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.

 

[521AB]

slavag, it is a 3-phase fault. Only one time the wire!
Anyway, if you take 2 times you are on the safe side!

I am in IEC land...

 

[521AB]

CT: 5P50 60 VA Rct = 0,7 ohm 600 A / 5 A
Cable: 400 m 10 squared mm
Relay burden: 15 VA
Ishort three-phase = 3000 A

Let’s go.

E2max = 775 V (almost knee point voltage)
Because:

What is the voltage across the CT when it delivers accuracy limit current (50 * In = 50 * 5 = 250 A) on the nominal burden of 60 VA?

Go back inthe resistive world:

CT is a current generator of 250 A, in series with Rct, in series with Rburden.

Rburden = 60 VA / In^2 ? Rburden = 60 / 25 = 2.4 ohm.

E2max = 250 A * (0,7 ohm + 2.4 ohm) = 775 V.


What is the voltage across the current transformer whe it delivers the fault current of 3kA (25 A secondary)?

The same as before: The CT is a current generator (25 A) in series with Rct, Rcable, Rburdenrelay.

Rcable = rho * LENGTHcable / AREAcable = 0,022E-6 * 400 m / 10E-6 = 8,8 ohm

Rburdenrelay = 15 VA / Irated^2 = 15 VA / 25 A^2 = 0,6 ohm

E2 = 25 * (0,7 + 8,8 + 0,6) = 25 * 10.1 = 250 V

250 V is much less then 775 V.

If we consider a multiplication factor of 3 for the fault current (the factor that should take in consideration dc current and its time constant --> read CT saturation), we go to 750 V, very close to E2max.
3 is a factor that can be used for numerical relays, I wonder if mechanical relays do not require more, maybe 5 or 6? Then the CT IS saturated.

Slavag, do you follow me?

 

[slavag]

Yes, of course, but you also don't believe in 3ph fault.
I'm on 99% sure it's start as LN or LLN fault and finished as 3ph fault.

 

[DanDel]

drchaos, either the relay picked up or it didn't; if it picked up, then the CT is working.

 

[slavag]

521AB my yes connect to your previos post.
I only read new.
Regards.
slava

 

[slavag]

Sorry 521AB, check again your calculation 0.022 x 400/10=0.88Ohm (or I'm in blackout).
In all case in two way we have same result.
Nice work with you.
BTW for this case I prefer use 5 or 6 factor.

 

[slavag]

DanDel,you are right, if relay pick-up, CT works, but:
1. What is a reason for pick-up?
2. What is accurace of CT.
More or les we try closed noe all option, you checked options of HW faults ( relay, CB, etc), according to drchaos are OK, will try next options.
Regards.
Slava.
OR THIS CASE IS DEAD CASE (I don't know if it's right expression)

 

[521AB]

Thanks slavag, it looke a little bit too much, I've done it 2 times ;-(
But you are right.
This means that the voltage is much less:

E2 = 25 * (0,7 + 0,88 + 0,6) = 25 * 2 = 50 V
Even considering a factor of 6 we go to 300 V, when we have "750 available".
CT cannot be blaimed. I think we can say that.

DanDel, good observation (relay started), but what does start mean? Is it just a very large starting zone, which doesn't trip. Relay can start even if CT saturates, if the zone is very large, but then the relay will underreach a lot (measure higher impedance then tha actual one) because current is much less.

Anyway I think you can't blaim the current transformer, and it seems they knew what they did 25 years ago, when they put a 10 squared mm cable there!

If CT is ok. can be too low resistive reach setting? Can you try just to inject the steady state values of the fault quantities in the relay, to see if it trips?

 

[DanDel]

The first thing to do is to determine if the relay flag indicates just pickup or pickup and trip. Typically, it indicates that the relay called for a trip.

If there were never any flag, then I would say to check out the relay, CTs, etc. The flag indicates that the relay and CTs work. The fact that the flag is there with no C/B trip indicates that something in the trip circuit or C/B is not working correctly.

It's a simple test to trip the C/B from the relay (the problem is scheduling the outage). This will check out the complete trip circuit out from the relay.

 

[slavag]

Drchaos, you see, we need more information about this relay.
BTW, Are you have BFP in this substation?
I back to Q: what was clean this fault?
DanDel, we are base on: system tested from end to end.(that mean from CT/VT up to CB), not only relay with test plug. If not...
Drchaos, please provide small test ( if possible), inject 5A
from CT terminals and meas. voltages on them.

 

[slavag]

In additional.
meas also what current you see on relay in injection time ,and sorry forgot about it, please provide test of this cable insulation ( 20-30 years of cable).

 

[drchaos]

Guys thanks for all your great work here. If I can answer some of your questions (BTW I haven't been on site - we have guys in the field who do all the tests):

The line was being fed from one end only, and the fault was definitely 3 phase, as I've seen the disturbance recorder (not at that substation, maybe next one along.) A digital relay at the next substation also picked up and finally cleared the fault (in its zone 2 or 3 or overcurrent).
When I say the relay picked up, this relay has an overcurrent starter element which picked up, but the distance element did not send a trip signal. The trip circuit has been tested and is good.
The CT kneepoint voltage is 609V. The techies used one of the Doble kits or whatever they are called to test it.

The fault current started out around 3kA (AC) and 3kA DC, but actually increased to 5kA AC, as there was a simultaneous fault elsewhere on the system - when this cleared, the fault current obviously increased. Where we are(in western europe), the fault levels are not particularly high as the system is quite weak in this area.

 

[slavag]

Hi Drchaos.
I'm hope that we cancelled all option of HW problem (trip circuit, current circuit, etc.).
This EM relay with O/C starter and I'm assume with single meas.system.
I don't work with EM relays many years, but as I remember this type of relay have some "problems"
1. "Long or slow time" of operation in the 1-st zone (ten's miliseconds).
2. Not realibility in case , as 521AB say, fault evolves.

I will ask my friends, that have very good excpirence with
EM distance relay, maybe them send me some tip.
Regards.
Slava

 

[521AB]

Dr chaos,

the measured knee point voltage of 609 V indicates that probably our calculations on the CTs are correct. We got E2max = 750V. Considering E2max = 1.2 Vknee we are almost there. This is good.

Even with a fault current of 5 kA, I don't think we can blame the CT.

Do you have the possibility to show us the disturbance recorder? Do you have it on paper or on file?

There are not so many possibilities now.
You have the disturbance recorder. You can read the pre-fault voltages and currents. Get them and inject them in the relay.
Then you can read the fault voltages and fault currents (steadu state, I would get the 3kA level for the beginning). read them and inject them into the relay.
Will the relay behave as during the fault? (Only start indication)?
Hopefully yes. Then we know it is probably a setting problem. You mentioned MHO characteristic. MHO characteristic are not sensitive for fault resistance, close to the reactive reach (reactive border).
If the relay trips correctly, with that injection, you should try to inject also the transient dc quantity. If the disturbance recorder is in form of COMTRADE file, DOBLE is able to play-out it again.
Will the relay now behave as when in service (just trip indication), probably the dc component created some problems (to the main CT itself or to the relay itself).
CONSIDER that the disturbance recorder DOES NOT show WHAT THE RELAY has measured (that's why digital relays with disturbance recorder included ARE important).

Try with the first test (steady state quantities). Hopefully this test gives some answers. Then let us know.

AND PLEASE: can you show us the disturbance file?

 

[drchaos]

The fault was close in to the substation, 2 or 3 kms.
I haven't got the disturbance recorder graphics unfortunately - it was looking at another line in the same substation - so those currents are estimates only. Do you think the relay maybe saw something else? Maybe bigger currents?
I really need to get my hands on the manufacturers CT curves, and I will probably try and model the setup in EMTP or something.
I suppose if the relay was working, it just didn't see the fault, and so that must have been due to the CT output.

I have a final question - when a CT saturates, will that increase the fault impedance seen by the relay?

 

[521AB]

If CT saturates, in principle it delivers less current than the current it should deliver. In reality is even more difficult because the delivered current contains only some ”pieces” of the 50 Hz waveform. Anyway, in principle if CT saturates, the relays measures a bigger impedance (underreaching is called).
Numerical relays may manage this situation by recognizing the waveform, and can try to reconstruct the waveform that had been generated if the CT had not saturated, but electromechanical relays just get what they get….

Yoe mentioned a radial feeder, and said that a remote numerical relay tripped. If the feeder is radial also for the remote relay, the fault current measured by that relay must not be far away from the fault current measured by YOUR mechanical relay. Do you follow me?

For the voltages it is a little bit more complicated, but let’s try to get the fault current level..
And one suggestion: don’t promise any result. No disturbance recorder available, we can in practice just speculate.

Anyway, let’s see if we can get something from the remote relay.

 

[slavag]

Hi.
521AB, I'm agree with all.
Drchaos, please check with your team what is a type of
distance characteristic in this relay?
It's more and more seems as high resistive fault.

 

[slavag]

Hi
I disscaused with my several friends and all of them saied
same:
All distance protection (digital, numerical, static and EM)
have same problem:
Not enough sensetive to high resistive fault ( high DC component add problems).
For increase sensetive we add SDEF function.
It's all time game between sensetivity and stability of distance protection.
Only before two weeks (just now I recieved this information)
one of numerical protection send trip according to second zone, but fault was in first zone and near to SS. Same problem : ARC and high resistive fault.
More from this are possible that your relay see this fault as overload ( are possible also in the digital relay).

And I would ask again what is atype of grounding in this system???
But as saied 521AB, w/o fault data, all of our disscasion is only speculate!!!!!!
Regards.
Slava

 

[Bluezee]

Here are my thoughts on what could be happening here. There could be two reasons why the distance elements did not operate.
1. During the fault the voltage collapsed near zero. Most impedance relays do not operate when the voltage is zero to guard against miss operation due to PT fuse failure.

2. Because the fault was 3 phase the relay might have see it as a power swing, therefore blocking all distance elements form operating.

Bare in mind that I am not at all familiar with that particular relay so take from this what you may.

 

[drchaos]

Bluezee, thanks for your reply. I don't think this relay is clever enough to do the things you said! It's prob 20-30 years old.

Slavag, the fault was on 110kV system, which I think is solidly earthed at trafo neutrals.

We are blaming CT saturation at the moment. I will try to model the situation using digsilent or emtp soon.