Power Factor Testing: Some Unknown Facts

Dinesh Chhajer
Megger
Dallas, Tx
dinesh.chhajer@megger.com

Abstract: Power Factor (PF) testing has been a very critical to make correct decisions when it comes to
common insulation diagnostic test and is frequently estimating an assets remaining healthy life.
used to evaluate the insulation condition of electrical Electrical insulation has been a very common reason
equipment. This paper focuses on some unknown and
for most of the electrical equipment failures. A lot of
not so popular facts that get overlooked easily. PF
time and effort have been put forward to diagnose
testing has been standardized at 10 kV and readings are
trended over a period of time at that voltage level or
and better estimate the condition of insulation as it
with values equivalent to 10 kV. Is PF a function of ages. Deterioration of insulation is dependent upon
voltage? Why is only 10 kV the standard value? various factors like ageing, mechanical stress,
Excitation current measurements taken at lower voltage thermal and chemical stress, over loading and
levels are normalized to 10 kV using a linear V/I varying environmental conditions. Dielectric strength
relationship, whereas excitation or magnetization curve of insulation can be assessed by various off-line and
exhibits a non-linear relation. PF testing of low- online testing techniques.
capacitance specimens, at times, will reflect negative PF
values. How can PF readings be negative? PF values of Power Factor (PF) test is a traditional and reliable
all insulation materials vary with temperature. Use of way of estimating the condition of insulation of
average temperature correction factor tables for
electrical equipment such as power transformers,
specimens subjected to different loading, ageing
circuit breakers, generators, cables and other
acceleration and stresses only gives an approximate
value and not accurate temperature compensation for electrical assets. PF values, trended periodically, can
PF values. help in detecting problems like contamination, high
moisture content and the presence of voids in
This paper is intended to explain these unusual PF specimen insulation. Excitation current tests, along
testing facts which include PF as a function of voltage, with PF tests performed on power transformers, can
negative PF values, non-linearity in excitation current
help in detecting turn to turn insulation failure.
measurements, use of average temperature correction
factor values and some other PF facts that are often Diagnostic tests, along with best testing practices as
overlooked.
per internationally accepted standards, provide
Keywords- Power factor (PF), excitation current, valuable information that can prevent unplanned and
temperature correction, dielectric frequency response unexpected failure of the insulation system.
(DFR) Maintaining good quality of insulation helps in
preventing failures and outages that jeopardize the
I. INTRODUCTION reliability of the overall electrical system operation.
Extending the life of substation assets result in less
With an increasing failure rate of substation electrical outages, higher reliability, system stability and a
equipment, utilities and large industry are faced with substantial financial benefit.
a challenge to focus on preventive and predictive
maintenance to ensure the power systems integrity II. Power Factor vs. Voltage
and stability. Precise and accurate measurements are
PF tests are usually performed at 10 kV or the Test instruments with very high noise suppression
readings are obtained as 10 kV equivalent. The capability are required for measurements in high
appropriate voltage level at which to perform the voltage substations where electrical noise and
power factor tests is a frequently debated topic. In interference level can be as high as 15 mA. Since,
todays world, the available instrumentation and signal current is very low when performing AC
measurement techniques provide flexibility in insulation test, a SNR of 1:20 can be expected in
performing the same PF test with a variable test worst case scenarios.
voltage. With the test instruments available, PF can
be performed at any voltage ranging from 27 V to 12 With the new developments in testing technologies,
kV. What test voltage is good enough for accurate we can expect to see testing being performed at lower
and reliable measurements? The answer resides in voltages with better filtering and noise rejection
what kind of specimen is being tested and under what capabilities. Additionally, we can expect the size of
conditions it is being tested. the portable test sets to reduce further in coming
years.
Most of the power transformers have oil-paper type
insulating systems. This type of insulation exhibits a III. Negative Power Factor Values
flat response (no change) when PF is measured as a
function of test voltage. On the other hand motors For an ideal insulation, the value of PF should be
and generators typically have dry or solid type zero. For all practical purposes, any specimen
insulation that exhibits PF values as a function of test exhibiting a value close to zero is considered to have
voltages. Its value increases with the increase in test a good quality of insulation. PF test sets always try to
voltage due to the presence of voids. Almost all solid measure one single capacitor. If it is not just a
insulation materials have voids present in it. The capacitor, the results look strange. This should give
number of voids present will vary based upon the an indication to the user about the characteristics of
geometry, age, construction and design of the the specimen. When performing tests on bushings,
insulation system. PF is a measure of losses taking three winding transformers or inter-phase insulation
place in the insulation system. The amount of of rotating machinery, the PF values will sometimes
increase in PF value as a function of voltage turn out to be negative. What does that mean? PF is a
corresponds to the degree of ionization taking place measure of the amount of watts loss taking place in
in voids. the insulation. Negative PF corresponds to watts
generation as opposed to watts loss. Obviously,
One of the reason industry has standardized on 10 kV insulation cannot generate it, which proves that
as a typical test voltage is for immunization against negative PF values are not real.
electrostatic interference. Power transformers
operating in high voltage substations are subjected to
a lot of electrical noise and interference around the Specimen
nominal frequency. A HV test signal provides better
signal to noise ratio (SNR) for precise and accurate
measurements. Table 1 below shows a SNR ratio of
1:2 in 765 kV substation on a UST (Ungrounded
Specimen Test) measurement.

Table 1: Four repeat UST measurements on a

grading capacitor in a 765 k V Substation

Figure 1: Specimen in UST mode with surface loss

components
Negative phantom values appear for some specimens Specimens with low capacitance values (smaller
that have higher surface leakage loss. As shown in current) are more susceptible to this surface loss
Fig 1, phantom circuits introduce a current Is which current and can lead to negative PF values as shown
would cause change in the phase angle of measured in Fig 3.
test current (IT). The surface loss current (Is) is
dominated by mainly a resistive component (Rs) and It is important to understand where negative PF
has a very small phase angle with respect to the values come from. For some specimens it is just by
applied voltage. Capacitive coupling (Cc) may be the virtue of apparatus design. For example, the
present as a result of this parallel path of Rs to main presence of electrostatic grounded shield between the
insulation under test. Smaller phase angles for inter-windings of a transformer. Another example of
surface loss current (Is) can lead to negative PF the potential for negative PF values would be the
values. Measured test current (Is) is vector difference inter-phase insulation of motors because of the way
of total current (INET) and surface loss current (Is) as end windings are terminated. In other cases, where
shown in equation 1 below. negative values are encountered users should
consider eliminating all external effects by following
best testing practices such as verifying proper
IT = INET - Is (1) grounding circuits, cleaning external bushing
In UST or GST (Grounded Specimen Test) mode surfaces , avoiding unfavorable weather conditions
configurations, this surface loss makes the measured (especially high humidity) and effective use of guard
test current (IT) phase angle greater than 90 degrees circuits. Repeated negative values even after applying
as shown in Fig 2 and this results in negative PF these best testing practices could point towards some
values. contamination or a bad insulation system.

IV. Excitation Current vs. Voltage

Excitation current test is a very commonly performed
test along with traditional PF testing. It is a voltage
dependent test and is always performed in UST
mode. Like PF tests, all the excitation current
readings are normalized to 10 kV equivalent values
as well. A linear approximation is used to determine
the excitation current value at 10 kV when these tests
are performed at some voltage lower than 10 kV.

Figure 2: Vector diagram with different Is phase 20 mA @ 5kV is 40 mA@ 10 kV equivalent.

angles

Figure 3: Vector diagram with different magnitude of Figure 4: Typical magnetization curve of a power
current INET and purely resistive loss current Is transformer
When dealing with specimens that are highly When performing excitation current measurements
inductive, such as power transformers, voltage and on delta windings, it is very important to ground the
current do not exhibit a linear correlation as shown in third leg of the delta configuration in order to obtain
Fig 4. correct readings as shown in Fig 5.

The relationship between V & I is dependent on core Since the excitation current is a UST test, grounding
characteristics and the material used. As applied the third leg eliminates the current flowing in the
voltage is increased, inductance of the transformer other two windings from the measurement circuit.
will change and therefore the excitation current will Pending the inductance and resistance of each
change too. Since, change of inductance is not linear winding, if third leg is not grounded, the results
with respect to voltage; change in current is also non- would be approximately 30 to 50% higher than true
linear. Assuming a linear relation to determine 10 kV readings as shown in Table 3.
equivalent excitation values would only give the
approximate values. It is very important to perform Table 3: Excitation current test performed on delta
tests at the same voltage if excitation current winding with and without third leg grounded
historical data need to be trended and compared with
present values. Test performed at 2 kV and then
corrected to 10 kV may not be comparable to test
performed at 10 kV as shown in Table 2. The ability
to trend and compare data is critical when trying to
find out problems associated with turn to turn A transformer with magnetized core can cause
insulation. erroneous readings in excitation current
measurements. The core may have residual
Table 2: Excitation current test performed at magnetism as a result of heavy fault seen by a
different voltages on a single phase Transformer transformer or because of application of DC
current/voltage to it. Presence of residual magnetism
would result in higher excitation current values than
normal. IEEE 62-1995 section 6.1.3.4 states that If a
significant change in the test results is observed, the
only known reliable method of excluding the effect of
residual magnetism is to demagnetize the transformer
core. It is recommended that the dc measurements of
the winding resistance be performed after the
exciting current tests.

Excitation current measurements can be affected by

various different factors as discussed above and
should be kept in mind before performing the test.

V. Temperature Correction
Factors for PF readings
PF values are highly dependent on temperature. IEEE
C57.12.90 section 10.10.4 Note 3 (b) states that
Experience has shown that the variation in power
factor with temperature is substantial and erratic so
Figure 5: Excitation current measurement on a Delta that no single correction curve will fit all
winding with third leg grounded cases.Correction factor tables have traditionally
been used to bring all the recorded data to a common in the second half of the life cycle. In this second
base of 20 C. half, correction factors would be larger because of the
increased effect of temperature on condition of the
It is imperative to only compare a specimens PF insulation. Using average temperature correction
values that are either taken at a similar temperature or values can lead to incorrect trending of results and
corrected to the same temperature accurately. inaccurate estimation of the remaining healthy life of
Different specimens behave differently to the effects the object.
of change in temperature on PF values. Additionally,
the same specimen will become more temperature IEEE Std. 62-1995 states that Testing at
dependant as it ages, Ref [3]. temperatures below freezing should be avoided, since
this could significantly affect the measurement.
Each specimen is unique in its construction, design Among the primary reasons for performing this test is
and ratings. They are each subjected to different the capability of detecting moisture in insulation. The
kinds of stresses, loading and different environmental electrical characteristics of ice and water are quite
conditions. A transformer, with 20 years of service different and it is much more difficult to detect the
operating at 100% of its capacity in weather presence of ice than it is to detect water; sometimes it
conditions of 45 C would have a totally different is impossible. Measuring PF at too high or too low
level of stress on insulation than another transformer of temperature conditions can introduce error in PF
with 10 years of service running at 75% of its measurements. IEEE recommends performing the PF
capacity in an environmental condition of 30 C. tests at or near 20 C and repeated periodically for
Figure 6 below shows four transformers with varying correct trending.
degree of temperature dependence based upon their
age, rating and moisture level, Ref [3]. However, in reality its not always practical to cool
down or heat up the transformer or any other test
specimen to 20 C for the purpose of performing
these measurements.

With the advancement of newer technologies, it is

possible to accurately correct the PF values to 20 C
without the use of average correction factor tables.
Using dielectric frequency response (DFR), one can
determine the unique temperature correction factor
for each individual test object. The method is based
Figure 6: Temperature dependence of different on the fact that a power factor measurement at a
transformers with different levels of ageing certain temperature and frequency corresponds to a
measurement made at different temperature and
Temperature correction factor values are highly frequency. By measuring power factor at different
dependent on insulating material, their structure, frequencies and at any given insulation temp, one can
ageing, presence of moisture or contamination and determine PF at any temperature [5-50 C] and at
various other factors. The available temperature
nominal frequency. With technique like this,
correction factor data is based upon the average
individual temperature correction factors can be
value. Since, each test object is unique; using the
determined and values can be accurately corrected to
average correction would introduce some error into
20 C. The big advantage of this technique is that PF
the corrected value. New transformers have relatively
can now be measured at any insulation temp. [5-50
weak temperature dependence and use of standard
C] and still can be corrected to 20 C accurately and
tables would over compensate as shown in Fig 6
precisely. Reference [3] talks in detail on the
above. The measure of error becomes more
technique and method used to determine the
predominant (in other direction) as the object ages.
individual temperature correction factors.
Its a known fact that insulation deteriorates with
aging. Trending of PF values becomes more critical
VI. Conclusion [9] Alex Rojas, Power factor testing in transformer
condition assessment- Is there a better way?, 2006
Electric apparatus have failed and will continue to IEEE PES Transmission and Distribution Conference
fail because of insulation deterioration. A proactive and Exposition Latin America, Venezuela
and smart approach is cardinal to diagnosing and
checking the integrity of the insulation system. Power
factor diagnostic test is an important tool in
determining the quality of the insulation and
estimating its remaining healthy life.

Power factors readings are dependent on various

different factors and it is very important to
understand the effect of these factors. Test voltage,
electrostatic interference, temperature, humidity,
surface losses and various other parameters can
greatly influence the PF readings. A better
understanding of the impact of these varying
parameters would help in obtaining accurate
measurements. Those correct PF measurements
would then certainly help in taking a definitive and
rightful approach in the decision making process.

VII. References
[1] IEEE Standard Test Code for Liquid-Immersed
Distribution, Power, and Regulating Transformers
and Guide for Short-Circuit Testing, IEEE
C57.12.90-2006
[2] IEEE Guide for Diagnostic Field Testing of
Electric Power Apparatus; Part 1: Oil Filled Power
Transformers, Regulators, and Reactors, IEEE 62-
1995
[3] P. Werelius, M Ohlen, Dielectric Frequency
Response and Temperature Dependence of Power
Factor, Electrical Insulation (ISEI), Conference
Record of the 2010 IEEE International Symposium
[4] D Chhajer, Power Factor Tip up Test
Techniques & Concerns, paper presented at NETA
2010
[5] Doble manual Power Factor Testing of electric
power apparatus insulation with Doble type M
instruments, Doble engineering company, 2002
[6] Megger, Instruction manual for Delta 2000- 10
kV automated insulation test set Rev. B 2001
[7] P. Werelius, thesis Development and
Application of High Voltage Dielectric Spectroscopy
for Diagnosis of Medium Voltage XLPE Cables
[8] Megger, Application guide for power factor
testing of power & distribution transformers