Partial Discharge Theory and Applications to Electrical Systems

Gabe Paoletti, P.E. Alex Golubev, PhD

Division Application Engineer Manager, R&D, Predictive Diagnostics
Cutler-Hammer Engineering Services Cutler-Hammer Engineering Services
Pennsauken, NJ Minnetonka, MN

Abstract - Partial discharge monitoring is an effective Table 1 – Motor Electrical Failure Causes
on-line predictive maintenance test for motors and IEEE Study EPRI Study
generators at 4160 volt and above, as well as other Failure % Failed %
electrical distribution equipment. The benefits of on- Contributor Component
line testing allow for equipment analysis and Persistent 4.2 Stator 23.0
diagnostics during normal production. Corrective Overloading Ground
actions can be planned and implemented, resulting in Insulation
reduced unscheduled downtime. An understanding of Normal 26.4 Turn 4.0
the theory related to partial discharge, and the Deterioration Insulation
relationship to early detection of insulation Bracing 3.0
deterioration is required to properly evaluate this Core 1.0
predictive maintenance tool. This paper will present a Cage 5.0
theory to promote the understanding of partial Total 30.6 Total 36.0
discharge technology, as well as various
implementation and measurement techniques that The IEEE publication under development, “IEEE P1434
have evolved in the industry. Data interpretation and - Guide to Measurement of Partial Discharges in
corrective actions will be reviewed, in conjunction Rotating Machinery” [4] also identifies similar failure
with comprehensive predictive maintenance practices causes for motor insulation systems. These include
that employ partial discharge testing and analysis. thermal, electrical, environmental and mechanical
stresses. These factors correlate to the two studies, since
I. BACKGROUND they result in the stator ground insulation and turn
insulation failure (EPRI Study); as well as can be
Reliable manufacturing operations will always be interpreted as normal deterioration (IEEE Study).
concerned with process production motors.
Comprehensive programs to maintain electrical The next section provides a review of partial discharge
equipment for peak performance have been theory. It is interesting to note that over 25 years ago,
recommended and implemented at various plants [1]. large motor manufacturers recognized the need for partial
Detailed motor failure analysis has been completed; discharge testing in the slot area between the winding
resulting in the identification of approximately 30% of insulation and the iron [5]. The testing was called the
failure causes being related to electrical failures [2]. A “Slot Discharge Test” and involved applying a test
summary of the IEEE transaction entitled: “Report of voltage while observing the waveform on an
Large Motor Reliability Survey of Industrial and oscilloscope. At that time only minimal partial discharge
Commercial Installations, Part I [3] included both the measurement technology was available, therefore
results of an IEEE survey and an EPRI survey. The two limiting the wide spread use of such testing.
sources of information proved extremely useful since the
IEEE survey identified the “Failure Contributor”, and the II. PARTIAL DISCHAGE THEORY
EPRI survey identified the “Percentage Failure by
Component.” The IEEE survey includes an objective Partial discharge theory involves an analysis of
opinion, whereas the EPRI survey includes actual failed materials, electric fields, arcing characteristics, pulse
components. The summary of the electrically related wave propagation and attenuation, sensor spatial
causes of the two studies is shown in Table 1, and will be sensitivity, frequency response and calibration, noise and
referred to, when discussing root cause failures related to data interpretation. It is obvious from the above that most
partial discharge test results. plant engineers will not have the time, or available
energy, to pursue such a course of study.
Presented at the 1999 IEEE IAS Pulp and Paper Industry Conference in Seattle, WA: © IEEE 1999 - Personal use of this material is permitted.
In an effort to promote a better understanding of partial Copper Conductor
discharge (PD), this paper attempts to provide simplified
Contaminated
models and relate the characteristics of these models to
the interpretation of PD test results. Insulation cracks
Surface
First, we will present a few technical concepts relating to discharge
partial discharges. Partial Discharge can be described as through air Contaminated
an electrical pulse or discharge in a gas-filled void or on
a dielectric surface of a solid or liquid insulation system. insulation
This pulse or discharge only partially bridges the gap Iron Core surface
between phase insulation to ground, or phase to phase
insulation. Figure 2 – Surface Partial Discharges
These discharges might occur in any void between the
copper conductor and the grounded motor frame A. Insulation System Model
reference. The voids may be located between the copper
conductor and insulation wall, or internal to the A simplified model of an insulation system can be
insulation itself, between the outer insulation wall and represented by a capacitance and resistance in parallel
the grounded frame, or along the surface of the [6]. This is the concept employed in the use of power
insulation. The pulses occur at high frequencies; factor testing of insulation systems. The leakage current
therefore they attenuate quickly as they pass to ground. is split between the resistive and capacitive paths. The
The discharges are effectively small arcs occurring power factor is the cosine of the phase angle between the
within the insulation system, therefore deteriorating the total leakage current and the resistive component of
insulation, and can result in eventual complete insulation leakage current [5].
failure.

The possible locations of voids within the insulation

system are illustrated in Figure 1.
Simplified
Insulation
Copper Conductor Model
Void between
Insulation copper and
insulation
Void internal Figure 3 – Simplified Insulation Model and
Model for an Electronic Attenuator
to insulation
Void between
The above model is also used for attenuator circuits in
Iron Core insulation and electronics [7]. Signal attenuation results in reducing the
Iron Core amplitude of the electrical signal. This underlies the
Figure 1 – PD within Insulation System
problem with partial discharge detection. The insulation
medium, which is being exposed to the partial
discharges, acts to attenuate the signal, therefore
The other area of partial discharge, which can eventually weakening this damaging signal which we are trying to
result, is insulation tracking. This usually occurs on the identify at our sensor locations. In addition, the
insulation surface. These discharges can bridge the attenuated partial discharge signal can be masked by
potential gradient between the applied voltage and sources of electrical noise, which shall be reviewed later
ground by cracks or contaminated paths on the insulation in this paper.
surface. This is illustrated in Figure 2.
The above concept of the insulation system being an
The above can be illustrated by development of a effective attenuator circuit gives rise to critical detection
simplified model of the partial discharges occurring issues, such as:
within the insulation system. • Sensor locations and sensitivity
• Measurement system response to attenuated signals
• Noise detection and elimination
B. Partial Discharge Void Model arc. Even though the arc is small, severe local heating
results. The intense heat of the leakage current arc is
Simplified models of the area of the void have been sufficient to cause a molecular and chemical breakdown
described as consisting of capacitors only [8]. A review of the underlying insulation. On organic materials, a
of the progressive failure mode of these voids indicates frequent by-product of arcing is carbon.” The above
an additional resistive component in parallel with the “tiny arc” along the insulation surface can be
capacitive component. An electrical equipment design represented by partial discharge activity. Figure 5
handbook [9] states: “Discharges once started usually illustrates the failure mode of deteriorated insulation
increase in magnitude with stressed time, but discharges related to the intensity of partial discharge
can become short circuited by semiconducting films measurements.
inside the void and discharging is terminated.” The
referenced semiconducting films can also consist of Failure
carbonization of the organic insulation material within
the void due to the arcing damage. Therefore the model
of the partial discharge void is similar to that of the
insulation medium itself.

IntensityPD PD drop before

final failure
Simplified
PD Void
Model Insulation Deterioration

Figure 5 – PD versus Insulation Failure Mode

Figure 6 illustrates a circuit breaker bushing which as

Figure 4 – Simplified Partial Discharge Void
progressive tracking highlighted for presentation
Model with Internal Resistive Leg purposes. At the point near eventual failure, the tracking
and resistive component of the insulation have increased
Actual failure modes have indicated a drop in partial to the point where partial discharges have been reduced,
discharge intensity shortly prior to complete failure. This since the “tiny arcs” have caused the carbonization and
would occur when the internal arcing had carbonized to the tracking, therefore providing a direct path for current
point where the resistive component of the model was low flow. At this point, evidence of insulation deterioration is
enough to prevent a build-up of voltage across the void. usually detected by traditional methods of insulation
This new low resistive component would also allow higher resistance, or megger testing. For the above reason,
current flows, and additional heating and resultant partial discharge on-line testing and traditional insulation
insulation damage. The above model, including the resistance testing are complimentary. On-line partial
resistive component correlates to the actual failure mode of discharge testing can detect insulation in the progressive
a partial discharge void, with the resistive component phases of deterioration, with trending identifying
passing more leakage current as the partial discharges problems long before eventual failure. Traditional
increase with time. One form of this resistive component is insulation resistance testing provides a “current-state” of
visible tracking on the surface of insulation. An the insulation system.
explanation of tracking, and how surface partial discharges
are related to the development of tracking follow [5]:
“Tracking damage has been traced entirely to the locally
intense heat caused by leakage currents. These currents
flow thru any contaminated moisture film on the bridging
insulating surface. As long as this film is fairly broad and
continuous, the heat associated with the leakage current is
spread over a wide area and is dissipated. However,
heating promotes film evaporation. This causes the film to
break up into small pools or islands. Each break in the film
tends to interrupt a segment of the leakage current, causing
a tiny
Figure 6 – Bushing with progressive tracking
 Copper
Conductor

Copper-
Insulation Void

Copper Conductor Void between

copper and
insulation
Void internal Internal Insulation

to insulation
Insulation
Void between Void
insulation and
Iron Core Iron Core

Insulation-Iron
Void

Iron Core
Figure 7 – Insulation System Partial Discharge Model
With the development of the above models, we can positive waveform cycle, a discharge, or a partial short-
illustrate a complete model of the various insulation circuit, results in a negative, downward oriented pulse.
system discharges represented in Figure 1. This is referred to as a partial discharge with a negative
polarity, and occurs during the first quarter-cycle of
Figure 7 is be used to provide an understanding of partial increasing positive voltage applied to the void. During
discharge activity. the third quarter-cycle, a partial short-circuit results in a
positive, upward oriented pulse. This is referred to as a
C. Partial Discharge Concepts partial discharge with a positive polarity and occurs
during the third quarter-cycle of the increasing negative
The first concept to review is the characteristic trait that voltage applied. These partial discharges, which are
partial discharges occur only during the first and third measured as a high frequency change in the power signal
quarter of each cycle. This is the initial rising positive in millivolts to a few volts, can not be observed with a
signal, and the initial rising negative signal. Effectively, standard scope; therefore they are exaggerated in Figure
during the initial rising positive signal, all of the 8 for illustration purposes.
capacitive components are being charged until the partial
Positive Charging &
discharge inception voltage is reached across each
specific void, and partial discharges commence. When Exaggerated Negative
the positive wave cycle begins to decrease the positive Polarity Pulses Measured
voltage across each void is reduced, since some
capacitive charge remains. Some level of charge must
exist since the voltage across a capacitor can not be
changed instantaneously. During the first quarter cycle
we are creating a positive charge and the resultant partial
Negative Charging
discharges. During the third quarter cycle, this positive
charge is effectively reversed, resulting in a positive & Exaggerated Positive
charge in the reverse direction, and the resultant partial Polarity Pulses
discharges. Measured

The second concept to review is that partial discharges Figure 8 – Exaggerated Positive & Negative Polarity
are measured as voltage pulses; therefore, during the Pulses for Illustration Purposes
As stated, since the pulse of voltage change is being This interesting phenomenon is related to the applied
measured, the negative polarity pulses occur during the voltage level to the void, the void’s geometric shape and
first quarter cycle, or during the rising positive cycle of the specific materials that are acting as the anode and
the wave; and conversely, the positive polarity pulses cathode. The critical material is the cathode, since the
occur during the third quarter cycle, or during the rising cathode supplies free electrons to allow the partial
negative cycle of the wave. discharges to continue. As illustrated in Figure 9 the
various anodes and cathode materials are shown for the
When viewing the results of partial discharge signals, the rising positive and negative parts of the AC cycle, which
above will be illustrated, on a three dimensional graph, is the two areas where discharges are measured.
with two critical measurements plotted in relation to the
360 degrees of a typical cycle. The 360 degrees is Depending on the part of the power cycle, the material
usually split into four segments, therefore the level of representing the cathode differs. The cathode material is
partial discharge in the first quarter cycle, or negative most important since the cathode will supply the
polarity discharges, can be compared to the third quarter electrons to support partial discharge activity. The
cycle, or positive polarity partial discharges. The characteristics of copper and iron are defined in their role
differentiation of positive versus negative polarity partial of a cathode, related to their conducting characteristics.
discharge pulses will be related to a probable root cause When the insulation becomes the cathode, and a partial
and corrective actions. The two measurements illustrated discharge occurs at the surface of the insulation, the
on the two dimensional graph are partial discharge characteristics of the insulation create a plasma. A
Maximum Magnitude, usually represented in millivolts, plasma is a very good source of free electrons to promote
and Pulse Repetition Rate, represented by the number of partial discharge, and in addition, the discharge area is
partial discharge pulses during one cycle of an AC extended by the nature of the plasma area. The result is
waveform. This is discussed further in Figure 12. that a greater tendency of partial discharges will occur
when the insulation is in the cathode role.
The partial discharge magnitude is related to the extent
of damaging discharges occurring, therefore related to For the negative polarity pulses, occurring in the first
the amount of damage being inflected into the insulation. quarter cycle, the insulation acts as a cathode across
The pulse repetition rate indicates the quantity of voids in the copper conductor-to-insulation space (Fig.
discharges occurring, at the various maximum magnitude 9– A). During these negative polarity pulses, a greater
levels. Both play a role in determining the condition of tendency of discharges will occur in this area near the
the insulation under test. Whereas seldom possible with copper conductor. Therefore if negative polarity pulses
on-line motors, the maximum magnitude level should be greatly exceed the positive polarity pulses, then the root
calibrated to reflect the actual charge, measured in pico- cause is considered to be voids in the copper conductor-
coulombs. The benefits of such calibration are offset by to-insulation area.
the relative comparison of similar motors, and more
importantly by trending of the partial discharge activity For the positive polarity pulses, occurring during the
over time. On-line partial discharge testing allows for third quarter cycle, the insulation acts as a cathode across
such trending and analysis of the electrical equipment. voids in the insulation-to-iron space (Fig. 9 – C). During
The illustration of the partial discharge activity relative these positive polarity pulses, a greater tendency of
to the 360 degrees of an AC cycle allows for identifying discharges will occur in this area near the iron. Therefore
the prominent root cause of partial discharges, therefore if positive polarity pulses greatly exceed the negative
appropriate corrective actions can be implemented. polarity pulses, then the root cause is considered to be
voids in the insulation-to -iron area, or in the area of
The third concept to review is the effect of high negative surface tracking since this also bridges the outer
polarity pulses, occurring during the first quarter cycle of insulation wall to the iron.
the positively rising wave, in relation to the high positive
polarity pulses, occurring during the third quarter cycle; Also note that when the voids are prevalent internal to
and vice versa. It as been found that if the positive the insulation material itself (Fig. 9 – B), then for both
polarity discharges exceed the negative polarity the positive polarity and negative polarity pulses, the
discharges then the probable root causes are either voids cathode remains the insulation itself. In this regard, when
between the insulation and iron core (slot discharges), or positive and negative polarity pulses are equally
at the winding end-turns, or surface partial discharges. It prevalent, then the root cause is considered voids within
as also been found that if the negative polarity discharges the insulation material itself, and not between the
exceed the positive polarity charges then the probable insulation and either the copper conductor, or the iron.
root cause is voids in between the copper conductor and
insulation.
 Positive Polarity
Pulses
Relationship between Positive &
Negative Pulses and Insulation
Negative Polarity acting as the Cathode
Pulses

Copper
Anode - COPPER - Cathode Conductor

A Copper-
Insulation Void

Cathode - INSULATION - Anode

Anode - INSULATION - Cathode

Internal Insulation
B
Insulation
Void
Cathode - INSULATION - Anode

Anode - INSULATION - Cathode Insulation –Iron

Void
C

Cathode - IRON - Anode Iron Core

Figure 9, 9-A, 9-B and 9-C – Relationship between Positive & Negative Pulses and Insulation acting as
the cathode at the regions of greater “electron” flow, or greater measured partial discharges
The above-simplified modeling attempts to provide an traditional testing methods: Insulation Resistance Testing
understanding of the measurement results of partial or “Megger Test” which is at a reduced DC voltage,
discharges, and their interpretation related to corrective Polarization Index Test (1 and 10 minute readings of the
actions. The following section shows the relationship to insulation resistance test to equalize the effects of
traditional testing methods and details the results, and humidity and temperature) and High-Potential Testing
associated corrective actions. (higher DC voltage test with leakage current monitored).
The fifth column includes the expected results from
Partial Discharge Testing.
III. PARTIAL DISCHARGE TESTING RELATED TO
TRADITIONAL TESTING METHODS For insulation considered “Good” or “Marginal” the
results are similar for all test methods. For insulation
Table 2 illustrates the relative relationships between the which is “Dry but insulation delaminated”, traditional
results of partial discharge testing and traditional testing test methods will provide a false sense of a “Fair”
methods. The insulation model, contained in the first condition; whereas partial discharge testing indicates the
column, illustrates the internal copper conductors, the presence of internal insulation voids. “Poor” or
outer insulation surface and various formations of voids “Unacceptable” insulation conditions can not be
within the insulation. The second column states the differentiated with traditional testing methods; whereas
insulation condition. The third, forth and fifth columns partial discharge testing identifies the regions of
indicated the expected results from the following insulation voids, and the appropriate corrective actions.
 Table 2 – Partial Discharge Testing related to Traditional Testing Methods
Insulation Insulation Megger Polarization High-Pot Test Partial Discharge
Model Condition Test Index Test Testing
Good High Good Linear leakage Unmeasurable

current partial discharge

vs. voltage is activity
minimal
Marginal Fair Fair Linear leakage Minimal discharge
current activity, balanced
vs. voltage is both positive and
stable negative discharges
Dry False False Fair False linear Partial discharges
but Fair value leakage current observed, therefore
insulation Result vs. voltage accurately showing
delaminated insulation problems
which are missed
by traditional tests
Poor High leakage High positive

- Cleaning current. May be polarity discharges

or required to indicate probable
Overhaul limited test surface tracking
Required Low Poor voltage.
Unacceptable High negative
- Major Potential failure polarity discharges
Repair or during testing indicates internal
Rewind voids near the cop-
Required High leakage per conductor.
Near-Failure Minimal partial discharge
condition Very Very low current and activity. Partial discharge
- PD arcing as low probable failure arcing as progressed to
caused carbon the point where
tracking during testing permanent damage
(tracking) as occurred.
Internal copper conductor

Insulation void experiencing internal partial discharge

Outer insulation surface Insulation Model

Internal copper conductor Descriptions

Surface tracking resulting from partial discharges

Outer insulation surface

For “Near-Failure” conditions, partial discharge arcing motor failure modes. One specific weakness that was
may have progressed to the point where permanent identified in medium voltage motors was the end-turn
damage, or tracking, as occurred, therefore the level of insulation, specifically at the junction where the motor
partial discharges as decreased. This is also illustrated in winding extended from the iron core. This region is
Figure 5. During this condition, traditional test methods shown in Figure 10. After several motors had failed, it
more accurately reflect the insulation condition, whereas was decided to review and ensure that the originally
a High-Potential traditional test may cause insulation installed differential motor protection was operating
failure during the test period. For this reason, trending is properly. It was found on several 13.8KV motors,
recommended for the first year of partial discharge equipped with operable differential motor protection, that
testing. this specific end-turn area, exactly at the point where the
winding extended from the iron core, was the root cause
of repeated failures. These motors were operating in a
IV. DATA INTREPRETATION AND CORRECTIVE wind-tunnel application, therefore were subjected to
ACTIONS OF MV MOTORS repeated starts, stops, and thermal loading swings. It was
determined that the cycling was causing excessive stress
Table 3 is used to summarize data interpretation and on the end -turn windings, therefore causing small cracks
corrective actions, but first a discussion is presented and voids to the insulation at this point where the
concerning viable corrective actions. winding extends from the iron core.

Based on the partial discharge characteristics and various This damage went unnoticed by traditional testing
root cause analysis, we can begin to identify deficient methods, therefore resulted in the ultimate failure of
areas of an on-line motor. The presence of relatively high large and costly machines. Only due to the presence of
positive polarity partial discharges indicates potential the differential protection relaying, was the root cause
problems at end-turns, surface discharges or tracking due identified, since the motor had been de-energized
to contamination, or voids between the outer insulation immediately after the initial fault. Detailed inspections,
wall and the iron core. The end-turn potential problems and attempts at patch-repairs to this junction of the
and surface discharges can be effectively addressed, winding and core-iron proved unsuccessful, resulting in
thereby mitigating any additional insulation complete rewinds of the entire motors. Following the
deterioration, and possibly providing extended costly rewind of three units, it was determined to remove
equipment life. the remaining nine units and complete a cleaning of the
winding, baking, dipping with new varnish and a final
bake cycle. In addition, the end-turn support rings were
reinforced, therefore minimizing the mechanical stress on
the insulation at this junction, due to the cyclic
operations. The remaining nine motors operated
satisfactory for the next eight years, until the wind tunnel
was decommissioned. If on-line partial discharge
technology was utilized at that time, we would have
expected to see an increasing trend of the positive
polarity discharge pattern, which would have identified
the weakened end-turns, and resulted in a considerable
savings in repair costs, as well as operating up-time. This
facility had the personnel and funding to complete such
on-line testing, but partial discharge technology was not
yet widespread.

Referring back to Table 1, which summarized the results

of an IEEE and EPRI study: Stator Ground Insulation
accounted for 23% of the failures, while Bracing
End turn stress where the winding extends
accounted for an additional 3%. This total of 26%, may
from the iron core.
be related to end-turn damage to some extent.
Figure 10 – Motor Stator End Turns
In contrast, if on-line testing indicated a negative polarity
partial discharge pulse rate to be more prevalent, this
The author, having over ten years of association with motor would indicate voids between the copper conductor and
repair shops, was witness to many medium voltage the insulation, which could be evidence of possibly poor
impregnation during construction, or a recent rewind. simultaneously, and the resultant waveforms are
Partial discharge testing after the rewinding of medium compared. If turn-to-turn problems are clearly
voltage motors is highly recommended to provide base- identified, this may limit the effect of the cleaning
line data, and possibly uncover potential quality process since internal turn-to-turn faults can not be
problems involving the rewound motor. Partial accessed by external cleaning.
discharges between the copper conductor and insulation
result in limited low-cost corrective actions, since access 5. Retest of the following after initial cleaning:
to the problem areas is not possible, even if the motor is a) Partial Discharge
removed from service. In these cases, repeated b) Insulation Resistance and optional power factor
monitoring should be maintained, and the unfortunate testing
budgeting for a major future repair and the associated c) Polarization Index
downtime. One operating advantage is the possibility of d) Three-phase surge test if available and required
completing the necessary major repairs, during other based on initial test results.
major plant improvements. It would be most
embarrassing to allow a large potential cost item to go 6. Repeat Item #5 as cleaning progresses.
unapprised, and then to possibly fail after a plant as been
restarted following other major improvements. 7. Retest as in Item #5 before unit reinsulation.

The last alternative, is an equal balance between positive 8. Retest as in Item #5 after unit reinsulation.
and negative pulses. Based on the theoretical discussions
presented, a balance would indicate either an equal 9. Final pre-energization tests as in Item #5, after
intensity of voids at the inner copper conductor- complete baking of new insulation applied.
insulation interface versus the outer insulation wall and
iron (slot discharge) or surface tracking related voids; or NOTE: Temperature and humidity should also be
most likely, that the majority of pulses are emulating recorded each time Item #5 testing is repeated. During
from voids internal to the insulation, as illustrated in on-line testing, loading levels should also be recorded.
Figure 7 and 9-B. In either case, repeated on-line
trending would be recommended to identify a difference Results have shown that some insulation systems can be
in polarities. If no difference is evident, then the improved by cleaning and reinsulating, but it depends on
conclusion can be made that the voids are within the the extent of existing insulation damage. This effort is
internal insulation. Corrective actions are similar to usually much less costly than a complete rewind,
negative polarity pulses, since the internal voids can not therefore is worth considering, even though positive
effectively be repaired without a complete motor tear results can not be guaranteed.
down and rewinding
The previous theory, and data interpretation can be
Other viable solutions now include effective field applied to the following two illustrations of actual
cleaning of large machines involving either corncob measured partial discharge activity.
materials, CO2 or traditional hand cleaning. In all such
cases, the following testing protocol is recommended: First, Figure 11 illustrates the two critical parameters,
partial discharge Maximum Magnitude and Pulse
1. Review of on-line PD measurements for the last six Repetition Rate plotted separately for the first half of the
months. sine wave, 0 to 180 degrees, and also shown for the
second half of the sine wave, 180 to 360 degrees. Keep
2. Off-line PD testing, with additional sensors installed in mind, as illustrated by Figure 7, that the negative
where possible. polarity pulses will be represented in the first half of the
sine wave (0 to 180 degrees); while the positive polarity
3. Pre-cleaning off-line insulation resistance testing pulses will be represented in the second half of the sine
and polarization index determination. Power factor wave (180 to 360 degrees).
testing is another tests that can be completed, if
available. The ‘y’ axis indicates Pulse Repetition Rate, shown in
pulses per cycle and the ‘x’ axis indicates the Maximum
4. Since the unit is out-of-service, it is also Magnitude, shown in volts. By plotting this relationship
recommended to complete a three-phase surge test as two separate curves, for each polarity, we can begin to
to determine if any turn-to-turn problems may exists determine the possible root cause for the partial
at this time. During this off-line test a periodic high discharge activity.
frequency signal is injected to all three phases
 100 Table 3 summarizes the data interpretation and
_ _ _ _ _ _ 0 - 1 8 0 , _ _ _ _ 1 8 0 - 3 6 0 recommended corrective actions. The first column
10 includes the partial discharge results. This is followed by
the possible root-cause, based on the partial discharge
Pulse per Cycle

levels and the regions of associated insulation voids. The

1 next two columns include the short-term and long-term
0,1 recommendations. The root causes vary from normal
partial discharges to significant regions of voils within
the insulation system. Concerning recommendations,
0,01
trending is recommended within a 3 to 6 month period at
the first indication of substantial partial discharges. In
0 0,02 0 , 0 4 [V ] 0 , 0 6 0,08 0,1 most cases the root cause and partial discharge activity is
comparative except for the situation when the insulation
is old and shows signs of external wear, or if there is
Figure 11 – Positive & Negative Polarity evidence of surface tracking. These situations may
Partial Discharges indicate insulation at a “near-failure’ state where the
partial discharge arcing as progressed to the point where
permanent carbonization, or tracking, as occurred to the
The example illustrated in Figure 11 shows the positive insulation system. In this case it is recommended to
polarity pulses exceeding the negative polarity pulses. schedule an outage for traditional insulation resistance
After further on-line monitoring and trending the root testing, and possible installation of permanent partial
cause may be considered the interface between the discharge sensors for improved on-line measurements.
insulation and iron, or surface tracking contributing to
surface partial discharges. In this case, repeated
observations and trending are required, and external Incremental Testing can help further identify, or clarify
cleaning and reinsulating should be considered if the possible root causes. The first Incremental Test is the
partial discharge trend increases. “Temperature Variation Test”. In this test, you start the
motor at as close to full load as possible, thereby
maintaining the load as constant as possible, during the
test. The voltage should also be constant for this test.
Record partial discharges as the temperature increases. If
the positive polarities increase, then the problem maybe
3 related to slot discharges or end- turn tracking. If the
per Cycle

2 negative polarities increase, then the root cause maybe

related to the copper conductor- to-insulation area. The
1 second Incremental Test is the “Load Variation Test”. In
this test, you attempt to start lightly loaded and record
Pulse

0,003

0 0,009
partial discharges as the load is increased. The voltage
0,030 and temperature should remain constant, since the
0
90 0,094 [V] increase in load should be completed in a relatively short
180 time period. If the positive polarities increase, then the
[Deg.] 270 root cause is most likely loose windings, or end-turn
360
tracking. This is another area where a cost-effective
repair is possible, by having the motor removed and re-
wedged, or other winding tightening techniques applied.
Figure 12 – Partial Discharge – 3D View
This approach is still much less costly than a complete
rewind. Off-line partial discharge testing would involve
Figure 12 illustrates the above two critical parameters, applying a voltage to the motor and recording partial
with the full power cycle degrees (0 to 360) as the third discharge activity. This testing may not be possible
axis; thereby providing a two dimensional presentation without a variable voltage supply. Lastly, during repeated
of the partial discharge activity. This figure also on- line monthly testing, for cases under investigation,
illustrates the higher level of partial discharges during the humidity and temperature should be recorded. If the
the rising negative half of the power cycle (180 to 270 partial discharge activity substantially varies with
degrees), or as previously discussed the higher level of humidity, then the cause may be surface tracking.
positive polarity discharges.
 Table 3 - Motor Partial Discharge Data Interpretation & Corrective Actions
Partial Discharge Possible Root- Short Term Long Term Corrective
Results Cause – PD Site Corrective Actions Actions
Moderate to low • Normal Partial At first indication, If trending is level,
partial discharge Discharge repeat on-line testing extend on-line testing to
Magnitude and in 3 months. 6 months, or as
Repetition Rate • Beginning of PD scheduled.
activity If insulation is old and If insulation is near

• Insulation near shows signs of failure, traditional testing

failure external wear, or any should indicate low
evidence of surface insulation resistance
tracking, then values.
schedule outage for
traditional insulation
resistance testing.
Trending indicates • Slot / surface Repeat on-line -If positive polarity,
increasing partial tracking PD testing in 1 to 3 schedule field or shop
discharge activity months, depending cleaning and reinsulating
• Internal insu- on the severity of the with end-turn bracing.
lation voids increase. -If negative polarity,

• Winding loose- If the trend increase budget for major rewind

ness if indicated is substantial, and schedule outage.
by the “Load schedule outage and -If winding loose-ness
Variation Test” test monthly until
(Increase in outage. indicated, schedule for
positive polarity removal and shop
pulses with Add permanent rewedging.
increased sensors if required to
loading). improve PD testing.
During outage,
complete Off-line /
Incremental Testing
and traditional
insulation resistance
testing.
Positive Polarity • Voids in the slot Same as above Schedule field or shop
pulses prevalent between insu- cleaning and reinsulating
lation and iron, with end-turn bracing
• Surface tracking

at end-turns
Negative Polarity • Voids at inner Same as above Budget for rewind, and
pulses prevalent copper / insu- schedule off-line test, in
lation interface. hope of minor problems.
Balance of Positive • Voids internal to Same as above Budget for rewind and
and Negative Pulses insulation major outage.
system
 V. PULSE MEAUREMENT ISSUES illustrates discharges identified by RTD’s sensors, which
were not detected by RFCT’s (radio frequency current
As discussed in the theoretical review, partial discharges transformers) on the ground circuit of a surge capacitor.
are high frequency pulses originating at various sections In addition to identifying discharges further within a
within an insulation system. These pulses generate a winding, the other major advantage is obtaining
voltage and current signal into the insulation, returning discharge data from RTD’s can be completed on-line,
through a ground path. There are three partial discharge without an outage as shown in Figure 13. By this
measurement methods actively being applied in the field method, motors equipped with RTD’s can be analyzed
today. now, and a preliminary evaluation completed.

The first involves sensing the pulse voltage signal of the

partial discharges. This pulse voltage signal does
attenuate rapidly as it is transmitted away from the
discharge site. One method is to apply coupling
capacitors at the motor terminals directly [10]. Since
these sensing units are at the motor terminals, they may
not identify partial discharges within the winding depth,
which have attenuated to the a level of background noise.
They would identify discharges at the terminal
connections and at the end of the winding where the
higher voltage stress is present. These sensors do require
an outage for installation; therefore no partial discharge
testing can be implemented until the sensors are
installed.

Using these types of sensors allows for elimination of

noise, and partial discharges external to the motor based
on the rise-time of the signal. The assumption is that
noise external to the motor will have a slower rise-time, Figure 13 – RTD’s Connections for Partial
therefore can be eliminated without further Discharge Detection – On Line
consideration; and that motor internal discharges will
consist of high frequency, high rise-time signals. One
problem with this approach is that brush excitation, or One recommendation may be to install permanent
brush sparking can produce high frequency, high rise- sensors on critical motors, which have evidence of high
time noise inside of the machines, therefore resulting in a partial discharges. Any type of permanent sensor will
false indication of partial discharges. These false require an outage for installation, but will allow for more
indications of internal discharges can be eliminated if accurate on-line partial discharge measurements. As with
properly detected, identified and eliminated from the all partial discharge readings, trending is critical and the
signal under analysis. Terminal mounted coupling use of RTD’s also allows for ease of data collection, and
capacitors also may miss the internal discharges that minimum investment to start a partial discharge
have a slower rise-time due to attenuation within the predictive monitoring program.
insulation, since the low rise-time noise elimination
technology which accompanies the coupling capacitor One issue of obtaining discharge measurements from
approach will eliminate this signal. As with all discharge RTD’s is that noise must be properly identified and
measurements, trending results with coupling capacitors eliminated. Advanced noise and discharge monitoring
will identify deteriorating conditions. techniques, using eight channel data collection units,
have resulted in adequate elimination of noise, and have
Another method of obtaining a level of the partial improved detection of partial discharges. This is not
discharge pulse voltages is to attach special sensors to possible using traditional coupling capacitors connected
existing RTD’s external wiring connections. The RTD at motor terminals.
wiring within the motor is exposed to the partial
discharge pulse traveling through the insulation. In The other method of measurement is applying RFCT’s
addition, motor manufactures apply RTD’s at locations on the ground circuit of motor surge capacitors, or on the
of greatest thermal stress in the winding. The result is the cable shielding grounding conductor. These are generally
identification of discharges that are closer to the root more sensitive than coupling capacitors, since the
cause. Field data has been complied [8] which capacitors attempt to measure the attenuated voltage
of the discharge signal; whereas the radio-frequency 3) PD identification through attenuation analysis. Since
current transformers (RFCT’s) have a greater zone of PD sensors detect pulses originating close to it as
sensitivity to discharges further into the winding depth. well as pulses coming from different locations, this
Referring back to partial discharges being high- attenuation can be observed, thereby rejecting cross-
frequency signals necessitates the need for RFCT’s, that coupled pulses, and recording only signals
are designed to have a high-frequency response band. originating from a particular sensor. This further
Installation does require both an outage, and if not eliminates false PD activity measurement, and
already installed, the addition of an insulated surge allows for improved identification of PD activity.
capacitor unit. On motors with an existing surge
capacitor unit, a difficulty may arise since the frame of 4) Pulses width validation. Pulses with a width
the units may be grounded. The RFCT must be installed considered to be noise, and not high frequency
on the ground connection between the insulated surge partial discharges, can be quickly identified and
capacitors ground and the earth ground; or as an alternate eliminated.
on the supply cable shielding ground connection.
5) Experience and the effective use of eight channels of
Insulation of an existing surge capacitor base may be
input data, as shown in Figure 14, allow for
necessary; therefore an inspection is required for each
motor to determine the best sensor application. The comparisons to external signals to further improve
addition of surge capacitor units on large critical motors the PD data collection process.
has much merit in itself, since all plant switching surges,
or disturbances shall be shunted via the surge capacitor,
VI. APPLICATIONS TO OTHER ELECTRICAL
and not induced into the motor winding.
SYSTEMS
Partial discharge testing as been used primarily for
predictive maintenance of medium voltage motors.
Obtaining measurements on other electrical systems such
as switchgear units as identified severely deteriorated
insulation, approaching eventual failure. A post-failure
root cause analysis of a failed transformer bushing
showed evidence of partial discharges occurring internal
to the bushing porcelain. If partial discharges were
monitored these could have been found, and corrective
actions may have prevented a complete bushing failure.
The following Figure 15 illustrates insulation
deterioration found, resulting from on-line measurements
of partial discharge activity within energized 15KV
Figure 14 –Low Level PD Identification switchgear cubicles. An outage was scheduled and an
& Noise Detection & Elimination internal inspection performed. Severe insulation
deterioration was found at the junction between the two
cubicles which had high partial discharge activity.
With available measurements of lower frequency partial
discharges, originating closer to the root cause, via
RTD’s and via RFCT’s, the issue of noise and partial
discharge identification will be reviewed. With the use of
eight channel recorders and advanced software and
analysis techniques [11], data processing consists of the
following areas.

1) Background noise detection. With the additional

channels, noise originating from specific sources
such as brush sparking can be targeted and
eliminated from the signal analysis.

2) Elimination of synchronous noise, or noise

occurring in a cyclical pattern, every cycle.
Examples of sources are thyristor-firing circuits.
Figure 15 - Bus Insulation Deterioration found by
partial discharge measurements
Figure 16 indicates a side view of a failed transformer power distribution equipment. A RCM study can be
bushing, which resulted in substantial equipment and completed for the electrical distribution system within a
environmental damage, with considerable downtime. plant, thereby providing the following benefits:

1) Identification of critical electrical equipment and

recommended predictive and preventive maintenance
on each device. A RCM study will identify which
equipment warrants the installation of on-line sensors
and other technologically advanced predictive
maintenance tools. A cornerstone of RCM is on-line
predictive measurements.

The partial discharge discussion concerning motors,

within this paper, ideally satisfies these
requirements. In addition, continuous predictive
maintenance methods of measuring the change in
transformer bushing capacitance and power factor
on-line are also available, with periodic partial
Figure 16 – Failed porcelain bushing, with
discharge testing being completed without an
sections removed for evaluation outage. In addition to motor and generator insulation
systems, switchgear, circuit breakers, bus duct,
Figure 17 illustrates the internal examination of the medium voltage cable systems and instrument
above failed transformer bushing porcelain. Evidence transformers can be adapted for on-line partial
exists of carbonization and partial discharge activity, discharge measurements. A recent partial discharge
which if identified could have prevented a complete test of switchgear shown in Figure 15 identified high
bushing failure. levels between two specific cubicle. A scheduled
outage and internal switchgear inspection identified
badly deteriorated bus insulation at the interface to
the bus supports between cubicles. The small air-gap
between the bus insulation and these through-
cubicle bus supports creates an effective void for
partial discharges to develop.
2) Identification of specific preventive maintenance
functions to be completed on a periodic basis, and
more specifically during unplanned outages. This
allows for effective maintenance, on critical
equipment, to be completed rather than randomly
selecting equipment to be maintained. During an
unplanned outage, the RCM study would direct
which equipment should be serviced and a specific
workscope. Part of the implementation of the RCM
study would be on-site training of plant personnel to
Figure 17 – Evidence of internal partial ensure they can properly complete all maintenance
discharges occurring within the failed tasks in a safe and effective manner.
transformer porcelain bushing
3) Identification of auxiliary spare equipment such as
feeder, or main circuit breakers, which should be on-
hand, related to the critical nature of the equipment,
VII. RELIABILITY CENTERED MAINTENANCE rather than past practices of purchasing only less
AND PARTIAL DISCHARGE PREDICTIVE costly spares.
DIAGNOSTICS
4) Identification of equipment which would be
Most manufacturing plants are being forced to operated maintained at a minimum level, thereby redirecting
at greater efficiencies at lower costs [12] and [1]. A new these maintenance dollars to more critical
process called “Reliability Centered Maintenance” equipment.
involves the review of plant processes and supporting
 VIII. SUMMARY [6] D. Fink, H. W. Beaty, Standard Handbook for
Electrical Engineers, Pages 4-117, 118, McGraw-
Partial discharge monitoring is an effective on-line Hill Book Company, 1987
predictive maintenance test for motors and generators at [7] J. Millman, H. Taub, Pulse, Digital and Switching
4160 volt and above, as well as other electrical Waveforms, Pages 50-54 McGraw-Hill Book
distribution equipment. Partial discharges in 2300 volt Company, 1965
equipment can also be observed depending on the [8] C. Kane, B. Lease, A. Golubev, I. Blokhintsev,
equipment design, level of partial discharge activity and “Practical Applications of Periodic Monitoring of
sensor placement. The benefits of on-line testing allow Electrical Equipment for Partial Discharges,”
for equipment analysis and diagnostics during normal NETA Conference, March, 1998
production. Corrective actions can be planned and [9] C.H. Flurscheim, Power Circuit Breaker Theory
implemented, resulting in reduced unscheduled and Design, Pages 556-557, Peter Peregrinus Ltd.
downtime. on behalf of the Institution of Electrical Engineers,
1985
Understanding of partial discharge theory allows for [10] G. Stone, J. Kapler, “Stator Winding Monitoring,”
improved interpretation of results, and the benefits of IEEE Industry Applications, Sept./Oct., 1998
such measurements. Data interpretation and corrective [11] Z. Berler, A. Golubev, A. Romashkov, I.
actions can be clearly identified with cost effective field Blokhintsev, “A New Method of Partial Discharge
corrections implemented, prior to further equipment Measurements,” CEIDP, Atlanta, GA, Oct. 1998
deterioration. Advanced noise analysis techniques and [12] J. Moubray, Reliability-centered Maintenance,
new diagnostic measurement methods using existing Industrial Press Inc., 1997
RTD’s rather than permanently installed sensors, allow
for the implementation of a partial discharge predictive
maintenance program with a small initial investment. Authors:
Partial discharge monitoring technology fully satisfies
the cornerstone of a maintenance program designed to Gabriel J. Paoletti, P.E. received a B.S.E.E degree from
address the critical process support equipment, which Drexel University, Philadelphia, Pa. in 1976. Mr. Paoletti
can be identified by a Reliability Centered Maintenance has over twenty-three years of engineering service
study. experience with Westinghouse, ABB and Cutler-
Hammer Engineering Service. His electrical distribution
The technology as advanced, with improvements equipment experience includes field testing, predictive
resulting in a minimal initial investment, thereby and preventive maintenance, applications engineering,
allowing for partial discharge testing to become a part of failure analysis, and power systems studies. He has
everyday predictive maintenance. design experience with vacuum circuit breaker
modernization, low voltage circuit breaker cell-retrofits;
References: and motor and transformer repair experience. Mr.
[1] P. Roman, “Maintaining Electrical Equipment for Paoletti is a Registered Professional in the States of
Peak Performance, “ IEEC Conference, Sept., 1997 Pennsylvania and Delaware. He is currently Division
Applications Engineer for Cutler-Hammer Engineering
[2] G. Paoletti, A. Rose, “Improving Existing Motor Services.
Protection for Medium Voltage Motors,” IEEE
Transactions on Industry Application, May/June, Alexander Golubev, PhD has PhD’s in both
1989 Mathematics and in Physics from the Moscow Physical
[3] Motor Reliability Working Group, “Report of Technical Institute. Dr. Golubev has extensive
Large Motor Reliability Survey of Industrial and experience in research and design in Laser Beam
Commercial Installations, Part I,” IEEE Generation, High Voltage Plasma Coatings and Electron
Transaction on IA, Vol. IA-21, No. 4, Beam Generation techniques. Since 1990 Dr. Golubev
July/Aug.,1985 has devoted his research energy to the field of Partial
[4] Draft of the IEEE P1434 Standard, “Guide to Discharges. He is currently Manager of Research for
Measurement of Partial Discharges in Rotating Cutler-Hammer Predictive Diagnostics, an operating
Machinery,” 1996, 1997 group of Cutler-Hammer Engineering Services.
[5] Westinghouse Electrical Maintenance Hints, Pages
19-14 and 15, and Page 7-23, Westinghouse
Electric Corporation Printing Division, Trafford,
Pa, 1976