Protective Relay Testing - Insurance requirement 2

[gokulkrish2]

Hi Everyone,

I am working for a mining company and barely starting. We have quiet a handful of substation and a decent number of SEL microprocessor relays. Our insurer has advised that we have to do testing on those relays every 3 years.

They recommend us for

1. Insulation Resistance Testing and
2. Secondary current Injection testing to check the functions.

Now my question is

1. What is insulation resistance testing. Is that an insurance criteria for relays across the industry?

2. I am not a fan of bench testing microprocessor relays. I may be wrong but doesnt the microproccessor relays have more reliabity than 3 years? And what are we testing?

Please let me know wat is the industry standard for testing relays......

Thanks

Gokul

 

[Zogzog]

Check the ANSI/NETA testing specs for relay testing specs.

http://www.netaworld.org

 

[cranky108]

With self-checking relays we monitor the alarm contacts, and test inputs and outputs (parts that are not self tested) every 6 years.
We also don't bench test relays. Microprocessor relays are tested through test switch's in the panels.
If you remove a relay from a panel, you can create a problem with the connections (and there by should prove your connections).

Insulation resistance sounds like a hipot test. DON'T HIPOT MICROPROCESSOR RELAYS.

 

[slavag]

Hi.
Intresting treads on the topic:
thread238-182620
thread238-211068
thread238-209938
thread238-184847

Best Regards.
Slava

 

[gokulkrish2]

Sorry for getting back too late... Was in a long vacation.

I did go through the links provided by salvag, but it didnt discuss the anything about insurance requirements.

I want to argue with the insurance people on why they mandate these testing to be done.

Also i have a question to cranky108

"With self-checking relays we monitor the alarm contacts, and test inputs and outputs (parts that are not self tested) every 6 years."

What do you mean by Self Checking? I have heard people say microprocessor relays have self testing mechanisms.. How does that work?

Thanks again for all you to have taken time to go through this discuss...

gokul

 

[marks1080]

I am guessing you didnt install Flexi test switches for analog inputs to your relays. That unfortunate, because if you had there would be no need to remove the relay to test it.

The relay should be tested at a regular interval IMHO.

marks1080

 

[marks1080]

Perhaps instead of just testing the relay itself you should function test the entire scheme. If you have redundent schemes it would be very easy and not only would a complete function test prove your relay to be in good working order but it would also prove all wiring and other devices are working properly as well.

 

[TurbineGen]

The insulation resistance test I think would be a Megger test. Just a DC voltage test to insure that the insulation is intact. You'll want to diconnect your relays for this test.

------------------------------------------------------------------------
If it is broken, fix it. If it isn't broken, I'll soon fix that.

 

[Eleceng01]

gokulkrish2, I just recently read a paper from a NERC (North American Electric Reliability Corporation) member that addressed this topic (for the utilities), but you may be able to use it to make your argument with the insurance people.

NERC's recommendation is the similar to cranky108's and should be published soon in IEEE Std. C37.233.

 

[cranky108]

Simply, a microprocessor device can test all the digital circuits, by compairing memory data, with that of an identical memory data. The microprocessor should hit a location, called a watchdog circuit, at certion intervals.
Thus it would be difficult for you to test the SEL relay the way it tests itself.
What the relay can't test is the analog input filters, the optisolators, and the output relays.

So ask yourself, what are you testing for? To satify the insurance company, or are you testing to see that it works correctly.

You can test the untested parts, or test that the relay, and logic operate as you expect. But if you are testing to see if the relay responds to the curve in the book, you are wasting your time (my thinking).

If you look at some of the NERC interpations, you will find they accept the self testing argument, and accept the testing of inputs and outputs.

As mentioned above, if you don't have test switches, or redundit relays, you have a reliability/testing issue. Test switches and redundit relays are uncommon in industrial plants, but most utilities insist in them. This is a matter of viewpoint from the construction perspective. But you have to live with it, so use this as a learning oppertinity.

 

[gokulkrish2]

Please forgive for my ignorance. I have not been in the field and i am barely seeing things.

I see rectangular black boxes, All ABB's below all the relay in the panel. I assume that those are the test switches. If they are, then we do have the test switches.

So according to you guys, i jus have to test those inputs and output relays?

Do you guys recommend any good testing company which does these things accurately and cost effectively??

Also, marks1080....

What does IMHO stand for?? &

What is function testing of a relay mean? How are they different from the conventional test?? How are they done?? I see that you say it has a lot of advantages which is good. but do you know if this test is a must during commission?? Can we rely on commission data and take it granted that this would have been done during commissioning..

I really appriciate the quick and appropriate answers you guys are providing me.... They are absolutely helpful and i am learning a lot.

Thanks once again.

gokul

 

[davidbeach]

IMHO = In My Humble Opinion

Go to

https://www.netforumondemand.com/eWeb/DynamicPage.aspx?Site=NETA&WebCode=OrgSearch

and find a NETA member organization to help you with your testing.

 

[marks1080]

Those ABB block might in fact be test switches. But without seeing for myself its hard to say. If they are you wouldnt test those switches they are used to inject (either voltage and/or current) into the relay. The beauty of test switches is they allow you to 'break' into the circuit with out affecting downstream devices (typically CTs and PTs). You need to make absolutly sure that these are test switches when dealing with CT circuits. If you open a CT circuit your boss will very likely get mad at you, and probably fire you. A current test switch will have a make-before-break shorting device that ensures that the CT won't be open circuited after you open the switch.

IMHO - in my humble opinion.

A full barage of tests should have been done on any relay and associated circuit during commissioning. I am not in any position to say whether the commissioning done on your relays is enough or not.

There are many different philosophies when it comes to testing/maintenance of protective relays. Depending on your set up some may be more favorable than others. To test a relay you need to have an outage on the associated protection and possible the element it is protecting. If you have two redundant protective systems than it is easy to have an outage on one, as the redundant system will provide protection and the element can remain in service. If you do not have redundant protections than the element will have to come out of service. You can leave the element in service while testing the protections but nobody who has half a clue about this stuff would ever recommend this, you would create a potential for sever equipment damange and a huge safety risk.

I have not read the latest NERC document recommending 6 year testing cycles, but I'm sure I will soon as I am in a NERC zone. I can't offer an opinion on the testing cycle for your industry, but I would think six years to be much too long an interval for high voltage protections.

Most micropro relays now do have self checking elements. I do not trust them 100%. I have seen weird things happen in some of these relays that are sometimes unexplainable.

You also do not want to limit yourself to testing just the relay alone. You want to test the entire protection scheme to ensure the system as a whole works properly. You may test a relay to be in good working condition, but if the cables connecting its inputs/outputs have been damaged than it doesn't matter anyway. If a certian operation of the relay is going to open a breaker, make sure the breaker opens. If an input is supposed to read the position of an element, trigger that element and make sure the relay sees its position change. This is function testing.

When testing a micropro relay focus on logic checking it. You should have a logic diagram of some sort that outlines all the functions programmed in that relay. Remember, these are more or less versatile computers. You may have two identicle relays but they could be programmed to do different things.

Also remember things can change over time with these relays. We bought a whole batch of SEL relays a while ago that all started failing at once. Finally the manufacturer admitted that batch of relays had bad memory installed in them.

 

[cranky108]

Good point about testing the whole scheme, and NERC apperently agrees.

It is possible the have rodents, or other things damage cables, and create circuits you never expected.

It also never hurts to exercise breakers, as some can become stuck.

 

[SHR1969]

1. Do not do an insulation resistance (hy pot) test on microprocessor based relays.
2. What type of SEL relays do you have? I assume they are distribution and motor protection type relays. To take the mystry and misery out of testing the SEL relays you might want to look into a fully automated test system that downloads the settings from the relay automatically and then tests the relay to the settings. You can find one at megger.com look under relay test equipment.

 

[cranky108]

How do you expect to test a self testing relay in a way that is not already self tested? (no answer expected)

Yes you need to test the inputs and outputs, but do you really need to test the settings?

I could see you testing the settings upon commissing, and settings changes. But just to verify your expetations of the relay settings operating as expected. (find logic errors etc.)

The relay should self check itself better than you ever could. (provided you monitor the alarm contacts)

 

[pwrtran]

Add couple comments here:

1) as people already addressed, do not try to megger any digital relay, just tell the insurer it doesn't work out that way. Let them megger their plasma or LCD big screen TV if they insist and they will agree with you.

2) check with the manufacture to see if there is any newer firmware upgrade available as this is the perfect time to do so.

3) verify your relay logic and pay attention to the word bit mapping for SEL relays