The group met on Wednesday, April 20, 2005 at 10:15 AM with 38 members and guests.

Participants	Affiliation
Densley, J.	ArborLec Solutions
Petzold, F.	SEBA KMT
Schlesinger, R.	R. Schlesinger Consulting
Smalley, M.	We Energies
Koch, F.	PSE
Landers, G.	Von Corp.
Lanz, B.	IMCORP
Levine, J.	Kinectrics
Bernstein, B.	B. Bernstein Consulting
Russwurm, D.	Specialty Instrument Service
Fryszczyn, B.	CTL
Goodwin, C.	HV Diagnostics
Ainscough, J.	Xcel Energy
Hinkle, J.	PPL Electric Utilities
McDermid, W.	Manitoba Hydro
Von Hermann, M.	The Von Corp.
Mashikian, M.	IMCORP
Grodzinski, C.	EHV Power
Parker, N.	Puget Sound Energy
Sikora, D.	Duquesne Light
Hunt, J.	Salt River Project
Zemyn, N.	PPL Electric Utilities
Lindler, C.	Pirelli
Brown, K.	TVA
Walton, M.	General Cable
Rickmann, J.	Phenix Technologies
DiGuglielmo, F.	PSE&G
Patterson, R.	Intl. Electrical Testing
Buckweitz, M.	General Cable
Jasek, M.	PNM
Hans, J.	Exelon-ComEd
Hinkson, J.	Westar Energy
Malia, M.	T&B Elastimold
Cunningham, R.	Robert Cunningham
Graham, S.	Duke Power
Bush, R.	T&D Magazine
Merando, J.	Bechtel Corp.
Jaroslavsky, V.	CTL
Kopchick, T.	First Energy

The chairman stated that the guide IEEE 400.2 was available from IEEE at the end of March 2005.  Working Group members will receive an electronic version from IEEE.  As the group does not have a PAR, it is considered a Discussion Group.

The purpose of the Discussion Group is to gather data on test voltages and durations, practices, and results to determine if a revision of the Guide will be needed.  In particular the values of the test voltages for sinusoidal and cosine rectangular waveshapes need to be clarified, for example, should the rms or peak values of the two waveshapes be equivalent as in Tables 4 and 5 of the present version of the guide.

There were three presentations:

1.             "High Potential Testing " Voltage Withstand vs. Voltage Step at 0.1 Hz, Effect of rms vs. peak Test Voltage- by Frank Petzold.

2.             "MV Cable Testing with VLF at TVA- Nuclear" by Kent Brown (see Appendix C18D-2).

3.             "Summary of Test Data at Exelon-ComEd" by John Hans (see Appendix C18D-3)

In his presentation Frank Petzold stated that the purpose of withstand testing is to detect and eliminate local degradation and weak spots in the cable insulation while the purpose of step testing was to determine the global or integral aging condition of the cable insulation.  There was considerable discussion on this issue without any definite resolution.  Data was presented that showed that the electrical tree growth rates from needle electrodes in XLPE cables were similar when the rms values of the cosine-rectangular and sinusoidal waveforms were the same but that the cosine-rectangular waveform showed a faster tree growth rate if the peak values are the same. Data from over 15 years of testing showed that 97% of XLPE cables that passed a 3Vo withstand test for one hour are reliable for more than 2 years without cable failure.  The overheads used in the presentation are shown in Appendix C18D-1. Dr. Petzold was asked if he could provide more data about the numbers and types of cables tested.

Kent Brown reviewed the characteristics of the medium voltage cables operating in some nuclear stations.  Some of these cables are between 25 and 35 years old.  A failure of an XLPE cable in 2002 revealed significant water treeing which led to a VLF test program.  Both withstand (up to 3Vo according to an earlier version of 400.2) and VLF tan delta tests (at 0.5, 1.0, 1.5 and 2.0Vo) were carried out.  Two additional failures occurred in the withstand testing.  A conclusion from the VLF tan delta tests was that the results showed good correlation with service history (not all cables are continuously energized and some cable lengths are known to operate under dry conditions.  Some of the cables that failed the VLF tan delta test were removed and subjected to ACBD tests.  The data is presented Appendix C18D-2.  K. Brown raised several important issues regarding VLF testing.  These were:

• Acceptance criteria for EPR cables (VLF tan delta criteria will not the same as that for XLPE cables).
• Variation of VLF tan delta with time of voltage application
• VLF tan delta of some types of accessories can mimic degraded cable behavior
• In the published version of 400.2 the withstand test voltage is about 2Vo, the same value as the maximum VLF tan delta value.

John Hans presented a brief summary of VLF test data obtained at ComEd using a sinusoidal waveform.  299 withstand tests have been carried out, about 75% using a test voltage of 3Vo for 15 minutes, and about 25% using a test voltage of 2.2 Vo (both test conditions were in different drafts of the 400.2 guide).  Of the cables tested at 3Vo for 15 minutes, 49% failed during the test and a further 7% of the surviving cables have failed in service.  Of the cables tested at 2.2Vo for 30 minutes, 40% failed during the test and a further 8.7% of the surviving cables have failed in service.  About 95% of the cables that failed were rated at 15 kV.  The number of failures of each type of cable insulation was not given.  The data are presented in Appendix C18D-3.

A template of a survey form to collect data from utilities, originally prepared by R. Schlesinger, was circulated to some members of the group for comments.  Fred Koch made some changes and the revised version was circulated to members for comments.  Attendees were asked to send their comments to the Chair and Vice-Chair by the end of April so that the information can be compiled and distributed to WG members for further comments

 

The group met on Wednesday, November 2, 2005 at 10:15 AM with 35 members and guests.

The two scheduled presentations, by John Hans and Craig Goodwin, could not be given as the authors could not attend the meeting.

The following items were discussed:

Test voltage and duration:  One of the most critical areas to be resolved in the revision of IEEE 400.2 is to make clear to users the magnitude of test voltages and the duration of the tests.

The most experience with the application of VLF testing for power shielded cables comes from Europe (primarily Germany, Austria and the Netherlands) where normal test voltages were 3xU_o (rms) and test duration was 1 hour.  These parameters are not easily accepted in the US, which has lead to the Guide stating a voltage range from 2 to3 times U_o  depending on the voltage rating of the cable under test and a test duration between15 minutes and 60 minutes, with a recommended time of 30 minutes.

The issue is to determine a test voltage and time duration that will be severe enough to fail marginal cables without doing harm to good ones and select a time that is (economically) effective for the user.

A questionnaire was prepared since the last meeting to gather information from user of VLF, but after further review of it, it was decided that it required more time effort from the user than they would probably provide, and would result in minimal feedback.  It was decided that the questionnaire would be sent to utilities but would be used as a basis to collect information by personally contacting users.  The user names and companies would be obtained from a list provided by the VLF manufacturers.

S. Wright referenced a paper he presented at the T&D Conference in Chicago in 1994 covering cable performance after testing.  He will make that paper available to John Densley.

It was suggested that we post on our website a "VLF Peer to Peer User Group" so users could discuss issues and problems and issues they were encountering.

M. Baur (who was not in attendance) sent some details of a German utility's with VLF testing.  The testing was done at 1.7 U_o because there were some problems with the integrity of the cables.  The data clearly showed that the number of service failures decreased by more than 50% after VLF testing was introduced.  About 68% of the failures occurred in the first 10 minutes of testing and 83% within the first 30 minutes of testing.  Depending on the duration of the VLF test on a circuit, no additional VLF testing would be carried out on that circuit for periods up to two years (up to 1 year after a 15 minute VLF test, between 1 and 2 years after a 30 minute test, and at least 2 years after a 60 minute test.

Mr. Moh from Malaysia, also not present, offered some written comments.  He has performed a large number VLF tests and his present test voltage level is 2.2 Uo for one hour.  His failure statistics are:

                67% failed within 12 months

                21.5% failed in13-30 months

                5.9% failed within 31-45 months

                4.9% failed within 46-60 months

There was discussion regarding the question of the test duration for a failed cable circuit on retest after repair.  Should it be tested for the remainder of the time duration, or re-started at time zero?  IEEE 400.2 specifies that the test should be repeated at time equal to zero.

VLF testing of EPR cables: M. Baur noted that he was unaware of any VLF tests carried out on EPR in Europe.  There appears to be little information regarding VLF testing of EPR cables.  Information will be sought when utilities are contacted about their VLF test data.

VLF diagnostic tests

An example of diagnostic testing was discussed, which included:

                TDR testing of the neutral to determine the extent of the corrosion.

                Dissipation Factor (0.1 Hz or 60 Hz.)

                PD testing

This was thought to be a comprehensive and a suggested set of tests to be recommended.

There is a need to collect data on VLF diagnostic testing (dissipation factor and PD) to establish test levels and criteria to assess the condition of cable circuits.

Action Items:

• Contact VLF test equipment manufacturers to see if they will provide names and contacts of utilities using VLF test equipment.
• Contact users and obtain information on test procedures and failure statistics using the questionnaire as a guide.
• Compile this information for discussion at our next meeting
• Investigate setting up a "VLF Peer-to-Peer User's Group" on the ICC web site.