CIRED2017 - 0049 - Final OFFLINE PD DIAGNOSTICS USING SEVERAL EXCITATION VOLTAGES
CIRED2017 - 0049 - Final OFFLINE PD DIAGNOSTICS USING SEVERAL EXCITATION VOLTAGES
CIRED2017 - 0049 - Final OFFLINE PD DIAGNOSTICS USING SEVERAL EXCITATION VOLTAGES
Paper 0049
Paper 0049
Paper 0049
Figure 5: PD mapping comparing 3 different wave shapes at U0 and 1.7Uo. Top VLF Sinus 0.1Hz, middle DAC and bottom VLF CR 0.1Hz/
50Hz slope technology (x-axes cable length, y-axes PD level).
In addition for both PD Level and PD intensity only small Case Study 2:
differences have been observed at Uo and 1.7Uo. The second practical case study is a measurement
Looking to the Partial Discharge Inception Voltage performed on pure 12/20 kV service-aged XLPE cable.
(PDIV) of all excitation voltages, this was exactly the As of failing information the number and position of
same for all wave shapes. However some major joints was unknown. The cable has a length of 1202m
differences could be observed between the VLF Sinus and is composed out of three single cores, see Figure 6.
0.1Hz excitation voltage and DAC/ VLF CR 0.1Hz
excitation voltage. At operating voltage only 1 PD weak
spot have been identified with VLF Sinus 0.1Hz whereas Figure 6: Schematic display of the 1202m long XLPE cable.
2 weak spots with the other excitation voltages. In
addition the PD intensity is far lower at VLF Sinus Figure 7 shows the individual PD mappings of the PD
compared to the other wave shapes. This taken into diagnostic measurements performed with VLF Sinus
account that the time of data capturing has been equal. 0.1Hz, DAC and VLF CR 0.1Hz. On the left hand side
Only at elevated voltages VLF Sinus 0.1 Hz was able to the PD mappings at Uo are shown, on the right hand side
detect the second weak spot inside the cable. PD intensity the PD mappings at maximum test voltage 1.7Uo are
of the first weak spot hardly increased with increasing shown. The frequency of the DAC excitation voltage and
test voltage for VLF Sinus 0.1Hz. With the VLF CR the frequency of the polarity reversal from the VLF CR
0.1Hz and DAC voltage it did change more. PD level of 0.1Hz excitation voltage was 294Hz.
VLF Sinus 0.1Hz has also been lower compared to DAC
and VLF CR 0.1Hz.
Paper 0049
Figure 7: PD mapping comparing 3 different wave shapes at U0 and 1.7Uo. Top VLF Sinus 0.1Hz, middle DAC and bottom VLF CR 0.1Hz/
50Hz slope technology (x-axes cable length, y-axes PD level).
Comparing the PD mapping at Uo measured with VLF weak spots have been detected within the cable
Sinus 0.1Hz with that from DAC or VLF CR 0.1Hz insulation/ joints using VLF CR 0.1Hz and just 4 PD
major differences can be recognized. With VLF Sinus weak spots using VLF Sinus 0.1Hz. Three out of the four
0.1Hz excitation voltage there was no PD measurable localized PD weak spots identified with VLF Sinus 0.1Hz
whereas with either DAC or VLF CR 0.1Hz two weak are matching with the DAC and VLF CR 0.1Hz voltage.
spots were identified. One weak spot in Phase L2 at One PD weak spot positioned in Phase L1 at 993m,
position 155m and one in Phase L1 at 649m. The partial which incepted at 1.3Uo, has only been recognized using
discharge inception voltages measured with DAC and VLF Sinus 0.1Hz. On the other hand four PD weak spots
VLF CR 0.1Hz of both weak spots are identical and far identified with DAC and VLF CR 0.1Hz have not been
below operating voltage. See Table 1 for the comparison identified with VLF Sinus 0.1Hz.
of all excitation voltages. With VLF Sinus 0.1Hz
excitation voltage even up to maximum test voltage TABLE 1: PD INCEPTION VOLTAGE AND FAULT POSITION OF
1.7Uo no PD has been detected in both of these two INDIVIDUAL EXCITATION VOLTAGES.
localized weak spots which had a PDIV below Uo.
L1 L2 L3
Looking to the PD mappings at maximum test voltage 1.3Uo 1.3Uo 1.2Uo
1.7Uo further differences can be recognized. DAC and VLF Sinus 0.1Hz
(993m) (394m) (394m)
VLF CR 0.1Hz PD mapping are still both comparable
with only some minor differences in PD level and 0.7Uo 0.6Uo 1.1Uo
DAC
intensity. The higher PD level and intensity are with VLF (649m) (155m) (0m)
CR 0.1Hz excitation voltage and is most probably caused 0.7Uo 0.6Uo 1.1Uo
by the 5 second DC stage at each period. In total 7 PD VLF CR 0.1Hz
(649m) (155m) (0m)
Paper 0049
SUMMARY REFERENCES
Summarizing both case studies it can be recognized, like [1] F.H. Kreuger, “Industrial High DC Voltage”,
other studies, that both DAC and VLF CR 0.1Hz show Delft University Press, 1995
comparable results, either related to PDIV or fault
positions of the identified weak spots. VLF CR 0.1Hz [2] IEEE 400-2001 “IEEE Guide for Field Testing
PD measurements results differ from the DAC results and Evaluation of the Insulation of Shielded
only in PD intensity and level, which is most probably Power Cable Systems” IEEE guide
caused by the 5 second DC stage in each half-period. A
higher PD level with VLF CR 0.1Hz compared with [3] G. Voigt, “VLF-TE Messungen an
50Hz power frequency has also been recognized in a betriebsgealterten Mittelspannungskabel”,
study from Elben [7]. Abschlussbericht FH Konstanz, Germany 2002.
Considering the results gathered with VLF Sinus 0.1Hz
compared to excitation with DAC and VLF CR 0.1Hz [4] N. Jäverberg, H. Edin, “Applied Voltage
than it can be concluded that there can be major Frequency Dependence of Partial Discharges in
differences. In most cases the PDIV at VLF Sinus 0.1Hz Electrical Trees” Proc. IR-EE-ETK, Stockholm,
was higher, which is similar reported by other studies [8]. Sweden, 2009.
In case study two although the PDIV of one localised
defect was far lower than DAC and VLF CR 0.1Hz, also [5] D. Götz, F. Petzold, H.T. Putter,
this has been seen in other research studies. This was “Zustandsbestimmung und Qualitätskontrolle
mainly observed on cold shrink accessories and may be von Montage an Mittelspannungskabeln unter
related to the physical function of refractive stress dem Aspekt zunehmend großer Kabellängen“,
grading [9]. In the last case study far more defects have VDE ETG Fachtagung, Fulda, Germany, 2012
been recognized by DAC and VLF CR 0.1Hz compared
to VLF Sinus 0.1Hz excitation voltage. Overall it can be [6] F. Petzold, H.T. Putter and D. Götz,
concluded that PD measurements performed with VLF "Combination of VLF and Resonance Principle
Sinus 0.1Hz excitation voltage show lower PD levels. for Withstand Testing of long Cable Length",
unpublished, presented at IEEE / PES ICC
CONCLUSIONS Meeting Sub F, Pittsburgh, USA, May 2013.2.
PD measurements as part of a commissioning test or as
[7] A. Elben, W. Kalkner, R. Plath, „Alternative
part of condition monitoring are proven to be useful in
Prüftechniken für lange AC-
identifying bad quality of workmanship or ageing type of
Hochspannungskabel Abschlussbericht“, TU
related defects. For offline PD measurements on MV
Berlin, Germany, 2015
cables 3 types of excitation voltages are established, VLF
Sinus 0.1Hz, DAC and VLF CR 0.1Hz. Several research
[8] H.L. Halvorson, “Condition Assessment of
studies have shown that PD measurements performed
windfarm MV cable joints ” NTNU Trondheim,
with DAC and VLF CR 0.1Hz voltages show comparable
Norway, 2012
results, like seen in these both case studies as well. The
PD results obtained with VLF Sinus 0.1Hz however
[9] F. Mauseth, “PD and Diel. Response
differ with that from DAC and VLF CR 0.1Hz excitation
measurements service aged cable joints” CIGRE
voltage. It can be concluded that in most cases the PDIV
D1.48 Meeting, Paris, France, 2014
is higher at VLF Sinus 0.1Hz compared with the PDIV
from DAC or VLF CR. Moreover the PD intensity and
level are lower at VLF Sinus 0.1Hz, using exactly the
same data capturing time for VLF Sinus 0.1Hz and VLF
CR 0.1Hz. Although there has already been a lot of
research performed in this field, further research
specifically regarding different type of PD defects is
recommended. This to confirm which defects incept with
VLF Sinus 0.1Hz excitation voltage at lower voltages and
which at higher voltages compared to operating
frequency. It is known that interfacial discharges incept at
a higher voltage at VLF Sinus 0.1Hz compared to 50Hz
or DAC/ VLF CR 0.1Hz, this as interfacial discharges are
mainly depending on the voltage slope dU/dt.