???????? ??????? ????????? ??????????? ???? ??? ?????????? ????????
???????? ??????? ????????? ??????????? ???? ??? ?????????? ????????
Author
Udo Ranninger | [email protected]
Date
December 1, 2021
Application Area
VLF medium-voltage cable diagnostic, partial discharge measurement and localization
Keywords
partial discharge, very low frequency (VLF), medium-voltage cable system, onsite PD testing, TDR, sTDR,
PD cable localization
Version
v1.0
Abstract
Energizing power equipment, such as power cables, can be challenging due to the enormous capacitive
load. Reducing the test frequencies down to 0.1 Hz or even lower is one approach to overcome this
challenge. Performing a partial discharge (PD) measurement under VLF sinusoidal voltages leads to
different requirements for the test system and how the results are analyzed and interpreted. This application
note provides insights into how the MPD Suite software supports and simplifies a PD measurement under
sinusoidal VLF test voltages.
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 undertaken 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, 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 2021. All rights reserved. This Application Note is a publication of OMICRON.
NOTICE
Carefully read and understand the content of this Application Note as well as the manuals of the
systems involved before taking them into operation.
Please contact OMICRON support if you have any questions or doubts regarding the safety or
operating instructions.
Follow each instruction listed in the manuals, especially the safety instructions, since this is the
only way to avoid the danger that can occur when working on high voltage or high current
systems.
Only use the equipment involved according to its intended purpose to guarantee safe operation.
Existing national safety standards for accident prevention and environmental protection may
supplement the equipment’s manual.
Before starting a test always check that the test signals are suitable for your system under test.
Only experienced and competent professionals that are trained for working in high voltage or high current
environments may implement this Application Note. Additionally, the following qualifications are required:
• Authorized to work in environments of energy generation, transmission or distribution, and familiar
with the approved operating practices in such environments.
• Familiar with the five safety rules.
• Good knowledge/proficient in working with the MPD 800 system.
A very low frequency (VLF) test voltage is commonly used to perform on-site diagnostics
such as dissipation factor or partial discharge (PD) measurement on medium-voltage
cable systems. Commercially available VLF voltage sources have the advantage of being
small and lightweight compared to most resonance test systems, as well as having the
ability to control the test voltage amplitude which is essential when it comes to
characterizing PD phenomena.
PD measurement on cable systems not only allows user to assess the integrity of
insulation, but also to localize a potential defect which is often linked to an accessory
installed on-site.
Depending on the type of defect, the test results (PRPD pattern, PD inception voltage
(PDIV) and PD extinction voltage (PDEV)) at voltages of power frequency or at very low
frequency may not be directly comparable [1, 2, 3]. Studies have shown that the
development of electrical trees strongly depend on the frequency of the test voltage.
Individual discharge events can influence the properties in the discharge branches, such
as temperature, pressure and conductivity. If the pressure cannot dissipate in between
consecutive PD events, it is likely that a new discharge channel/branch will develop. At
lower test voltage frequencies, the pressure inside a discharge branch has a higher
probability to dissipate due a typically higher time difference between consecutive pulses,
leading to fewer branches and a straighter tree structure. Figure 1 highlights that an
electrical tree at power frequency will look more bush-like compared to a tree developed
under VLF voltage stress.
Figure 1: Growth of electrical trees in polyethylene at different test voltage frequencies and durations [3]
In the default mode, the MPD Suite software can only detect AC signals with frequencies
between 5 Hz and 10 kHz. Activating the VLF mode in the VLF control panel tab extends
the frequency range down to 0.01 Hz. In this mode, the MPD Suite does not automatically
search for the AC frequency, instead the user must specify and enter the target voltage
frequency to perform an accurate and reliable measurement.
A state-of-the-art VLF voltage source allows users to either select a fixed output frequency
or use the automatic mode. The automatic mode aims for the highest output frequency for
each voltage step and automatically adjusts the frequency if the power consumption
reaches its limit. To ensure a stable PRPD pattern as well as charge and PD reading, it is
recommended to have the same test frequency throughout the entire voltage sweep.
Figure 2: VLF control panel only requires two actions. Activate the VLF mode and enter the VLF target frequency.
Based on the entered frequency, the MPD Suite starts to detect the actual applied
frequency which must be within 1% of the entered value to ensure an accurate voltage
and current measurement.
The entered frequency also has impact on the time constant which is used by the AC and
PD pre-amplifiers. A low PD repetition rate will not lead to a constantly changing gain
level, even when the fully automatic AC and PD gain mode is used.
The apparent charge defined in the IEC 60270 [5] is a weighted value, based on the
repetition rate of the recurring PD events with the highest magnitude within a certain time
frame. Table 1 shows the allowed reading R of a PD measuring system depending on the
PD repetition rates that are related to the absolute time and not normalized to a single test
voltage cycle.
N [1/s] 1 2 5 10 50 ≥100
Rmin [%] 35 55 76 85 94 95
Table 1: Pulse train response of a PD instrument according IEC 60270: 2000 + AMD1: 2015 [5]
The lower the repetition rate, the lower the corrected apparent charge value must be. As
an example, the charge value of a PD event with a repetition rate of 1 pulse per second
must be reduced to 40% of its original value. The above weighting table attenuates
disturbance pulses with repetition rates below 50/60 Hz. For pulses with repetition rates
greater to or equal to 50, the pulse is weighted with 100%. Most of the PD defects occur in
one or both test voltage halfwaves (N>=100). In other words, the authors of Table 1
assume a test voltage frequency of 50/60 Hz and do not consider VLF test voltages. The
scope of IEC 60270 includes alternating voltages up to 400 Hz and direct voltages. Direct
voltages are specifically excluded by the above table (Note 4 of Table 1 in [5]) which does
not apply for VLF.
With VLF, the AC cycle is much longer than with 50/60 Hz measurements, and thus using
the table above would lead to dramatically lower PD readings for PD events recurring at
every VLF cycle, severely limiting the usability and comparability of the reading.
The dynamic weighting feature is automatically activated when starting the VLF mode and
adjusts Table 1 for frequencies other than 50 / 60 Hz. The actual measured AC frequency
is used to derive the PD repetition rate at which down-weighting should occur. For PD that
recurs with half the measured AC frequency, the charge level will be weighted down the
same way as a 50 Hz PD signal would be without dynamic weighting. A PD signal with a
repetition rate of 10% of the AC frequency will experience the same down weighting as a
10 Hz PD signal would see with dynamic weighting disabled. As a result, the weighted
charge value does not alternate like it would when using the default QIEC weighting.
The dynamic weighting feature is a unique feature in the MPD Suite software and can be
considered as being standard compliant as the weighting of pulses for VLF frequencies is
implemented in the adaption of the logic of Table 1 in [5].
Statistical values such as Qmax over a test voltage cycle are available in the measured
values overview and can be plotted in various graphs.
The AC / DC setup settings within the SETUP & CALIBRATION tab need to fit the actual
test setup to ensure a correct voltage and frequency measurement. For most cases a
coupling capacitor in combination with the internal quadripole is used for the PD as well as
VLF voltage detection. The system automatically compensates AC phase shifts based on
the detected test frequency, as long as no third-party impedance is used. A Phase offset
caused by a third-party impedance can be identified by a corona test and compensated
with the phase offset setting. The CPL1 / 2 is not designed to measure DC or VLF signals.
It is recommended to either use the internal quadripole of the MPD 800 or an external
resistive divider to measure the applied test voltage equivalent.
Figure 4: Settings to synchronize to the applied test voltage Figure 3: Progress bars indicating when the reading will
using a 1 nF coupling capacitor be updated
500 nA is the minimum RMS signal at the PD input, that can be used to measure and
synchronize to the test voltage. This current is reached at 800 V when using a 1 nF
coupling capacitor and a default test voltage frequency of 0.1 Hz.
The system requires the test voltage to be applied for minimum duration of 3.5 times the
AC period before it displays any readings. This ensures an accurate AC reading as well as
a stable PRPD display over the entire test cycle. After the initial measurement, the
readings are updated at least every 5 seconds. To get the MPD system synchronized with
the applied test voltage frequency before performing test voltage sweep, it is
recommended to start with a voltage level where no PD activity is expected.
The MPD Suite software provides helpful information bars and user prompts to indicate
when the next VLF measurement reading is available. Figure 4 shows some examples of
the VLF-related readings / messages.
Figure 5: Left – PRPD pattern with sliding needle view of current voltage / PD stage
Right – same PRPD pattern with dithering enabled
The left part of Figure 5 shows a snapshot of the PRPD view. The events happening at the
point in time are displayed as single needles within a sliding window 100 ms. This allows
users to easily follow the level of the test voltage and it indicates whether or not there is
PD activity.
The dithering function located in the PRPD & AC tab can be used to significantly increase
the visibility of low repetitive PD events as shown in the right part of Figure 5.
Qmax, Qwtd or Qavg can be plotted in different diagrams and linked to other values, such as
test voltage Q(V) or time Q(t).
The ASSESSTMENT tab allows for settings automated detection of PDIV and PDEV or
simple pass / fail statement based on a set charge threshold(s).
Additional diagrams, such as the PD scope or the PD localization diagram, are used for
fault localization and are explained in the case study.
Figure 6: PD test setup for a sensitive PD detection under VLF voltage stress
The overall PD sensitivity in the field is often heavily influenced by external noise. The
MPD 800 with its freely-selectable digital band pass filters allows a high degree of noise
reduction. Cable length, wave-propagation characteristics, amount and type of joints,
insulation material, mix-cable systems and obviously the ambient noise, can have
significant influence on the detection sensitivity as well as localization accuracy. A detailed
list of influencing factors is described in [1].
The detection sensitivity is typically expressed by the minimum apparent charge
magnitude that a PD system can display when a defined charge signal in pC (charge
calibrator) is injected at the far end of the cable system [1]. A calibration of the propagation
speed based on the known cable length is performed using the different localization
methods highlighted in the corresponding chapter.
Figure 7: PD sensitivity check from the far end of the cable shows a minimum detectable charge level of ~ 2 pC
Based on Figure 7, one could assume that the detection sensitivity of the test setup is
approx. 3 pC, however this is only true when the used VLF voltage source does not cause
any additional noise signals, which can be more dominant than the impact of the blocking
impedance. The IEEE Guide [1] highlights that some commercially available VLF voltage
sources include a mechanical switch in the high-voltage circuit and thus are not PD free.
Figure 8: PRPD pattern at 20 kV / 0,1 Hz with 3 different VLF voltage sources; no blocking impedance used
Figure 9: PRPD pattern at 30 kV / 0,1 Hz with 2 different VLF voltage sources; no blocking impedance used
The impact of the noise signals caused by the VLF voltage source on the actual PD
reading depends on various factors, such as type of VLF voltage source, maximum test
voltage in relation to maximum device specification and measuring frequency as well as
the influence of the connected load. The examples in Figure 9 show that without the usage
of a high-voltage filter, disturbance patterns of several 100 pC or a few nC are present and
significantly reduce the sensitivity for the PD measurement.
To ensure a sensitive PD measurement, it is highly recommended, or possibly even
required, to use a proper blocking impedance. Such a filter not only limits the high
frequency noise from the VLF voltage source to be measured by the PD detector, but it
also ensures a defined recharging loop. Only currents that will be recharged by the
coupling capacitor will be measured, which will result in a higher sensitivity.
The MPD system supports four methods to perform a fault localization on a power cable
system. Highest accuracy can be achieved by a distributed PD measurement using the
time-of-flight or statistical time-of-flight methods between two or more detection points.
For onsite medium-voltage cable PD measurements, where VLF test voltages are used,
the setup typically consists of a single-end measurement. Therefore, often the basic time
domain reflectometry and the more advanced statistical time domain reflectometry are
used. Additional information about distributed PD measurements and localization can be
found in the application note [6].
4.3.2.1 Principle
Figure 10 to 13 represent the basic TDR working principle as well as the signal distribution
on a cable system with limited or no impact of signal reflection at joints, attenuation and
dispersion.
Figure 11: Signals will be reflected at impedance changes, such as the transition from cable impedance to joint impedance or at cable
terminations.
Figure 12: The PD scope shows the direct as well as reflected pulse detected at the near end of the system.
Figure 14: The type of joint can have significant impact on the complexity of the diagram.
Figure 15: PD Event trigger window used to trigger the PD scope; fault location from near end is 900 m – 251 m = 649 m
Figure 17: sTDR fault localization on a 900 m power cable; fault location is 653 +-0,95 m seen from the MPD measurement device.
Figure 18 shows the PRPD pattern of a cavity below the stress control tube of a medium-
voltage cable joint at different test voltage frequencies. All patterns are recorded over an
interval of 39 test voltage cycles. The repetition rate of the given fault strongly depends on
the VLF voltage frequency. The overview indicates that the shape of the pattern gets
distorted with decreasing output frequency but remains similar to some degree.
10 mHz, 65 min
Figure 18: Comparison of internal PD fault at different VLF test voltage frequencies
[1] IEEE Std. 400.3-2006: “IEEE Guide for Partial Discharge Testing of Shielded Power Cable Systems in
a Field Environment”
[2] E.Ildstad / F. Mauseth: “Voltage and Frequency Dependence of Electrical Treeing in XLPE Cable
Insulation”, Nordic Insulation Symposium 2009
[3] D. Pepper: “Grundlagenuntersuchung zum Teilentladungsverhalten in kunststoffisolierten
Mittelspannungskabeln bei Prüfspannungen mit variabler Frequenz und Kurvenform”
[4] IEEE Std 400.2-2004: "IEEE Guide for Field Testing of Shielded Power Cable Systems Using Very Low
Frequency (VLF)", New York, USA, 2005
[5] IEC 60270: "High-voltage test techniques – Partial discharge measurement", Version 2000, 3rd Edition
[6] OMICRON application note: Distributed partial discharge measurement and localization using GPS
synchronized datasets
© OMICRON
www.omicronenergy.com Subject to change without notice.