Transformer Primary Side Relay Setting 4

[tmaly1]

Folks,

I'm hoping you share your opinion with me on this.

Do you consider the Minimum Utility Fault Current on the transformer primary side when setting the instantaneous element for the protective relay (115 KV side) or you need to be concerned only with the Maximum Fault Current??

Should the instantaneous setting pickup be below the maximum and above the minimum or as long as it is below the maximum, that's good enough?

Thanks for your feedback in advance.

 

[dpc]

Any instantaneous element on the transformer primary side must be set high enough to NOT operate for a fault on the secondary side of the transformer. It should be set to pickup just above the maximum low side fault current (referred to the high side), possibly allowing for asymmetrical current if necessary, depending on relay response. In this way, it will not miscoordinate with the low side protection but may operate for a severe internal transformer fault.

This is basically a function of the transformer impedance and has not much to do with the system impedance.

 

[riteshbharat]

I agree, I generally set it to 1.3 times the transformer full load current divided by percentage impedance.

 

[sslobodan]

Well yup, overload current of LV side should not trip protection on High end side unless the LV side loses its protection integrity, so You mast watch few things
1. What is equivalent maximum current from LV side on HV side?
2. What is the maximum current that overload transformator isolation can whitstand?
3. Is those currents larger than maximum current that HW side can withstand (usually not)

Short circuits are passing bu the transformer on either side, so you must make sure that all eqipment must be protected. That means if you have short circuit on HW side it must not damage transformer, nor cable, also when it passes over transformer on LV side it mast not damage LV cable due to an overcurrent (LV side protection should turn it off). same aspect should go in your calculations from LV side. That will give you eivalent short circuit currents transformed for both voltage sides and that is data that must be lover or equal on:
1. HV end withstanding current max.
2. HV Isolation of transformer max withstanding current.
3. LV end transformed HV current < than max withstanding current
.
and

1. LV end HV max current eqiuvalent < LV max withstanding current.
2. LV end max. withstanding current
3. LV max current.
When protection is set that no short circuit can damage either sides, you have done your job.
That is why setting the relay protection is the hardest part of EE.

 

[stevenal]

Is this element part of a microprocessor transformer protection relay that includes differential protection? If so, instantaneous primary OC is redundant and less selective. I generally do not back up one element with another within the same box under the theory that if the relay fails, the failure is unlikely to involve only one element. (Anyone disagree here? Please jump in)

If the the instantaneous OC element is in a second box for local backup, DPC has the right idea.

 

[jghrist]

stevenal,

I generally use the instantaneous overcurrent element in the transformer protection relay, even if there is a separate overcurrent backup relay. True, it is redundant and less selective, but it's faster. Example: SEL-387 unrestrained element pickup time (Min/Typ/Max) = 0.8/1.0/1.9 cycles; restrained element pickup time (Min/Typ/Max) = 1.5/1.6/2.2 cycles; instantaneous oc element pickup time (Typ/Max) = 0.75/1.2 cycles. Not really enough faster to worry about, but it's free.

 

[tmaly1]

dpc, sslobodan, stevenal, and jghrist,

Thank you all for your valuable input. I applied the guidelines you provided me and it seems very proper to use. I confirmed that as I plotted (using the software i use) the TCC for the relays versus the TX thermal damage curve, TX inrush current curve, and TCC for downstream breakers. All seem to be coordinating well. THANKS AGAIN FOR UR INPUT

 

[dpc]

stevenal,

I can't see much downside in using the instantaneous element if it is there. You're right - if the relay fails, the instantaneous won't back up the differential, but I don't see that as a reason not to use it. If you have enough outputs, you can wire it to a separate output contact since most common failure in digital relays is the output contact. I generally also use a separate overcurrent relay as a backup to the transformer protection relay, but I like to wear a belt and two pairs of suspenders, too.

 

[slavag]

Hi dpc.
In additional to your recommendation ( I think is must used two trip output contacts from relay), we also have mechanical protection, like to gas relay, as back up protection.
I think used or not two different boxes is transformer rating request (of course cost).

 

[521AB]

"Is this element part of a microprocessor transformer protection relay that includes differential protection? If so, instantaneous primary OC is redundant and less selective. I generally do not back up one element with another within the same box under the theory that if the relay fails, the failure is unlikely to involve only one element. (Anyone disagree here? Please jump in)..."

mmmh.. interesting point but I think you are wrong.
What is a relay failure in the numerical technology?
- it can be wrong settings (usually it is), so, having back-up in the same box helps.
- it can be "software bug" in the application (for instance differential protection". In this case, probably the same bug is not in th eovercurrent application. Back-up in the same box helps.
- it can be "hardware failure" affecting protection functionality. Self supervision usually blocks everything that "trips".

Anyway your point has a logic. just one suggestion: if you do not want to trip with the integrated back-up protection, set it and send its trip to internal disturbance recorder. Its start and trip signals might help you in case of fault analysis.

 

[stevenal]

We can get carried away with these things. If one instantaneous is good, why not three? Or three times three for phase, ground and -sequence. They're not really free, each setting must be derived, set, tested, and documented. And when an event occurs, more data must be sorted. No reason to use separate outputs for different elements, each output can utilize the same long trip equation. In my experience, bad trip relays indicate bad auxiliary switches, so paralleling them only ensures more outputs will get burned.

An extra element cannot always help with wrong settings. Tripped on REF from a through fault a few weeks back. Had the neutral polarity reversed. No number of additional instantaneous elements could have fixed it. And if the wrong differential settings were do to an incorrect transformer model, the instantaneous setting is likely to be wrong also.

 

[slavag]

Hi.
Stevenal, you are right, number of functions and outputs must be logical. Any extra function it's problem.
Of course don't need separate trip outputs to several functions. But you "must" use two separate common trip
outputs DC1 and DC2 (or AC) and maybe next one to LOR.
But, from my point of wiew two stages of O/C protection are
important: one inst for diff protection back-up and second
IDMT for LV faults clear.
Please pay attention again on answers of 521AB, only one problem we have in case of "all eggs in one box", relay fault ( some HW problem). In this case you have signal from IRF (self-supervision system)and as back-up protection Bucholtz relay.
In additional, REF it's very problem function and you don't have any back-up function for it.
Regards.
Slava

 

[ScottyUK]

How did you find that fault Steven? I had exactly the same problem on a couple of generator transformers a few years ago. In that case they were Hi-Z electro-mechanical relays: no digital event recorder to take a snapshot at the time of trip, just some seemingly inexplicable transformer trips when a ground fault occurred out on the network. I ended setting up a primary injection test to prove the REF scheme was unstable - not a trivial matter on a 170MVA GSU transformer. It was an interesting task though!


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[stevenal]

Scotty,
Your GSUs are probably serving a balanced load. We had the advantage of having unbalanced distribution loading. All the metering data was available in the relay, we just neglected to look at it. I have a commissioning check list now for next time.

When replacing relays at an industrial site that had balanced loading, and adding REF, we carefully checked the polarity on the four units with externally mounted CTs. On the fifth unit with its bushing CT, we used primary injection as you did. A rolled out breaker provided the access to the transformer terminals while the other side remained energized. This was before the arc flash standards were in effect.