Use of Power Factor Insulation Tests

Why Power Factor Tests?

Scope of Field Testing Technique : Maintenance Testing of Electrical Apparatus.

The power-factor insulation test is being used for the routine maintenance testing of the following types of electrical apparatus:

    • Bushings (breakers and transformers)
    • Bus supports (one-piece and multi-piece, in place)
    • Cables (in lengths up to approximately 500 feet-longer lengths require special test equipment)
    • Cable Joints (requires removing 3″ section of lead sheath at both ends of joint)
    • Capacitors (carrier and power-factor correction)
    • Circuit Breakers, Air (insulators, barriers, operating rods)
    • Circuit Breakers, Oil (bushings, lift and guide rods, tank insulation, oil)
    • Hot-Line Tools and Switch Sticks
    • Insulators (suspension, post, pin and stack types)
    • Lightning Arresters (gap and valve units)
    • Liquid Insulation-Oil and Askarels (for presence of moisture and contaminants)
    • Potheads (can be tested without disturbing the cable sheath)
    • Rotating Machinery (large units require special equipment)
    • Transformers, Instrument (bushings, windings, barriers and oil)
    • Transformers, Distribution (bushings, windings, barriers and oil)
    • Transformers, Power (bushings, windings, barriers and oil)
    • Voltage Regulators (bushings, windings and oil)

Advantages of Power Factor Tests:

The power-factor insulation test is effective in locating operating hazards in electrical power apparatus, before failure, because it has the following advantages:

    1. It can detect the presence of “bad” insulation even though there may be a layer of “good” insulation in series with the “bad” insulation; do tests usually show infinity or high insulation resistance under this condition. Many forms of insulation deterioration involve layers of “good” and “bad” insulation in series, for example, sections of the insulation of a compound embedded bushing which have been unevenly exposed to moisture entering through the gaskets or a crack in the weather shed.
    2. It provides a measure of the overall operating condition of the insulation in terms of ratio (the power factor of the insulation) which is independent of the amount of insulation being tested , example, modern, oil-immersed, paper-insulated power transformers. in good condition will have a power factor less than 1.0% at 20°C regardless of make size and voltage the power factor of an insulation increases when the insulation becomes deteriorated, experience . shown the abnormal power-factor values which indicate failure hazards.
    3. It provides a measure of the overall condition of ac apparatus insulation under simulated- normalfrequency operating conditions which is independent of the duration of the test, do tests require more time, do not simulate normal insulation operating conditions, except for do apparatus, and the results are affected by the duration of the tests.
    4. It provides data for a comprehensive analysis of the nature of insulation deterioration and operating hazards m terms of dielectric watts loss and charging current, at selected test voltages, from which power factor, capacitance, ac resistance, and the presence of ionization (corona) can be determined. All of these measurable characteristics of insulation are used in the routine analysis of field test data.
    5. It provides data which permit an orderly grading of the serviceability of apparatus insulation by comparison with the correlated results of many thousands of tests and investigations made on similar apparatus insulation, since 1929.
    6. It provides a more comprehensive picture of the overall operating condition of an insulation than any other single type of test.

Status of Power Factor Tests:

Power-factor insulation tests were first used in this country about 1917 by cable manufacturers for research and control of manufacturing operations. Power-factor tests were first applied to the field testing of apparatus insulation about 1929 by the Doble Company.

All large manufacturers of bushings make routine power-factor tests on new bushings at their factories; bushing manufacturers’ literature recommends periodic field power-factor tests and contains typical powerfactor data.

All ASA Standard bushings rated 25 kV and above have a power-factor test tap which permits testing bushings without disconnecting windings, etc, from the bushings.

Most of the large manufacturers of transformers make routine factory power-factor insulation tests on new power transformers and on oilimmersed instrument transformers; all of the large transformer manufacturers are prepared to make powerfactor insulation tests at their factories and are familiar with the advantages of these tests.

Thousands of tests have been performed in the field on rotating-machinery and cable insulation. Special equipment is available for tests on large rotating machines and on long cable lengths. Some of the large manufacturers of lightning arresters recommend the routine field use of ac dielectric-loss tests to detect presence of moisture and corrosion in arresters, and mechanical defects such as broken gap shunting resistors.

Field power-factor tests on oil and askarel are made to detect the presence of moisture and other contaminants, and to determine whether the condition of the oil is responsible for the high power factor of bushings, circuit breakers, transformers, etc.

Test Procedure

Power-factor insulation tests on electrical power apparatus are made with the apparatus out of service. In many cases it is necessary merely to open the disconnect switches to prepare the apparatus for the initial tests.

A few lets, made by applying an ac voltage to the conductors of the apparatus in its assembled position and measuring the dielectric losses to ground, will indicate whether or not the insulation is in a normal condition.

If the initial tests indicate an abnormal condition, the test engineer has a number of special tests which may be applied to localize the deterioration; for example, HotGuard tests, Hot-Collar tests, Ungrounded-Specimen tests (UST), cross-check tests, oil tests, dielectric-loss tests on wood members, etc.

The Hot-Guard tests is used for an overall insulation test on a transformer bushing of the draw-lead type, without lowering the oil to disconnect the bushing from the transformer winding. Collar tests are often used for testing transformer bushings having a solid conductor, without disconnecting the transformer wording.

The Hot-Collar test is unique in that it makes possible the detection of moisture and other forms of deterioration in a section of a bushing or pothead without disturbing the main connnections This is done by applying a collar-shaped electrode over the section of the bushing to be investigated and applying an ac test potential between the electrode and the bushing conductor; the dielectric losses in the bushing insulation under the collar can be measured then in the usual manner.

The UST measurement is used on bushings equipped with capacitance or power-factor taps to test the main insulation of a bushing exclusive of connected windings, interrupters, etc. It is used also for direct measurements on interwinding insulation of transformers and rotating machinery.

One of the important points which should be kept in mind in the application of an overall insulation test is that the ability to detect localized faults diminishes as the amount of insulation under test is increased. It is therefore desirable to isolate the apparatus under test, as far as feasible, from other connected apparatus.

However, as previously stated, this may not be necessary for the intitial tests in those cases where but a single disconnect and a short section of copper are connected to a bushing under test Additional examples of cases where tests can be made on apparatus without removing connections are bus and disconnect insulators, potheads, bushings, etc., which can be tested in place by the Hot-Collar, Hot-Guard or UST method.

The test voltage normally used for the field testing of bushings and insulators with the Doble Type MH and M2H test sets is 10 kV at 60 cycles. In the case of transformer windings, the test voltage may range from 2.0 kV to 10 kV, depending upon the rating of the transformer. The Doble Type MEU and M2E 2500-volt test sets are normally used at 2500 volts.

Interpretation of Test Data

The interpretation of insulation test data involves the use of elementary principles in the dielectric behavior of insulating materials and the use of correlated test data acquired by experience since about 1929.

Tables for grading the serviceability of insulation based on the correlation of field test data and investigations will be found in Doble Instruction Manuals.

In many cases failure hazards of apparatus insulation are expressed in terms of maximum allowable power factor values; however, changes in the “normal” watts loss, ac resistance and capacitance are also used for indicating hazards in apparatus insulation, depending upon the type of insulation being investigated. Several bushing manufacturers have published factory and operating limits in terms of power factor.

The “normal” test values for various types of apparatus insulation have been obtained by testing thousands of similar pieces of apparatus in the field and in the factory.

The “remove” test vanes have been obtained from a correlation of known test values at which apparatus insulation has failed in service.

These data have been supplemented by many investigations, in the field and in the laboratory, of apparatus having abnormal test values.

Frequently, the abnormal test values obtained during the initial tests are due to accumulations of carbon deposit or bad oil which may be cleaned up without permanent injury to the main insulations.

These conditions may represent serious operating hazards which require immediate attention. The ability of the test to detect this type of operating hazard before a visual inspection is made of the interior of the apparatus, facilitates a more efficient use of maintenance manpower and savings in maintenance costs.

Apparatus with a deteriorated insulation may sometimes operate for a long time without a failure, depending upon its exposure to over-voltages and short-circuits; however, deteriorated insulation creates a definite operating hazard which must be recognized and corrected if needless interruptions to service are to be eliminated.

Deteriorated insulation, if removed before failure, may often be reconditioned and restored to service, with substantial savings in material costs.

Dielectric Loss

All commercial solid and liquid insulations have some measurable dielectric loss at normal operating voltage and frequency, since the term “non-conductor” is but a relative one.

However, these losses are usually very small and vary approximately as the square of the applied voltage.

An appreciable increase in the normal dielectric loss is one of the first indications of deterioration and operating hazards.

Gaseous insulations such as air do not have a measurable loss until they become overstressed and ionized.

Solid insulation, unless thoroughly impregnated or immersed in liquid insulation, usually has voids containing air. If these voids become overstressed and ionized during the application of a test voltage, the resulting dielectric losses in the voids will be added to those in the solid insulation; hence, the losses will increase at same power higher than the square of the voltage.

The presence of voids that become ionized at normal line-to-neutral operating voltage may cause carbonization or radio interference; hence, apparatus-insulation designers endeavor to eliminate voids wherever feasible.

The dielectric losses of most insulations increase as the temperature of the insulation increases. In some cases, insulations have failed due to the cumulative effect of temperature; i.e., a rise in temperature causes a rise in dielectric loss which in turn causes a further rise in temperature.

The earlier cable insulations had a relatively high dielectric loss which increased rapidly with temperature.

However, modern paper-insulated cable has a very low dielectric loss which is not appreciably affected by rise in the cable temperature.

It should be noted that it is not necessary to have a direct metallic connection in order to measure ac dielectric losses m an insulation, but merely a capacitance coupling with the source of test voltage and measuring device.

Extraneous insulation that is in an electrostatic field set up by test conditions may cause losses which will be measured by the test equipment unless means are taken to shield out this extraneous insulation.

This phenomenon makes it possible to detect deterioration in insulations not directly coupled to the energized condutors: for example, wood angle braces in a breaker.


The elementary principles of an electrical condenser or capacitor are generally known; however, a few of the principles are reviewed here because the behavior of an insulation in service can be more easily understood when the insulation is considered as the dielectric of a capacitor.

The energized conductors may be considered to form one plate and the grounded apparatus frame the other place of the hypothetical capacitor.

Some of the characteristics of an electrostatic or dielectric field also are of interest in connection with testing of insulation. It will be recalled that the strength of a magnetic field is a function of current, whereas the strength of a dielectric field is a function of voltage.

The lines of magnetic flux around a conductor take the form of concentric circles, whereas the lines of dielectric flux around a conductor are radial. Bath electromagnetic and electrostatic fields are required in order to transmit power.

The dielectric constant (SIC or K) of an insulation is an indication of the ability of the insulation to pass dielectric flux through it, with air as a standard of comparison.

An insulation having a high dielectric constant will pass more dielectric flux than an insulation having a low dielectric constant, under the same electrical conditions.

The dielectric constant of some of the more common insulations, together with their power factors, are listed later.

In order to maintain an alternating voltage and a dielectric field through an insulation or between the plates of a capacitor, it is necessary for a charging current to flow.

The charging current of an insulation is somewhat analogous to the magnetic excitation current of a transformer. The amount of charging current taken by a given capacitor of insulation is a function of the impressed voltage and frequency, and the capacitance of the capacitor of assembled insulation.

The capacitance of a capacitor depends upon the area of its plates, the spacing between the plates and the dielectric constant of the dielectric medium (insulation) between the plates.

The voltage drop through a series of capacitors varies inversely as the capacitance of the individual capacitors.

Since capacitance is a function of the dielectric constant of the dielectric medium, designers of electrical apparatus select, where possible, insulation with dielectric constants which give the desired voltage distribution through the sections of insulation, or change the dimensions of the plates of the capacitors for the same purpose as in the case of the well known condenser bushing.

The dielectric constants of most commercial insulations range from 2.0 to 7.0. However, the dielectric constant of water is 81; hence, when an insulation becomes wet, it is apparent that the capacitance is increased as well as its dielectric loss.

Changes in the normal capacitance of an insulation indicate to the test engineer such abnormal conditions as the presence of moisture, short-circuited condenser sections in a bushing, or breaks in the ground shields of bushings.

The capacitance of a dry insulation is not appreciably affected by an increase in temperature; however, in the case of wet insulation, there is a tendency for the apparent capacitance to increase with temperature.

The charging current of an air capacitor, which normally has no losses, is a pure capacitive current leading the impressed voltage by ninety degrees.

The charging current (Iz) of commercial insulation, which has some loss, is the resultant of the capacitive component (Ic) and energy component (Ir).

However, for good insulation, Ic and Iz are of substantially the same magnitude since the energy component is very small.

The approximate capacitance of a condenser may be computed from its charging current at a given voltage and frequency.

Power Factor

The power factor of an insulation is the cosine of the angle between the charging current vector and the impressed voltage vector. In other words, it is a measure of the energy component of the charging current.

The amount of charging volt-amperes and the dielectric loss in watts, at a given voltage, increase with the amount of insulation being tested. However, the ratio (power factor) between the charging volt-amperes and watts-loss remains the same regardless of the amount of insulation tested, assuming that the insulation is of a uniform quality.

This basic relation eliminates the effect of the size of electrical apparatus in establishing “normal” insulation values and thus simplifies the problem for the test engineer.

The Doble power-factor insulation test equipment measures the charging current and watts-loss from which the power factor, capacitance and ac resistance can be easily computed at a given test voltage.

Dielectric Absorption

The exact cause of the phenomenon of dielectric absorption is not fully understood by scientists; however, the effect of this phenomenon is well known.

All electrical men know that when a do insulation resistance test is first applied to a transformer or generator, the initial resistance is low and gradually increases as the duration of the test is lengthened.

It takes energy to establish an electrostatic field in an insulation; however, once the field has been fully established, the charging current drops down to a value which is a function of the continuous leakage through the insulation.

The energy required to “charge” an insulation is usually referred to as the dielectric-absorption loss. The losses from dielectric absorption may be on the order of one hundred times the losses from continuous leakage through an insulation.

When an alternating current is applied to an insulation, the effect of the phenomenon of dielectric absorption greatly predominates over that of leakage or conductivity, because the dielectric field never becomes fully established with one polarity before the alternating charging current is reversed and starts to build up with the opposite polarity.

For all practical purposes this makes constant frequency alternating-current measurements of dielectric-abosrption loss independent of the duration of the test potential, provided the insulation has not reached an unstable position with respect to temperature effects.

Dielectric-absorption losses are very sensitive to small changes in the moisture content of an insulation and to the presence of other impurities; i.e., a small amount of moisture will cause a large increase in dielectric absorption.

The fact that ac dielectric losses are due almost entirely to the phenomenon of dielectric absorption makes the ac dielectric-loss and power-factor test a very sensitive test for moisture.


Since practically all insulators are conductors of electricity to some extent, it appears more logical to use the term conductivity that resistivity in discussing this component of insulation losses. The conductivity or leakage of an insulation may be considered as the reciprocal of its volumetric and surface resistivity.

The final resistance value obtained from a dc insulation resistance test is a measure of the continuous conductivity of the insulation.

In cases where an insulation consists of several layers or laminations, it is possible to obtain an erroneous indication of the breakdown strength of an insulation with do insulation-resistance tests; for example, if there is a layer of high-resistance insulation in series with one having a low resistance, the overall do insulation resistance will be largely determined by the high-resistance layer.

     In effect, a layer of good insulation interrupts the do test circuit except during the initial transient charging stage. The tact that a high-resistance layer of insulation is frequently found in series with a low-resistance layer of insulation in the service deterioration of electrical-apparatus insulation is one of the chief reasons for the development of the ac dielectric-loss and power-factor test.

The ac test is not hindered by a “good” layer of insulation in series with a layer of”bad” insulation, because ac merely requires a capacitive coupling to pass charging current and permit the measurement of the total dielectric loss of an insulation. The “Hot-Collar” measurement is a good illustration of a capacitance coupling.

By applying ac potential to a metallic band wrapped around the outside of a porcelain bushing, with the bushing conductor grounded, losses resulting from damp or carbonized insulation contained within the porcelain can be measured. This cannot be done with commercial do insulation test equipment.

     The dc test does, however, have an advantage over an ac test in those cases where it is desired to separate the conductivity losses from the dielectric-absorption losses; for example, to detect more easily a localized fault where there is a continuous conducting path to ground.

However, a large majority of the cases of deteriorated insulation, as found in the maintenance of electrical apparatus, involves discontinuous paths to ground and requires an ac test to locate.

     In general, the dc test is largely a measurement of surface resistivity unless a complete dielectricabsorption test (resistance vs. time) is made, whereas the ac test provides a measurement of surface and volumetric resistivity, dielectric absorption, and the presence of ionization.

AC Resistance

Dielectric loss may be expressed in terms of ac resistance by dividing the square of the test voltage by the watts loss.

The insulation quality of sections of wood members, such as lift rods, may be judged more easily and more comprehensively in terms of ac resistance per unit length than in terms of power factor, because of the variation in capacitance during tests.

Any section of a wood member may be easily tested in place by the application of three electrodes. The two outer electrodes are usually grounded and the middle electrode is energized. DC tests on wood members are almost entirely confined to a measurement of surface leakage.

Transformer Excitation Current

Excitation-current tests may be used to locate certain types of faults in transformers, such as defects in the magnetic core structures or insulation failures which have resulted in conducting paths between winding turns.

Either type of fault increases the apparent reluctance of the magnetic circuit and may be recognized by the abnormally high excitation current required to force a given flux through the core.

Certain other problems, such as incorrectly connected windings and defective tap-changing devices, may also be detected by excitation-current tests.

Excitation-current tests m the field have been developed as an adjunct to Doble insulation power-factor tests, conveniently utilizing the same test equipment and transformer outage for both tests.

The limited voltage and current capability of a typical insulation test set necessitates the measurement of excitation currents at a fraction of operating voltage, generally at 2.5 or 10 kV.

Valid comparison of the excitation currents of similar transformers may be made, however, by taking the measurements on corresponding windings and at the same test voltage.

The basic procedure is to measure excitation currents in individual phases of the high-voltage winding in a transformer bank energized or excited by the Doble test set.

For routine tests on single-phase and wyeconnected three-phase transformers, each phase is energized separately and measured; for deltaconnected three-phase transformers, the windings are energized in pairs and the exciting current in one winding of the pair is measured.

Since the excitation current may not vary linearly with voltage, particularly at relatively low voltages, the same test voltage should be used for each phase to permit proper comparisons and interpretations of test data.

Leave a Reply

Your email address will not be published. Required fields are marked *