Beruflich Dokumente
Kultur Dokumente
Author
Udo Ranninger | Udo.Ranninger@omicronenergy.com
Daniel Gebhardt | Daniel.Gebhardt@omicronenergy.com
Date
Sep 30, 2016
Application Area
MV and HV cable
Keywords
High Voltage Test, Cable, Fault Localization, Coupling Methods
Version
V1.2
Document ID
ANP_16004_ENU
Abstract
This application note gives an overview of PD measurement on medium- and high-voltage cables and their
accessories using different coupling methods. Basic theoretical information and fault localization techniques
are also explained.
General information
OMICRON electronics GmbH including all international branch offices is henceforth referred to as
OMICRON.
The product information, specifications, and technical data embodied in this application note represent the
technical status at the time of writing and are subject to change without prior notice.
We have done our best to ensure that the information given in this application note is useful, accurate and
entirely reliable. However, OMICRON does not assume responsibility for any inaccuracies which may be
present.
OMICRON translates this application note from the source language English into a number of other
languages. Any translation of this document is done for local requirements, and in the event of a dispute
between the English and a non-English version, the English version of this note shall govern.
All rights including translation are reserved. Reproduction of any kind, for example, photocopying,
microfilming, optical character recognition and/or storage in electronic data processing systems, requires the
explicit consent of OMICRON. Reprinting, wholly or partly, is not permitted.
OMICRON 2016. All rights reserved. This application note is a publication of OMICRON.
Working on high-voltage assets can be extremely dangerous. Consequently, only personnel qualified, skilled
and authorized in electrical engineering and trained by OMICRON are allowed to handle the OMICRON
measurement equipment. Before starting to work, clearly establish the responsibilities. Personnel receiving
training, instructions, directions, or education on the measurement equipment must be under constant
supervision of an experienced operator while working with the equipment. The operator is responsible for the
safety requirements during the whole procedure.
The user manuals of OMICRON measurement equipment or alternatively the e-book has always to be
available on the site where the equipment is operated.
The users of OMICRON measurement equipment must read this manual before installing the equipment and
they must observe the safety, installation, and operation instructions therein.
The OMICRON measurement equipment and its accessories may be used only as described in this user
manual. Any other use is not in accordance with the regulations. The manufacturer and the distributor are
not liable for damage resulting from improper usage. The user alone assumes all responsibility and risk.
Opening OMICRON measurement equipment or its accessories invalidates all warranty claims.
1.4 Disclaimer
If the equipment is used in a manner not specified by the manufacturer, the protection provided by the
equipment may be impaired.
2.1 Overview
Due to different application areas and historical reasons, several types of MV and HV cables, each with their
advantages and disadvantages are in use out in the field. The most common types are impregnated paper
three core single sheath (MV); belted (MV); external gas pressure; internal gas pressure and thermoplastic-
(XLPE, EPR) cables, where the core conductor and sheath exist in different designs and materials, including
single or multi core, cooper, aluminium, lead, etc.
EPR
Very flexible 1-110 kV
(tan ~ 0,5%) Low tensile strength
Large permissible temperature range Special cases
(r ~ 3)
Flexible
Multi core oval conductor Current suppression External gas pressure
Simple setup
Sheath
The metallic sheath of an MV or HV cable is used as the return path for the capacitive load and fault current.
It is a protection against accidental contact, electromagnetic shielding and it builds a basic mechanical
protection. Beside the electrical sheath, an additional armor is often used to protect the cable from external
mechanical forces (Except when using the steel tube technique).
A layer sheath is a combination of wires and conductive tape.
Depending on the cable type, additional layers build the total cable construction. As PE for instance is not
resistant to water, the cable sheath should be totally water tight. Often there is an additional fleece
preventing longitudinal water distribution.
Light
Cooper wire sheath
Small cable diameter Low mechanical protection PE - cable
Aluminium layer sheath
Longitudinal watertight
PE - cable
No corrugation essential
Heavy Single wire of external
Lead sheath Small cable diameter
Pressure protection (oil cable) gas pressure cable
Longitudinal watertight
Low pressure oil cable
Light PE - cable
Big cable diameter due to
Good mechanical protection corrugation Single core gas
Aluminium wave sheath
pressure cable
High short current Complex longitudinal water tight
conductivity Low pressure oil cable
Figure 5: Common faults causing PD activity within a cable, electric field lines are shown in blue
Figure 6: Electrical tree growing from the main conductor towards the counter electrode. For the photo the XLPE has been heated to
become clear (around 115C).
PD signals in general are very broad-band signals. The unavoidable low-pass characteristic of a cable
influences the signal. Higher frequencies will get damped while the signal propagates from the fault to the
sensor. The amount of damping as well as the frequencies that are damped depend mainly on the distance
between fault and sensor.
A high frequency PD impulse travels in both directions between the core conductor and the sheath. The
recharging current, which represents the PD impulse can therefore be measured either at the core conductor
or the sheath of the cable. An inductive PD coupler (HFCT) should not cover both conductors adjacent to the
insulation material, as the signal would be eliminated due to invers polarity.
Figure 7: Simplified cable circuit diagram, low pass behaviour depends on core conductor, semi conductive layers, insulation material
and cable length
Figure 8: With increasing distance from fault location, the PD signal loses its broad-band characteristic due to the low-pass
characteristics of the cable under test. (Colour of the frequency curves correspond to the impulses in Figure 7).
Using a higher frequency range often allows a less disturbed measurement, while it limits the range from a
sensor in which faults are detectable. Vice versa, lower frequencies are often more disturbed but allow for a
'deeper' look inside the cable.
Figure 9: Damping of a calibration impulse along a HV cable in dependence of the measuring frequency - each joint represents 500m.
The MPD allows selectable center frequency as well as the bandwidth for the PD measurement. For a first
measurement, a bandwidth of 300 kHz with a center frequency below 1 MHz (e.g. 400 kHz) is a good
starting point. Depending on the disturbances and the cable length between consecutive sensors, these
settings have to be adapted.
"Visibility" of PD
The low-pass characteristic of the cable damps higher signal frequencies. Higher measurement frequency
reduces the "visibility". It is not possible to measure PD far away from the measuring point at high
frequencies. Therefore the center frequency should be selected as low as possible to see far inside the
cable. On the other hand, using a high center frequency allows for a very sensitive measurement that
focuses on the point of measurement, which is often the termination. Care needs to be taken when not
following the above mentioned points, as resonance effects in the cable or on the termination might lead to
changes in sensitivity.
Disturbances Noise
Disturbances and noise are not equally distributed over the frequency range. There might be certain
frequencies which are noisier than others. Very often, noise is dominant in the low frequency range. It might
be possible to avoid disturbances and to get a better signal-to-noise ratio by carefully choosing the
measuring frequency and bandwidth.
Figure 11: Disturbances are dominate in the low frequency range up to 1,5MHz the measurement filter should be set next to the noise
hump but as low as possible
Figure 12: TDR example joints can be seen as impedance changes where HF signals are partly reflected.
Figure 13: Influence of wide band filter compared to narrow band filter.
PD occurring with a very high repetition rate, like corona, might blind the system so that other PD events
cannot be measured, respectively the sTDR will not work.
It is beneficial to use a wide filter bandwidth as positive superposition, negative superposition or even a total
cancellation of the PD signal caused by reflections and its resonances can be avoided. Resonance might
be seen in the FFT view.
One would think that the highest filter bandwidth leads to the best result, however there is one drawback. A
low-pass damped PD impulse is low in amplitude but relatively long in time. The peak detector will not
recognize the signal as a PD impulse in case the rise time is longer than the filter dead time. In such a case
the statistical localization might not work, thus the bandwidth should be reduced.
The statistical fault localization can be done using a recorded measurement, however it is recommended to
perform a quick localization onsite, allowing the shift of the measuring frequency if required.
Cable type
Cable rating
Cable length
Single phase or 3 phase
Type of end termination
Joints How many joints and their exact position
Reason/expectation of measurement
Test setup
Coupling method
Surge arrester
Sheath grounding
Online or offline test (power frequency or VLF)
Test procedure
Relevant standard
Test reports of previous measurements
Weather (Indoor/Outdoor)
End terminations are used to manage the electrical field at the end of MV or HV cables, similar as bushings
are used at transformers. Different solutions exist due to the variety of insulation materials. The most
common techniques are geometric (HV/MV cable) and refractive/resistive field control (MV cable).
Figure 15: Cable end without field control Tangential concentration of the electric field at the surface of the insulation material. The
inception voltage can be estimated by knowing the dielectric strength, insulation capacitance and surface capacitance. =
Figure 16: Geometric field control Deflector attached to the semi conductive layer increases the insulation capacitance which is
directly related to the PD inception voltage. In addition, the electric field is displaced into the solid insulation (higher r). There is no
tangential concentration of the electric field at the surface of insulation anymore. Most common solution for HV cable termination/joints.
Figure 17: Resistive and non-linear field control A semi conductive tube or tube with microvaristors (siliciumcarbit, zinkoxid) is
attached to the outer semicon layer of the cable. The higher the electric field, the more conductive the tube gets, thus the field is
automatically controlled especially in the areas of high field strength.
Figure 20: Capacitive field control - Defined voltage drop across the grading layers reduces the local electrical field.
DANGER
Death or severe injury caused by high voltage or current
Switch off the high voltage. Always obey the five safety rules and follow the
detailed safety instructions.
When mounting Measurement Equipment and its signal cables, please observe the
high-voltage safety distances between the Measurement Equipment (incl. cable)
and HV, as well as the phases and phase to ground. Always adhere to
international standards and local safety regulations.
Ensure that the rating of the Measurement Equipment fits the requirements of the
grid or test procedure.
All capacity (coupling capacitor or DUT) can store and recover dangerous charge
(and high voltage). Therefor ground all capacities carefully between the HV tests.
Figure 21: Basic connection diagram of a coupling capacitor at a cable end termination.
HFCT
The HFCT has to be installed at the grounding wire between the cable sheath and the end termination.
If more than one grounding wire is used, it should be considered to extend the length of one of these wires
to feed both grounding wires through the HFCT. Otherwise, the HFCT will measure only parts of the high-
frequency signals. The percentage of the signals measured is defined by the high-frequency impedance of
the wires. If the impedance of both wires would be similar, a maximum of 50% of the signal will be detected.
The test or grid frequency which is required for a phase-synchronized measurement can be transferred by
the sensor (HFCT) in case the current is high enough. An air gap should be added in case the current is
higher than 15 A (for MCT 120) to avoid saturation effects (7,5 A for MCT 100). A Rogowski Coil around the
MV or HV cable (connected to the V input of the MPD) might be used if the current through the HFCT is too
low.
The grounding of the cable sheath heavily influences the signal quality, thus there is a difference if the
sheath is connected directly to ground or via SVL.
DANGER
Death or severe injury caused by high voltage or current
Switch off the high voltage. Always obey the five safety rules and follow the
detailed safety instructions. Keep safety distance to high voltage.
Do not mount the HFCT to any grounding cable or electrical conductor where the
occurrence of high currents is possible during mounting. Due to electromagnetic
force, the split-core elements could be contracted, resulting in possible injuries
(trapped fingers) or damage to equipment.
Never open grounding connections where the occurrence of any currents is
possible.
Figure 24: An HFCT installed at the sheath grounding wire of a HV cable termination.
Figure 27: HFCT installed at the sheath of a HV cable close to the end termination.
DANGER
Death or severe injury caused by high voltage or current
Switch off the high voltage. Always obey the five safety rules and follow the
detailed safety instructions.
When mounting Measurement Equipment and its signal cables observe the high-
voltage safety distances. Always adhere to international standards and local safety
regulations.
Ensure that the rating of the Measurement Equipment fits the requirements of the
grid or test procedure.
The UHF approach ensures a very sensitive (local) PD measurement in environments with high disturbance
levels.
Figure 28: Basic connection diagram of a UCS1 sensor at different cable end termination designes.
A current flow in the core conductor of an MV/HV cable causes a magnetic field which induces a voltage into
the sheath of the cable. The longer the cable, the higher and more dangerous the voltage-to-ground on the
sheath can be. The sheath has to be bonded to ground potential in order to limit the voltage to an acceptable
level. If the sheath is grounded at both ends or at multiple points, then the induced voltage causes a
circulating current in the sheath. This current will cause additional heating and results in a reduction of the
cable carrying capacity (losses). To minimize such losses, different bonding techniques like single point (only
for short cables), both ends (additional cable for circulation current) or cross bonding can be used.
Losses in long 3-phase HV cable systems can be reduced best by using the cross bonding technique. The
length of the cable is divided into three approximately equal sections, where the screen is cross bonded in
each joint (link box) and connected to ground at the end of the circuit. The vector sum of the induced voltage
is practically zero in case of a symmetrical operation (same current in all phases, 120 phase shift). [2]
Figure 29: Principle cross bonding setup, sheath voltage distribution above the connection diagram.
Build up from a stack of varistors (e.g. a metal-oxide varistor MOV) causes the voltage above a certain
response voltage to be reduced to a pre-set voltage level. Because the construction is made of multiple
single varistors, difficulties during PD measurement might occur. During normal operation, each inner
varistor pill is on floating potential and might produce partial discharges.
Figure 31: Principle drawing of a sheath voltage limiter. The arrester pills are shown in red.
HFCT
HFCTs can be used to decouple the high-frequency PD signals at the cross-bonding connections, as these
impulses travel between the main conductor and the cable screen. The signal propagation can be very
complex, therefore it is beneficial to have a connection straight through, especially during tests with an
external voltage source. The HFCTs have to be connected around the cross-bonding connections.
DANGER
Death or severe injury caused by high voltage or current
Switch off the high voltage. Always obey the five safety rules and follow the
detailed safety instructions. Keep safety distance to high voltage.
Note: The cable sheath may lead dangerous high voltage.
Do not mount the HFCT to any grounding cable or electrical conductor where the
occurrence of high currents is possible during mounting. Due to electromagnetic
force, the split-core elements could be contracted, resulting in possible injuries
(trapped fingers) or damage to equipment.
Never open grounding connections where the occurrence of any currents is
possible.
Figure 32: HFCTs installed in a cross-bonding box, on every phase of a three-phase cable system.
3.3 Joint
The acceptance level of a newly-manufactured MV and HV cable is less than 10 pC and therefore often not
an issue within the first years of service. End terminations and joints have to be built onsite, hence the risk of
failures at those components is much higher. Inadequate field control or voids (e.g. cavities, dirt, etc.) due to
the onsite fabrication can lead to an inhomogeneous electric field that can cause PD activity.
Different cable types and voltage levels require different joint techniques, therefore several solutions do
exist.
Figure 34: Component joint, two component material (silicon rubber) shrinks and applied pressure on the single parts.
HFCT
HFCTs can be connected around the bypassed sheath at the cable joint to decouple PD signals. The
physical principle is the same as when measuring at cross-bonding connections.
The IEC 60270 states that a coupling capacitor has to be used when performing a PD measurement. Such
an approach provides information about the voltage, which is needed for a phase-synchronized PD
measurement. OMICRON offers standard coupling capacitors from 12 kV up to 100 kV with and without
external measuring impedance (quadripole / CPL 542).
The more energy that is saved in the capacitor, the higher the recharging current is in case of a PD event,
and the sensitivity is higher. Therefore a capacitor value should be as high as possible.
DANGER
Death or severe injury caused by high voltage or current
Switch off the high voltage. Always obey the five safety rules and follow the
detailed safety instructions.
When mounting Measurement Equipment and its signal cables, please observe the
high-voltage safety distances between Measurement Equipment (incl. cable) and
HV, as well as the phases and phase to ground. Always adhere to international
standards and local safety regulations.
Ensure that the rating of the Measurement Equipment fits the requirements of the
grid or test procedure.
All capacity (coupling capacitor or DUT) can store and recover dangerous charge
(and high voltage). Therefor ground all capacities between the HV tests carefully.
Figure 37: Typical cable PD test setup according to IEC60270 measurement on ground potential.
When using a coupling capacitor (CC) without an integrated measuring impedance, the low side of the
coupling capacitor has to be connected to the input (red socket) of the CPL542 (basic test setup with
measurement on ground potential). Connect the PD output of the CPL to the PD input of the MPD and do
the same for V. The black socket or grounding screw has to be connected via a short, low inductive cable to
ground potential. Please note that the measuring impedance and acquisition unit can be placed on various
positions, such as on HV potential or within the test object path due to the fiber optic approach. Therefore
different setups have their benefits.
The same setup but usually a higher charge value might be used to calibrate the propagation speed or cable
length. To do so, use the Measure- / Time- (TDR) or Localization- / Loc.- (sTDR) tab. Detailed information
can be found in the TDR and sTDR chapters.
Figure 40: Software settings PD tab External quadripole, preamplifier auto gain checked, trigger settings.
The influence of using short or long banana cables connected to the calibrator is shown below. It is highly
recommend to use short and low inductive connections to avoid oscillations and resonances. The reflection
might not be visible in case of an oscillating signal. Respectively a localization might not be possible.
Figure 43: Software settings V tab Trigger Source MPD 600, Select for display, V calibration.
It is recommended to double check the software calculated divider factor. As an example, a 2 nF coupling
capacitor in combination with a 30 F CPL leads to a divider factor (test voltage) in the range of 1:15500
Figure 45: Omicrons split core MCT 100 and MCT 120 HFCT inductive sensors.
PD causes electromagnetic signals. Inductive sensors pick up the magnetic part of the electric signal by the
same principle as a "real" transformer. HFCTs are often used if a ground/earth connection is available.
Therefore the HFCT is placed around these connections and transfers the high-frequency impulse to a
secondary winding. The main benefit of using HFCTs is the possibility to measure PD pulses not at high
voltage potential but at grounding connections without opening them.
DANGER
Death or severe injury caused by high voltage or current
Switch off the high voltage. Always obey the five safety rules and follow the
detailed safety instructions. Keep safety distance to high voltage.
Do not mount the HFCT to any grounding cable or electrical conductor where the
occurrence of high currents is possible during mounting. Due to electromagnetic
force, the split-core elements could be contracted, resulting in possible injuries
(trapped fingers) or damage to equipment.
Never open grounding connections where the occurrence of any currents is
possible.
The MPD contains an internal quadripole. This has to be enabled in order to separate and detect the 50 Hz
components of the coupled signal directly from the PD input path. Select the PD tab in the OMICRON
software. Set all units where a HFCT is connected to internal quadripole and do not use the V port. The
HFCT couples inductive with the current to be measured. In order to gather stable PD patterns, it will be
sufficient when the current has the same frequency as the voltage. However there will be a phase shift to the
normally used voltage signal. This shift can be compensated in the Q tab of the software (). See the
chapter phase synchronization for more information.
All other software settings remain as shown in the chapter coupling capacitor.
Table 5: Typical divider factor (charge) values of MCT 120 with and without air gap.
The amount of primary turns has to be considered by entering the corrected value in the calibration text field
of the Q tab. For example, a calibrator value of 100 pC injected over 5 turns to the HFCT means 5x 100 pC
= 500 pC.
Figure 49: Calibration impulse (10nC) measurement with a heavily saturated HFCT load conditions with 200A through the HFCT.
Beside the saturated voltage waveform (green), the actual current through the HFCT was a sinusoidal
current a line can be seen in the histogram. This line was generated by constant calibrator pulses of 10nC.
Due to the core saturation, the true PD value could not be measured!
The effect of saturation can be reduced by introducing an air gap to the HFCT. Shown below is the same
test setup as above but with an air gap.
Figure 50: Calibration impulse (10nC) measurement with a heavily saturated HFCT after introducing a 0.2 mm air gap load conditions
with 200A through the HFCT.
The air gap will not only avoid saturation at higher currents, but it also influences the calibration factor and
the dynamic range of the HFCT. The standard airgap foil provided by OMICRON is 0.1 mm.
The UCS1 has a signal side on top and a ground side next to the TNC connector. The ground side must
always have a solid, low inductive grounding, whereas the signal side has to be connected to the cable
sheath. The coaxial cable towards the UHF converter must have a certain distance from the voltage lead
(signal side) due to safety reasons. The distance shouldnt be less than 15 cm. The UHF converter attached
to the MPD 600 has to be connected to the TNC connector of the sensor. Besides the safety distance, the
cable should be as short as possible to ensure a sensitive measurement.
Due to safety reasons, the senor should not be connected while the cable is energized.
DANGER
Death or severe injury caused by high voltage or current
Switch off the high voltage. Always obey the five safety rules and follow the
detailed safety instructions.
When mounting the Measurement Equipment and its signal cables, please observe
the high-voltage safety distances. Always adhere to international standards and
local safety regulations.
Ensure that the rating of the Measurement Equipment fits the requirements of the
grid or test procedure.
In the PD tab, select the UHF preamplifier, deactivated Auto gain and set the gain regulator to position 6-8.
In case the signal is very low, the gain level might be changed. The UCS1 is not designed to transfer the
50/60 Hz signal content, thus an external voltage trigger (phase synchronization) or additional channel has
to be used. Prior to starting the UHF sweep in the UHF tab, the start and stop frequencies should be
selected. Typical measuring frequencies are 300 MHz up to 600 MHz. The upper limit of the sensor is
1000 MHz. It is recommended to perform several sweeps, as partial discharge might not be active at the
time of a single sweep. After a few sweeps, the sweep feature has to be deselected and the measuring
frequency (V_UHF tab) has to be tuned to a proper frequency range to ensure the highest signal-to-noise
ratio. The lower curve in the UHF view represents the noise and continues the signals, whereas the upper
curve contains the noise, continues signals as well as sporadic PD impulses (maximum value).
Using the UHF 620, several narrowband filters, a 70 MHz medium band filter and a 1.9 GHz wideband filter
can be selected. This allows a precisely tuneable measuring frequency. Please note that in case of using a
narrowband filter, the center frequency might have to be shifted during the measurement.
Figure 54: UCS Software settings UHF tab Set start and stop frequency, hold Min/Max selected, activated and deactivate f-sweep.
HFCT
In case the current through the HFCT is high enough, the pattern can be triggered using an internal
quadripole, as mentioned in the HFCT chapter.
Figure 56: Example setup of an HFCT and a special current clamp connected to an MPD.
Even when using an external current clamp, the PD pattern will stay phase-shifted in relation to the test
voltage. This phase shift will most probably be dependent on the load conditions and will not be constant
over the time! A phase shift in the PD pattern can be compensated by entering the angle in the Q tab.
The only way to get the real test voltage in on-line measurements is to either have an MPD connected to a
voltage measuring device (e.g. capacitive divider, secondary side of a VT) or to know or estimate the phase
difference between voltage and current. In general, the additional current clamp can be connected around
the sheath or the whole MV/HV cable. A rough estimation in no-load situations is a phase shift of 80 to 90.
Figure 57: MPD software settings using a light source as trigger Select Line Trigger 50/60 Hz as Trigger source.
Voltage divider
In general, all kinds of sinusoidal voltage signals below 60 V can be connected to the voltage input port of
the MPD 600. This signal might be gained form a LV side of a VT or 230 V to e.g. 50 V transformer.
Internal Trigger
If none of the above-mentioned solutions is feasible, the internal trigger of the MPD 600 might be used.
Select the trigger source Internal Trigger and change the Int. trigger freq. until a stabile PRPD can be
obtained.
6.1 Corona
Figure 58: Corona visible in both half waves Source is most likely on high voltage potential.
How to identify/distinguish:
o Inception (not applicable during online measurement)
Starts in the negative half wave of the test voltage if corona is on high voltage potential or in
the positive if it is on ground potential.
o No change in amplitude on change in voltage level (not applicable during online
measurement).
o If discharge patterns appear in both half waves, they will have very different amplitudes
while being 180 apart.
o Reflections: Full reflections can be seen at short cable lengths.
o Partial reflections can be created on joints, which are installed in the cable under test.
Corona
First reflection
Second reflection
Figure 59: Corona on ground potential Measured at a cable with reflections visible.
Due to the change of the characteristic impedance on joints and on the end of a cable, a PD impulse
propagating in the cable will be partially transmitted and partially reflected. The ratio of transmission to
reflection depends on the size of the change of characteristic impedance.
Due to the damping of the signal, the reflection is always much lower in amplitude than the original signal.
Thus if the cable or the distance between joints gets too long, the reflection is not visible anymore.
How to identify/distinguish
o Two clouds are visible (almost symmetric in size and shape), which have phase difference
of 180 to each other.
o The clouds are displayed in both half-waves.
o Sometimes the amplitude of one cloud is higher than of the other one. But the density acts
the other way around.
o Dominant in the first (0 to 90) and third quadrant (180 to 270). This applies only, if phase
correction was carried out
How to identify/distinguish
o The discharge pattern will always appear around the zero crossings of the test voltage.
o It will grow to both sides on increasing test voltage (not applicable during online
measurement).
o The dumpling shape pattern will appear on both half waves with comparable amplitude
(attention logarithmic scaling).
Figure 62: Developing void fault in XLPE cable Approximately one minute measuring time.
Figure 63: Developing void fault in XLPE cable Approximately two minutes measuring time.
Figure 64: Developing void fault in XLPE cable Approximately three minutes measuring time.
In the PRPD view, a potential fault within a surge arrester with more exact floating potential will be located
close to the zero crossing of the test voltage.
Compared to contact discharge, the repetition rate is usually lower and the PD pulses do not disappear with
time.
Udo Ranninger
7.1 TDR Time Domain Reflectometry
Udo Ranninger
Figure 66: TDR schematic A single PD impulse causes traveling waves on the cable in both directions (green and blue).
Udo Ranninger
Figure 67: TDR schematic The impulse traveling to the near end (green) gets measured directly at the time t 1.
Figure 68: TDR schematic The impulse traveling to the far end (blue) gets reflected and is measured at the near end at the time t 2.
The time difference t is directly proportional to the distance of the fault to the far end.
A single PD pulse on an expanded test object (cable) causes traveling waves in both directions. One pulse
will directly be measured (green in the example) at the near end. The other pulse reaching the far cable end
will be reflected. It will also be measured at the near end after the time t. The time delay of these two
pulses is directly proportional to the position of the PD fault. If the propagation speed is known, the fault
distance from the far end can be calculated easily by: =
2
As TDR is a well-known method the OMICRON software, as most solutions, expects a propagation speed of
to be entered to return correct results.
2
OMICRON 2016 Page 52 of 70
Software Settings TDR
Prior of using the localization tool, the PD impulse has to be triggered. Cursors can be activated in the
Measuring tab allowing a precise time measurement. Enter the propagation speed and set the cursors to the
beginning or peak of the original impulse and its reflection. As mentioned earlier, the shown distance
represents twice the signal path and therefore has to be divided by two. Discharges along a cable lead to
results seen from the far end. This approach can be used to calibrate the cable length propagation speed
respectively or to locate discharges.
Figure 69: Determination of total cable length (400m) using TDR. In addition a joint can be seen at 2.9 s with respect to the near end
of the cable (point of measurement).
Besides the well-known TDR, there exists a second way for fault localization on cables. Statistical TDR often
allows for good localization results in on-line/field measurements as it is able to scope better with a higher
noise level.
As sTDR is a partly automated statistical approach, the real propagation speed v is used to calculate the
distance from the source to the far end. For the best results, use a high bandwidth filter as mentioned in the
chapter measurement frequency selection.
The sTDR functionality is shown in the following two examples:
Example 1
Given is a cable consisting of three parts with two joints in between. The propagation speed on the cable is
taken to be 200 m/s. In a real cable the propagation speed is lower, depending on the type of cable.
There is an HFCT sensor around the grounding of the sheath at the near end. The far end is left open. A PD
source in the first section of a cable emits a signal that is propagating along the cable in both directions.
Near end
PD
Far end
source
Joint Joint
Measuring point
Figure 70: PD signal propagating from the source in both directions along the cable.
While one signal (green) can be measured directly on the near end, the other signal (blue) has to travel
along the whole cable and through both joints before getting reflected at the far end.
PD
source
Joint Joint
2 s 3 s 1 s
Figure 71: The signal propagating into the cable (blue) is partly reflected and partly transmitted on every joint along its propagation path.
On the way along the cable, the signal will get partly transmitted and partly reflected on every change in
characteristic impedance. The most common change is on joints in the cable. In this example, 80% of the
signal is transmitted and 20% is reflected.
10 pC
4 s 10 s 12 s Time
0 s
Figure 72: Scope view of the main signals and the noise for the given example.
On the near end, the original signal, as well as all reflected signals will be measured after a certain time, as
long as the damping along the propagation path is not too high. The figure above shows a schematic scope
view of the signals being measured at the near end. For the reason of simplification only the first reflection is
shown.
Amplitude Amplitude
10 pC
100%
0 4 10 12 Time 0 5 10 Time
[s] [s]
Figure 73: Scope view and schematic of signal correlation for sTDR.
At this point the sTDR kicks in. The user selects a certain area in the PRPD view. In this example, the user
selects only the PD source in the cable (green in the scope view). Other signal sources like possible corona
or noise are neglected.
Every single event in the selected area is correlated to all other signals measured. The results of these
correlations time difference between signals and relative amplitude with respect to the starting signal are
now plotted into a histogram view.
As noise events are not phase stable and happen with a statistical manner, they result in a homogeneous
distribution of correlations. The signals from the PD source (green in the scope view) in the cable and the
corresponding reflections are recurring events, which are happening always after the same time distance
and with the same amplitude. So they will pile up in the histogram view of the sTDR, generating clusters.
As stated at the beginning, the software will calculate the real distance for a selected cluster to the far end of
the reflection path.
12s 200 ms
= = = 1200m
2 2
PD
source
Joint Joint
2 s 1 s 1 s
Figure 74: The signal propagates in both directions (green and blue). It is partly reflected and partly transmitted on every join along its
propagation path.
As in the example above, the selected signals are correlated to every signal measured and a histogram view
of the amplitude and time difference correlations is generated.
Amplitude Amplitude
10 pC
100%
0 4 8 12 Time 0 5 10 Time
[s] [s]
Figure 75: Scope view and schematic of signal correlation for the sTDR of the second example.
The highest reflection can be seen at = 4 s, therefore the distance of the PD source to the far end will be
4s 200 ms
= = = 400m
2 2
By using the statistical fault localization mode, it is possible to perform a dual-end measurement. For fault
localization, the dual-end measurement is more accurate than the measurement with one channel. The
original impulse is measured as always. This advantage can be used for measurements on short cables or
for measurements on joints and terminations.
Please note that the scope function (TDR, Time view) of multiple channels is not synchronized, therefore a
dual-end measurement is not possible using TDR.
The main software and filter settings have to be chosen, such as using SDTR with a single channel.
After selecting the discharge pattern in the PRPD view, the second MPD unit has to be selected. Now the
marked discharges of MPD unit 1 (Measuring point 1) are correlated with the signals of MPD unit 2
(Measuring point 2). The shown distance is calculated from the middle of both units and has to be divided by
2. If the distance is positive, the position of the source is towards to the trigger unit. If the distance is
negative, the second MPD channel is closer.
An example of how to use the MPD software in terms of dual-end measurements is shown in the next
chapter Measurement example.
PD source
Measuring Measuring
point 1 Center position point 2
of both sensors
Figure 77: Dual End measuremt
The test object is a 1200 meter long MV cable with five joints and shrink-end terminations. To demonstrate
different types of coupling methods, respectively fault localization techniques, a coupling capacitor and two
HFCTs are used.
As there are five joints, multiple reflections can be expected. The graphic below shows the complexity of the
signal propagation within the cable. Please note that only a few scenarios are shown. It is shown that
different reflections lead to the same reflection time. This will form a single cluster in the STDR view,
resulting in a complex interpretation. Measuring at a joint for instance can lead to at least three clusters,
representing the distance of the unit to the cable end, the distance to the beginning of the cable and the fault
position.
The background noise is dominant in the lower frequency range. Therefore a starting center frequency of
3 MHz and a bandwidth of 1 MHz (high bandwidth filter) was selected.
The PD calibration should be performed for multiple filter settings (fcentre, f) to ensure a calibrated
measurement, as a frequency shift could be essential after applying the test voltage. A calibration signal of
1 nC was injected between the main conductor and the screen of the Test Object. The point of injection was
at the coupling capacitor side.
The divider factor (charge see Q-tab) of the coupling capacitor unit (at 3 MHz) was 3.501. A factor of 5 was
selected for the HFCT units.
Divider factor for a coupling capacitor:
2 MHz = 3.48
3 MHz = 3.50
4 MHz = 2.82
5 MHz = 4.13
6 MHz = 5.67
7 MHz = 5.67
As expected, multiple reflections within the cable, respectively between the joints, can be seen. The length
of the cable is 1200 m, thus the resulting propagation speed is 163 m/s.
The voltages slowly increased to 11 kV. This level also represents the PD inception voltage.
Due to the high noise level in the lower frequency range and the relatively low discharge level, the TDR
approach could not be used to locate the fault position.
Figure 88: STDR fault localization - Result has to be seen from the far end of the reflection path (cable end).
Figure 89: STDR fault localization - Result has to be seen from the far end of the reflection path (cable end).
As the result is 400 m from the cable end, the fault is most likely in the joint marked red.
Figure 91: Dual End settings for the trigger unit (HFCT2)
The distance shown of 562.9 m has to be divided by 2. The signal source/fault has to be closer to the unit
HFCT2 as the result is positive. The resulting fault position is 520 m + 281 m = 801 m seen from the near
end of the cable.
Special Bonding of High Voltage Power Cables, Cigre, Working Group B1.18 2
OMICRON
www.omicronenergy.com
Subject to change without notice.