IDC - HV Conf - Perth 2015 - Condition Monitoring of High Voltage Switchgear - KH

Session 5: Condition Monitoring of High Voltage Switchgear
Session 5:
Condition Monitoring of High Voltage Switchgear
Karl Haubner
Applications Engineer, Doble Engineering
Part 1
CONDITION MONITORING OF METAL-CLAD SWITCHGEAR
Karl Haubner, Doble Engineering
Introduction
Metal Clad Switchgear is one of the key assets of the electrical distribution system.
An in-service failure has significant consequences on the reliability of supply.
In the case of a failure large amounts of energy are released. Faults on the board
can cascade to other compartments causing collateral damage within the switchgear
and consequent loss of adjacent circuits and in the worst case scenario loss of the
whole board. This is particularly true for switchboards with oil filled circuit breakers,
resin bonded paper insulated bushings and compound insulated busbars.
Modern switchboards are designed to eliminate the potential to cause injuries by
containing the arc by-products but older boards still in service are not designed to be
arc resistant and any failure there is associated with a high risk of personal injury.
This paper describes some common and not so common condition monitoring
techniques to assess the insulation condition of switchboards.
1. The need for Condition Monitoring

Industry and most utilities are experiencing increased pressure to maximise the
economic return on investments and are forced to operate aging switchgear.
The age of the switchboard population in industry and utility is increasing which
means that an effective condition monitoring program is gaining importance. At the
same time experience indicates that whilst newer arc resistant switchboards are
much safer, the margin to withstand electrical stresses are reduced and many
quality control issues have been detected early using effective commissioning
testing
Age by itself is not a good indicator of estimated remaining life and performance.
Insulation does deteriorate due to electrical, thermal and mechanical stress,
chemical attacks and environmental contamination.
A condition-monitoring program to assess the condition of the switchboard is
required to:
1. Assess the risk of failure.
2. Identify deteriorated components and initiate maintenance or refurbishment (if
possible). There is no point replacing oil circuit breakers with modern vacuum
High Voltage Conference 2015 IDC Technologies
breakers if the condition of the busbar or CT chamber insulation has deteriorated

to an unacceptable degree.
3. Identify the priority for replacement or refurbishment to ensure that limited
resources are distributed effectively.
An effective program also needs to take the type of insulation and arc interruption
method employed into account. The test method best suited to detect insulation
defect in a minimum or bulk oil circuit breaker will not be the same as for SF6
insulated GIS.
Apart from human error, problems with the mechanism (which will not be covered in
detail in this paper) and operating outside of the rated capability, the main electrical
insulation failure modes of metal clad switchgear are:
Failure of the insulation system due to elevated temperatures leading to

thermal runaway conditions.
Surface tracking mainly associated with moisture, dust and contamination on
insulating surfaces. For air insulated equipment this is the most common
failure mode. In many cases tracking is triggered by other discharge
phenomena such a corona or sparking type PD activity that as a by-product
generates corrosive gases that deteriorate the insulation surfaces.
Partial Discharge activity in components such as epoxy insulators, VTs CTs,
bushings etc.
Partial discharge activity from unscreened cables in switchgear or insufficient
clearances such as voltage transformers to frame
Failure of cable terminations not part of but connected to the switchgear
A range of techniques and tools are available to assess the condition of

switchboards. The first distinction must be made between on and off-line testing.
2. Tests to determine the condition of the insulation

Off-line testing is usually employed during the commissioning of new switchgear or
during major outages. Once in service, the asset owner is somewhat reluctant to
give access to equipment so the initial tests are important to provide background
data and to verify the design, material and workmanship during the assembly of the
board on site. In addition the switchboard can then be visually examined which is
just as important as electrical testing.
3. Commission and Maintenance off-line testing of Switchgear

insulation
3.1.
Insulation Resistance Measurement
The Insulation Resistance (IR) measurement is the

oldest, most commonly used and at first glance simplest
test. The Insulation Resistance tester generates a dc
voltage typically between 0.5 to 10kV.
The small current flowing through and over the
insulation under tests is measured and the insulation
resistance is directly indicated on the display.
Temperature has a pronounced effect on the insulation
resistance of insulation material. The value decreases
Figure 1 Commonly used rapidly with an increase in temperature. When individual
values are to be compared with each other it is
IR tester
important that both measurements are taken under
similar conditions or readings are normalised to 20 C. Temperature correction
factors for different switchgear types and makes are generally not available and
generic correction tables are not applicable across all different insulation system.
Here the application of the polarisation index (10 min reading/1 min reading) or the
polarisation factor (60 sec reading/15 sec reading) test which is largely independent
of temperature can improve the meaning of the measurement. However, many
switchgear components do not return significant polarisation currents and the
obtained PI results are meaningless.
Insulation Resistance results are also time dependent and when comparing spot
measurements it is important that the measurement times are the same.
The significant contributing factor to Insulation Resistance results is humidity. The
higher the moisture content in the air the lower the IR values, in particular if surface
contamination across any component of the insulation system is present. No
repeatable results can be obtained if the humidity exceed 70-80%. Temperature &
humidity should be part of the test record.
The test voltage (e.g. 5kV for 11 & 22kV Boards and 10kV for 33kV boards) is
applied between one phase and the other phases connected to earth with all
breakers closed and in the racked in position. Voltage application is typically for 60
seconds. The same procedure is then repeated for each phase. In addition the IR
value across the open contacts of the Circuit Breaker is determined.
IR values vary widely depending on the insulation system employed. Oil filled circuit
breakers or compound insulated busbars can be as low as 5000 M without being
defective but GIS and air insulated switchboards are expected to return IR values in
the high G range if the humidity is low.
One of the major disadvantages of the Insulation Resistance test is that defects in
condenser type multi-layer insulation (which is common in older style paper
insulated bushings used on oil circuit breaker trucks from 6.6 to 33kV and even
some older 33kV busbars) cannot be detected using DC methods.
Figure 2: Failure of paper insulated condenser type circuit breaker bushing. Failure
occurred when racking a spare breaker into the board. The breaker passed a
simple IR test.
3.2
High Voltage Power Frequency Withstand Test
As part of the commissioning process it is common to conduct a Power Frequency

withstand test in accordance with the relevant standards. This test, also called HighPotential test, is made at voltages above the normal system voltage for a short
duration, such as 1 minute.
Test voltages in the field are typically reduced to 80% of the listed values:
Rated voltage
kV
3.6
7.2
12
17.5
24
Rated power frequency
withstand voltage
kV
10
20
28
38
50
36
70
The Switchgear has passed the PFWT test if the test voltage can be maintained for
1 minute without puncture, flashover or disruptive discharge.
Observe and record:
Noticeable rise or decrease in test transformer primary & secondary currents

Large fluctuations of the applied voltage
Audible discharges
Visible discharges
Tripping of HV test set circuit breaker
In essence the PFWT is a go-no-go test and whilst it is essential that a newly
assembled switchboard passes this test, little information about the quality of the
insulation system is obtained. There are many examples of switchgear having
successfully passed a HV withstand test having serious assembly issues.
Figure 3 Example of a switchboard having passed the HV withstand

test successfully but subsequent PD testing identified an incorrect
assembly of VT connections resulting in insufficient clearance
between phases.
A High Voltage Withstand test can also indicate a loss of vacuum in VCBs. A
healthy vacuum interrupter must withstand the applied test voltage above across the
open contacts.
3.3
Dielectric Dissipation Factor (DDF) also called Dielectric Loss Angle
(DLA) tan or Power Factor Measurement
DDF is the ratio between the resistive power loss and the reactive power loss of the
insulation material. This is equal to the tan of the angle and typical displayed as %
Dissipation or Power Factor.
The DDF measurement is one the standard methods to characterise the condition of
insulating materials. The value can vary from 0.1% up to 15% depending on the
insulation system used. The method is most effective for older type switchgear in
particular when condenser foil type stress control methods are employed.
Impregnated paper insulation as found in older Metal-clad switchgear is prone to
absorb moisture and the Dielectric Dissipation Factor is a significant indicator of
moisture ingress. Many cases have been encountered where circuit breaker
bushings returned acceptable Insulation Resistance readings but the DDF
measurement revealed poor insulation values.
The following sketch -figure 4- indicates how the DDF measurement can detect
partial degradation within a multi-layer insulation system.
Figure 4 = Detection of defects in multi-layer insulation.

Older type oil filled circuit breakers have a number of graded electrostatic screens
through the insulation structure of the bushing. A DC test performed on the same
insulation structure is not capable to detect a defect that is still surrounded by good
insulation.
The DDF can be measured using a differential transformer ratio arm bridge, a
schering bridge or instrumentation which measures total current and watt loss using
a reference resistor.
Figure 5 Modern insulation analyser used to conduct DDF tests

As is the case for insulation resistance measurements, DDF values need to be
corrected for temperature for bitumen and oil filled switchgear. At the same time the
variation of the dissipation factor with temperature can also be a significant indicator
of the insulation quality. The temperature dependence of the insulation is more
pronounced with poor insulation quality and the possibility of insulation breakdown
due to thermal runaway is higher. Thermal runaway occurs in unstable insulation
at high temperatures when the high dielectric losses heat the insulation further which
increases the temperature further, losses increase again, etc., until complete
insulation failure occurs.
For modern oil epoxy or gas insulted metal clad Switchgear no correction is
required.
The voltage dependence of the insulation tested is the other important criteria. Good
insulation will show very little increase in DDF value up to 120 % of operating volts.
As the capacitance of the insulation under test increases, the ability of the DDF
measurement to detect incipient localised faults is reduced. When testing, for
example, the busbar insulation a large number of spout bushings are measured at
the same time and the results indicate the average DDF. The true DDF of e.g. a
single bushing with high DDF will be masked by the low DDF of the large number of
good bushings. Thats why on a circuit breaker each bushing is tested separately.
3.4
The DDF measurement is sensitive to moisture ingress and as this is a typical

defect found in older metal clad switchgear components, the testing technique is
considered a significant indicator of deterioration and can detect defects in
condenser type stress grading insulation systems.
The ability to detect localised insulation defects is reduced with increased
capacitance of the test sample.
The DDF tip-up (voltage dependence) can be another indicator of insulation
integrity.
A pass/fail criteria needs to take the type of insulation system employed into
account. Older switchgear can have DDF values above 10% and is considered
poor but still serviceable whereas modern air, gas or epoxy insulation systems
have values below 0.1%.
For modern air or gas insulated boards with Epoxy Cast Resin insulated
structures the absolute value of the DDF test is not as important as the DDF tipup.
Partial Discharge Measurement (PD)
Partial Discharge (PD) activity is produced by incipient faults in HV insulation and is

regarded as one of the best indicator of insulation condition providing an early
warning against insulation faults. By definition PD represent discharge events that
are limited to only a small portion of the dielectric and only partially bridge the
insulation between the electrodes. PD usually occurs due to local electric stress
concentration at defects and hardly within a homogenous insulation structure were
the electric field is a function of the applied voltage and the distance between the
electrodes and is designed with some margin above the breakdown strength of the
material. However in defective insulation systems there can be many locations
where the electric field is not uniform. For example a void inside an otherwise
homogenous dielectric will cause a localized field enhancement across it and if the
dielectric strength of the gas or air inside the void is below the breakdown voltage
determined by Paschens law, a partial discharge pulse is generated. Over time the
fast moving electrons or ions can cause deterioration of the surrounding void wall
insulation resulting in chemical decompositions of the material. This can lead to a
complete breakdown of the insulation. PD testing, either as a factory test to ensure
that the equipment is designed and build not to exceed permissible discharge levels
or as a condition monitoring tool to trend the transition from acceptable to critical
levels, is gaining importance.
In principal the type of discharge can be categorized as external and internal
discharges.
Figure 6 types of Discharge activity

In most cases partial discharge activity has a detrimental effect on the insulation
material. The spectrum of damages due to PD activity include heating, oxidation,
chain scission of polymer molecules, stress cracking due to UV light, surface
erosion, build-up of aggressive gases (e.g. Ozone) which in turn cause corrosion of
metal surfaces, delamination, etc.
Detection Circuit
Although the following simplified a-b-c model depicting a single void surrounded by
insulation was rejected by researchers because a cavity cannot be represented by a
capacitance, it is often used to explain PD activity and how to detect it. The
capacitance of the void and the series and parallel capacitances of the surrounding
insulation is shows as a, b, c
Figure 7 simple a, b c model and current flow

If a voltage is now applied across the test sample a proportional voltage will now
appear across the void
Cc: Capacitance of void
Cb: Capacitance of solid in series w/void
Ca: Capacitance of the rest of the solid
Va: Applied voltage of solid
Vc: Voltage across void
V+/V-: Inception volt. for PD in void
Gro
ups of discharges originate from a single vo
Figure 8
Distribution across void surrounded by insulation.
When the breakdown strength of the gas or air inside the void is exceeded the void
capacitance is temporarily shorted so the voltage across the void drops to 0 V and a
charge transfer from one side to the other side of the void occurs. Ca now needs to
supply the lost charge to Cb to compensate. This is associated with some current
pulses between Ca and Cb/Cc.
Figure 9
Typical current transients associated with a Discharge from a single void
Figure 10 Current flow in coupling capacitor

If now a capacitor Ck with a series impedance is now connected across the circuit,
Ck will supply some charge to Cb/a for Cc. This is detected as a current by the
series impedance in series with the coupling capacitor. (Figure 10)
This circuit is basically described in IEC60270 and represents the standard test
circuit showing the test object, the coupling capacitor, the measurement impedance,
the HV supply and the instrumentation.
The coupling capacitor provides the compensating charge and functions as a High
Pass Filter i.e. blocks power frequency but lets the HF PD signals through. These
decoupled current pulses are converted by the measuring impedance into equivalent
voltage pulses. The pulses are integrated to provide a measure of charge involved in
the PD event at the point the sensor is connected too. As per IEC60270 the
apparent partial discharge level is expresses in pC (pico Coulomb) Those pulses
have very short time parameters in the ns range and are low in magnitude. For
example for an apparent charge of 0.1 pC the corresponding voltage magnitude is
about 100 V, if a time constant of 50 ns for the PD pulse & a 50 Ohm impedance is
assumed.
In many instances the background interference from non-PD sources is much higher
than this.
Figure 11 Standard PD measurement circuit.

Apart from replacing the coupling capacitor with for example a HF-CT there are
several options as far as the PD instrumentation is concerned such as narrow band,
wide band and ultra-wide band detectors going up to hundreds of MHz with the
ability to resolve features of individual pulses. Each system has their own merits and
disadvantages. The most common method employed nowadays is the wide band
technique described in IEC 60270. With modern digital detectors the detection
center frequencies and bandwidth are freely adjustable permitting investigation of
the pattern a different frequencies and superior noise suppression. However it
should be noted that the frequency range used is outside of the IEC guidelines and
results expressed in pC may not be comparable with factory data or measurements
taken at other detection frequencies.
Input Unit
DAQ
vPD
PD-Signal
DSP
Filtering
Detection
Postprocessing
A
D
vU
Voltage Signal
PC
A
D
Computer
Software
User/ Panel
Figure 12 Block Diagram of modern Digital DP detector

With these digital detectors every PD event is captured and recorded permitting
advanced post recording waveform analysis.
The test voltage applied is increased to up to 120 to 130% of line to ground voltage.
10
Figure 13
PD Test Circuit
In addition to the magnitude and type of discharges the other important parameters
such as PD inception voltage, PD extinction voltage and effect of time are recorded.
All these parameters need to be considered before a valid judgment can be made.
Displaying the PD data using the Phase Resolved Partial Discharge (PRPD)
patterns also called phi-q-n diagram is recognized as the one of most effective
methods to display the characteristics of PD signals and is used to classify different
types of PD and interference signals. The method produces patterns that can be
directly related to gas discharge events in the dielectric. The PD pattern reflects the
sum of all individual pulses correct in time (phase position) on the X axis, magnitude
on the Y axis and repetition rate which is color coded collected during a preselected
acquisition time. The user can also select to display the pulse polarity. From this
information the generation of 2 or 3 dimensional graphs is possible.
Figure 14
corona.
Phase resolved pattern from sharp point at HV potential producing
11
Figure 15
3.5
Sparking type discharge from floating bushing screen in switchboard.
Location of partial discharges
After a discharge has been detected, its location is beneficial. If, for example, a
single component can be identified for replacement without sectionalising the
switchboard, significant time & cost savings are possible. Location of PD is possible
with non-electrical and electrical means. Electrical measurements usually involve
comparing arrival times between different sensors (they may be TEV sensors or
Coupling capacitors) using a high speed oscilloscope. Non electrical methods uses
several acoustic sensors.
Figure 16 Location of PD source using time of flight measurement
12
Figure 17 Non electric location methods use microphones to detect the panels that
return the loudest indication of PD activity.
If panels can be removed during off-line testing a corona camera can be helpful
determining the origin of the PD activity.
Figure 18 Defect visualised using a corona camera during off-line testing
4. Non-intrusive on-line tests

An effective condition monitoring regime combines regular non-intrusive survey type
measurements with less frequent off-line assessments.
Often the initial survey test results trigger more detailed testings including monitoring
over a longer period or initiating some off-line testing.
13
Figure 19 Condition Monitoring using a combination of Survey, focused monitoring

and intrusive testing
Whenever there is a PD event different Energy forms are released that can be
detected using different techniques and instruments.
Figure 20 Energy Forms released by Partial Discharge Activity

One of the most commonly used method is the detection of electromagnetic waves.
A PD current pulse comprises moving charges which results in an emission of
impulsive electromagnetic waves that travels in the surrounding media. This
radiation can be detected using specialised spectrum analysers in the RFI range.
4.1
Partial Discharge Monitor using TEV Principles
When a partial discharge inside the switchboard occurs, electromagnetic waves
propagate away from the source and are capacitively coupled to the inner surface of
the metalwork.
Figure 21 TEV Transient Earth Voltage Method
14
These pulses travel due to the skin effect on the surface of the metal and are able to
propagate through an opening in the metal cladding (such as a gasket) onto the
outside surface of the switchboard. The travelling steep current pulse can generate
a transient earth voltage on the metal surface, which can be measured by attaching
a capacitive probe to the metalwork.
This capacitive plate sensor probe is often integrated in handheld detectors
permitting quick scanning of the switchboards.
Figure 22 Commercial handheld TEV detector to survey cable box of switchboard

The TEV detectors have a frequency range of around 2 80MHz and
measurements are unavoidably influenced by other narrowband and background
broadband noise signals in a substation environment that are in the same frequency
range. These external interference signals may originate from air-conditioners,
lights, radio and communication signal, power electronics such as variable speed
drives or other HV plant such as overhead lines and substation equipment in close
proximity to the board and are often of high magnitudes compared to the target
signal leading to false positives. Its important to always compare to background
readings taken from panels that cannot be associated with discharge activity such
as LV boards instead of just reacting on absolute dB values.
Surface PD activity which is the major failure mode in air
insulated switchgear has less energy and emits
electromagnetic waves outside of the detectors
frequency range and can in general not be detected
using TEV principles. Detection sensitivity is naturally
much lower than off-line PD testing and depend on the
propagation path so that a direct correlation to IEC type
measurements in pC are not possible.
Figure 23 Example of significant defect that was undetectable using TEV
techniques
Despite these shortcomings the non-expert ease, speed of use and the ability to
detect many types of high level PD activity non-intrusively makes the TEV method a
popular choice.
15
TEV Sensors can also be used with more sophisticated instrumentation such as
oscilloscopes, recorders or spectrum analysers.
Using an optimised spectrum analyser as shown in figure24 it is possible to display
the captured signal in both the frequency domain and in a time resolved mode.
Figure 24 PD Instrumentation & TEV sensor on switchgear
RFI Frequency Spectrum from

above instrumentation. (Black is
Background), the PD causes an
uplift compared to the
background across a wide
frequency range
Display of the TEV/RFI signal in the time domain

at a spot frequency (6MHz BW) selected from
above spectrum. This guarantees the best signal
to noise ratio.
The trace shows clear phase correlation an
indication of true PD activity but also two types
of PD.
Similar to the phase resolved pattern from the
off-line PD tests characterisation indicating the
type of PD is possible
Figure 25 RFI spectrum between 50 & 1000MHz and time resolved trace with phase
synchronisation @ 620 MHz
16
Some instruments have additional post processing features that characterise the
type of PD activity. Algorithms automate the PD characterization process by analysis
in the time domain extracting the features of each PD event such as Pulse rise and
decay times, pulse width and other essential frequency components. These features
are determined by the type of discharge
Figure 26 Example of automatic software based classification from on individual PD

pulse implemented in a commercial detector
To further investigate the influence of humidity and load conditions on Partial
Discharge activity several manufacturers offer permanent monitoring solutions
combining sometimes TEV and acoustic sensors.
Figure 27- Example of permanent monitoring system installed

Another method is on-line monitoring via permanent installed PD coupling
capacitors. This technique is not popular in Australia as retrofitting is intrusive and
can compromise the BIL rating of the switchboard. As modern flash hazard rated
boards are relatively sealed often having a double skin resulting in reduced or no
effectiveness of TEV and Acoustic methods this policy may need to be re-evaluated.
In theory possible but often subject to significant system interference is the
monitoring via the existing build-in capacitive voltage indicators.
17
Figure 28- Example of PD detection via Voltage indicators

4.2
Ultrasonic Detection
4.2.1 Airborne type Ultrasonic Detectors.

Surface discharge activity, corona, and tracking are best detected with high
sensitivity using an ultrasonic listening device. Ultrasound easily passes through air
but is readily blocked by a solid surface. The detector is a directional microphone
with amplification in the sonic and ultrasonic range. Typical detection frequencies
are 30-500 kHz. The key to a successful ultrasonic survey with a high confidence
level is to find opening in the switchboard were sound can escape.
There must be an uninterrupted air path between the discharge site and the
instrument to allow the airborne ultrasound waves to be detected. Some of the older
switchboard have ventilation vents or other small gaps between panels or covers
suitable for surveying.
18
Figure 29- Examples of suitable opening to conduct ultrasonic survey

Where possible a small gap can be created at switchgear panels permitting effective
access to the airspace to permitting more sensitive acoustic measurements. Another
approach where access to the airspace for the ultrasonic inspection is limited is by
installing acoustic windows as shown.
Figure 30- Examples of acoustic windows retro fitted to Metal Clad switchgear
Instruments provide indications of the acoustic magnitude in dB or dBV but in

practice this level can alter significantly just by changing the angle of the sensor by a
small degree and classification purely on dB results are not conclusive. The operator
can differentiate from the sound what type of PD activity such as corona, contact or
tracking is active. In general acoustic measurements are not effected by electrical
interference (but can be masked by excessive vibration in heavy current
applications) and any acoustic emissions should be monitored and investigated
regardless of the dB V level.
Insulation breakdown in the solid section of the switchboard (e.g. insulated busbar,
internal discharge in instrument transformers or bushing insulation) cannot be
detected using ultrasound.
4.2.2 Contact Type Ultrasonic Detectors.
On well-sealed boards a contact type probe can be used. These have excellent
sensitivity detecting PD on Gas Insulated Switchboards. On Air Insulated
Switchboards contact type sensors have reduced sensitivity when compared to
airborne Acoustic Emissions type and should only be used if there are openings.
19
Figure 31- Contact Type Ultrasonic sensors in use

Some instruments further visualise the acoustic activity by displaying the pattern as
a phase resolved trace in the time resolved mode which permits further classification
and recognition of the type of PD. Synchronisation to the power cycle can be either
via an internal time clock or truly phase locked via a mains outlet.
Figure 32- Phase resolved Acoustic Signature

4.3
Other on-line Tests
External thermo-graphic testing although sometimes used to detect OCB faults, is of

limited use in metal clad switchgear. The problem needs to be in a very advanced
stage to raise temperatures of the steel covers. Some asset owners install Infra-Red
inspection windows (which can also be combined with acoustic windows) increasing
the effectiveness of infra-red surveys on switchboards considerably.
5 Verification of on-line techniques compared to off-line tests
Appendix 1 shows results from tests conducted in the laboratory and provides a
comparison in terms of sensitivity of the most common on-line techniques in use
with reference to off-line PD measurements.
From the results obtained from this controlled study and based on the field
experience of both authors the following conclusions can be made:
20
The Laboratory tests indicate that by far the most sensitive method and the only
way the PD activity can be accurately characterised, the degree of severity
quantified and compared to standards is an off-line PD measurement in
accordance with IEC60270.
However surface PD activity can be detected with very good sensitivity using
Acoustic techniques. Here the airborne type acoustic sensors achieve a much
better sensitivity than the contact type sensors. The dedicated ultrasonic detector
used in the trial with the inbuilt variable frequency filter returned the best
performance in terms of clarity and using the filter being able to focus on different
types of activity and noise suppression.
The same surface PD activity can in general not be detected using methods that
rely on the transmission of electric magnetic radiation (TEV) unless PD activity is
above 1000pC.
TEV/RFI methods can be strongly effected by external interference sometimes

totally preventing meaningful measurements. If the interference frequencies are
narrow banded it can be beneficial to focus on frequencies outside of the
interference band using filtering and to check at spot frequencies if activity is
phase related. In case of broad banded interference such as seen in installations
with variable speed drives, these methods are not effective. Ultra-wide band
measurements can be even more subject to interference with algorithms in most
cases not being able to differentiate between interference and true PD activity.
Additional add-on filters can assist in these cases.
Strong Void PD activity, Sparking type PD and severe PD activity from cable
terminations can be detected using TEV techniques. Representing the signals in
the time domain (phase resolved) can assist in separating interference from real
PD.
Internal PD such as voids inside instrument transformers and insulators cannot

be detected using ultrasonics so a combination of both TEV/RFI and ultrasonics
is essential to cover all defects.
Switchboards where no interference is present can be easily surveyed using

simple hand held detectors combining peak TEV signal indications or LED and
ultrasonic tool. In case interference is present more sophisticated
instrumentation reduces the number of false positives.
Considering that the main failure mode in air insulated epoxy resin based
switchboards is surfaces deterioration and subsequent tracking much more
emphasis should be placed on effective acoustic measurements. This can be
achieved by installing acoustic windows.
The corona camera is not useful for survey type assessments with Panel covers
fitted but is a good investigative pinpointing tool during off-line testing with covers
removed.
21
6
Test Criteria
It is impossible to establish a universal heath index or acceptance criteria that
covers the whole range of switchboards and insulation system employed.
Very often guidelines are only applicable to a specific type, manufacturer and
voltage range.
Modern Metal Clad Switchgear utilising gas or air insulation and epoxy resin
structures have negligible losses and Dielectric Dissipation Factor Measurements
(DDF) are not a significant indicator and not commonly applied but are still listed in
the table.
Some simple guide lines have been established using the hand held detectors.
Figure 33- Some Guide Lines Handheld TEV detectors

Based on experience the following guide lines are proposed:
Acceptance Criteria Modern Metal Clad Switchgear
Test Method
Acceptance Criteria
Insulation
Resistance
> 20 000 M
Partial Discharge
Off-line IEC
values
Ultrasonic
TEV
DDF
(typical values found are > 100G)

< 50 pC @ 120% Uo but take into account type of PD from
Phase Resolved Signature
No AE signal unless identified as non-relevant such as
vibrations
No signals exceeding 20dB, Pulse count <20
Unless identified as background
If DDF is used typical values are < 0.4% and there should
be no increase in DDF with an increase in voltage up to Uo.
22
Figure 34 Acceptance Criteria Modern Air/Vacuum/Gas insulated with Epoxy Resin

support
Example of Criteria for older 11kV bitumen insulated boards with Oil Circuit Breakers
established earlier by Australian utilities.
Test Method
If insulation indicators are worse than levels below classify

into bad category
Insulation
Resistance
Partial Discharge
Off-line IEC
values
> 200 M
< 100 pC @ 110% Uo but take into account type of PD from
Phase Resolved Signature
DDF
< 10% & no significant increase between 2 kV and system

voltage
Ultrasonic
No AE signal unless identified as non-relevant such as

vibrations
TEV
No signals exceeding 35dB, Pulse count <50 per cycle

Unless identified as background
Figure 35- Acceptance Criteria older Oil/ Bitumen insulated boards with Paper
insulated Graded OCB bushings
For results falling in the bad category, remedial action is recommended, access to
the board is restricted and detailed contingency plans in case of failure are
prepared.
If the test results fall between a fair and bad category the time period between retests is shortened or if possible some permanent monitoring is utilised. For
switchgear older than 20 years with results in the good and fair category, a
combination of off- and on-line re-tests typically every 4 years are scheduled.
The value in most insulation diagnostic testing lies more in the survey of a trend than
in the measurement of absolute values. No decision should be made based on the
results of a single test (unless very high readings are discovered). For older bitumen
compound insulated boards, DDF and insulation resistance varies considerably with
temperature and humidity and only trending over a number of tests can provide a
realistic assessment if the insulation is stable or deteriorates further.
It is important to collect benchmark data early in life of the switchboard so that
effective trending is possible.
The data from these diagnostic condition-monitoring tests presents the Asset
Manager with a powerful tool that can help in the decisions making process to
accept, refurbish, replace or retain the new and ageing switchboard. Warranty
issues or defects can be addressed early before the switchboard is placed in
service.
23
Figure 37- Condition Monitoring Maintenance Scheme
24
7.
Case Studies
Case Study 1
11kV Switchboard with defective bus bar support

This case study presented shows how the described tests helped
to determine defects on bushings / barrier boards on an 11kV
busbar in modern air and epoxy insulated switchgear. These tests
were conducted after some noise was heard with the new
switchboard in service coupled with a strong ozone smell.
Initial testing using TEV and RFI principles did not return any
indications of PD activity with typical values equal to background
readings. The modern board is well sealed but both contact and
airborne acoustic sensors placed at small gaps on the arc chute
confirmed PD activity.
Figure 38- Acoustic tests showing PD activity

Whilst online PD methods relying on electromagnetic radiation failed to detect any
PD an off-line test confirmed relative high levels of surface PD activity up to 2000pC
on all phases.
Figure 39- Test Set-up
Figure 40- Phase Resolved PD Signature

25
Switchgear panels were then removed and the location of discharge activity was
exactly pinpointed using ultrasonics and a corona camera.
Figure 41- Pinpointing techniques using both acoustic and corona camera with
covers removed
The PD activity originated from the busbar support fibre glass rings that were not
sealed and had absorbed moisture at some stage. There was also evidence of white
power at the rings and the heat shrink / rubber interface typical for crystallisation
associated with PD activity. The fibre glass material itself showed discoloration.
Figure 42 showing defective support fibre glass board
26
Case Study 2
Multiple PD Sources 22kV Switchboard
This case study portrays the findings on a 22kV switchboard which was removed
from service due to audible discharge and a strong ozone smell emanating from the
switchboard. Testing was requested by the Asset Manager in order to determine the
levels of partial discharge present before performing fault finding to determine the
source of the audible discharge.
Figure 43 Switchboard tested
Figure 44 Phase Resolved PD Signature showing several types of PD

The audible discharge was clearly identifiable to originate from one end of the
switchboard therefore after initial testing which confirmed that the discharge activity
was originating from the bus section and not the circuits; inspections were focused
in this region. Inspections revealed extensive corrosion to copper components,
severe surface deterioration of resin components and signs of localised heating on
heat shrink. This inspection confirmed the offline measurement which indicated that
multiple PD sources were present with different partial discharge characteristics.
27
Figure 45 Busbar connection at rear of spout with cover removed
Figure 46 Surface condition of spout

Testing was repeated with the end busbar panel removed in order to utilise an
ultrasonic detector to pinpoint exact sources of discharge but unfortunately this was
unsuccessful due to limited access whilst maintaining safe clearances, although
during these attempts visual discharge was observed. The lights were then turned
off and visual discharge was very clear and easily distinguishable to be originated
from between the spout and the surrounding frame. Closer inspection of these areas
identified a build-up of discharge powder which had bridged the gap between the
two as well as corrosion on the metal frame.
Figure 47 Tracking point between spout and frame of switchboard

28
The discharge in this case was clearly occurring over an extended period of time
which went undetected until discharge levels were severe enough to be audibly
heard during general substation inspections. This confirms the requirement for
regular monitoring of electrical apparatus both online and offline which helps to
minimise repair costs through early detection or in a worst case scenario save the
apparatus from catastrophic failure.
Case Study 3
Loose mounting bolts on spout flange
This case study focuses on an older bitumen/pitch filled switchboard (oil circuit
breaker). During a scheduled outage on a single CB, audible discharge was heard,
upon racking down the CB. Testing staff were called to the site and through the use
of an ultrasonic detector were able to pinpoint the discharge to the Red phase Front
Bus spout. Based on the testers onsite assessment and recommendations, the
Asset Manager decided to de-energise the Front Bus in order to perform offline
testing and subsequent fault finding.
Figure 48 Switchboard tested showing Bitumen bus bar insulation & oil circuit
breakers
29
Figure 49 Phase Resolved PD pattern

During initial inspections it was noticed that the shutter was fouling and not
functioning correctly and therefore had sustained damaged but at that stage this
was not seen as significant. During offline testing, severe audible discharge could be
heard when a voltage of only 1kV was applied to the busbar. The rear mounting
bolts of the Red Front Bus spout were found to be loose which confirmed the
floating potential pattern obtained with offline testing. After tightening the loose bolts,
not only did the discharge activity disappear but the shutter no longer fouled and
therefore operated smoothly.
Rear flange mounting bolts
Figure 50
PD Signature Loose rear mounting bolts on spout flange
30
Figure 51
Front Bus shutter in fouled position
The switchboard was successfully returned to service, with follow up online testing
and acoustic surveys confirming that the discharge activity has not returned.
Case Study 4
Secondary wiring of CT touching body of another CT
This case study details the findings as a result of a switchboard which was deenergised after general substation inspections identified audible discharge and the
smell of ozone within the switch room. Offline testing of just the busbar was
discharge free but severe audible discharge was present with all CB closed (thus
including all circuits in the test). Simply closing one CB at a time allowed the
discharge activity to be attributed to a single circuit on the switchboard.
Figure 52 Phase Resolved PD signature

The circuit portion consisted of a spout, one stand-off insulator, one neon pickup
insulator and two sets of CTs in series (mounted one below the other.)
31
Initial inspection of the circuit portion of the switchboard revealed extensive

corrosion and surface deterioration of components. Parts deemed as unserviceable
were replaced before performing follow-up offline testing which yielded still very high
levels of discharge. Components of the circuit were segregated and tested
individually (although still mounted in position) with no discharge present but when
the two CTs were tested together the high level discharge was present as well as
able to be heard audibly. Closer inspection between the two CTs identified severe
surface deterioration of the CT body as well as damage to secondary wiring.
Figure 53
Surface deterioration found when Top CT was removed
The CTs were replaced and secondary wiring damage repaired before retests were
completed but large level discharge was still measured. Further investigation
identified that the close proximity of the secondary wiring of the Top CT to the body
of the Bottom CT. The discharge activity was successfully rectified by simply
increasing the clearance of the secondary wiring from the CT body. It appears that
only the combined field strength of both CTs being energised together was enough
to create the discharge that was measured.
Figure 53 Secondary wiring in close proximity
32
Case Study 5
Defective CT busbar section
This case study presented shows how the described tests helped to determine
defects on a CT busbar in an air and epoxy insulated 22kV switchgear.
Figure 54 Switchboard tested
Routine off-line testing of a section of the switchgear returned the following results:
DC Step
voltage test
(Ileakage @
30kV) A
Suspect
Section
Normal
Section
Dielectric
Insulation
Dissipatio
Resistanc
n Factor
e @ 5kV
%
G
@10kV
PD
(pC)
PD (pC)
@
12.7
@ 15.3
L
1
<0.2
0.8
>1000
<10
<10
L
2
1.37
260
510
600
L
3
<0.2
0.89
>1000
<10
<10
L
1
<0.2
0.88
>1000
<10
<10
L
2
<0.2
0.85
>1000
<10
<10
L
3
<0.2
0.91
>1000
<10
<10
Figure 55 Test Results Off-line test
33
Panels of the board were removed and PD activity was located (whilst the busbar
was energised from the test transformer) using a PD probe. It was interesting to
observe that no ultrasonic discharge activity could be detected as the defect was
internal to the busbar insulation. A visual inspection of the L2 CT busbar also
revealed pitting and burn marks of the screen conductor. It was concluded that the
PD activity already caused further deterioration of the insulation evident from the
reduction in insulation resistance and increase in DDF. The CT busbar was removed
and was retested in the Laboratory. Similar high levels of partial discharge activity
were found. The area of highest PD intensity was located to be at the end of the
internal screen.
Location of highest PD
activity
PD signature @ 16 kV
PD Pulse Rate - Phase
5
PD Number/Second
0
0
45
90
135
180
225
270
315
360
Degree
Figure 56
Test of faulty Busbar section once removed from switchgear
Case Study 6 - Significant deterioration of a 30 year old 11 kV switchboard

The following test results indicate a significant deterioration of the rear busbar. Key
indicators of the deterioration are high DC leakage, low IR (in comparison with other
phases and other air-insulated switchboards), high DDF and DDF tip-up, and very
high partial discharge activity. In particular partial discharge results did indicate a
critical deterioration. The PD signature suggested an arcing in air type discharge.
With the side cover removed the discharge source was located (using ultrasonics) to
originate from the bus coupler panel. Visual examination revealed signs of severe
discharge activity between the compound filled bus joint and the panel as per photo.
Although the white phase also displayed signs of previous corona activity between
the panel and the bus joint no partial discharge activity was detected during the tests
probably due to the low humidity present during the tests. This further highlights the
fact that often more than one type of test is required for a comprehensive analysis of
the condition.
34
Rear
Busba
r
DC Step
voltage
test
(Ileakage
@ 20 kV)
A
Dielectric
Dissipation
Factor
Insulation
Resistance @
5kV
@7.5kV %
PD
(pC)
PD
(pC)
PD
@
6.3
@ 7.6
Inception &
Extinction
L1
6.8
4.42%
93.1
<20
<20
>7.6kV
L2
7.4
5.75%
45.2
<20
<20
>7.6kV
17.7
1500
1800+
8.38%*
L3
50
also high
DDF
2.4 kV inc
2.5 kV ext
tip-up
Figure 57 Test Results Off-line test
Figure 58
Discharges between bus bar joint and panel
Case Study 7
Commissioning Issues on AIS 22kV Switchboard
Commission testing showed unacceptable high levels of PD activity in the nC range.

The cloud type pattern suggests a sparking through air type PD activity.
Figure 59
Test Set-up & Phase resolved PD pattern
35
By sectionalising circuits the activity was localised to one panel and a visual
examination showed insufficient spring contact between the busbar and the screen
of the bushing leaving it floating. After rearranging the spring acceptable discharge
levels were obtained.
Figure 60
Insufficient spring contact of screen to HV conductor
Case Study 8 -
Commissioning Issues on 6.6kV AIS Switchboard
During Commissioning of the new Switchboard the L3 (C) Phase showed elevated
PD levels. Although the PD levels were relatively low the phase resolved PD pattern
suggested an internal void discharge with good symmetry between positive and
negative half cycles
Figure 61
Phase resolved PD pattern & off-line test results
Acoustic location was not possible but by switching off sections of the switchboard
and removing links and components the PD was pinpointed to originate from a faulty
voltage indicator. Separate tests at component level confirmed the defective part.
Figure 62
Defective Voltage Indicator
36
Case Study 9 - Commissioning issues on 33kV GIS Switchboard

During commissioning of a new 33kV GIS Switchboard one Phase showed high PD
levels only slight above system voltage. The PD patterns suggested a sparking type
discharge from floating components.
Figure 63
Test Set-up & Phase resolved PD pattern
A subsequent examination of components showed several parts of the switchgear

were not grounded. (Installation error)
Figure 64 Showing some of the components that were not connected to ground
during the assembly.
37
Part 2
CONDITION MONITORING OF GAS INSULATED SWITCHGEAR GIS
1. Introduction
GIS can be defined as switchgear where the conductors and contacts are insulated
are by pressurised Sulphur Hexafluoride gas (SF6). The SF6 gas is used both as
switching and insulation medium.
Due to their compactness, immunity against environmental conditions such as
pollution, their very high reliability, low maintenance and long service life they are
now a popular choice for both distribution and transmission voltage levels.
There are now several GIS installations in Western Australia and the monitoring of
these installations in service is covered in this second part on Switchgear
monitoring.
Figure 65
Outdoor GIS substation in Perth WA
38
Gas insulated Switchgear is considered as overall very reliable however failures still
occur.
Figure 66 Source: Cigre TB 513 Final Report of the 2004 - 2007 International
Enquiry on Reliability of High Voltage Equipment Part 5 GIS 2012
Figure 67 Source: Cigre TB 525 Risk Assessment on Defects in GIS based on PD

Diagnostics 2013
The GIS Equipment is often installed at strategically important locations in the
network and since the time to repair Gas Insulated Switchgear can be considerable
longer than with conventional switchgear any failure can have serious
consequences.
2. Failure modes in GIS
Despite stringent QA procedures during manufacturing and assembly failures still
occur.
39
Figure 68 Failures in GIS

Typical defects in GIS are:
Protrusions
Any sharp points on earthed and live parts in the GIS in the vicinity of on electric
field causes a field enhancement. This reduces the ability of the GIS to withstand
Switching and Lightning Surges. Any sharp protrusions exceeding a size of 1-2mm
are harmful. Protrusions may originate from production, transportation and from
assembly.
Particles
This is a common problem and exists in almost every GIS
Particle with several mm length can reduce the AC withstand level considerably.
If they settle on epoxy spacers and insulators surface treeing my result leading to
flashovers. Particles can be introduced from contamination during manufacturing or
assembly on site or from metal abrasion due to vibrations or from moving parts from
e.g. switching operations
Floating Components
Floating Electrical and mechanical loose shields or other metal parts not bonded to
either HV or ground potential can generate large discharges that degrades the SF6
gas.
Internal PD
Internal PD from Voids and defects in spacers and epoxy components such as
insulators are due to manufacturing issues and factory routine tests should detect
these before shipment.
40
P article on s pacer
Electrically floating
shield
Protrusion
Void in
s pacer
Protrusion on
earth potential
Free particle
Figure 69 Type of failures

3. Monitoring Methods
The majority of users conduct a High Voltage withstand test combined with a Partial
Discharge test during commission of GIS. These measurement sensitivity during
these tests should be in the order of 5pC or better when using a conventional
IEC60270 complied PD measurement system. This is hard to achieve using a
conventional test system with open air connections and its preferable to use a bolt
on totally enclosed transformer to maintain a screened measurement circuit.
Figure 70 Frequency tuned resonant bolt on test sets for on-site testing of GIS
(Source: Siemens)
Once energised the following monitoring options are available:
3.1 Chemical Analysis of SF6 gases
Permanent installed SF6 gas gauges usually monitor the Gas Density which
provides an indication that a sufficient quantity of gas is present in the chamber.
Monitored. These are typically equipped with relay contact that are connected to the
SCADA system to signal the state of the Gas pressure/density.
Figure 71 SF6 density gauge installed on GIS CB chamber

41
In addition it is common to conduct annual SF6 quality check using portable gas
analysers.
Figure 72 Example of SF6 analyser commonly used in Australia

Parameters measured using include
Humidity/Moisture and by-products
Moisture and SF6 are normally non-reactive with each other but in the presence of a
high temperature from arcing the hydrogen and oxygen of the water vapor may react
with the sulfur and fluorine of the SF6 to create hydrofluoric acid (HF), sulfuric acid
(H2SO4, and sulfur dioxide (SO2)
These by-products reduce the dielectric breakdown strength inside the GIS across
insulating surfaces and causes corrosion.
The amount of acid is related to the amount of moisture.
SF6 must have a certain level of purity to ensure safe operation of the GIS.
Purity
Safe limits for moisture and purity are included in guidelines and standards but vary
between companies.
42
Figure 73 Example of limits sets by one utility considering various sources.

3.2
Partial Discharge Measurements (electrical)
Onsite Partial Discharge Measurements fall into two categories.

Conventional PD measurement in the low frequency (<1MHz) according to the IEC
60270 standard. Ion general some direct connection of the PD sensor (Coupling
Capacitor or HF-CT) to the test circuit is required. This is covered in part 1 of this
paper.
3.2.1 Non-conventional measurements in the UHF frequency range
When partial discharges occur they generate electromagnetic waves that propagate
through the switchgear. In Gas Insulated switchgear the transients produced by the
defects have a very short rise time producing Electro Magnetic Radiations in the
upper MHz to low GHz range.
Gas Insulated Switchgear is fully shielded and no magnetic radiation can escape
easily. To detect this EMI it is common to install during manufacture antennas (UHF
sensors) inside the switchgear.
Figure 74 Internal GIS UHF sensor

The most common option to monitor the Partial Discharge Activity is via a spectrum
analyser and an IEC type Partial Discharge analyser or a UHF PD detector with
integrated spectrum analyser.
43
Figure 75 UHF Test Circuit

To maintain good measurement sensitivity a high bandwidth across the whole
measurement path is desirable. The typical frequencies of interest for GIS is
between 250KkHz to 1.5GHz. Due to the near perfect shielding and a detection
frequency of > 200MHz, away from most interference signals, this method permits
the most sensitive PD measurement. A calibration of the UHF method in terms of pC
is not possible although parallel measurements have confirmed sensitive levels of
<2pC. It is recommended to perform a sensitivity check to determine attenuation
through the GIS. There is little attenuation through straight section but GIS
components such as C/B or isolators can attenuate the Signal considerable.
Figure 76 Sensitivity Check on GIS

Many older GIS systems dont have integrated Disk or Cone UHF sensors. Then the
epoxy spacers between the compartments and inspection windows provide the only
aperture where Electric Magnetic radiation can escape.
Specialised external UHF sensors can be fitted to these apertures in the shielding to
detect the EMI waves.
44
Figure 77 External Window and spacer sensor.

Measurements with the external spacer or spacer sensors are not as interference
immune as the fully shielded measurement with internal integrated UHF sensors
UHF Measurement results
Measurements are taken initially in the spectrum analyser mode and if possible
compared to background or historical readings.
With test voltage applied,
large uplift across wide
spectrum from
background response
Background without test

voltage, no fault
Figure 78 UHF signal detected by Spectrum Analyser Circuit

From the spectrum view a frequency of interest is selected and a time resolved or
Point of wave measurement is taken to identify the type of fault. The pattern is
synchronised to the mains frequency via a wireless adapter.
Figure 79 UHF Signal in Point of Wave/ Phase synchronised view

Various knowledge rules are available to categorise the PD patterns based on
shape, statistical parameters and frequency content
3.2.2 Acoustic PD measurements
A valid alternative or a complementary technique to assess the GIS for any defects
is the acoustic measurement already covered in part 1 of the paper.
45
Acoustic measurements are performed by using

an external piezoelectric sensor and a portable
instrument as shown.
The sensor picks up the acoustic waves/sound
that propagate in the enclosure due to emitted
acoustic signals from defects inside the GIS. Each
section of the GIS enclosure separated by flanges
should be tested with at least one measuring point.
The acoustic signals from the described defects
are generally wide
the range of 10several MHz
defects may vary
signals from
signals from for
The shape of the
the type of source,
signal and the
banded - partial discharges in

100 kHz and particles up to
The acoustic signals from the
widely from continuous
corona to pulse shaped
example moving particles.
acoustic signal will depend on
the propagation path of the
sensor characteristics.
The measured parameters of the acoustic signal can be displayed and evaluated in
three different measuring modes the continuous mode, the particle mode and the
phase resolved mode.
Figure 80 Knowledge rules to identify defects based on the acoustic signature

The measurement is immune against most types of external interference, has good
sensitivity for detection of the most common types of defects. Once a defect is
detected the origin can be localised to at least between two flanges and a risk
assessment based on source characterisation is possible.
The following UHF and corresponding signals were detected using a portable device
that combines both UHF and acoustic modes.
46
UHF Mode- Electrical
UHF Spectrum up to 1GHz shows

PD activity indicated by the large
diversion from Background and
high Peak to Average ratio
In Time /Phase resolved mode at

711.2MHz 6 MHz BW high
signals but poor relation to power
cycle
Acoustic Mode
Both Bar and Oscilloscope

mode show high signal
levels with poor power
cycle correlation
47
Phase synchronized mode

shows some phase
correlation
Interval/particle mode
clearly shows a bouncing
particle inside the GIS
Figure 81 UHF and Acoustic Signatures for same defect

The particle plot (signal amplitude vs. flight time) shows whether the signal is due to
a particle.
The flight times provide information on whether the particle moves into the high-field
region. Signal amplitudes provide information on the particle length.
Based on the parameters of the signal some estimation about size and risks can be
made.
48
Part 3
Time, Motion and Travel Testing of switchgear
Whilst so far the discussions have been limited to the assessment of the insulation
quality of Switchgear it should be noted that the circuit breaker and its operating
mechanism itself is basically a mechanical device. Life expectancy of HV circuit
breakers is at least 30 years but under normal operating conditions the breaker will
operate less than 10 minutes and under fault conditions it will operate less than 1
minute throughout its service life.
During the Time and Motion test various timing and motion measurements are be
made. After the test is complete the results are compared to the manufacturers
specifications.
During the test we may identify problems in the:
Mechanism, Linkages & Shock Absorbers the most common problems
discovered are lubrication issues, mechanical binding, or mechanical
interference. These types of problems may be identified by abnormal timing,
travel, or velocity results.
Main & Resistor Arcing Contact Systems these types of problems are most
frequently identified by excessive contact bounce on the contact timing plots.
Control Circuit these types of problems can be found directly or indirectly
during the motion and timing test. Most frequently they are first identified
indirectly, where the problem in the control circuit causes the CB to fail to
operate or to fall outside the manufacturers timing specifications. During the
subsequent investigation problems in the control circuit can also be identified
directly using the instruments Analog or Auxiliary channels.
49
Figure 82 Typical setup for timing without motion transducers

The above circuit just measures the contact state of the break per phase but in
addition the trip and close currents are measured to obtain the timing.
To further analyse the breaker other parameters are recorded via Auxiliary channels
that monitor the state of contacts in the circuit breakers control circuit whilst
analogue channels can measure any voltage or current of interest such as DC
supply voltage or current and voltage during dynamic resistance tests.
These are common timing parameters recorded for such a test.
Opening Time (Trip Test O)
Closing Time (Close Test C)
Reclosing Time (Reclose Test O-C)
Reclose Open-Close Time (Reclose Test O-C)
Trip-free Dwell Time (also known as Close-Open Time) (Trip-free Test CO)
Most instrument provide the initial data from each test in graphical format but in
addition tables are available indicating numerical data
Figure 83 Plot & data from a TDR test showing parameters of all connected
channels
The numerical data for each test can be shown and if manufactures specification
are available in the test plan automated pass/fail indications are given.
For detailed analysis of circuit breakers Motion Channels are used to measure
motion of the circuit breaker using transducers. Motion transducers that
temporarily connected to predetermined points on the mechanism convert
measured mechanical motion into an electrical signal that is communicated to
Instrument.
the
are
the
the
Figure 84 Motion transducers temporarily connected

50
Including motion measurement in the measurement protocol provides Travel Plot

and Velocity Plots. The Travel Plot shows the movement of circuit breaker over time.
The Velocity plot shows the velocity over time. From these Travel and Velocity Plots,
travel and velocity parameters are calculated.
Average Velocity
Total Travel
Overtravel
Rebound
Contact Wipe
Figure 85 Travel and Velocity Plots and table

Contact Resistance
The contact resistance tests using a 4 wire DC method injecting typically at least
100A to 200A are essential to identify any contact deterioration.
A popular requirement now is the so called Dynamic Resistance measurement that
monitors the voltage drop across the contact continuously. Either a DC source is
integrated in the instrument or a separate DC source such as a battery or the output
of the Resistance Meter can be used whilst utilizing analogue channels to measure
voltage and current. The actual value of the calculated resistance is of no
importance but the plot can identify additional features. This is a complimentary
measurement to the above test but is required if graphite arcing contacts are used.
Voltage drop measurement results depends on breaker contact system design and
analysis is best done by comparison with previous tests.
Figure 86 Dynamic Resistance (Voltage drop) measurement.

Minimum voltage required to trip the breaker and oil tests are common but exceed
the scope of this paper.
Whilst it is essential for Transmission type Circuit Breakers to monitor travel &
velocity by attaching motion transducer to the mechanism these test are usually not
conducted on distribution type breakers although timing measurements should be
conducted at least during commissioning and major maintenance activities.
51
References:
1. Balcombe, H. The assessment and management of older oil-filled switchgear,
Power Engineering Journal December 1997
2. Bradwell, A. and Bates, G. Analysis of dielectric measurements on switchgear
bushings in British Rail 25 kV electrification switching stations, IEE Proceedings,
Vol.132, Pt.B, No.1, January 1985
3. Brown, P. Non-intrusive partial discharge measurements on high voltage
switchgear
4. Caviagelli, G., Kopaczynski, D., Lachman, M. and Levi, R. AC Power Factor
Versus DC Insulation Resistance Measurements
5. Douglas, R. and Booth, N. Testing Ageing Switchgear, IE Australia Electric
Energy Conference, 1992
6. EA Technology PDM3, Ultratev+ User Manual
7. HVPD PDsurveyor Air & Longshot User Manual
8. Ultra probe 2000 User Manual
9. Hilder, D. Partial-discharge measurements for insulation quality, Power
Engineering Journal, March 1992
10. James, R., Phung B. and Blackburn, T. Partial Discharge Phenomena
Characteristics, Interaction With Materials, Interpretation, ESAA Short Course,
Brisbane, April 1999
11. Koenig, D. and Y. Narayana Rao Partial Discharges in Electrical Power
Apparatus VDE Verlag
12. Lachman, M., Doble Engineering , 1998 Seminar Course Notes
13. Lemke Diagnostics GmbH Differential PD-probe LDP-5 Users Manual
14. Lundgaard, L. Partial Discharge Part XIV: Acoustic Partial Discharge
Detection Practical Application, IEEE Electrical Insulation Magazine,
September/October 1992-Vol.8, No.5
15. Mettam, J. Insulation Aging In 12kV Switchgear A Users Perspective,
Proceedings of the 4th International Conference on Properties and Applications
of Dielectric Materials, July 3-8, 1994, Brisbane Australia
16. Neil Davies, Simon Goldthorpe Testing distribution switchgear for partial
discharge in the Laboratory and the field, Paper 804, CIRED Prague, 2009.
17. M. Boltze, S. Kornhuber, Various Methods of the partial discharge detection at
switchgears, 2011 Doble Engineering Company -78th Annual International
Doble Client Conference
18. Doble Lemke PD smart User Manual
52
19. Tettex Instruments, Information No. 21, Teilentladungs-Messtechnik

20. Tettex Instruments, Information No. 23, Teilentladungsmessung
21. TransiNor AIA User Manual
22. Doble PDS100 & DFA300 User Manual
23. A. Nesbitt, B.G. Stewart, S.G. McMeekin, Substation Surveillance using RFI and
complementary EMI detection techniques, 2011 Doble Engineering Company 78th Annual International Doble Client Conference
24. Sumereder C, Muhr, L Applied Risk Analysis for High Voltage Equipment,
Proceedings of the 9th International Conference on Properties and Applications
of Dielectric Materials, July 2009, Harbin, China
25. U. Schichler, E. Kynast, High-voltage Tests and Measurements during the Life
Cycle of GIS, HIGHVOLT KOLLOQUIUM `07
26. R.G.A. Zoetmulder- S. Meijer- J.J. Smit, Delft University of Technology.
Conditional Based Maintenance with On-line Partial Discharge Measurements of
HV and MV Switchgear Systems, C I R E D 17th International Conference on
Electricity Distribution, 200
27. HV PD Course Notes
28. Transpower NZ, Metal Clad Switchgear Maintenance TPSS02.26 2002
29. Bernhard Fruth, PD Interpretation Scheme, PDTech 1999
30. E. Kuffel, W. S. Zaengl, and J. Kuffel, High Voltage Engineering: Fundamentals,
2nd ed Butterworth-Heinemann, 2000
31. F. Kreuger, Partial Discharge Detection in High Voltage Equipment, Butterworth
& Co., Kent, 1989
32. IEC Publication 60270, 2000, Partial Discharge Measurements
33. R. James, Q Su, Condition-Assessment-of-High-Voltage Insulation in Power
System Equipment, IET Power and Energy Series 53, The Institution of
Engineering and Technology UK, 2008, ISBN 978-0-86341-737-5
34. E. Lemke, S. Belijn, E. Gulski, M. Muhr, E. Pultrum, T. Strehl, W. Hauschild, J.
Rickmann, and G. Rizzil, Guide for partial discharge measurements in
compliance with IEC 602
35. Cigre Technical Brochure 513 - Final Report of the 2004 - 2007 International
Enquiry on Reliability of High Voltage Equipment, Part 6 Gas Insulated
Switchgear (GIS) Practices (2012)
53
36. Cigre Technical Brochure 556- SF6 Analysis for AIS, GIS and MTS Condition
Assessment (Feb 2014)
37. Cigre Technical Brochure 525- Risk Assessment on Defects in GIS based on PD
Diagnostics (2013)
38. Doble TDR 900-TDoble Manual.
39. Levi Jozef, Motion measurement and use of transducers for the Detection of
circuit breaker characteristics 2011 Doble Engineering Company -78th Annual
International Doble Client Conference
40. Levi Jozef, Radenko Ostojic Use of Micro Ohm Meter as a Power Source for
DRM testing of Dead Tank Circuit Breakers 2013 Doble Engineering Company 80th Annual International Doble Client Conference
Biography
Karl Haubner joined Doble Engineering in 2004 and is employed as the High
Voltage Test Application Engineer servicing the Asia-Pacific region. Prior to his
appointment with Doble Engineering he worked for the Utility Western Power. As the
Superintendent of the HV Test Laboratory he was responsible for all technical
aspects of the test group such as development and introduction of new condition
monitoring techniques on distribution, transmission and generating plant. In addition
to his employment with Doble Karl also provides testing and consultancy services to
the industry via his company High Voltage Solutions. He is the author of several
technical papers on condition monitoring of HV assets and cable fault location and
has delivered short courses on testing techniques at University and Industry level.
Karl has a Diploma in Electrical Engineering and is member of Cigre D1, AS
Committee EL-007, and of VDE.
54
Appendix 1
Verification of effectiveness and limitation of various on-line techniques used by conducting parallel on and off-line measurements using a range
of commercial detectors.
A ring main unit (RMU) that had been removed from service due to PD activity (signs of surface PD activity at the fuse holders evident) was
energised from a HV test transformer. The RMU is gas insulated and all defects (existing or simulated) relate to the exterior connection points
and not the busbar. The level of Discharge Activity present was measured via a 20nF coupling capacitor and an IEC 60270 type compliant PD
detector (Doble-Lemke PDsmart) operating in the frequency range between 100-500 kHz as prescribed in the standard. The total test circuit was
calibrated by injecting 10pC across the switchgear. The RMU was first energised as found and then measurements were taken at several
voltage levels. Several defects were then introduced to simulate other faults than the ones present already on the switchboard but commonly
found in switchgear
Whilst the off-line PD measurements provide more statistical parameters then the pC value shown below for the purpose of this exercise only
IEC pC values and the phase resolved PD patterns are listed. Phase synchronisation for the off-line measurement was via the coupling
capacitor (true phase position) whereas synchronisation for the hand held RFI/EMI Spectrum Analyser and the Ultra-Wideband TEV analyser
was from the Mains. Humidity and temperate conditions were recorded during the tests but are not shown as the aim was to compare
measurement techniques.
The transmission path from the PD source to sensor the metal cover was direct with the defect always in close proximity to the enclosure. This
represents an ideal situation compared to example busbar defects in Metal Clad Switchgear were the transmission path may be longer and
obstructed by other barrier material or additional covers.
55
Off-line IEC 60270
TEV
Acoustics
Airborne Acoustics
Contact
RFI using unidirectional Acoustics (Phase

resolved)
Antenna
Acoustic
Contact type Sensor
s
Airborne
RFI/EMI
UHF/TEV sensor
Full spectrum & point
on wave (phase
resolved)
Corona Camera
Ultra-wide Band
TEV Sensor
Test
(Background~15dB)
Measurement ~15dB
30pC
1-2dB
BG 14db
As found
at 3kV
Clearly
Audible Audible
audibl
17dbm 15dm
e
None
detected with
covers on
~15dB
150-200pC
As found
10kV
as found
@ 14kV
500-600pC
4dB
Clearly
Audible Audible
audibl
19dbm 20dbm
e
None
detected with
covers on
13dB
Pulse/
Cycle:
Clearly
Audible Audible
audibl
25dbm 20dbm
e
None
detected with ~15dB
covers on
56
1.5
Severity
5
With covers
removed
~15dB
800-1000pC
as found
@ 14kV
B Phase
13dB
Pulse/
Clearly
Cycle: Audible Audible
audibl
0-3
46dbm 31dbm
e
Severity
1
~ 15dB
Void
Discharg
e insert
insulator
with
voids
800-1200pC low
pulse count
24dB
Pulse/ -3dBV -3dBV Not
Cycle: 0 Not
Not
audibl
Severity audible audible e
0.16
N/A
57
700-800pC
Voltage
indicator
O/C
+2000pC
Cable PD
(Stress
Control
removed)
33dB
Pulse/
-3dBV -3dBV Not
Cycle:
Not
Not
audibl
56
audible audible e
Severity
2.08
25dB
Pulse/
Cycle: Audible
5.5
25dbV
Severity
84
~ 26dB
N/A
~
30dB
Audible
Clearly
fair to
audibl
poor
e
19dbV
58

IDC - HV Conf - Perth 2015 - Condition Monitoring of High Voltage Switchgear - KH

Uploaded by

Copyright:

Available Formats

IDC - HV Conf - Perth 2015 - Condition Monitoring of High Voltage Switchgear - KH

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

IDC - HV Conf - Perth 2015 - Condition Monitoring of High Voltage Switchgear - KH

Uploaded by

Copyright:

Available Formats

Session 5: Condition Monitoring of High Voltage Switchgear

Applications Engineer, Doble Engineering

1. The need for Condition Monitoring

Session 5: Condition Monitoring of High Voltage Switchgear

breakers if the condition of the busbar or CT chamber insulation has deteriorated

Failure of the insulation system due to elevated temperatures leading to

A range of techniques and tools are available to assess the condition of

2. Tests to determine the condition of the insulation

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

3. Commission and Maintenance off-line testing of Switchgear

Insulation Resistance Measurement

The Insulation Resistance (IR) measurement is the

Session 5: Condition Monitoring of High Voltage Switchgear

High Voltage Power Frequency Withstand Test

As part of the commissioning process it is common to conduct a Power Frequency

Noticeable rise or decrease in test transformer primary & secondary currents

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Figure 3 Example of a switchboard having passed the HV withstand

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Figure 4 = Detection of defects in multi-layer insulation.

Figure 5 Modern insulation analyser used to conduct DDF tests

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

The DDF measurement is sensitive to moisture ingress and as this is a typical

Partial Discharge (PD) activity is produced by incipient faults in HV insulation and is

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Figure 6 types of Discharge activity

Figure 7 simple a, b c model and current flow

ups of discharges originate from a single vo

Session 5: Condition Monitoring of High Voltage Switchgear

Distribution across void surrounded by insulation.

Typical current transients associated with a Discharge from a single void

Figure 10 Current flow in coupling capacitor

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Figure 11 Standard PD measurement circuit.

Figure 12 Block Diagram of modern Digital DP detector

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Phase resolved pattern from sharp point at HV potential producing

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Sparking type discharge from floating bushing screen in switchboard.

Location of partial discharges

Figure 16 Location of PD source using time of flight measurement

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Figure 18 Defect visualised using a corona camera during off-line testing

4. Non-intrusive on-line tests

High Voltage Conference 2015 IDC Technologies

Session 5: Condition Monitoring of High Voltage Switchgear

Figure 19 Condition Monitoring using a combination of Survey, focused monitoring

Figure 20 Energy Forms released by Partial Discharge Activity

Figure 21 TEV Transient Earth Voltage Method