Electromechanical relay failing to trip

[drchaos]

Hi, I've been asked to look into why an R1Z23B electromechanical distance relay failed to react to a 3phase fault on a 110kV line, leading to a fairly big blackout in the affected county. The relay is 600/5 A, 60VA, and connected via a 25VA cable to a burden of 15VA. The fault had a significant DC component of about 3kA, and the AC component was also about 3kA. My question is: would that DC component have saturated the CT, or would the problem lie with the extra burden of the cable? The relay was tested and was working fine.
I'm not a protection expert, I know a bit, but any pointers would be greatly appreciated. Thanks.

 

[slavag]

Hi Drchaos.
Please provide more information:
What is a CT: type, burden ( 60VA?), ratio(600/5A?)?
What do you mean cable 25VA to burden 15VA?
Regards.
Slava

 

[drchaos]

Hi slava, sorry that I confused you.
CT - current transformer with ratio of 600/5 amps. It is connected via a long cable to the relay in the control room. The burden of the cable is 25VA and the burden of the relay itself is 15VA. I've been told that the current transformer is rated 60VA. So the total burden on the CT is 40VA. We know the relay picked up, but never sent a trip signal. Do you think the burden on the transformer was too much, or would the large DC component of the fault current have caused the transformer to saturate?

 

[slavag]

Hi Drchaos.
Burden of CT, 60VA is OK. what is type 5P10/20/30 or ANSI type.
What is a lengh of cable?. 25VA that mean 1Ohm, for my pinion is so much.
Problem is, I don't know this relay, what is requirements to
CT. Please check what is n'> requirements for this protection.
It's new installation, it's first fault on this line?
Regards.
Slava

 

[drchaos]

Sorry, but I don't know the accuracy class of the CT.
This is an old installation. The relay has probably been in place for 20 or 30 years, but has received regular testing and maintenance. I will check the settings of the relay - its a distance relay, not overcurrent.

 

[slavag]

I don't know exctly, but something not seems good with CT secondary wiring.
60VA CT, 600/5A, long cable, we put with this data digital relay with 0.3, 0.5VA current input, here is 15VA!!!!
From other side 20years, 110kV line w/o faults (also possible???).
Second option what was fault impedance, maybe via arc and not eq to pick-up.
Of course with CT was in saturation, this influence on dist protection.

In all cases, we don't have any recorders of fault and very dificalt say,what is a problem.

Check operation of relay from CT, not from terminals into cubicle.

 

[drchaos]

The fault was due to a lightning strike. Flashover at a tower i guess. You think that maybe the tower had a lot of resistance? The fault current was large, consistent with short circuit levels in the area.

I think we will add a digital relay as main distance protection now. Should have been done years ago!

 

[slavag]

Add digital relay is good and right solution.
Yes, it seems as fault via high resistance ( arc resistance, not tower). This type of fault with high level of DC component is also problem for digital protection.

3kA is not so high level of fault for 110kV (I assume you have 3-ph fault around 30kA in the system).
I'm also assume, that fault was in end of first zone of dist protection.
Regards.
Slava

 

[DanDel]

How do you know that the relay picked up, but never sent a trip signal? I would check the C/B, including a test of the trip circuit.

 

[drchaos]

I think on these relays a flag appears when it picks up, and another will show that a trip signal was sent. So it did detect something abnormal, but not enough to send a trip signal.

 

[DanDel]

Typically, a flag appears when the relay picks up and closes (or opens) the trip contacts. This should be easy enough to check with a relay test. If the test shows that the relay picks up and doesn't activate the contacts, then something's wrong with the relay. If it picks up and does activate the contacts, then I would check the C/B and trip circuit.
The easy way to test everything is to test it as a system; with the relay in service, do a trip test with secondary injection and allow it to trip the C/B.

 

[drchaos]

Thanks DanDel. I'm not in work today, but as far as I know
the relay and various components have been tested and are working normally. The finger of blame is being pointed at the CT - but I am wondering has anyone come across a similar failure to trip because of a large DC component to the fault current.

 

[521AB]

It appears strange to me the "three phase fault". They usually do not exist, unless they are the evolution from a phase-ground or 2-phase-ground faults into a multiphase fault (3-phase in this case). If this can be assumed (if you have an external disturbance recorder you might see this), then I believe that the relay picked up (starting flag shows this), probably for a phase-ground initial fault. Maybe not in the instantaneous zone, so the relay started the timer. During the time counting, the fault evolved into a multiphase fault. If the relay is of a switched type, it cannot see it because it was told to calculate the distance on the phase-ground loop, and now it should use instead another loop (like phases 1-2 for instance). The relay will with most probability not trip for that evolving fault.

Possibility n. 2. The fault was not of evolving type. The relay picked-up, on th enon instantaneous zone. During the timer the arc "gets longer", i.e. the fault resistance gets higher, larger than the setting. Old relays do not allow to have different fault resistance settings (only one setting for all zones), so when the fault resistance went out from the relay zone, it went out from all the zones. Hopefully some other back-up relay (in remote circuit breakers) tripped for that fault.

 

[drchaos]

Thanks 521AB, that's very interesting. I will look into it - I think there may have been a disturbance recorder looking at the fault.
So you are saying the fault appears initially in the zone 1 R-X circle, but then moves back out again as the fault evolves and the resistance grows.

 

[slavag]

Hi. Right reasons!
521AB , you are right, I know about only one case of 3-ph fault(one person closed ES on live line in 161kV). Protection usally started on 2ph-fauly or I think 90% faults are phase to earth(ground) fault.
Drchaos
What is a company this relay?

 

[521AB]

dr. chaos: It's difficult that the fault moves out from instantaneous zone1. If it s in zone 2 (400 ms delay time) or other zones it is more probable.
For the moment I would more think about evolving fault, if the relay is a switched relay, and not a full scheme one, which is quite possible if the relay is old. Relay picked-up for a phase-ground fault (LN), but when it started to measure the distance, the fault was 2-phase-ground (LLN) and the measured distance was longer than the actual one, because it was measured with the single phase loop.
But this is just an assumption. Try to get the disturbance recorder. You can't say anything without it.
Remember that if you get the disturbance recorder (probably it will be in form of a file: comtrade standard), it is possible to play-out again the waveforms from the modern test equipments, so you can inject again the waveforms into the relay and verify the "missed trip".'

 

[drchaos]

I've just been told that it was definitely a 3-phase fault initially - they are very common where I am (no earth-shield wires). The CT is a 5P50. On the face of it, there should not have been a problem. The cable running from the CT to the relay is 400m long and is a four-core cable with each phase having a 10mm cross-section. Do you think this could cause a problem?

 

[521AB]

Ok, interesting that you had the 3-phase fault!
You are asking about CT check.
First of all, as Slavag wrote: do you have any CT requirement for the relay manufacturer?

The CT i 5P50. Hoe many VA? What about CT resistance?

To calculate the voltage across the CT during the fault, not considering CT saturation due to dc-component (this should available in the formulae for CT sizing given by the manufacturer), do like that:

E = Ishort * (Rct + Rwire + Rrelay)

Rct is the CT resistance (ohms)
Rwire = resistivity at 75 deg * length / area = 0.022E-6 * 400 / (10E-6)
Rrelay = relay burden / Irated

As I said, there should be some coefficients mutilplying Ishort, considering the dc-component (or the time constant of the fault). But let's assume 1 for the moment (it can also be 3 or 4..).

You should compare the calculated voltage with the so called Accuracy Limit Voltage (E2max).
E2max can be calculated from CT data (accuracy limit factor, 50, Rct ?? and the nominal buden), but you can assume E2max = 1.2 Vkneepointvoltage (according to British Standard).

If E > E2max (by using 1 as coefficients multiplying the symmetrical steady state fault current), for sure the CT has to be considered saturated.
If not we don't know, because we don't know the coefficients in the formula.

I hope this helps.

 

[slavag]

5P50 60VA 600/5, 400m of 10mm^2 current cable, 15VA!!! of relay burden. I assume for calculation that Rin of CT about 0.7 Ohm.
521AB small correction, you need use in calculation 2 x R wire.
Drchaos please see link
thread238-194167
calculate n'=50x(60+0.7x25 )/(0.7x25+15+0.019x800x25/10)=55
not bad.
In all case check CT and CT wiring and realy mnf requirements, but n'=55 I think is good data ( of course for digital relay, but I think same for EM relay too).
Not correct setting???? dificult say.
It's short or long line?
I back to 521AB very important Q, what clear this fault?
3ph fault with 3kA current level, seems very small.
What is a type of neutral grounded in your system: solid grounded, resistive, "ungrounded???"?
W/O DR data and event list of relay or system are dificult say, but try help.

 

[slavag]

Hi 521AB.
Were are you : IEC or ANSI land?
Regards.
Slava