A Study on Transformer Oil (Mineral Oil)

Prepared By

Sunil Nannaware

R&D- ITR

Definition:

A mineral oil or liquid petroleum is a liquid by product of the distillation of petroleum to

produce gasoline and other petroleum based products from crude oil. A mineral oil in this

sense is transparent, colorless oil composed mainly of alkenes (typically 15-40 carbons)

and cyclic paraffin’s, related to petroleum jelly (also known as “white petroleum). It has

density of around 0.8 g/cm3. Mineral oil is available in light & heavy grade.

Mineral oil application in the electrical equipments falls under the category of liquid

di-electrics

Mineral oil is commonly referred as ‘Insulating oil’ or ‘Transformer Oil’. Though it

is widely used not only in the transformers but also in variety of electrical equipments it

is still referred as transformer oil. Because of this reason referring electrical quality

mineral oil as transformer oil may lead to confusion. But this terminology is so

commonly used in the electrical industry that veterans will find this bit funny. In order to

avoid any confusion regarding the application of mineral oil we will refer mineral oil as

insulating oil.

In addition to the mineral oil there are other liquids which are used in the electrical

equipments for the same application. We will group all this under one umbrella of

insulating liquids.

Organic Inorganic

- pure water

- -liquefied

gases

Mineral oils Synthetic oils

-Silicone oils

-Chlordiphenyles

-high molecular

weight hydrocarbon

-Tetrachlroethylene

Insulating liquids

Transformer Oil: Properties of Naphthenic and Aromatic Fractions

Functionally most electrical insulating fluids are considered to be equivalent and they are

handled as such.

Chemically most electrical insulating fluids are not equivalent. While the differences

normally do no defeat the prescribed functions of the fluids they do affect the way they

function. Transformer fluids vary in composition from nearly pure compounds to mixture

that is too complex to fully describe. Due to this complexity we will first understand

about functional group of the transformer oil. Such as properties, processing, purification

& transport of insulating oil. Then in next section we will under stand chemistry behind

the transformer oil.

Physical properties:

Aniline point: The aniline point (temperature) of a insulating oil indicates the solvency

of the oil for some materials that are in contact with oil. A high aniline point indicates the

aromaticity and a lower solvency for some material (e.g. rubber)

Test method: ASTM D-611-01b

Relative density (specific gravity)

The relative density of oil is the ratio of the weights of equal volumes of the oil and

water, tested at 15°C. the relative density is significant in determining the solubility for

use in certain applications, in cold climates, ice may form in equipment exposed to

temperatures below freezing. When considered along with other oil properties, relative

density can be an indicator of the quality of the oil.

Discussion:

Relative density (RD) or specific gravity is dimensionless quantity, as it is the ratio of

either densities or weights.

RD = ρ substance / ρ reference

Where RD is relative density, ρ substance is the density of the substance being measured

and ρ reference is the density of the reference (by convention ρ, the Greek letter rho,

denotes density)

If the substance relative density is less than one then it is less dense than the reference

and vice versa. If the reference is water than substance with relative density (or specific

gravity) less than one will float in water. For example an ice cube with relative density of

about 0.91 will float. A substance with relative density greater than one will sink.

The specific gravity (relative density) of insulating oil influences its heat transfer rates. In

extremely cold climates, an upper limit of 0.895 is placed on the specific gravity (relative

density) of insulating oil. If oil contains any moisture, the resulting ice from the freezing

of the moisture in the oil filled apparatus should not float on the surface (specific gravity

of ice is approximately 0.915). Floating ice on oil can cause a flashover of conductors

extending above the oil level. Oils of different specific gravity (relative density) may not

mix when added to each other, and precautions are taken to guard against this possibility.

Test Method: ASTM D-1298-85

Scope

1.1 This practice covers the laboratory determination, using a glass hydrometer, of the

density, relative density (specific gravity), or API gravity of crude petroleum,

petroleum products, or the mixture of the petroleum and non petroleum products

normally handled as liquids.

Values are measured on a hydrometer at a convenient temperatures, readings of the

density being reduced to 15°C, and readings of the relative density (specific gravity)

and API to 60°F, by means of international standard tables. By means of these same

tables, values determined in any one of the three systems of the measurement are

convertible to equivalent values either of other two so that measurement may be made

in the units of local convenience.

Terminology:

Density: for the purpose of this practice, the mass (weight in vacuum) of liquid per unit

volume at 15°C. When reporting results, explicitly states density in mass & volume along

with reference temperature.

API Gravity:

A special function of relative density (specific gravity) 60/60°F, represented by

API Gravity, Deg = (141.5/sp.gr 60/60°F) - 131.5

No statement of reference temperature is required, since 60°F is included in definition.

Color:

Insulating oil should have a light color and be optically clear so that it permits visual

inspection of the assembled apparatus inside the equipment tank. Any change in the color

of oil overtime is an indication of deterioration or contamination of oil.

The color test is performed by visually comparing the color of the oil to a color chart

provided by most oil manufacturers. The scale on these charts range from 0.5 to 8.0 with

new oil has color of 1.0 or less. When exact match is not found and sample color falls

between two standard colors, the higher of the two colors is reported. New oil will appear

to be clear to light straw color, while red to black oil indicates sludge or other

contamination.

Test method: D1500

This test method covers the visual determination of the color of wide variety of petroleum

products such as lubricating oils, heating oils, diesel fuel oils and petroleum waxes.

Significance of use:

Determination of the color of petroleum products is used mainly for the manufacturing

control purpose and is an important quality characteristic since color is readily observed

by the user of the product. In some cases the color may serve as an indication of the

degree of refinement of the material. When the color range of the particular product is

known, a variation outside the established range may indicate possible contamination

with another product. However color is not always a reliable guide to product quality and

should not be used indiscriminately in product specifications.

Interfacial tension:

The interfacial tension of mineral oil insulating fluid is related to deterioration of the

sample. The mineral insulating oil is essentially a non-polar saturated hydrocarbon;

however, when the sample undergoes oxidative degradation there are formed oxygenated

species such as carboxylic acids, which are hydrophilic in nature. The presence of these

hydrophilic materials in the insulating fluid can affect the chemical (acidity), electrical

(dielectric strength), and physical (interfacial tension) properties of oil. In this procedure

one measures the surface tension of the oil against that of water, which is highly polar.

The more nearly the two liquids are alike in their polarity the lower the value of surface

tension between them. Thus the higher the concentration of hydrophilic materials in the

insulating fluid, the lower will be the interfacial tension against water. One method

measures the size of a drop of water that is formed below the surface of the oil (ASTM D

2285) however, if one desires more accurate values it is recommended that ring method

described below used.

The details of the entire procedure for determining the interfacial tension (IFT) of oil

against water by the ring method are given by ASTM 971 standard and are only briefly

mentioned here. The device used to determine the IFT is a tension meter based on Du

Nuoy principle. A sample of oil is carefully floated on the top of a layer of water and

force necessary to pull a platinum ring upward from below the water level through the oil

is measured by using a calibrated torsion wire. The force is measured at the point at

which the ring breaks free of the water layer as it is being pulled upward through the oil

layer. The platinum ring is made to precise dimensions. Since this test is very sensitive to

trace contaminates, one must be very thorough in the handling the sampling device, the

sample, and the instrument. There are correction factors that have to be taken in to

consideration relating to the dimensions of the ring and the densities of the water and

sample. The results are given in units of dynes/cm or the numerically equivalent units of

milliNewtons/meter (mN/m).

Significance:

The magnitude of the IFT is inversely related to concentration of hydrophilic degradation

products from the deterioration of the oil. Since the hydrophilic materials are usually

highly polar and thus not very soluble in the non-polar oil, the presence of these species

can results in the sludge formation. These materials that remain dissolved in the oil can

affect the desired electrical properties of the oil. There is usually inverse relationship

between the neutralization number of an oil and its IFT. As an oil sample undergoes

oxidative degradation, its neutralization number will increase while its IFT value will

decrease. It should also be recognized that a decrease in the IFT does not imply that the

acidity must also be high, since there are other non-acidic contaminants that could be

present in the oil that are hydrophilic and will lower the IFT but not raise the acidity. An

example of such situation might be that of a free breathing transformer near salt water

where a salt water mist might be able to enter the unit. Such event will not affect the

acidity but would affect IFT & dielectric strength of the oil.

Quality index system:

Dividing the interfacial tension (IFT) by the Neutralization number (NN) provides a

numerical value that is an excellent means of evaluating oil condition. This value is

known as the oil quality index (OQIN) or Myers Index Number (MIN). a new oil, for

example has OQIN of 1500.

OQIN= IFT/NN. 1500= 45 (typical new oil)/0.03 (Typical new oil)

Flash Point:

The flash point is the minimum temperature at which heated oil gives of sufficient vapor

to form a flammable mixture with air. It is an indicator of volatility of the oil.

Test Method: ASTM D-92 (Cleveland open cup method)

Significance:

1. Flash point measure the tendency of the sample to form a flammable mixture with

air under controlled laboratory conditions. It is only one of number of properties

that must be considered in assessing overall flammability hazard of material.

2. Flash point is used in shipping and in safety regulations to define flammable and

combustible material.

3. Lower flash point can reveal the presence of highly flammable substances in the

oil.

4. Flash point assumes significance after an arcing fault inside the transformer.

Test Equipment for Flash Point:

1. Cleveland Open Cup:

2. Approximately 75 ML oil needs to be taken in Cleveland cup

3. Prepare the sample

4. Prepare the Kit with appropriate settings

5. The flame is dipped for every 3 degree (or any value) starting from 105°C till the

flash point occurs. The flash is then recorded.

S K value:

Nominal Value: -4 to 8.

S K value is defined as the increase in the volume of concentrate sulphuric acid on adding

a given test sample. It signifies the degree of the refinement of oil. This test is under

consideration.

The effect of S K value on other characteristics properties like resistivity. The dissipation

factor, stability test like accelerated ageing and oxidation stability, have been studied in

various oil refinery laboratories. However there exists no correlation between an

increases in S K value, but is still lower than that under specification. A similar effect of

SK value on the stability characteristics does not specify any co-relating trend. These

results have been collected by research laboratories.

The present trend in the manufacture of transformer oil is based on the achieving the best

performance and the stability characteristics specified in above. The indigenous TOFS

available are paraffinic in and nature and require greater refining. This results in

eliminating some anti oxidants in oil.

Pour Point:

The pour point is the lowest temperature at which oil will just flow. A low pour point is

important, particularly in cold climates, to ensure that the oil will circulate and serve its

purpose as an insulating and cooling medium. It may be useful for identifying the type

(naphthenic, paraffinic) of oils.

Testing Method: D 97

Nominal Values: -5°C,-30°C, -40°C.

Viscosity :

The viscosity of insulating oil is measured by timing the flow of a known volume of oil

through a calibrated tube. Viscosity is not significantly affected by oil contamination or

deterioration, but may be useful for indentifying certain types of service aged oil.

Viscosity has an important influence on the heat transfer characteristics of oil. High

viscosity decreases the cooling efficiency of the oil. High viscosity will also affect the

movement of parts in electrical equipment, such as circuit breaker, switchgear, tap

changers, pump and regulators. Viscosity is the factor in determining the conditions for

oil processing & cellulose Impregnation time.

Testing method: ASTM D88-94, ASTM D-445, ASTM D- 2161

Poiseuille’s Law:

Where

V = volume of oil within the capillary.

r = average equivalent capillary.

P0 = external pressure Pe + pressure created by capillary action.

Pi = Internal pressure.

L = the depth of impregnation increasing with time.

η = Viscosity.

Capillary equitation:

Where:

Θ = Contact angle of oil.

r = average equivalent capillary radius.

T= Surface tension of the oil.

η = Viscosity.

Oxidation stability:

Uninhibited oils must be free of additives, either natural or synthetic those are used to

improve oxidation stability. This includes but is not limited to 2-6 ditetiary phenol, 2-6

diteriary-butyl cresol, or metal deactivators such as benzotriazole and its derivatives.

Inhibited oils are insulating oils, which have been supplemented with either 2-6 ditetiary

butyl phenol or 2-6 ditetiary-butyl cresols or any other specified and acceptable oxidation

inhibitor. If an additive other than 2- 6 ditetiary-butyl phenol or 2-6 ditetiary cresol is

used, appropriate limit values for oxidation stability tests should be agreed to by the

purchaser and seller. If more than one additive is added than many stringent limits for

oxidation stability would apply.

Pour point depressants, gassing tendency improvers, additives for corrosive sulfur and

static electrification (metal passivators), antifoaming agents and other additives.

All additives should be specifically identified or at minimum identified by class of

compounds such as metal passivators if the specific information is proprietary.

Oxidation stability (ASTM D2112-01a):

This test method is rapid test for evaluating the oxidation stability of new insulating oil

that contains the synthetic oxidation inhibitor 2-6 DBPC or 2-6 DBP. The test measures

the length of time required for the oil sample to react with a given volume of oxygen

when a sample of oil is heated and oxidized under test condition.

Oxidation stability (ASTM D2440-99)

This method determines the resistance of mineral insulating oils to oxidation under

prescribed accelerated aging conditions. Oxidation stability is measured by the propensity

of oils to form sludge and acid products during oxidation. The test method is applicable

to new oils, both inhibited and uninhibited.

Anti oxidants:

Mineral oil insulating fluids undergo oxidative degradation in the presence of oxygen to

give a number of oxidation products. The final products of oxidation are acidic materials

that can affect the characteristics of the insulating oils as well as cause damage to

components of the electrical units. Oxygen is a diradical species and the reactions of the

oxidative process are complex but they do involve free radical reaction. One way to

prevent these types of the reaction is to incorporate an oxidation inhibitor that will

interrupt and terminate the free radical process oxidation. Phenolic materials are quite

good for this purpose and two most commonly used inhibitors are 2-6 di tertiary-butyl

phenol (DBP) and 2-6 ditetiary butyl – 4 – methyl phenol or 2-6 di tertiary-butyl-Para-

cresol (DBPC).

Natural Inhibitors:

New insulating oil as normally refined contains small amounts of certain chemical

compounds that act as oxidation inhibitors. These naturally occurring materials retard oil

oxidation until such time as they expended. The rate at which the inhibitors in the oil are

used up is dependent upon the amount of oxygen available, soluble contaminants in the

oil, catalytic agents in the oil and temperature of the oil. In modern transformers, either

sealed to exclude air and moisture or protected by inert atmosphere, the benefits of the

inhibitors can be extended over many years. As the inhibitors are exhausted, the rate of

oxidation and the deterioration of oil increase. Reclaiming processes, such as acid

refining or clay treating can restore the oil so that it has most of its original

characteristics, but this has no effect upon restoring the usefulness of the natural

inhibitors occurring in the oil. In fact, both fuller’s earth and activated alumina remove

the natural inhibitors and the reclaimed oil has no resistance to oxidation. To overcome

this undesirable condition, synthetic oxidation inhibitors are used to extend the life of

reclaimed oil.

Significance:

The presence of inhibitors in the oil will increase the useful life of the oil with respect to

oxidative degradation in the presence of oxygen. As the oil is exposed to this type of

oxidative degradation, the oil is protected as long as there is inhibitor present. However,

as the process proceeds the inhibitors will be used up and when it is gone the oil will

degrade at much higher rate. Thus the determination of the amount of inhibitors present

in the oil can be used to estimate useful life of the oil. It can also be used to determine

whether or not new oil has been properly inhibited prior to its use. As inhibitor is used up

its concentration can be monitored and additional inhibitor added as needed to maintain a

proper concentration in the unit. Typical values for fresh oil are in the range of 0.25 to

0.35% DBP or DBPC by weight.

Due to chemical structure transformer oil oxidized easily. The oxidation process involves

chemical reaction between oil, oxygen and metallic compounds. The results of these

reactions are the formation of oil degradation by products, mainly acids, which may

affect the dielectric properties of the oil. Conservator type or breathing transformers

allow oxygen present in the ambient air to be absorbed by the oil. This oxygen and high

temperatures facilitates the oxidation process within the oil.

Antioxidants slow down the oxidation process by trapping the oxidation byproducts such

as free radicals and stop the spreading of the degradation of the oil. During this process,

the inhibitors are consumed, to the point where all the remaining inhibitors have trapped

free radicals. At this point the reaction will once more continue freely.

Methods to determine inhibitor content in the oil.

1. Infrared spectrophotometery (IR)

2. Gas chromatography (GC)

3. High performance liquid chromatography (HPLC)

Common antioxidants:

Butylated hydrooxtoluene (BHT), also known as butylhydroxytoluene, is lipoholic

(fat-soluble) organic compound that is primarily used as anti oxidant.

BHT is prepared by the reaction of p-cresol with isobutylene (2-methylpropene)

catalyzed by sulfuric acid.

CH3 (C6H4) OH + 2CH2 = C (CH3)2 – ((CH3)3C) 2 CH3C6H2OH

Alternatively, BHT is prepared from 2-6 ditetiary butyl phenol by hydroxymethylation or

aminomethylation followed by hydrogenolysis.

The species behaves as a synthetic analogue of vitamin E; primarily acting as a

terminating agent that suppresses auto oxidation, a process whereby unsaturated (usually)

organic compounds are attacked by atmospheric oxygen. BHT stops autocatalytic

reaction by converting peroxy radicals to hydro peroxides. It affects this function by

donating one hydrogen atom.

RO2. + ArOH – ROOH + ArO.

RO2. + ArO. - - NON RADICAL PRODUCTS

Where R is alkyl or aryl and where ArOH is BHT or related Phenolic antioxidants. One

BHT consumes two peroxy radicals.

2-6 Di-tert-butylphenol

2-6 di-tert-butylphenol is an organic compound with the structural formula 2-6

((CH3)3C) 2C6H3OH. This colorless solid alkylated phenol and its derivatives are use

industrially as UV stabilizer and antioxidant for hydrocarbon based products.

Oxidation stability determination:

There are three ASTM tests that will help in establishing the performance of transformer

oil relative to oxidation stability: ASTM D 1934, D2112 and D2420

ASTM D1934

In the procedure found in ASTM D1934, a 300 mL volume of oil, contained in the 400

mL beaker is aged for 96 hours in circulating-air oven controlled at 115°C with or

without presence of catalyst. It is particularly useful as a check on the continuity of

production. It is applicable to unused oils. The properties of acid content (D974),

dissipation factor (D924), and resistivity (D1169) are commonly used as criteria for

judging the oxidation stability of relative samples.

ASTM D2112

ASTM D2112 is intended as a rapid method for the evaluation of the oxidation a stability

of new mineral insulating oils containing synthetic oxidation inhibitors. This test is

considered of the value in checking the oxidation stability of new mineral insulating oil

contained 2, 6- diteriary-butyl Para cresol or 2,6 diteriary –butyl phenol in order to

control the continuity of this property from shipment to shipment.

The test specimen is agitated by rotating axially at 100 rpm at an angle of 30° from the

horizontal. Under an initial oxygen pressure of 90 psi in a copper bomb with a glass test

specimen container and copper catalyst coil in the presence of water at a bath temperature

of 140°C. The time for an oil to react with a given volume of oxygen is measured;

completion of test is indicated by a specific drop in pressure.

An acceptable value for oils run per ASTM D2112 is 195 minutes for new oil which

when inhibited by supplier is inhibited to a level of 0.3% DBPC.

ASTM D2420

Oxidation stability is measured by the tendency of oils to form the sludge and acid

products during aging. The test method is applicable to new oils, both uninhibited and

inhibited.

ASTM D2440 involves aging of 25 ml of oil in the presence of 300 mm of polished

copper wire. Oxygen is bubbled through the oil in the glass container which is in an oil

bath at 110°C. Two oil samples are oxidized for 72 and 164 hours or 500 hours

respectively. After prescribed aging periods, the samples are filtered to collect the sludge

formed. The sludge is weighed and the acid content of the filtrate measured.

Stages Of oil Oxidation:

Water Content (ASTM D1533):

Water in oil appears as an unwanted substance. It is accepted that water in microscopic

amounts & not liters is cause of more electrical break down than any other impurity.

Moisture causes hazard not only to the insulating qualities of oil but also to the

insulations that are immersed in oil.

Water may be introduced in the oil by leaking gaskets, poor handling or form the

insulating paper and oil degradation. As the paper degrades, it produces carbon dioxide

and water and as the insulating degrades, water acids, sludge and other polar compounds

are formed. So its presence is inevitable in the normal service of life of the transformer.

Water is polar liquid having high permittivity or dielectric constant it is therefore

attracted to areas of strong electric field. This sees the internal moisture distributed not

uniformly, but in fact potentially concentrating in most dangerous parts of the system.

The dielectric strength of the paper insulation decreases significantly when its water

content increases. Similarly dielectric breakdown voltage of the oil is also affected by

relative saturation of the oil.

It is important to note that water is continuous state of movement between oil and the

paper insulating system. This is caused by internal temperature variation due to load and

ambient conditions.

Maximum loading while retaining the reliable operation is function of the water content

in the oil. For example, dry transformer (<0.5% moisture) is less susceptible to water

vapour bubbles formation. In this case emergency loading at hot spot temperatures below

180°C is possible with little risk of bubble formation. In contrast a wetter transformer

with 2% moisture in the paper runs the risk of bubble formation with hot spot

temperature as low as 130°C under the same conditions. A more long term problem is

that excessive moisture accelerates the aging of the paper insulation almost

proportionally. Insulating paper with 2% moisture ages three times faster than paper with

the 1% moisture.

The destructive effect of the water:

• Expansion of the paper insulation, altering the mechanical pressure on the

component.

• Loss of insulating ability.

• Accelerating paper aging i.e. triggering decomposition of the fibers in the paper.

• Increased corrosion of the core & tank.

• Progressive consumption of oil additives.

This may occur due to:

• During the high load currents and at high ambient temperatures, dielectric

breakdowns can results from the reduced oil strength with high absolute amount

of water.

• With sudden high loads, water can boil of conductor surface and the vapour

bubbles can cause failures as they rise to the top.

• During the cool down period after high load, the relative saturation of the oil will

increase. At its extreme at 100% relative saturation, water will precipitate out &

greatly reduce dielectric strength of oil.

If the oil is oxidized to any extent, any water coming in to the transformer will partially

be absorbed in to the oil decay products (it is this fact which causes old or highly

oxidized oil to dissolve more water than the new oil). As the decay products build up in

the oil, the surface tension of the water or the interfacial tension between the oil & the

water is lowered dramatically. This heavier decay molecule will then recirculate

throughout the entire transformer and will find its way in to the paper insulation, or in to

areas of high electrical intensity thus reducing the insulation strength. The water saturated

oil decay molecule has a preference for the coolest parts of the transformer (bottom

leading to corrosion) and areas of highest di-electrical stress (leading to arching)

Water may be present in four possible forms they are,

• Free water – that is water that has settled out of the oil in a separate layer. It is

water indicated by low dielectric breakdown voltage.

• Emulsified water- or water that is suspended and has not yet settled out in to free

water (indicated by “caramel” colored oil. A high power factor value indicates

the possible presence of this suspended water trapped in oil decay products.

• Dissolved in the oil.

• Chemically bounded water- water which is chemically attached to the insulating

paper and which is released when oxidized.

Detection of water in mineral insulating oil:

The detection of the water in oil is performed in the laboratory is most commonly

performed by analytical technique called Karl Fischer titration described in ASTM 1533

or IEC 60184. Both methods employ coulometric titration technique involving the

reduction of an iodine-containing reagent. This methods are used to determine the

amount of water in oil sample on weight-to-weight basis or more popularly known as

PPM (parts per million).

The concepts of solubility and relative saturation can some times be difficult to

understand, but it is an important concept when accessing the dryness of the transformer.

Solubility is defined as total amount of water that can be dissolved in the oil at specific

temperature. The solubility of water is not constant in oil but changes due to temperature.

As temperature increases, the amount of water that can be dissolved in the oil also

increases. The increase is not linear but exponential in function. For example, at 10°C

36ppm of water that can be dissolved in the oil, where as when temperature is 90°C,

amount of water that can be dissolved in the oil in increases to 600 ppm. Following table

shows the list of calculated solubility limit for that particular temperature. If the water

content is more than the limit at that particular temperature than there is greater

possibility that free or emulsified water is available in the oil.

The solubility of the mineral oil can be calculated using.

Log S0= -1567/K+7.0895

Where: S0 is the solubility of water in mineral oil.

K is the temperature in Kelvin (°C+273)

Relative saturation (RS) is the actual amount of the water measured in the oil in relation

to the solubility level at that temperature. Relative saturation is expressed in units of %, is

the concentration of the water (WC) in the oil relative to the solubility (S0) or

concentration of the water the oil can hold at the measurement temperature.

RS = Wc/S0 (100%)

Where: WC is in ppm.

S0 is in ppm.

A simple example illustrates that the dielectric breakdown voltage of the insulating oils

is proportional to the relative saturation of water in the oil than it concentration in the

ppm. The humidity is controlled in this example so that concentration of the water is held

constant at 30 ppm. The first dielectric breakdown is measured at 100°C. at this

temperature the solubility of the water is 772 ppm. The relative saturation of water in oil

is therefore about 4% (30ppm/772 ppm*100), and the dielectric breakdown voltage of

well filtered oil is quite high. The temperature is now reduced to room temperature or

about 22°C. The solubility of water in the oil is about 60 ppm. And relative saturation is

50%. The dielectric breakdown voltage would be expected to be about half of what as

when relative saturation was very low. If the temperature is cooled to 0°C, the results of

the dielectric breakdown voltage should be quite low as solubility of the water at 0°C is

22 ppm. As the water content is higher than this water in oil starts to emulsify & begins

to condense. During all this water content is not changed.

We will now discuss the above point along with the relative saturation & its effect in

details.

Break Down Voltage (BDV):

The dielectric break down is the minimum voltage at which electrical flashover occurs

in the oil. It is a measure of the ability of oil to withstand electrical stress at power

frequencies without failure. A low value for the dielectric breakdown voltage generally

serves to indicate the presence of contaminants such as water, dirt or the conducting

particles in the oil.

Before moving ahead one should really understand that how break down happens in

the oil? In next few paragraphs we will try to explore this based on the work done by M,

Butcher.

Over the years, many models have been proposed to explain the mechanism of liquid

breakdown. Establishment of clear and effective model is difficult in part due to various

characteristics of liquid dielectrics. The physical nature of the liquids (it is extremely

difficult to get pure liquid), high density, viscosity, quality, thermal and electrical

properties, compared to gases adds multiple dimensions to problems associated with

developing model.

Many models begin with some form of charge injection at the electrode /liquid

interface where current injection at an area of fields leads to the formation bubble or low

density region. Inside the bubble, electron avalanches occurs producing current pulses, if

they fail to completely cross the gap; multiple impulses eventually lead to final

breakdown. How the formation of a bubble or low density region in the liquid develops is

still not completely understood since there are several ways to generate a bubble.

Explanations include pre-existing bubbles or dissolved gas, local heating of the liquid due

to energy injection, cavitations caused by fluid motion and electrical stress on the liquid

molecules.

Theories of liquid breakdown:

Liquid breakdown does not consist of a single comprehensive theory for breakdown that

is unanimously accepted for describing dielectric liquid break down phenomena. There

are several reasons for lack of a single theory. The complex nature of liquids makes the

theoretical analysis more difficult than gases. Liquid quality is a critical issue. It is very

difficult to create a pure liquid compared to pure gas. It is generally accepted that liquid

purity plays a very important role in the development final breakdown. The predominant

of liquid breakdown is the bubble theory, followed by the suspended particle and

electronic theories.

Bubble theory:

The bubble theory is based on a combination of gaseous and liquid components for

breakdown. In general it requires the formation of a bubble near the electrode tip. Then

after a bubble is present an electron avalanche or amplification process begins. The

avalanche is triggered either by the injection of more charges in to the vapour phase or by

field ionization occurring in the bubble due to the field enhancement caused by a region

of lower permittivity.

Several methods of bubble formation such as local heating, cavitation, or electrical

stress may occur simultaneously to one degree or another in liquids. The mechanism that

dominates is controlled by the properties of the fluid at the time of the test. The primary

mechanism discussed is localized energy injection. Partial discharges inject “hot”

electrons into the liquid and cause local heating and vaporization. Hot electrons are

electrons with energy in excess of the thermal energy. The vaporization phase change

consumes most of the energy of the pulse.

T.J Lewis put forth a mechanical model for bubble formation. His model for the

formation of a bubble or low density region comes from mechanical properties of the

fluid.

Fig. Illustration of crack development at the cathode surface.

a) Pre breakdown conditions.

b) “Crack” develops at the surface of the cathode

c) Jet of electrons from the cathode surface act as an extension of the cathode

and severely distorts the local field.

d) An array of the secondary cracks develops in the radial field.

Figure illustrates the process of mechanically creating a low density/bubble region in a

liquid dielectric. The mechanical stress placed on liquid molecule by very large electric

fields near break down 108 to 10

9 V/m, enhances the thermal generation ruptures or holes

throughout the liquid not just at the electrode/liquid interface. These holes in the liquid

will provide weak spots in the molecular structure of the liquid and allow a “crack” in the

liquid to form at the cathode surface. The crack is now low density or even vacuum

region, occasionally referred as a vacuole where electrons can be injected. The vacuole

does not require a vaporization of the liquid to form, it formation due to mechanical stress

instead of energy injection. Vacuole allows breakdown to develop in low density region,

the same as bubble created by another process. When electron is injected in crack it

becomes the extension of cathode, similar to newly formed micro protrusion. The

enhanced field will repeat the process.

It was experimentally observed that in many liquids, breakdown strength depends

strongly on applied hydrostatic pressure, suggesting that a change of phase of the medium

is involved in the breakdown process, which in other words means that a kind of vapor

bubble formed is responsible for breakdown. The following processes have been

suggested to be responsible for the formation of vapor bubble.

a) Gas pockets at the surface of the electrodes.

b) Electrostatic repulsive forces between space charges which may be sufficient to

overcome the surface tension.

c) Gaseous products due to the dissociation of liquid molecules by electron

collisions; and

d) Vaporization of the liquid by corona type discharges from the sharp points and

irregularities on the electrode surface.

Once bubble is formed it will elongated (log and thin) in the direction of the electric

field under the influence of electrostatic forces. The volume of the bubble remains

constant during elongation. Breakdown occurs when voltage drop along the length of the

bubble becomes equal to minimum value on the Paschen’s curve for the gas in the

bubble. The break down field given as

( )( )

1

2

1 2 h0

1 2 0

2 2 V1E 1

(2rE )r

ε επσ

ε ε+

= −

−

(3.2)

where σ is the surface tension of the liquid, 1ε is the permittivity of the liquid, 2ε is the

permittivity of the gas bubble, r is the initial radius of the bubble assumed as a sphere and

bV is the voltage drop in the bubble (corresponding to minimum on the Paschen’s curve).

From this equation it can be seen that the breakdown strength depends on the initial size

of the bubble which in turn is influenced by the hydrostatic pressure and temperature of

the liquid. But this theory does not take into account the production of the initial bubble

and hence the results given by this theory do not agree well with the experimental results.

Later this theory was modified, and it was suggested that only incompressible bubbles

like water globules can elongate at constant volume, according to the simple gas law

pv = nRT. Under the influence of the applied electric field the shape of the globule is

assumed to be approximately a spheroid. These incompressible bubbles reach the

condition of instability when β , the ratio of the longer to the shorter diameter of the

spheroid is about 1.85, and the critical field producing the instability will be:

10

1 1 2

600 HE GR

επσ

ε ε ε

= −

− (3.3) where σ = surface tension, R= initial radius of

bubble, 1ε = permittivity of the liquid dielectric, 1ε = permittivity of the globule,

( )

1

122 2

1 cosh1

11

Gβ β

ββ

−

= − − −

And

1

2 32

1H 2 2 1β β

β

= − −

For a water globule having R=1 mµ with 143dyne/cm and 2.0σ ε= = (transformer oil),

the above equation gives a critical field cE 226 /Kv cm= which is approximately the

maximum strength obtained for commercial oils.

In the case of gas bubbles the equation for the critical field is rewritten as

( )( )

1

1

1 2 4 112 2

c

1 2 1 2

8A BE 600 G cosh

3R

πσε

εθ

ε ε β ε ε

= − − −

Where, ( )

2

3 2

1 2

2 1A= 1

B=2 1

β β

ε β ε β

− −

− −

G, σ and R are as above for liquid globules, and

( )1

35 2

1 2-1

3

271 PRcosh

3 2B

β ε εθ

σ

−

=

Where P is the hydrostatic pressure. (Equations 3.2-3.4 are in c.g.s. units). The

expressions are quite complicated, and the breakdown voltages were obtained using a

computer. Results thus obtained showed good agreement with the experimental results in

n-hexane. This theory suggests that sub-microscopic particles (diameter 100-250 A) and

bubbles greatly influence the maximum electrical strength attainable in commercial

liquids. The critical condition is reached when cavities are formed due to zero pressure

conditions given by

c vp es s hP P P P P ,= = + + (3.5) where,

c

vp

es

s

h

P coulombic pressure,

P vapour pressure inside the cavity,

P electrostatic pressure,

P pressure due to surface tension, and

P hydrostatic pressure.

=

=

=

=

=

From this condition, an expression has been obtained for the maximum breakdown

strength of pure liquids which was found to be in good agreement with the experimental

results.

In general, the cavitation and bubble theories try to explain the highest breakdown

strengths obtainable, considering the cavities or bubbles formed in the liquid dielectrics.

Thermal Mechanism of Breakdown

Another mechanism proposed to explain breakdown under pulse conditions is thermal

breakdown. This mechanism is based on the experimental observations of extremely

large currents just before breakdown. These high current pulses are believed to originate

from the tips of the microscopic projections on the cathode surface with densities of the

order of 21A/cm . this high density current pulses give rise to localized heating of the oil

which may lead to the formation of the vapor bubbles. The vapor bubbles are formed

when the energy exceeds 7 310 W/cm . when a bubble is formed, breakdown follows,

either because of its elongation to a critical size or when it completely bridges the gap

between the electrodes. In either case, it will result in the formation of a spark.

According to this mechanism, the breakdown strength depends on the pressure and the

molecular structure of the liquid. For example, in n-alkanes the breakdown strength was

observed to depend on the chain length of the molecule. This theory is only applicable at

very small lengths ( )100 mµ and does not explain the reduction in breakdown strength

with increase gap lengths.

Suspended particle theory:

In commercial liquids, the presence of solid impurities cannot be avoided. These

impurities will be present as fibers or as dispersed solid particles. The permittivity of

these particles ( )2ε will be different from the permittivity of the liquid ( )1ε . If we

consider these impurities to be spherical particles of radius r, and if the applied field is E,

then the particles experience of force F, where

( )2 13

2

1 2

1F= grad E

2 2r

ε ε

ε ε

+

− (3.1)

This force is directed towards areas of maximum stress, if 2 1ε ε⟨ , for example, in the

case of the presence of solid particles like paper in the liquid. On the other hand, if only

gas bubbles are present in the liquid, i.e. 2 1ε ε⟨ , the force will be in the direction of areas

of lower stress. If the voltage is continuously applied (d.c) or the duration of the voltage

is long (a.c), then this force drives the particles towards the area of maximum stress. If

the number of particles is large, they become aligned due to these forces, and thus form a

stable chain bridging the electrode gap causing a breakdown between the electrodes.

If there is only a single conducting particle between the electrodes, it will give rise to

local field enhancement depending on its shape. If this field exceeds the breakdown

strength of the liquid, local breakdown will occur near the particle, and this will result in

the formation of gas bubbles which may lead to be breakdown of the liquid.

The value of the breakdown strength of the liquids containing solid impurities was found

to be much less than the values for pure liquids. The impurity particles reduce the

breakdown strength, and it was also observed that the larger the size of the particles the

lower were the breakdown strengths.

Electronic breakdown theory:

A review of this theory is discussed by Lewis. He discussed the conditions necessary for

cathode and anode initiated process. In general, it is unlikely for electrons injected in to

the liquid to produce significant ionization in the liquid state to develop an amplification

process. Cathode initiation assumes that electrons are injected into the as a column of

electrons emitted from micro-protrusion on the cathode. The electrons will collide with

the molecules locally heating the liquid through collisional impacts. The heating of the

liquid reduces the density allowing future electrons to obtain more energy from the field

before colliding with another molecule. The impacts that ionize a molecule will leave two

slow electrons to drift in the field. The build up of low energy electrons at the front of the

streamer “marks the onset of spherical expansion under the columbic forces of the

consequent cloud of negative charge”. The streamer will expand and appear thick and

bushy as it propagates across the gap. The lower density of the region the faster electron

amplification will occur. The obvious extreme of this process is the formation of an

avalanche within a bubble.

Anode initiation of break down requires a rapid charge development mechanism.

Assuming that holes emitted from the anode via resonance tunneling from molecule to

molecule, requires that the intermolecular distance be as small as possible to achieve high

probabilities for tunneling. High density regions are required to maintain minimal

molecular spacing. Tunneling establishes a hole propagation pathway through the liquid.

These pathways will trace hole through the liquid to the electrode, and should appear as

thin filamentary structures in the liquid. The path left by the propagation of the hole

through the liquid will be ideal return path for a complementary electron to propagate in

the opposite direction. The electrons prefer the lower density region created by passing

hole through liquid. Energy transfer via collisions that heat the fluid will be detrimental

to hole propagation since the density of the liquid will be reduced, reducing tunneling

probabilities.

c. Stressed Oil Volume Theory

In commercial liquids where minute traces of impurities are present, the breakdown

strength is determined by the “largest possible impurity” or “weak link”. On a statistical

basis it was proposed that the electrical breakdown strength of the oil is defined by the

weakest region in the oil, namely, the region which is stressed to the maximum and by

the volume of oil included in that region. In non-uniform fields, the stressed oil volume

is taken as the volume which is contained between the maximum stress ( )maxE contour

and 0.9 ( )maxE contours. According to this theory the breakdown strength is inversely

proportional to the stressed oil volume.

The breakdown voltage is highly influenced by the gas content in the oil, the

viscosity of the oil, and the presence of other impurities. These being uniformly

distributed, increase in the stressed oil volume consequently results in a reduction in the

breakdown voltage. The variation of the breakdown voltage stress with the stressed oil

volume is shown in Fig. 3.4.

400

300

200

100

0

10-2 1.0 102 104 106

Stressed oil volume (cc)

Breakdown Stress (KV/cm)

Fig 3.4 Power frequency (50 Hz) a.c breakdown stress as function of stressed oil volume

one minute withstand voltage

With steady voltage rise

3.4 CONCLUSIONS

All the theories discussed above do not consider the dependence of breakdown strength

on the gap length. They all try to account for the maximum obtainable breakdown

strength only. However, the experimental evidence showed that the breakdown strength

of a liquid depends on the gap length, given by the following expression,

n

bV Ad= Where,

d=gap length,

A=constant,and

n=constant,always less than 1.

The breakdown voltage also depends on the nature of the voltage, the mode in which the

voltage is applied, and the time of application. The above relationship is of practical

importance, and the electrical stress of given oil used in design is obtained from this.

During the last ten years, research work is directed on the measurements of discharge

inception (starting) levels in oil and the breakdown strengths of large volumes of oil

under different conditions.

It may be summarized that the actual mechanism of breakdown in oil is not a simple

phenomenon and the breakdown voltages are determined by experimental investigations

only. Electrical stresses obtained for small volumes should not be used in the case of

large volumes.

Di electric dissipation factor:

This measure of the quality of oil. Low DDF indicates the good oil with low dielectric

losses and low level of soluble polar ionic or colloidal contaminants.

All kinds of dielectric or insulation materials and systems can be characterized by its

inherent polarization phenomena. Which in frequency domain can be expressed by a

capacitance ‘C’ and a magnitude of power dissipation (di electric loss) as quantified by

dissipation or loss factor tan delta. Where as these quantities within a wide frequency

range are of at most interest for new materials or even for the quality control of well

known materials on receipt of delivery. These tests are generally performed at the

frequencies for which equipment/product is designed. Such tests are in general made in

dependence of test voltage applied both magnitudes, capacitance and dissipation factor

shall be essentially constant with increasing voltage as insulation system are linear

systems and any ‘tip-up’ of the tan delta with voltage level called ‘ ionization knee’ is

preliminary indication of partial discharge within the system.

Transformer oil is a non-polar liquid with a capacitance that changes very little with

frequency. The relative permittivity, rε , of oil is usually in a range 2,1-2,3. The dielectric

loss of oil is mainly governed by its DC conductivity that varies considerably with the

quality of the oilError! Reference source not found. As a result, measured dissipation

factor, as a function of frequency will slope downward in a straight line as shown in

fig

10-4

10-3

10-2

10-1

100

101

102

0,001 0,01 0,1 1 10 100 1000

Oil conductivity = 10-12

S /m

Oil conductivity = 10-10

S /m

Frequency [Hz]

Transformer oilsDissipation factor

(Tan )δ

10-3

10-2

10-1

100

101

10-3

10-2

10-1

100

101

102

103

mc=1.5%, σ=2e-12 S/m

mc=3.5%, σ=2e-12 S/m

mc=1.5%, σ=2e-10 S/m

dis

sip

atio

n facto

r (t

an

δ)

frequency, [Hz]

50 Hz

60 Hz

High oil conductivityLower moisture content

in paper

Good oil conductivity

High moisture content in paper

Corrosive sulfur:

INTRODUCTION

Corrosive sulfur and the effect that it has in transformer systems can be significant. The

extent of the corrosion damage caused by sulfur, if left unchecked, can be so severe as to

cause failure of the apparatus. The problems with corrosive sulfur have been recognized

for quite some time. As early as 1948, F.M. Clark and E.L. Raab issued a report on the

subject for method development within what is known now as ASTM Committee D 27

and eventually became ASTM Method D 1275. Sulfur is found in many materials of

transformer construction including the copper, paper insulation, gaskets and oil. Not all

sulfur is considered corrosive but the tendency to operate transformers at substantially

higher temperatures can aggravate an already present corrosive sulfur condition or

convert stable compounds into reactive ones that will cause damage.

PRESENCE OF SULFUR IN MINERAL OIL

There are different types of sulfur compounds found in refined transformer oil but not all

types are considered to be corrosive or reactive. Elemental sulfur and sulfur compounds

in concentrations up to 20% are present in the crude oil used to make transformer oil.

There are five basic groups of sulfur and sulfur compounds found in crude oil (see Table

1).

Sulfur and Sulfur Compounds Found in Crude Oil

Sulfur is commonly found in crude oil sources as it comprises almost 0.05% of the

earth’s crust. As shown in Table 1, elemental sulfur and the sulfur-containing mercaptans

are very reactive followed by sulfides. Reactive sulfur is mainly in the form of organic

sulfur compounds like R-SH, where the sulfur is attached at the end of an organic

molecule. When the molecule is more complex, for instance when the sulfur is

surrounded or contained within the molecule then the sulfur compounds are more stable

and less reactive, like in R-S•S-R. Thiophenes are the most stable of all these sulfur

compounds. Some sulfur compounds can actually aid in the oxidation stability of the

transformer oil and may also act as metal passivators and deactivators reducing the

catalytic effect on oil oxidation in transformers. The goal of the refining process is to

either remove or convert many of the corrosive and reactive sulfur species (i.e. elemental

sulfur, mercaptans, and sulfides) to more stable compounds such as thiophenes in an

unsaturated ring and disulfides in a saturated form. The steps in the refining process that

aid in this effort are atmospheric distillation at various temperatures, vacuum distillation,

catalytic reaction, and hydro-treating and hydro-generation. It should be recognized that

the refining process is not always totally successful as incomplete refining may leave

small quantities of mercaptans behind or the hydrogenation process may produce

elemental sulfur as opposed to hydrogen sulfide.

After refining, there is some sulfur left but the total sulfur (comprised of the five groups

listed above) remaining in new oil product is expected to be from 0.02 % to 1%. This

information was slightly dated so the Doble Materials Laboratory analyzed several

samples from Doble annual Survey 93 and found that most oils had very low total sulfur

content as shown in Table 2.

WHAT IS CORROSIVE SULFUR?

Corrosive sulfur species are defined as all organic sulfur compounds that will react with

mercury to form sulfides, such as mercaptans. Elemental (free) sulfur is very reactive and

will react to form corrosive acids. It has been suggested that low elemental sulfur levels

(low ppm range) can cause a corrosive condition [4]. It has recently been suggested that

the definition of corrosive sulfur be narrowed to apply only to elemental sulfur whereas

those organo-sulfur compounds that react to cause a corrosive condition be termed

reactive sulfur.

REACTIONS OF CORROSIVE AND REACTIVE SULFUR

Corrosive and reactive sulfur compounds can react on contact with copper and other

metals. Copper is, by far, the least resistant metal to a sulfur attack. Effects of elemental

sulfur are even more problematic as its ability to combine with copper does not require

heat to promote the reaction. In oxygen deficient environments such as those found in

sealed, gas blanketed and sealed conservator transformers, corrosive and reactive sulfur

species combine with copper, aluminum, and other metals to form copper or cuprous

sulfide (Cu2S), aluminum sulfide (Al2S3) and other inorganic sulfides. Copper sulfide is

black, gray, green, blue, or violet in color and is sometimes confused with carbon.

Aluminum sulfide is a yellowish-gray material that can become very gray in the presence

of oxygen and water.

In the presence of an oxygen environment such as that available in sealed transformers

that have a significant leak, free breathing transformers, free breathing conservator

transformers, and other free breathing apparatus such as OCBs and LTCs, different types

of compounds are formed from the reaction of sulfur with metals. Oxygen can also

become available from the copper itself. The copper used for the windings is usually

CDA-110 (UNS-C11000) material for ETP copper which is termed electrolytic tough

pitch. This specification has a minimum copper purity requirement of 99.90% and it is

not considered an oxygen free material (<5 ppm). There is not an oxygen specification

for CDA-110 but it is usually contains about 500 ppm of oxygen or less. Copper winding

samples that the Doble Materials Laboratory has tested contained around 200-250 ppm of

oxygen. Different grades of copper can contain much higher amounts of oxygen that is

then available for use in a reaction with sulfur. Therefore, the manufacturer of the

transformer must be careful in selecting the correct grade of material for construction.

Reactions involving oxygen, sulfur, copper, aluminum, or other metals can produce

copper or cuprous sulfite (Cu2SO3), copper sulfate (CuSO4), aluminum sulfate

[Al2(SO4)3], and other inorganic sulfates. Copper sulfite is usually white to pale yellow

in color whereas copper sulfate is white or pale brown in color and aluminum sulfate is a

white material.

NONCORROSIVE TO CORROSIVE

One of the major questions is if noncorrosive sulfur species can be converted to corrosive

and reactive species in a transformer. Experience has shown that non-corrosive sulfur can

become corrosive after being exposed to elevated temperatures on hot metal surfaces and

thus produce metal sulfides. This attack would corrode the metal surfaces. To make

matters worse, the corrosion material could detach and become nuclei for discharge and

gas inception [3]. This may not be of concern with oils with low sulfur contents that pass

the corrosive sulfur test as the quantity of corrosive sulfur compounds produced may not

be sufficient enough to cause extensive damage.

The published literature does not detail if arcing in a transformer can change stable sulfur

species into reactive or corrosive forms. The Doble Materials Laboratory performed

experiments in which a voltage of 25 kV was applied to sustain an arc through Cross Oil

CrossTrans 106 transformer oil in a test cell equipped with a needle to sphere with a 0.1

inch gap. The CrossTrans 106 was found to be noncorrosive prior to testing. The arcing

produced the following gases:

The arcing did not reduce the total amount of sulfur left in the oil nor did it convert any

of the sulfur compounds already present into free sulfur. However, conversion of some of

the thiophene compounds did occur (see Table 6). The presence of sulfates and sulfites

cannot be determined by this method and it may be that those were the compounds that

were converted. This may explain the differences between the before and after test

results. What is of significant interest is the amount of unidentified volatile sulfur

compounds that were created through the arcing process. No mercaptans or sulfides

(corrosive or reactive sulfur) were formed. The lack of these sulfur species may be a

result of the energy applied through arcing that could have been so severe that any

reactive or corrosive species that were produced were instantaneously degraded and thus

none remain.

SOURCES OF SULFUR IN TRANFORMER SYSTEMS

Oil is not the only material that contains sulfur. Sulfur compounds are also present in the

gaskets, some water based glues, copper and paper insulation used in the manufacture of

transformers. Sulfur can also be introduced into the transformer through accidental means

such as through the use of incompatible hoses.

It is generally accepted that older gaskets used in transformer applications were made

from cork, cork/glyptal, and corkprene and then in more recent years, the most oil

compatible gaskets have been nitrile rubbers such as BUNA-N and a fluoroelastomer

such as VITON®. Properly made nitrile rubbers (butadiene acrylonitrile) and

fluoroelastomers (fluorinated hydrocarbon) are excellent gaskets for use in transformers.

In the manufacture of these materials, sulfur is used in the curing process when the

formulations are being developed into a hardened material. The curing process is

supposed to eliminate all sulfur from the finished product. Most gasket manufacturers

assume that the sulfur is eliminated after the curing process. In some cases, the

concentration of sulfur contained in the final gasket product is not monitored. Doble

Engineering performed scanning electron microscopy/energy dispersive x-ray analysis

(SEM/EDX) analysis on numerous gaskets taken from recently manufactured

transformers. Each gasket was prepared for analysis by cleaning the outside surface with

a sulfur-free hydrocarbon solvent. The gasket was then cut lengthwise to reveal the inside

surface. The outside and inside surfaces of each gasket were coated with evaporated

graphite. The samples were then subjected to SEM/EDX analysis in which an electron

beam of the scanning electron microscope enters the bulk of a sample producing an x-ray

emittance. The x-ray peak positions, along the energy scale, identify the elements present

in the sample and can provide the percentage concentrations of each of these elements

thus providing an elemental breakdown of the material or particles. Results from two

gaskets are shown in Tables 7 and 8.

The inside of the O-ring gasket had a slightly elevated background level which indicates

that there is some organic component as well as the elemental component. Both gaskets

contained a large amount of sulfur especially the O-ring gasket. The SEM/EDX analysis

was performed on the inside surface of the gasket as well to distinguish between the

outside surface possibly being contaminated with corrosive sulfur from the oil. It is clear

that sulfur is a component of both original gaskets. The original formulations for a nitrile

rubber, fluoroelastomer or a silicone rubber, which is a polydimethylsiloxane, do not

contain any sulfur.

In discussions with elastomer manufacturers it was found that very few if any

manufacturers, (except for E.I. Dupont) were performing any chemical testing on the

finished product to determine what amount of sulfur remained if any. There also does not

appear to be any standard on what percentage of sulfur should remain in the final product.

It then becomes obvious that the onus is on the final user of the material to specify sulfur-

free or low sulfur material for use or to test it prior to use.

In light of this information, additional SEM/EDX analysis was performed on gasket

material that was available in the Doble Materials Laboratory. One sample was a

fluoroelastomer and another was a nitrile rubber produced by Parker. The results are

shown in Tables 9 and 10.

As shown in the two tables above, the fluoroelastomer shows no sulfur on the inside

surface and very little on the outside suggesting that material was cured correctly. The

opposite is true of the Parker nitrile material which shows extremely high sulfur content

on both surfaces suggesting that the sulfur was not removed after the curing process.

Water-based glues, used to secure the paper insulation during manufacture, often contain

sulfur compounds. There has been at least one known instance in which the glue used in

the manufacture of the windings has contributed to a corrosive sulfur condition.

Most coppers used in manufacturing transformer windings contain some impurities and

sulfur happens to be one, along with silver, arsenic, phosphorous, tellurium and oxygen.

The amount of sulfur that is allowed in most of the electrical grades of copper is 15 ppm

or less. In the analysis performed at the Doble Materials Laboratory on copper samples

from windings, the sulfur contents were very low at 5 ppm or less. However, there still

has to be care taken in the selection of materials used in construction so that copper with

a high sulfur content is not used.

The pulping process for electrical Kraft paper converts the wood chips to cellulose by

removing the majority of lignin (95-98.5%) and other impurities [5]. There are two basic

processes. The sulfite process is considered an acidic process and uses sulfur dioxide,

sulfuric acid and calcium bisulfite. The main process used today and the one that is used

to produce electrical grade Kraft papers is the sulfate process which is also called alkaline

pulping. Sodium hydroxide and sodium sulfide is used in what is termed the cooking

process. The cooking process under conditions of heat, pressure and chemicals (pulping

liquors) removes the lignin and impurities from the wood chips in order that only

cellulose remains. The pulping liquor is removed and recycled for use again and the

remaining cellulose pulp is washed several times to remove as much as the pulping liquor

as possible from the cellulose pulp. The Kraft process is slightly different in that the same

chemicals are used but the pulp is intentionally undercooked and results in the darker

color of the paper as well as exceptional mechanical strength. The pulp fibers in the Kraft

process do absorb some of the sulfur compounds that cannot be removed via the

washing/rinsing process [5]. The Doble Materials Laboratory performed analysis to

determine how much total sulfur remains in the finished paper products. The first analysis

performed was SEM/EDX analysis of new Kraft and thermally-upgraded (TU) Kraft

from United States manufacturers. These results are listed in Table 11 and are for the

surface of the paper only.

In addition, several different samples of Kraft paper insulation were analyzed for total

sulfur and total sulfate content. The results are present in Table 12.

As shown in Table 12, the amount of sulfur varies between various manufacturers of

electrical paper and sometimes considerably. The amount of sulfur present is fairly

significant in most of the samples. The amount of reactive or corrosive sulfur in relation

to the total is unknown although it is assumed that the amount of sulfates in the sample is

at least the minimum amount.

Accidental contamination of the transformer oil with corrosive and reactive sulfur

compounds can occur by use of incompatible materials or contaminated processing

equipment to transfer oil. For example, hoses that are made from natural rubber or

gasoline hoses both contain high amounts of sulfur that are easily transferable to the oil

that is being pumped through them. Extra care must be exercised in the selection of hoses

so that no incompatibility exists. Oil processing equipment runs the risk of being

contaminated from processing a transformer with corrosive and reactive sulfur to another

transformer that does not. The best safeguard is to check the remaining oil left in the

processing equipment prior to its next use.

As described above, several materials in the transformer will contain sulfur such as the

copper, paper and oil. In some cases, the sulfur species in question are stable or are so

tightly bound in the material that they would not be available for reactions. In other cases

some of the sulfur compounds are corrosive or reactive. In these cases, appropriate

material compatibility testing should screen out these materials before they are used in

transformer construction.

EFFECTS OF CORROSIVE/REACTIVE SULFUR

Corrosion of the metal surfaces especially exposed copper surfaces is one of the primary

reactions of a corrosive/reactive sulfur condition. Figure 1 shows how exposed metal

surfaces can be worm-holed by a corrosive/reactive sulfur attack. Figure 2 shows NLTC

contacts that are overly contaminated with what appears to be a buildup of carbon but in

reality is copper sulfide.

The NLTC contacts shown in Figure 2 were discovered by accident. A maintenance

function involving a power transformer led to the belief that a piece of hardware had

fallen into the main tank of the transformer. A visual inspection was conducted through

the manhole of the unit and the blackened NLTC contacts were visibly apparent. This

caused the unit to be removed from the station, drained and inspected by entering through

a manhole. After draining and entering the unit, personnel wiped the NLTC contacts with

a cloth and discovered that the plating of the contacts was removed along with the black

material. This of course caused concern and an investigation was initiated in which it was

determined that copper sulfide, not carbon was the black shiny material found on the

contacts. This type of corrosion, as shown in Figures 1 and 2, could easily lead to

overheating and arcing of these components thus severely damaging or causing failure of

the apparatus. The effects of a corrosive/reactive sulfur attack on a metal conductor do

not always result in a black coating. In some cases, a silver coating or a silver

discoloration of the copper conductor surface occurs as in Figure 3.

The analysis on the paper surface shown in Figure 4 was performed by SEM/EDX as

previously described. In analysis performed on paper, the SEM/EDX analysis usually

indicates a very high organic content due to the cellulose composition. In this particular

case (Figure 4), the copper-sulfur contamination that had been transferred to the paper

insulation had been significant enough to mask out most of the organic component. In

analyzing this paper, a closer examination of the surface was conducted and SEM

micrographs were produced at a magnification of 200 times. Figure 5 is a SEM

micrograph of uncontaminated Kraft paper and crepe paper.

The individual paper fibers are clearly visible in both the Kraft and crepe papers in Figure

5 with no signs of foreign material present. Even the crimping of the crepe paper is

clearly visible. When copper/sulfur or other metal/sulfur contamination of the paper

surface occurs the results can be profound as shown in Figure 6.

All the fiber surfaces and the gaps between the fibers are encrusted with a contaminating

material in Figure 6A. The contamination present on the surface of the paper in Figure 6B

has not yet progressed to the same point as in Figure 6A but there is significant

contamination present and consists of mainly copper sulfides and sulfates. The small

spherical structures in the left-hand portion of Figure 6B are composed mainly of copper

and aluminum sulfides and sulfates. In an effort to evaluate the effect that this type of

contamination has on paper insulation, dielectric breakdown strength testing, by ASTM

Method D 149, was performed on the three layers of insulation that surrounded the

copper conductor. The first layer or the one closet to the conductor was the insulation

shown in Figure 6A. The other two layers of paper insulation were also contaminated but

not to the degree of the first layer. The results of this testing is shown in Figure 7.

Layer 1, shown in Figure 7, had an extremely low dielectric strength at 80 volts/mil. As

the contamination is reduced with each paper layer further out from the conductor, the

dielectric strength increases significantly to almost as new condition (1800 volts/mil oil

impregnated). Of significant interest is the fact that the mechanical strength was not

impacted by the contamination as the DP values for all three layers ranged from 903 to

938. The silvery appearance of the paper in Figure 4 is due to the alteration of the sulfur-

bearing compound due to exposure to higher temperatures, as copper sulfide is usually

gray-black in appearance. This reaction of copper and sulfur created deposits on the first

and second layers of paper on the same sample. Because of the deposition of the copper

and sulfur, the first inside wrap became a poor insulator and instead was more

conductive. When the corrosive sulfur contamination is this severe, a failure of the

transformer is almost inevitable as in this case.

Testing For Corrosive sulfur:

ASTM D 1275B, Corrosive Sulfur in Oil

This test ages 220 mLs of oil in contact with a copper strip in a sealed vessel for 48 hours

at 50C. The primary purpose of the test is to determine if any corrosive sulfur

compounds in the oil will react with the copper strip to turn it gray or black. It is a

subjective test in that there is a comparison of colors of the copper strip with some

colored standards and a table of descriptions listing what is corrosive and noncorrosive

(Figure 1A and 1B). The test was recently modified from just D 1275 in April of 2006 to

D 1275A and D 1275B. The B method is more rigorous then the old method D 1275 or D

1275A and was developed in response to the problem with corrosive sulfur. The issue

was that some of these oils met the requirement of the oil specifications using the old D

1275 (D 1275A) test yet became corrosive while in service and eventually caused failures

of transformers, LTCs, and reactors. ASTM D 27 committee recognized this flaw and

modified the test method to include D 1275B. This is a very routine test.

ASTM D 5623, Sulfur Speciation

The ASTM D 5623 analysis is performed by gas chromatography with sulfur selective

detection and covers the detection of volatile sulfur-containing compounds. The test

method will not identify all individual sulfur components. Detector response to sulfur is

linear and essentially quimolar for all sulfur compounds; thus, both unidentified and

known individual compounds are determined. However, many sulfur compounds, for

example, hydrogen sulfide and mercaptans, are reactive, and their concentration in

samples may change during sampling and analysis. This test can be used to isolate

specific sulfur compounds, some of which may or may not be responsible for corrosive

sulfur attack. One of the issues is that depending on the lab, the database of sulfur

compounds that can be analyzed can be small or large but usually not more than about 70

compounds. There are thousands of sulfur compounds, and this test does not have the

ability to cover them all. Certain types of sulfur compounds cannot be fully isolated and

identified. It is usually not considered a routine test and is mostly used for research

purposes.

ASTM D 4294, Total Sulfur in Oil

There are actually several ASTM tests that can be used for the detection of total sulfur in

oil. This just happens to be the one that is used by our laboratory. This test determines the

total amount of sulfur in the oil but does not determine whether or not the compounds

being detected are corrosive.

Transformer oil is made from petroleum crude that has naturally occurring sulfur.

Depending on where in the world the oil is from dictates the amount of sulfur in the

crude. The process of refining the crude to transformer oil and other products attempts to

remove reactive (corrosive) sulfur compounds by converting them to hydrogen sulfide

gas which is easily removed from the process. The refining process can also convert

some of the less stable sulfur compounds into more stable un reactive sulfur compounds

that can provide benefits in the final refined product. The amount of conversion and

removal depends on the crude and the refining process itself, as each refiner has its own

distinctive process. Most modern transformer oils have a final sulfur content of less than

1500 ppm, and some are even less than 10 ppm. This is a routine test, but it only provides

information on how much sulfur is in the oil, not whether it is deleterious or not.

ASTM D 3227, Mercaptans in Oil

Mercaptans are sulfur compounds that can be very reactive. This is a potentiometric

titration test that is used to determine the concentration of mercaptans as a class of

compounds in the oil, but it will not be able to identify specific compounds. This test is

used in the fuel and lubrication industries often and is starting to be used in the

transformer oil industry. One of the problems with mercaptans is that some of them are

highly volatile and thus not easily sampled for. It is recently not a routine test, but it is

being used more and more.

Doble Covered Conductor Deposition (CCD) Test, Doble Test

There are two other variations, one developed by ABB and the other by Siemens, which

preceded the Doble method. A variation of the Siemens method is being developed into

an IEC test. This new test is very important in that the purpose of the test not only

determines if the corrosive sulfur will attack the copper and form copper sulfide, but also

if copper sulfide formations will develop in the paper insulation. The test is conducted by

taking an abraded copper rod and wrapping new Kraft paper insulation around the rod.

Two rods are prepared for each sample, and they are placed in 20 mL headspace vials.

Oil is added and then the vials are sealed. A stainless steel needle of a particular diameter

is then pierced through the septum of one vial and left there to allow air ingress during

aging (see Figure 2A and 2B). The vials are then aged at 140C for four days. At the end

of the aging cycle, the vials are removed and the copper rods with paper retrieved. The

paper is then removed from the copper, and both are washed in a solvent to remove the

oil. The rod is inspected to determine if it has been tarnished by corrosive sulfur, and the

paper is inspected to determine if deposition has occurred. The presence of deposition in

the paper is important as most of the recent transformer failures have been due to copper

sulfide deposits in the paper causing a severe reduction in dielectric strength. The

resulting "deposition," whether heavy, moderate, or light, means that the interaction of

the copper and oil with the paper insulation over the duration of the test resulted in an

obvious deposit of copper sulfide, copper, or oil/paper degradation by-products onto or in

the paper insulation. This is becoming a very routine test.

Doble Covered Conductor Deposition (CCD+DT), Doble Test.

A CCD test with dielectric breakdown voltage is performed on the paper after aging.

During the research conducted for the development of the CCD test, it was noticed that

many oils, under the conditions of the test, actually formed deposits on and in the paper.

It was determined that if a dielectric breakdown voltage test was performed on the paper

from the copper rod, it could be used to qualify the deposits. Basically there are two

categories of deposits: those having no impact on the dielectric strength of the paper and

those that negatively affect the dielectric strength of the paper. Obviously the deposits

affecting the dielectric strength by lowering it are of concern. This is also becoming a

very routine test.

Dibenzyl disulfide (DBDS in Oil), Doble Test

DBDS is a sulfur compound found in certain transformer oils. DBDS is not thermally

stable at higher temperatures and breaks down into benzyl mercaptans which is very

corrosive and attacks the copper quite quickly. It should be remembered that DBDS is not

the only compound to cause severe copper sulfide deposition on the copper. There are

other compounds that cause corrosive sulfur attack but as of yet have not been identified.

This is now a very routine test.

Transformer oil analysis