Minutes
Chair, Jody Levine, (HydroOne), opened the meeting at 8:00 a.m. on Tuesday, March 23rd, with 54 ICC members and guests in attendance. One presentation was made.
“Application of On-line PD Technology to Transmission and Distribution Systems Asset Management” by Luigi Testa - TechImp
The presentation is included in the Appendices as Appendix F10.
Following the presentation was a brief question and answer session. John Cooper (Consultant) asked a question about the time-frequency mapping using in the Techimp technique. Steve Boggs (Institute of Material Science – UCONN) made the comment that he believed the time-frequency mapping gave you little more than the distance to the PD source inside the cable. He felt it was good for most systems but for cable he had his doubts. Jody Levine made the comment that the time-frequency mapping was most useful for separating noise coming from outside the cable. Sacha Markalous (Doble Lemke) asked about the bandwidth of the Techimp PD detector. Luigi Testa said the bandwidth was from 16 kHz to 35 MHz. He also stated that it was important to have sensor bandwidths that matched the bandwidth of the PD detector. Sacha Markalous asked what bandwidth was needed for the time-frequency mapping technique. Luigi Testa stated that at least 5 MHz was needed. Narrower bandwidths tend to cluster the pulse groups on the PD map. Ed Gulski (Delft University) asked about ultra high frequency PD detection. Luigi Testa stated that when employing ultra high frequency sensors, a frequency shifter device is needed. Ed Gulski stated that measuring PD in a cable is a lot different that measuring PD in a GIS termination and he asked Luigi how he performed these two types of measurements. Luigi Testa stated that he uses an ultra high frequency sensor on GIS terminations and a different type of sensor on cables. Sacha Markalous asked how the PD signal is identified in noise. Luigi Testa indicated that he used analog filtering. Steve Boggs made the comment that with a noise level of 2.5 mV RMS, a 1 to 1.5 mV RMS PD signal is detectable with matched filters. Ray Awad (Hydro Quebec) asked what length of cables could be tested. Luigi Testa indicated a cable length of over 20 km could be tested. Ray asked if any precision of the PD data was lost during wireless transmission. Luigi stated that digital data acquisition techniques was used and the quality of the information remained intact, but the amount of data to be transmitted had to be considered. John Cooper asked if it was typical to simultaneously monitor from multiple locations for off-line PD measurements. Luigi stated it was more typical to sequentially move the setup from one location to another during off-line PD measurements.
Following the question and answer session, Jody Levine led the discussion by reviewing the comments of the presenters from the last two meetings in the following table which summarizes the key issues on sensitivity determination:
Going forward, Jody encouraged others to think about the characteristics of PD sensors to be used in field PD measurements. The consensus from the discussion was that some sort of pulse injection is still needed for PD sensitivity assessments in the field. What should the pulse look like? Steve Boggs agreed to prepare a document that summarized his position on these questions and a scenario for how it could be used. This document will be circulated for comments prior to the next ICC meeting.
The meeting was adjourned at 10:00 a.m.
The minutes were prepared by Mark Walton.
Agenda: This meeting was convened to discuss and prioritize the wish list for the next revision of the IEEE 400.3 Guide for Partial Discharge Testing of Shielded Power Cable Systems in a Field Environment.
Minutes
Chair, Jody Levine, (HydroOne), opened the meeting at 8:00 a.m. on Tuesday, October 19th, with 46 ICC members and guests in attendance. Mark Walton (General Cable), the Vice-Chair, was unable to attend the meeting and sent his apologies. John Densley agreed to take notes.
Jody opened the meeting by reviewing the minutes of the Spring meeting held in Nashville. She focused on the Table that summarized the comments on sensitivity determination of ultrawide band (UWB) PD measurement systems, the main topic of the meeting. She concluded by saying that “something different needed to be done to “calibrate” UWB systems”. One presentation was made.
“Practical Experience in PD Sensing of Power Cable Accessories in the VHF/UHF Frequency Range”, by Eberhard Lemke (Doble Lemke) see Appendix F-6
Eberhard stated that PD measurements on cables in IEC standards 60270, 60885-3, 60840 and 62067 cover “conventional/classical” techniques with narrow or wideband measurements up to 1 MHz. The apparent charge magnitude measured by these techniques is a time integral of the current. On-site classical PD measurements were not recommended for field measurements as they did not have sufficient sensitivity to detect PD that could cause breakdown Actual PD pulse shapes have frequencies extending into the gigahertz range (VHF/UHF) so that UWB sensors have to be used to measure the pulse shape or to obtain the frequency content of the pulses. These measurements are covered in the new IEC 62478, presently in preparation.
Eberhard also stated that there was no correlation between the conventional pC measurement using the classical method and the mV measurement using the UWB approach. However, cable users were used to PD measurements given in pC and insisted that UWB test results be also given in pC. This has resulted in the development of sensitivity analyses in the laboratory where classical and UWB techniques are used simultaneously to measure PD at an artificial defect to establish an approximate scale factor between pC and mV. Examples showed that this scale factor is not constant because the pulse shape of the actual PD varies according to the type of defect and also with time for the same defect. Eberhard concluded his presentation with several case studies of VHF/UHF measurements in the field, e.g., using directional couplers and installing VHF/UHF sensors at GIS terminations. He pointed out that the measurements are sensitive to sensor type and position in the circuit.
Steve Boggs (Institute of Material Science – UConn) commented that (1) the apparent charge has to vary considerably in a termination due to its shape and thus has little meaning, and (2) the spectrum measured does not represent the spectrum of the PD but is a signal depending on the bandwidth of the swept filter.
Ben Lanz (IMCORP) asked if it was more important to have measurement of the sensitivity threshold or to know the actual charge. Eberhard Lemke replied that it is possible to find threshold value under specific conditions and that one can find high (requires immediate attention) and low (inspect after ~ 6 months) values. Steve Boggs commented that one has to relate what is being measured to the test object. Ben Lanz expressed concern about how the position of the sensor can affect the results. Eberhard said that this was the case for laboratory tests where the sensor position changed the scale factor but in the field the positioning of the sensors is limited. A comment was made that it is the energy that causes failure and not the apparent charge magnitude.
Ed Gulski (onsite.hv.solutions) asked how the pulse detection sensitivity issue will be addressed in the revision of 400.3 and stated that background level was important.
Jody stated that she wanted to have a discussion on apparent discharge magnitude, Qa, as it relates to UWB, including UHF, measurements. Usually relating Qa to UWB/UHF measurements was done to pacify customers but was not really applicable. When a PD pulse is generated in a cable, the pulse shape will change as it is transmitted along the cable in both directions, and the accessories as well as the detection circuit will also modify the pulse. Thus the measured pulse could be significantly different to the actual PD pulse As an example, Steve Boggs has calculated the PD pulse bandwidth and PD pulse width as a function of distance propagated along a cable, see Appendix F-7. Jody pointed out that it is not possible to calibrate a UWB/UHF PD detection system in an actual cable system as it is not possible to inject fast risetime pulses into the cable. However, a sensitivity check can be done in the laboratory where fast rise-time pulses can be injected directly into the cable keeping the connecting leads as short as possible as shown in Appendix F-8.
Jody then outlined what needed to be done but Ed Gulski stated that this work had already done by CIGRE WG D1.28 and will be reported in the F10D meeting on Wednesday, October 20 2010, see Appendix F-10. CIGRE Technical Brochures on the topic are also about to be published. He pointed out that the sensitivity check is used to characterize the PD measuring system and to make sure that it is working in the field. Steve Boggs cautioned that the characterization of medium voltage and transmission cables will not be the same due to their different constructions. It was pointed out that it would be necessary to look at different cables, e.g., taped shield construction to determine what pulses can be transmitted. been
Ed Gulski promised to send Jody details of the relevant CIGRE Technical Brochures.
It was noted that there also needs to be a discussion on free space PD detection.
The meeting was adjourned at 09:55 a.m.
The minutes were prepared by John Densley.
