbles
in a region ofhigh electrical stress. These bubbles may develop
in the paper insulation, when hot spot temperature rises suddenly
above a critical temperature.
2.2. bong-term risks
Cumulative thermal deterioration o(the mechanical properties
ofthe conductor insulation will accelerate at high temperatures.
If this deterioration proceeds far enough, it may reduce the
effective life of the transformer. Total owning cost (TOC) also
gets increased drastically.
Transformer losses are segregated as no-load loss and load loss;
and the total loss can be expressed by the equation (I).
P T - Total loss, watt
Pc - Core or No load loss, watt
P LL - Load loss, watt
Load loss is subdivided into FR loss (P.
R) and "stray loss".
The stray losses can be further divided into winding eddy
losses(p EC) and structural part stray losses (P OSl)' Winding eddy
losses consist of eddy current losses and circulating current
losses, which are all considered to be winding eddy current
losses. Other stray losses are due to losses in structures other
than windings, such as clamps, tank or enclosure walls, etc.
The total load loss can then be stated by equation (2)
P LL == P'
+ P EC + P OSl watt -(2)
No Load loss is subdivided into Hysteresis (PH) and eddy current
loss (P CEC); Hysteresis loss is a magnetic loss, whereas eddy
current loss is an electrical loss .
The total no load loss can then be stated by equation (3)
P Nll = PH + P CEC watt
. heating of the conductors, enclosures, clamps, bolts etc, thus
. reducing the efficiency of the transformer & accelerating the
loss of life of the insulation due to the additional heating of the
windings. These non-sinusoidal currents drawn by the load also
cause voltage distortion that results in increased core losses.
Th. is will lead to a reduction in expected life span of a distribution
transformer and the method of calculating the reduction in life
3.1. Harmonic effects on transformer losses
. span is clearly explained in IEC 60076-7: Loading guide for
. oil- immersed power transformers. It is clearly stated that at
The contribution made by harmonic currents to different loss
components of the transformer is described in this section. The
IV;);) "VlllPVllClIl;' <1UC"U';U UJ llll,; lIullllVl11" "W1Clll
are the 12R loss, winding/core eddy current loss. other stray losses
and hysteresis loss.
3.1.1 Harmonic current effect on PR loss
Ifthe nns value ofthe load current is increased due to hannonic
component. the PR loss will be increased accordingly.
3.1.2 Harmonic current/voltage effect on P EC
Eddy current loss (P r,.,) in the power frequency spectrum tends
to' be proportional to the square of the load current and the square
of frequency. It is this characteristic that can cause excessive
. winding loss and hence abnormal temperature rise of
transformers supplying load currents. This excessive loss can
be calculated using the below equation.
P EC: Winding eddy current loss
h = Harmonic order, 1,2,3, etc.
= The greatest hannonic order to be considered
Ih = Current at hannonic order h, amperes
I R = Rated current, amperes
P EC.R = Eddy current loss at rated current and frequency.
Eddy current loss in core also can be calculated using the same
formula by substituting voltages instead of currents.
3.1.3. Harmonic current effect on P OSL
It is recognized that other stray 10ss(P OSL) in the core, clamps,
and structural parts will also increase at a rate proportional to
the square of the load current However, these losses will not
increase at a rate proportional to the square of the frequency, as
in eddy current losses. Studies at manufacturers and other
researchers have shown that the other stray loss in bus bars,
connecting and structural parts increase by a harmonic exponent
factor of 0.8 or less.
This excessive loss can be calculated using the below equation.
11 = hmu; ,
'" (1b\-.r:.S
OSL = OSL -R \1,1 I;
h:] 1.
3.1.4. DC Components of load current
Harmonic load currents are frequently accompanied by a DC
component in the load current. A DC component ofload current
will increase the transformer core loss and will increase the
magnetizing current. Higher DC current components (greater
than the rms magnitude of the transformer excitation current at
rated voltage) may adversely affect transformer capability and
should be avoided.
Both No load & Load losses are affected by the presence of
harmonics in load currents and voltages.
Uti IUUU un: III mlny
standards. Unfortunately very few studies are done on the
variation in no load loss. Infact, increment in no load losses in
a distribution transformer due to harmonics also has a
significant contribution to the capitalization cost when
operating in longer term.
Hence the time has come to analyze the effects of harmonics On
core loss and to find the solutions at manufacturer send.
The frequency dependence is particularly strong in the core
losses and stray losses.
The design and operation oftransformers assumes that they wiII
be operated at one single frequency of sinusoidal excitation.
However the presence of one single frequency is an ideal case
which does not occur in practice. There will always be some
harmonic frequencies present in addition to the fundamental
Both the magnetic core and the current carrying conductors in
the windings of the transformer generate heat during transformer
4.1. Core Loss
Transformer cores are designed for a specific frequency of
operation. The core size, the peak magnetic flux density in the
core and the winding voltage per tum around the core are all
interrelated with the frequency of excitation. Thus, any change
in this frequency of excitation for specific cores will affect the
transformer operation. In particular the core losses will be
affected by a frequency change.
For a specific core, if the frequency is increased, then the losses
will be greater than designed for, While the percentage content
of the harmonics may be low, although they are increasing all
the time, the frequency scaling of their associated losses means
that they have a disproportionately high component of the total
For example a 5 % magnitude of fifth harmonic will increase
the overall losses in a core by about 40%. Prevention of
harmonic generation by loads is thus required to limit the
increased transformer energy loss that accompanies their
4.1.1. Hysteresis loss
Hysteresis loss results from friction as the magnetic domains
attempt to follow the change in the alternating applied magnetic
field, the faster and more often these reversals of directions
occur, the more heat is produced. Thus hysteresis loss is
dependent on the frequency of the applied magnetic field. An
increase in the frequency of the exciting current will increase
the hysteresis losses by giving more reversals per second.
The generally accepted relationship is that hysteresis losses scale
up linearly with the frequency, but there is some evidence that
the scaling rate of the losses is a little faster, probably varying
with frequency to the power of about 1.3.
" . .. 1.2. Eddy Current Loss
other energy loss mechanism occurring in core material is
that due to eddy current loss. Any electricalJy conducting
material which is' subjected to an alternating magnetic field wilJ
have eddy currents induced in it by the varying magnetic field.
These eddy currents then generate heat in the same way that
any electric current flowing in a conductor will, namely by ohmic
heating (FR) generation.
fnt: :::::.::=::! :;::!::!;: scales linearly with
the frequency ofthe magnetic field (from Faradays law) and as
the eddy current heating scales with the square of the current,
the eddy current heating/loss has an exact square law increase
with the frequency of the exciting current of the transformer.
From the above it is very clear that both the components of
core losses are affected by harmonics and increase with the
4.2. Analysis
It is well known fact that no load loss in amorphous transformers
(AMT) is 70% less when compared to CRGO transformers. In
fact, AMT savings are larger than expected during non linear
load conditions. Same was observed by Japan industrial users
Simple analysis is given to show the effects of harmonics on
core loss and the advantages of amorphous core during non
linear load conditions for conventional )00 kVA, 11/0.433 kV
transformer. In general, hysteresis and eddy current loss ratio
in CRGO is 50:50 where as in amorphous core is 70:30. Now
the calculations are as follows
Table - I
Linear load condition
R 0 Amorp ous
Core loss 250W 75.0W
Hysteresis loss 125 W 52.5 W
Eddy current loss 125 W 22.5 W
Harmonic loss factor (similar to K -factor in load loss calculation)
depends on harmonic spectrum of the load, but not on core
material of the transformer and hence the same factor can be
applicable for both the types of transformers
Increase in core losses, for Hysteresis loss factor 1.1 and Eddy
current loss factor 1.5,are as follows. (These factors depends
on harmonic spectrum of the load)
Table - 2
Core loss
Hysteresis loss
Eddy current loss
Linear load condition
CRGO Amorphous
325.0 W 91.50 W
137.5 W 57.75 W
187.5 W 33.75 W
1l S UJilL lUI un;;:,e;: Vc1IUlwUlc11 lU,:),:) l<llwLU1':'
the incremental core loss in eRGO transformer is 75 W where
as for amorphous transformer is only 16.5 W.
Graph-I: Loss comparison


350 ________________________________ _
linea r Non linea r lire ar
- - --- --- --
iCHyst _Eddy i
t.= __________
AM T-Non
U rear
Same analysis can be applicable for all the ratings of
transformers and for different harmonic loss factors.
4.2.1. Cost of core loss for the above example can be evaluated
as follows
a) Using an "A Factor" as normally used in Total Owning Cost
(TOC) calculations. A Factor is typically 190.
TOC = Initial Cost of Transformer + Cost of the No-load
Losses (A* No load loss) + Cost of the Load Losses (8*
load loss)
Where A = Cost per rated watt of no load loss
B = Cost per rated watt of load loss
b) Calculating the annual cost of losses as follows:Annual
Cost = Rs/Kwh. x Core Loss x Factor
Rs/Kwh is the cost of power in Rupees per kilo-watt hour
(Say Rs 4)
Factor = Hours per year/IOOO = 8.76
Then, Annual Cost = 4 x Core Loss x 8.76
= Core loss x 35.04
Cost comparisons using the above methods are shown in
Table 3 and 4. .
Table - 3: Core loss Cost using TOC Method
(A Factor = 190)
Sinusoidal Non-Sinusoidal
load (Rs) load (Rs)
CRGO 47,500 61,750
Amorphous 14,250 17,385
lable - 4 : Lore lOSS cost usmg Annual Lost
Method (RslkWh = Rs 4)
Sinusoidal Non-Sinusoidal
load (Rs) load (Rs)
1 Year 10 Years 1 Year 10 Years
CRG 8,760 87,600 11,388 1,13,880
Amorphous 2,628 26,280 3,206 32,060
The data in Tables 3 and 4 shows that there is asignificant
increase in core loss cost when the load is non-sinusoidal.
Many transformers designed to operate at rated frequency have
had their loads gradually replaced with non-linear loads that
inject harmonic currents and voltages. These harmonic currents
& voltages will increase load loss & considerable core loss and
hence causeabnonnal temperature rises which will decrease
the expected lifetime. Such conditions require either transformer
de-rating, as much as 50% capacity when feeding loads with
extremely distorted current waveforms, to return to the normal
life expectancy or upgrading with a larger and more economical
unit. Application of this rating system to specify a transformer
for a particular environment requires knowledge of the
fundamental and hannonic load currents & voltages predicted.
In almost all the cases field measurements are required to
diagnose problems at a specific location, by analyzing
distortions in load currents and voltages.
Before doing the above exercise one can look into the
advantages of amorphous core transformers and save the energy
during non linear load conditions.
111 Ulc; UC;:(lf lUlure;: au eXlc;llSIOU 01 UU:' :'lUUy WUl ve IUdue Wiln
enhanced loss separation data with various load conditions.
[1] IEEEStdC57 .11 0-1 998(R2004): Recommended practice for
establishing transformer capability when supplying non-
sinusoidal load currents"
[2) IEC 60076-7: "Loading guide for oil immersed power
[3] IEEE Std 519-1992 IEEE Recommendedpractices &
Requirements for harmonic control inelectrical power .
[4] IEC 60076: 2000 "Power Transformers", 2ndEdition
[5] Paul K. Goethe and William D. Goethe, Loss evaluation-
will non-utility transformers be next?, Coil winding show
- Rosemont, IL, 1995.
[6] Paul K. Goethe and William D. Goethe, payback times for
premiums paid for high efficiency dry-type transformers,
Coil Winding Show - Rosemont, IL, 1996.
[7] Nathasingh, David and Pruess, Chris, "Amorphous Metal
Transformers-Performance under hannonic conditions",
proceedings of Allied Signal Conference, New Delhi, India,
April 1999.
[8] "Impact of Amorphous metal based transformers on
efficiency and quality of Electrical power distribution" -
R.Hasegawa, Fellow, IEEE and D.C.Pruess
[9) ERDA report "Transformer performance under harmonic
conditions in industry", Electrical Research and
Development Association, Vadodara, India, July 2000
Aradhana Ray
Maik Koch
Abstract: Aging phenomenon in Power transfonner due to
moisture in the liquid and solid insulation decreases the dielectric
withstand strength , accelerates cellulose decomposition and
causes the emission of bubbles at high temperatures. By carrying
out extensive research, a novel method is developed for
dielectric response measurement to analyze the moisture content
in oil-paper insulated power transformers and it provides
accurate results of moisture content in the winding insulation.
For a safe service and low risk, enhanced diagnostic methods
are required to evaluate the ageing state of transfomler. One
ageing indicator is water content in the solid part ofthe insulation
(paper, pressboard). Water is an ageing product and accelerates
the f'.lrther deterioration of cellulose through depolymerisation.
Moisture increases the risk of dielectric failures and has a double
function: it accelerates ageing and also ageing generates it.
Additional moisture can penetrate from the atmosphere into the
tank. Independent of its origin moisture is absorbed into the
oil-cellulosic insulation system. Hence, measurement of
moisture in different stages of transformer's life became a
challenge for transformer experts.
Dielectric measurements are applicable to determine moisture
in power transformers. The dielectric response analysis method,
which is based on wide range measurements in time and
frequency domains, is a useful tool to evaluate the condition of
the electrical insulation systems. Dielectric methods applied on
power transformers measure a superpos ition of conductivity and
polarisation phenomena. Moisture in paper/pressboard is
obtained by a comparison of measurement data to measurements
on oil-paper insulations with known moisture content.
Nowadays two approaches are applied; the measurement of
charging and discharging currents in time domain (Polarisation
and Depolarisation Currents POC) and the dielectric
spectroscopy in frequency domain (Frequency Domain
Spectroscopy FDS). Recent publications proved the reliability
of the mentioned methods, but nevertheless reported weaknesses
as too high moisture analysis results at high oil conductivities
and at high insulation temperatures. Additionally the
measurement procedure in frequency domain takes very long
of up to 12 h.
Based on extensive research and practical experiences a new
dielectric response method was developed. Its methodology
. improves the measurement itself and also the moisture analysis.
4: ~ ~ now it is not possible to discriminate between the dielectric
influences of water and the dielectric influences of aging
byproducts on site. Thus an aged insulation will appear to be
more wet than it actually is. It also covers the whole temperature
range occurring at power transformers and effect of oil
conductivity [1).
The insulation system of the high-voltage -equipment can be
described by a dielec.tric model consisting of resistances and
capacitances in series and parallel, representing the polarization
and conductive losses in the insulation. The dielectric response
is a unique characteristic of the particular insulation system.
The increased moisture content of the insulation results in a
changed dielectric model and, consequently, in a changed
dielectric response. By measuring the dielectric response of the
equipment in a wide frequency range, the moisture content can
be assessed and the insulation condition diagnosed.
Since all of the dielectric spectroscopy methods are based on
measuring dielectric polarization response in time and frequency
domain, they should provide equivalent results in theory if the
dielectric material can be described as a linear system. rn this
case the measured data can be transferred from one domain to
another. This is important for high voltage insulation materials
because it is in many cases easier and more accepted to interpret
data in frequency domain than in time domain.
The used methods of quantifying the dielectric response in time
domain are polarization/depolarization current (POC) and in
frequency domain the loss factor tan'(E;) measurement known
as Frequency domain spectroscopy (FOS) [2].
I. Polarization/depolarization current (PDC) method:
In the polarization and depolarization current (PDC) method, a
DC voltage is applied to the insulation system under test for a
specific time and the polarization current is measured. After
then, the insulation system is shortened and the depolarization
current is measured. From the polarization and depolarization
currents the dielectric response is evaluated and the dissipation
factor frequency characteristic is calculated. The POC method
is much faster than the FOS at very low frequencies.
Measurement results are usually presented in a log/log scale
with charging and discharging current over time as depicted in
Figure I. According to the common interpretation guideline the
first I - 100 seconds are influenced by oil conductivity. The end
value of polarization current is determined by the pressboard
resistance and therefore by moisture.
moisture of
and aging

1000 10000
Figure 1 . Interpretation of PDe measurement data
Figure 2. show the polarization currents under the influence of
moisture in pressboard at an insulation temperature of 23C.
The currents are normalized to the applied voltage and
geometrical capacitance. A clear dependence on moisture
content is visible, the amplitude increases with moisture and
the currents reach their final value faster. Depolarization currents
show the same initial amplitude as polarization currents. They
decrease faster at samples with a higher moisture content.
.!!? 1

O,6 -#- Ipn 1.0
2,O --l
--I 4,0

1 10 100 1000 10000
Time Is
Figure 2. Polarization currents under the influence of
II. Frequency domain Spectroscopy (FDS) method:
Using the frequency domain spectroscopy (FDS), the dissipation
factor of the insulation system under test is measured by
frequency sweep. Frequency versus Tangent Delta
measurements are shown in Figure 3 and the derived
measurement method is called Frequency Domain Spectroscopy
(FDS). In this approach frequency range is much enhanced
especially to low frequencies.
The dissipation factor plotted via frequency shows a typical s-
shaped curve. With increasing moisture content, temperature
or aging the curve shifts towards higher frequencies. Moisture
influences the low and the high frequency parts. The middle
part of the curve with the steep gradient reflects oil conductivity.
Insulation geometry conditions determine the "hump" left of
the steep gradient. In order to determine the moisture content
of the insulation, the measurement should also provide data left
of the "hump", where the properties of the solid insulation
dominate (Figure 3).
.... 10 -'------------moistureof1.
j cellulose, !
5 aging i


0.0001 0.001 0.Q1 0.1 10 100 1000
Frequency (Hz)
Figu re 3: Interpretation of FDS measurement data
The FDS allows fast measurements at high frequencies but
requires long measurement times at frequencies down to 0.1
III. The new approach of combined FDS +PDC method:
The onsite measurement is considerably accelerated by a
combination of time and frequency domain The
frequency range from 5 kHz down to 0.1 Hz is measured in
frequency domain, whereas the range from 0.1 Hz down to 0.1
mHz or below is measured in time domain as shown in Figure
4. During the measurement the time domain data are
continuously transformed into frequency domain. This results
into a reduction of the time needed in the conventional
measurement in frequency domain only as shown in Figure 5.
0.001'--_____ .-
Time Is] 1000

formation 0
0,1 1000
Frequency 1Hz]
0,001 '----------.-
Frequency [Hz]
Figure 4. Combination of Time and Frequency Domain
1"\ f'\,.f
0,001 u.
Figure 5. Measurement time of various Dielectric
Two novel features improve moisture analysis: Firstly the
influence of conductive aging by products is compensated and
secondly the low frequency data are weighted. The moisture
analysis is based on the comparison of onsite measurements to
a modelled dielectric response <tltabase consists of dielectric
responses of pressboard and paper samples measured at various
temperatures and moisture contents.
The time dependent properties of 0 it conductivity are considered
too. This follows into a reliable moisture analysis, even at aged
oil-paper-insulations. Examples of onsite moisture
determinations conclude this investigation. The analysis results
are compared to other measurement methods like oil and paper
sampling with subsequent Karl Fischer Titration.
Figure 6 compares the moisture results of the different
measurement and analysis approaches. Karl Fischer titration at .
paper samples came to 2,6 % moisture by weight (KFT) [3].
The analysis results of the dielectric response methods differ
from each other: Two algorithms had no compensation for
conductive aging products and determined 3,8 and 4 % moisture
by weight (FOS, POC)o The new analysis software having a
compensation for conductive aging products indicates 2,9 %
tpoisturse relative to weight (Oirana).
I 6,0
Paper Drelectric
sample response
Oil sample
Figure 6. Moisture content in solid insulation as obtained
from various measurement method.
In an oil sample the moisture.saturation was measured directly
onsite and also the moisture content in ppm by Karl Fischer
titration in a laboratory. via an moisture sorption isotherm the
relative saturation reading lead to 2,5 % moisture in cellulose
(RS), which well agrees with the paper samples and the dielectric
response analysis compensating for conductive aging products.
A dielectric response measurement is a three terminal
measurement that includes the output voltage, the sensed current
and a guard as shown in Figure 7 .. The guarding technique
insures for an undisturbed measurement even at onsite
conditions with dirty insulations or bushings.
For two winding transformers, after disconnection from the
network, the voltage output can be coupled with the HV winding,
the current input with the LV winding and the guard with the
tank. It is unnecessary to wait until the transformer cooled down
or reached moisture equilibrium.
. ~ .
, Current &ense 1
Current :';81lSe 2
Figure 7. Schematic for a two channel measurement at a
three winding transformer.
Reliable information about the moisture content in the solid
insulation is valuable for residual life assessment of the
transformer also and it is discussed in next sections.
Moisture determination bases on a comparison of the
transformers dielectric response to a modeled dielectric
response. A fitting algorithm rearranges the modeled dielectric
response and delivers moisture content and oil conductivity.
A reliable moisture analysis of onsite measurements bases on
an exact data pool for the modeled dielectric response. The data
pool constitutes of measurements on new pressboard at various
temperatures, moisture contents and oils used for impregnation
. The dielectric properties of aged pressboard were investigated
as well in order to compensate for the influence of aging.
Consideration of Conductive Aging 8yproducts:
Aging of cellulose and oil causes conductive by-products as
carboxylic acids. These acids are deposited in the solid insulation
and dissolved in oil. Their DC-conductivity increases the losses
and thus "imitates" water. Figure 8 compares the dissipation
factor of aged material to that of new material. At similar
moisture content the losses in aged materials are much higher
than in the new material. Accordingly, a moisture analysis
algorithm that does not compensate for conductive aging
products will overestimate moisture content. This may mislead
to unnecessary drying of transformers; The newly developed
software compensates for the influence of conductive aging
byproducts, resulting in a more reliable moisture analysis at
aged transformers.


6 0.1
-e- 2.1 % aged
1.2% aged
..... 2,0% new
....;0- 0.8% new
0.001 -t----.---,-----,---,---,.-----r----i
lE-040.001 0.01 0.1 10 100 1000
Frequency (Hz)
Figure 8. Dissipation factor for new and aged pressboard
samples having similar moisture contents measured at
20C insulation temperature_
rhe curves indicated in Figure 9, I 0 and II represent different
;tages of transformer and their moisture content for new,
noderate wet and highly aged insulation. The trend is like it
;hifts to the right at high moisture content compared to dry
;ince the frequency points left of the "hump" reflect moisture
;ontent, they should be reached for later moisture detennination.
{equired measurement time to get the "hump" also depends on
he condition of the specific transformer.
1 ;
0.5 :
'm I
0.2 [
0.0001 0.001 0.01 0.1 1.0 10 Freq/Hz 1000
Figure 9_ Dissipation factor for new dry insulation
i 0.5 r
o 0;1 f
............. 44C

0.0001 0.001 0.01 0.1 1.0
10 Freq/Hz 1000
Figure 10. Dissipation factor for a moderate dry
Heavily aged
0.05 r
0.005 -
0.002 L ___ . __ _
0.0001 0.001 0.01 0.1
1.0 10 Freq/Hz 1000
Figure 11. Dissipation factor for aged insulation.
IEC 60422 categorizes moisture satorations of more than 6 %
as "moderately wet", which is equivalent to a moisture content
of approximately 2,2 % [4]. In this area the water molecules
become more and more active, causing the dangerous effects
of water. At this level, maintenance actions such as drying should
be considered taking into account the importance and future
operation of the transformer. Fig. 12 shows the relationship
between moisture content and moisture saturation which helps
to assess the condition and in analysis of the data.
.......... ,
> 30
Moisture saturation [0/,)
Figure 12. Moisture sorption isotherm for a cellulose
material elating moisture saturation to moisture content
with categories according to lEe 60422.
A (
an (
to (
in I
Various moisture measurement methods were applied on an aged
1 transformer to estimate its moisture content onsite. The
transformer was manufactured in 1967, had a rated power of
J33 MVA, a transformation ratioof230/115 148 kV and OFAF
A capacitive probe onsite the relative moisture
saturation in oil, the moisture. in cellulose ww; Ot;lin,J
advanced equilibrium diagrams. Beside this coulometric Karl
Fischer titration determined the moisture by weight (ppm) in
an oil sample and an conventional equilibrium diagram served
to derive the moisture in cellulose from 'this result. Ideally all
the methods should come to a similar moisture content in
cellulose but the data was indicating a different pattern as shown
in Figure 13.
Delec:ric reSOO;;se
o Mods
o W(RS)

Figure 13: Moisture content in the solid insulation as
obtained from another analysis software called MODS, the
new method FDS+PDC (Dirana) and via equilibrium
diagrams from the relative saturation of oil (W(RS and
from the moisture by weight in oil (WC(ppm.
The higher moisture content in the LV insulation agrees well
with the service conditions of the transformer: the LV winding
was notjn use. Cellulose at lower temperatures stores the water
in a transformer. Thus the dielectric methods allow for an
elementary localization of wet areas 'in the insulation [4].
Contrary to this the moisture content in cellulose as derived
from oil samples gives an average value. The result obtained
from the relative saturation in oil by advanced equilibrium
diagrams agrees well with the dielectric analysis.
However the conventional method of deriving the moisture in
cellulose from moisture by weight in oil (ppm) gives a too high
result. Aging of oil and paper makes the application of
equilibrium diagrams from literature sources impossible in most
cases. Hence, the dielectric methods allow for an elementary
21 I
localization of wet areas in the insulation. Contrary to this the
moisture content in cellulose as derived from oil samples gives
an average value.
Following points are the Key features of the new method
I. Aging of Cellulose and oil causes conductive byproducts
and these are deposited in the solid insulation and
alSSOlveo In OIL Lonsloeration of conductive aging
products are taken in analysis software. Up to now it is
not possible to discriminate between the dielectric
influences of water and the dielectric influences of aging
byproducts on site. Thus, an aged insulation will appear
more wet than actually it is.
2. Method is able to compensate for the influence of
insulation geometry, insulation temperature and oil
conductivity. Method show an apparently decreasing
moisture analysis result for increasing temperature and
an increasing moisture result for increasing oi I
conductivity .
3. A new software "Dirana'" was developed which bases on
a new data pool, measured at new and aged pressboard
with various moisture contents and oil impregnation.
4. The dielectric methods allow for an elementary
localization of wet areas in the insulation. Contrary to til is
the moisture content in cellulose as derived from oil
samples gives an average value.
This paper discusses approaches to measure moisture in power
transformers using dielectric response methods. Dielectric
diagnostic methods deduce moisture in the solid insulation from
dielectric properties like polarization depolarization currents
and dissipation factor vs. frequency. By comparing the dielectric
response with the previous measurements or with the modi!!
curves based on the insulation system's construction, the ne\\'
method called DIRANA provides an indication of the insulation
condition such as:
Moisture content in the oil/paper insulation
Faults in OIP, RBP and RIP high-voltage bushings
Faults in the generator, motor and cable insulation
Monitoring of moisture content in solid insulation during
maintenance actions such as On- site drying will be very
effective and should be considered as an important
To conclude, the findings at the very aged transformer show,
that a compensation for aging products is necessary both for
the measurement based on moisture equilibrium and them based
on dielectric properties.
1. M. Koch, S. Tenbohlen, T. Stirl: "Advanced Online
Moisture Measurements in Power Transformers" CMD
2006 International Conference on Condition Monitoring
and Diagnosis, Changwon, Korea, 2006.
2. W. S. Zaengl "Dielectric Spectroscopy in Time and
Frequency Domain for HY Power Equipment, Part /:
Theoretical CQllSiderations" IEEE Electrical Insulation
Magazine, Vol 19, No.5 pp.5-18, September I October
3. M. Koch, S. Tenbohlen, I. Hoehlein and J. Blennow'
"Reliability Improvements of Water Titration by the Kari
Fischer Technique" Proceedings of the XVth International
Symposium on High Voltage Engineering, ISH, Ljubljana,
Slovenia, 2007
4. M. Koch, S. Tenbohlen, M. KrUger and A. Kraetge: "A
Comparative Test and Consequent Improvements on
Dielectric Response Methods" Proceedings of the XVth
International Symposium on High Voltage Engineering
ISH, Ljubljana, Slovenia, 2007 '
H. N. S. Gowda
Progress in science has been always been linked to advances in
measurement capability. Measurement play an important role
in Man's life and also an effective tool oflearning. Measurement
ofan alternating current is one of the most frequent operations
not only because of its inherent interest but also because it is
necessary in determining the parameters of electrical circuits.
Instruments can be applied widely in the interest of safety to
personnel and equipments. As long as the current to be measured
are small and at low voltages, the direct method of connecting
the instrument and relays in the circuit is convenient and
adequate. Ifthe current exceeding several hundreds of amperes
have to circulate continuously through the instruments and relays
their design becomes more complicated and expensive. Still
difficulties are more, when the current to be measured in circuits
excess of 1,000 V. The best solution to these problems is to
obtain a replica of the current in the primary circuit and this is
achieved with the use of current transformers. Since, the relays
are the heart of the transmission and distribution system, it
receives commands from C.Ts to protpct the transformer to
operate relays I circuit breakers. This article highlights the salient
features' of C. Ts, Relays, and Transformers (Protection). It also
made an attempt to evaluate C.T's and relay's requirements for
the purpose of transformer Rrotection, since the transformer is
costliest and heaviest equipment in power system.
Current and voltages are the most commonly measured electrical
quantities. Alternating currents of one kA and alternating
voltages in excess of one kV are measured with the aid of
instrument transformer; Instrument transformers are special type
which converts currents and voltages into values
convenient for measurement. A transformer which is intended
to reproduce in its secondary circuit, in a definite and known
proportion suitable for utilization in measurement, control or
protective devices, the current (or voltage) ofits primary circuit,
with its Bhase relations substantially preserved. These
transformers are so important for both insulating and range-
adjusting purposes, that it is difficult to imagine operation of
anA.C power system without them. These equipments are widely
used and seldom discussed. These transformers are
fundamentally no different from ordinary transformers, but their
designs are different so that they will perform only a single
(unction well.
The operation ofC.Ts differ from power transformers in many
respects, some of them are:
The operation represents a short circuit condition in that
the secondary load is very low impedance;
rr The current in the secondary winding is determined by the
primary current and not bv the secondarv circuit
impedance (within practical limits);
rr C.Ts are used only for measurement and protection
purposes, their power rating is small generally between
15 to 200 VA;
rr In a power transformer the source is a voltage, where as
in C.T it is current;
rr In power transformer the exciting current remains constant
whereas in C.T goes on increasing with increase of .load.
There are three primary applications IT's used for:
rr Metering (energy billing and transaction purposes);
rr Protection control (system protection and protective
relaying purposes); and
rr Load survey (economic management of industrial loads).
Current Transformers:
An instrument transformer intended to have its primary winding
connected in series with the circuit carrying the current to be
measured. or controlled. It is a transformer governed by the law
of electro-magnetic induction. It operates Qn the ampere-tum
Primary Ampere Turns - Magnetising Ampere turns
= Secondary Ampere Turns
The C.T is not voltage dependent but it is voltage limited, this
current passes through impedance when a voltage is developed.
As this occurs, energy is depleted from the supply, thus acting
like a shunt. This depletion of energy results in the C.T errors.
As the secondary impedance increases the voltage
proportionally increases: Thus the limit of the C.T is magnetic
saturation, a condition when the core flux can no longer supports
the increased voltage demand. As this point, nearly all the
available energy is going into the core, leaving none to support
the secondary circuit. The C.T is designed to deliver from the
secondary winding a "current" is exactly as possible, an
integral fraction of the primary current. It is also designed so
that the phase angle between primary and secondary current is
as small as possible. To minimise magnetising current, and get
correct and reasonably constant ratio, the C.T is designed with
core of a best possible magnetic properties worked at a low
flux density. It is also relatively large values ofreactance.
The factors determining the choice of current transformers are
the values ofthe primary and secondary rated currents, the rated
outputs of different cores for a given accuracy class and the
rated over current factor. The rated current of the transformer
has to be adapted to suit the operating current. Specification of
the C. T's for a particular application must conform to standards.
IS: 2705 -1902 -General, Metering C.T's, protection C.T's, and
special purpose C.T's. IS: 2705 is in par with International
Standards (IEC).
Major information required for the selection are:
rr Ratio;
rr Number of cores;
rr Rated burden of each core;
rr Class of accuracy of each core;
c:r Rated short time current and it's duration;
rr Voltage class; and
rr Insulation level.
The purpose ofC.Ts is to transfonn the primary current in respect
of magnitude and phase angle within prescribed error limits.
The relays must have fairly accurate reproduction of the primary
current over the entire range of operation to ensure coordination
and system protection.
The requirements ofC.Ts for protective relays are quite different
from those of instrument C. Ts. A C. T used for instrumentation
is required to be accurate over the normal working range of
currents, whereas a C.T for protection is required to give a
correct ratio up to several times the rated primary current. It is
due to the fact that, the relay has to perform reliably at normal
currents as well as at faults current. Usually, fault currents are
many times more than the normal rated currents.
Generally instrument transfonners have a single current and
single rated voltage. Current transformers are used to reduce
the heavy current flowing in the element of a power system to
low values that are suitable for relay operations. The current
rating of a protective relay is usually 5 or I ampere. Besides
reducing the current level the C.T also isolales the relay circuit
from the primary circuit which is a high voltage power circuit,
and allows the use of standardized current rating for relays.
(Accuracy 0.1,
0.2, 0.5, I & 3)
Current Transformers
(5P, lOP, 15P & 20P)
Special Protection
& RCT)
Measuring C. Ts: The current transformer C. T is basically
applied to supply the current to the indicating instruments like
ammeters, integrating meters, wattmeters, energy meters, when
connecting to meters and measurements low saturation and high
accuracy is desirable. So, the windings used should be of better
steady state accuracy factor as compared to the transient
accuracy factor.
Protection Current Transformers: Transformer is applied to
supply current to protective relays in the circuit, the design is
mainly on the basis ofknee-point voltage (V k)' excitation current
at this knee point voltage should be suitable for both transient
and steady state condition for the operation of relays.
Special Purpose Transformers: The transformer is suitable for
relays used for special purpose applications like balance and
distance protection scheme. The knee-point voltage, excitation
current and secondary i"":-,,,rbnt ""I"
operation. These values should be within the limits for both
steady and transient condition. These C. Ts demands low
Causes of Errors in C. T's: As a result of physical limitations
inherent in electric and magnetic circuits of the trans fonner,
there are departures from ideal conditions and consequently
errors are caused. The reasons are:
r:r There is some _exciting mmf required by the primary
winding to produce flux and therefore, the transfonner
draws a magnetising current (I .,);
r:r The transfonner input must have a component which
supplies the core losses (Eddy current and Hysteresis
losses) and FR losses oftransfonner \vinding due to now
of current (10);
Therefore, the loss component Ie is required to feed the losses
associated with the flux and also the associated copper loss in
winding d,ue to flow ofI
rr The flux density in the core is not linear function of the
magnetising force i.e the transformer core becomes
r:r There is always a magnetic leakage and consequently the
primary flux linkages are not equal to the secondary tlux
Limits of Error and Composite Error for Accuracy
Class SP to ISP
Accuracy CurreD' Error Phase Displacement Composite Error
Class at Rated Primary at Rated Primary at Rated Primary
Current in % Current in % Current %
5P 1 60 5
lOP 3 - 10
15P 5 - 15
Cbuacteristics Protection Special Protection c.T's
C.Ts Ordinary for REF and Differential
Ratio 500.15 400/5
Burden 15 VA -
Class 5PIO or 10PIO PS or Class X
RCT - 0.25
Knee-point Voltage - 60 V
Magnetising Current - 250 rnA
C. T accuracy requirements: It is generally necessary to make
C.T accuracy calculations when applying power transfonner
differential relays. These calculations require knowledge of C. T
'characteristics either in the fonn of ratio-correction factor curves
or secondary excitation and impedance data. The relayi.ng type
bushing C.Ts may be operated on their lowest turns - ratio taps
makes it necessary that the rating of the full winding be higher
than if the full windings were used.
Meaning of 15 VA 5PlO means 15 VA burden not more than
5% composite error at 10 times rated current. When phase fault
stability and accurate time grading needed, class 5P is required
otherwise lOP is suitable.
Knee-Point Voltage: The knee point is the exciting characteristic
defined as the point at which a 10% increase in secondary e.m.f
produces 50% increase in exciting current. Knee-point voltage
V k is very important for specifying C. T for REF and differential
protection. The values of V k generally left to the manufacturers
with R
It is an usual practice the value of K is selected on the
basis of fault current If and in general is considered as 21,
Maximum Excitation Current (lm.g): C.T core needs to be
identified by certain parameters for "PS" class core,
magnetisation nature of this core is referred with the current
called maximum magnetization current at minimum level of knee
point voltage (V k). The value in practice shadd t-e less than the
maximum magnetisation current.
Where, P mA
pennissible magnetising current in rnA
Factor usually specified by the rel"y
manufacturer (2 or 4 depending on application).
Minimum knee point voltage is specified by a formula V k = K
Is (ReT + 2 R
Where,K A constant that depends upon the system fault
level and characteristics of the relay;
Secondary reflected current;
C.T secondary resistance at 75
One way lead resistance;
Minimum knee point voltage.
Secondary current --+
Figure: Variation of ratio R with
secondary winding current

'" .c

____ . "" L Luoww--lirnmppeeodClaTlncce

Figure: Variation of phase angle
with secondary winding current
Burdens: Burdens are the voltage requirements of the devices
fed by C.T.(say O.6V, 1 V, etc,. at rated current). The loads
connected on the C.T's secondary circuit are called as burden
and can be expressed in VA at rated secondary current with
some particular power factor or as impedance. All burdens are
connected in series. A C.T is unloaded ifthe secondary winding
is short circuited, under which the VA burden is zero because
voltage is zero. The rated output (VA) is an apparent power,
which the C.T can deliver to the secondary circuit at the rated
current and burden while maintaining its accuracy with in the
specified class. C. T burden impedance decreases as the
secondary current increases because of saturation in the
magnetic circuits of relays and other device.
Relation between Burden and ALF: The satisfactory operation
of the scheme for protection core depends upon the development
of voltage across the winding to drive the current through the
relays or instruments during the time of fault.
So the selection of burden andALF of the core is considered on
the basis of voltage developed during the fault condition. It can'
be expressed as follows:
V = (Burden x ALF) I Rated Secondary Current.
Calculation of Primary Current: Factors for calculation of
primary current are:
". Rated continuous thermal current;
". Maximum load on the system with 25% extra;
". Short time current allowable time;
". Suitable multiple fractions lQ, ) 2,.5,12, 20, 25, 30, 40, 50,
60 and 70 (Under lined values are preferred)
Characteristics curves for typical oriented steel is indicated as
in the figure. It is also called excitation curve of a C. T. The
excitation curve may be subdivided into main four reasons viz.: