5kV Bus Hipot Leakage Current 2

[oldbeau44]

An old Mechanical in need of help...

I have written a spec to cover inspection and DC hipot testing of 5kV switchger bus (excluding breakers.) 20 kV DC Hipot will only be allowed if Megger test confirms a minimum IR value of 100 Megohm at 40°C as per IEEE 43-2000.

Remaining issue I need help with is to define a limit on allowable leakage current during the hipot. From internet research it seems ths may be a bit of magic that is not well defined.

Can anyone suggest an appropriate upper limit on leakage current? Or a methd to determine a suitable max. current?

Thanks and Regards...

 

[Zogzog]

Won't find a set value, the key is to look for an "upward knee" in your leakage current during your hold period. What method are you using for the test?

 

[oldbeau44]

By 'method' I assume you mean the procedure. I have spec'd slow 2000 VDC steps with 20 sec hold at each increase, and then record leakage. Hold at max. 20kV for 5 min. and record leakage every 60 sec.

I am not aware of anyhting more specific, or an IEEE ref to apply. Your suggestions?

Thanks...

 

[rbulsara]

Look up NETA-ATS or MTS- Acceptance or Maintenance Test Specifications. ATS for new and MTS for old gear).

http://www.neta.org.

Depending on age and ambient conditions (moisture, temperature, etc) the acceptable values may vary and hence there is no fixed magic number.

Essentially it should be in several microamps and it should stabilize at some value, after rising for initial couple of min.

Just aside, I may be wrong, but 20kV sounds high for a 5 kV gear.


Rafiq Bulsara

http://www.srengineersct.com

 

[oldbeau44]

I will check NETS-MTS. The 20kV came from the ATS-1999, Table 10.2 for DC voltage. Now that you point out that ATS is for new gear, I would agree that 20kV is too high.

In "Electrical Power Equipment Maintenance and Testing" 2nd Ed. by Paul Gill, he suggests 20kV (Table 7.7) but does not indicate new or used.

He also suggests in section 2.3.2 the use of 65% of factory acceptance voltage (20kV.) So perhaps 13kV would be more appropriate.

As this hipot test is a client driven, annual requirement, I do not expect a test voltage equal to rated peak AC will be accepted.

 

[rbulsara]

If an annual requirement, the best is to compare the results with previous test results. Trending is the best tool to detect any deviations from the past. See what was done in earlier tests.

Rafiq Bulsara

http://www.srengineersct.com

 

[oldbeau44]

Agreed; trending will be be helpful. I'll include this in requirements.

Earlier ZogZog mentioned monitoring for "upward knee" in leakage current. Does this typically show up only during the max voltage hold?

While I’ve witnessed many motor & generator hipots, I have never seen a bus hipot. My concern is to prevent PD tracking damage should the leakage increase too rapidly. Am I just being overly cautious?
Thanks...

 

[oldfieldguy]

It is difficult to assign an absolute value as a no-go for the hipot testing, though. I've seen some very large differences having to do with age, number of cubicles, etc.

Try this on for size:

Do a megohmmeter test at 5 kV. Record the reading in megohms.

Using a DC hipot, apply voltage in 5kV steps to your final test voltage. (I don't have the book in front of me right now, but I seem to remember that the DC test voltage for 5 kV cals equipment is 17.5 kV.) At each step, hold the voltage for a minute. During that minute, calculate the resistance using the voltage and leakage current. If the calculated resistance is lower than your megohm reading from the megohmmeter test, don't go to the next voltage.

old field guy

 

[Zogzog]

Your procedure looks fine, but I suggest recording every 10 seconds and ploting them (Our test reports make a nice 3D graph), any increase in current at a hold point indicates the beginning of a failure or tracking so stop the test right away.

Gil's book is a good reference, but not really a standard. NETA MTS test voltages are usually lower than ATS value, so I would check that. You seem to have a handle on things though.

 

[oldfieldguy]

Good morning!

I checked NETA Maintenance Test Specification 2005 and the DC withstand voltage for 4.76 kV gear is 20 kV.

Zog--

I forgot to put a "if the current starts climbing during any stage of the test, shut the voltage down" statement in my little blurb. And if you have a test matrix set up in electronic form, such as a spreadsheet, where calculations of resistance based on a leakage current entry are instantaneously done, then fifteen seconds is good. I used to do this in the era before such things were common in the field.

old field guy

 

[Zogzog]

I checked the 2009 MTS this morning too, max of 20kVDC.

OFG, I attached an older spreadsheet I made (new power DB stuff we use does this all automatically), it paints a nice clear picture of what is going on.

 

[oldbeau44]

Gentlemen,
Thank you. Very helpful comments and data gathering techniques. If I may, a few final questions…

1. Failure Criteria: Is it fair to say the bus would fail the Hipot if the leakage current increases “exponentially” during any steady state voltage application? i.e. It indicates an ever increasing leakage; or the “upward knee” mentioned earlier.

2. As to the maximum DC voltage to be applied… This switchgear is 10 years old, in an air conditioned space on an offshore drilling rig, and will be inspected, cleaned and meg’d prior to the Hipot. After further consideration, I do not believe it wise to test at 20kV.
Do you feel 15kV is sufficient to prove, without risk of over stressing the insulation?

Best Regards...

 

[Zogzog]

1. Yes
2. The 20kV DC is the maximum recommend test voltage, the owner of the gear can choose any test voltage they want. I think 15kV will be fine.

 

[oldfieldguy]

I concur with Zog.

Further, if the test set operator notes the current starting to increase without the voltage being increased, he should zero the voltage, terminating the test before actual breakdown of insulation occurs.

old field guy

 

[oldbeau44]

Thanks oldfieldguy,
Adding "...without the voltage being increased" will help refine the cautionary note I've included.

Many thanks to all for your excellant advice. A very valuable forum which I really appreciate.

Just wish I'd found you a month ago!

 

[VLFit]

Your question illustrates the very reason why DC testing is not optimium. There are no, and can't be, standard leakage currents since they depend on so many variables. The same bus can be tested on different days and get different results. It realy should be an AC test application. The bus either withstands the test voltage or flashes over. There is nothing in between, no guess work or interpretation. It's good or bad with an AC test. If it holds the voltage, all is well. The DC leakage current is irrelevant. The same goes for many other loads, like cables, vacuum bottles, breakers, rubber gloves, etc.

 

[oldbeau44]

Thanks VLFit for your added comments. You raise an excellent point.

I had originally written this test procedure using AC for the same reasons you note. But others here with more field experience, advised the test gear is very large, and not always available in some of the Asian shipyards where these tests will be made. A large test unit might be difficult to get into the rather tight space available on the ship. Perhaps you would educate me further on the AC test & equipment.

1. If the modern test gear is actually of a reasonable size, and the power required is not too high, I would favor changing the Hipot to AC.

2. For this type MV metal-clad, non-segregated bus installation would you agree with the NETA recommended 14kV test voltage? Or lower, in light of the equipment's age?

3. And lastly, I understand that VLF AC testing reduces the power required. If that is more suitable to our needs, should I require a minimum frequency?

 

[VLFit]

Power frequency AC test equipment can get large if the capacitance of the load is high. It is not in a 5kV SG. You can use a 30kVac @ 2 - 3 kVA. very portable and inexpensive.

14kVac is the proepr voltage, regardless of age.

VLF hipots are used for cable testing mostly. A long cable can have 1, 2, 3uF of capactiance. To test at 60Hz would require a multiton test set: fine for the factory but not for the field. You don't need the extra cost and specialty of a VLF hipot for you applicaiton. Also, there are IEEE VLF standards written for cable and motor/gernearator testing only.

Good luck