HV Cable Partial Discharge Testing, Online V's Offline

[fz]

We are currently looking at introducing Partial Discharge Testing and Mapping for Cables in our distribution system. Can anyone help me better understand the advantages/disadvantages of the Offline (Cable out of service) PD tests V's the Online (cable in service) type tests, in terms of quality of data output? I know there are significant cost and outage time savings with the Online methods but are they as unreliable (due to noise, etc) as some believe? Any help would be appreciated!
thx.

 

[benlanz]

Medium and high voltage cable diagnostic technology can be divided into two categories, namely online and offline. Online refers to performing a diagnostic test to a cable while it is in service. Offline refers to performing a diagnostic test while the cable is not in service and has been isolated. Both of these technologies can locate defects on shielded power cable systems. However, both have limitations and benefits which make them better suited for different applications.

Online technology has the advantage of performing tests while the cable system is still in service. There are no switching costs and outages are avoided in the case of non- looped circuits. Defects in the cable system that produce partial discharge (PD is a sign of dielectric weakness) at operating voltage are recorded as signals and resolved to determine the point of origin. Online technology is limited by the physical nature of PD. PD in cable material (e.g. XLPE) can cause a failure, in a matter of minutes to days. For this reason online testing has to be done frequently to catch near term cable defects before they fail. Cable accessories are made with materials that can resist PD much longer before failure thus providing a larger window opportunity to observe and repair the defect. To date, case studies have demonstrated there is limited merit in this technique; however, no large scale studies have been performed that can point to reliability improvements or delays in capital expenditures.

Offline testing has the disadvantage of line switching costs. Offline testing requires the use of a mobile high voltage power source which in turn provides an opportunity to perform PD tests at higher voltages. Testing at higher voltages enables offline tests to catch defects at an earlier stage of development. Thus the offline test results can predict future cable performance for a much larger window of repair opportunity (years). Offline diagnostic technology has demonstrated in a number of large scale studies to correctly identify and locate 83-94% of all cable system defects.

Comparing the two techniques poses a question. Can the cable system owner afford to switch the cable out of service? Answering this question will clearly decide the test method that is best suited for the application. If switching is not the main concern, it is highly recommended to opt for the more thorough offline test.

The following table identifies the engineering tradeoffs for decision makers.


Online
Can test w/o switching out of service
Will detect and locate
Severe accessory defects
Some near term cable defects
Test sensitivity not calibrated
6 to 7, 3-phase cables per day
Large scale correlation data not available

Offline
Cable must be switched out of service
Will detect & locate
Near/long term accessory failures
Near/long term cable failures
Replicates Calibrated Factory Test
Repair window extended
3 to 4, 3-phase cables per day
Large scale correlation data shows 83-94% correlation

-cheers!

 

[Mstrvb19]

I have also heard some arguments for online testing due to the random nature of PDs. I thought I read that the regions of breakdown in the insulation develop in a random manner, and that often de-energizing the cable and then re-energizing will result in a different spectrum which may or may not be accurate. Is this true?

 

[benlanz]

I have heard the same argument! Theory is great isn't. We use theories to explain what we don’t have data to measure. If we make a few assumptions one can come to the conclusions that the cable will act in kinds of different ways. The problem is the use of the wrong assumptions!

The problem with the theory of spectrum change is that it doesn't hold much water in large scale studies. When scientists come up with theories that differ the only thing left is to perform statistically significant experiments. And that's what was done at a large utility in the southern US. The 3 year experiment is outlined in following reference.

thread238-98186
or

http://www.eng-tips.com/viewthread.cfm?SQID=98186&SPID=238&page=1

The experiment shows that off-line tests are very effective in determining defect sites with out further damaging the cable and that defects in cable insulation (PD sites) do not go rapidly failure regardless at which voltage stress they were observed.

1) 151 cables with failure rate 10/100miles/year
2) Cables tested each year for 3 years (2000- 2004)
3) 151 cable now have failure rate of 12/100miles/year
a) 51% of cables w/no PD sites fail. rate 4/100mi/yr
b) 49% of cable w/PD sites fail. rate 21/100mi/yr

I think the result speak for themselves. What ever the theory is, this method is obviously separating the good from the bad. This same utility has dropped its failure rate, according to their own database, by over 70% after 3 years replacing and repairing cable on the basis of the test results on their 4000miles of underground! Previously their failure was sky-rocketing even with their best 2 failures and you’re out replacement rule!

I would say on the basis of this experiment that the offline PD test is making accurate measurements and that the spectrum or karma or the spirits of the cable are left undisturbed!

According to a database of 2500km of extruded cable including 960 PD site in cable insulation (not accessories) only 3% of defects show up at operating voltage. Why? Extruded cable can not tolerate continuous PD and goes to failure in a very short time. This is the reason one needs to go to higher voltage stresses to find PD sites in cable insulation that have not failed yet!

I hope this clarify things.

Cheers