Differential Trip

[sanchezjl]

Hi..

I have a question for the relay engineers..

Recently we had a trip in one of our substations transformers where we are using a SEL-387 relay with an external single phase to ground fault.

One the event was downloaded; we sent it to SEL because it was not clear for us (operations put the transformer in service without diff protection and it worked, but today we took it out of service for inspections).

The thing is that they realize that the hi side C phase and the low side C phase (the faulted phase) where in phase and around the same magnitude before the fault, but during the fault it was a 50 degrees shift between those two.

Have any of you guys had something similar in the past?

I will appreciate your answers.

Thanks!

 

[ScottyUK]

Slow down a little:

What is the transformer vector group?
What current transformers are installed (ratio and class) and how are they connected?
What is the transformer ratio? Which currents were 'around the same magnitude'?
What order of magnitude was the earth fault?



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

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

 

[jghrist]

Check for multiple grounds on the CT circuits. This can do strange things during a ground fault because of stray paths for ground current.

 

[sanchezjl]

ScottyUK:

First of all, thanks for your response.

The transformer vector group is Dy1, but we have C phase on H1, B phase on H2 and C phase on H3. Therefore, for the relay is a Dy11 because the low side is leading the high side by 60 degrees. All our system is like this.

This transformer is a 69/24.94 kV, 33.3 MVA max. The CT's are C400 class on both sides and they are connected in Y on both sides, because we are using a microprocessor relays (SEL-387-5) where you can connect the CT's in Y.

The CT ratios are 300/5 in the 69 kV side and 800/5 in the 24.94 kV side.

When we put the transformer in service, we verified with the existing load that high side currents were leading the low side by 150 degrees, like the rest of our substations.

When we had a 1440 amps distribution fault on the C phase to ground in one of our feeders, we can see the pre-fault conditions (made by an special software that SEL has), the C phase in the transformer was in phase with the C phase in the low side of the transformer, but during the fault, the C phase on the high side was leading the C phase in the low side of the transformer by 50 degrees , when they should be in the same conditions like in the pre-fault for a trough-fault current in the transformer.

This is the information that I have until this moment. Today we are testing the transformer, wiring, relay, CT’s and grounding.

Do you have any suggestions?

Thanks!

 

[oldfieldguy]

sanchezjl-

Make sure that you only have single point grounding on your CT circuits. Schweitzer shows one ground for all input CT's It is not uncommon to find that the people doing the installation will have a ground for each set of CT's, but this is incorrect.

Assuming you don't find a physical problem in the CT connections and the transformer tests are satisfactory, then upon re-energizing the transformer, use the relay "METER DIFF" command to see if there is a significant differential current indication. This might indicate an incorrect CT connection or incorrect setpoint programming.

I have seen differential elements incorrectly set where normal load current was below the amount needed to operate the differential element, but on a through fault, the differential element operated. There was a dual problem: the differential pickup setpoint was too low and there was an icorrect application of the CT's. This transformer had been on line for ten years with improper relay setup. The problem only became apparent when the client had a differential operation for a through fault.

old field guy

 

[stevenal]

Your information is inconsistent. In the same post you claim load current on each side is leading the other. The shift should be 30 degrees, not 60. CT connections cause another 180 shift so they should be 150 out at the relay. Use the MET DIF command to see the operating and restraint quantities. IOP should be very small for a load condition, while IRT should be roughly equal on both windings and 180 out of phase.

 

[bacon4life]

old field guy,

Which SEL product requires only one ground?

Because each of the 18 current channels is independent, be sure to apply a ground to each set of three CTs forming the current input from each terminal.

 

[stevenal]

bacon,

Each set is grounded as the manual states, just make sure that they are all grounded to the same single point. The prefered location for grounding is at the relay. See IEEE C57.13.3.

 

[oldfieldguy]

Bacon--

I double-checked and was referring to CT connections only. The typical wiring diagram shows a single ground for ALL the CT inputs. Check figure 2.8 of the SEL-387 manual. There is a frame ground connection also.

What happens in many installations is that the installing technician will put a ground connection on each set of CT's, usually in the junction box of the transformer, but sometimes at a terminal strip in the relay house, because he "knows" that CT's are supposed to be grounded like that.

A good technician or field engineer will lift that ground connection and then check from the CT circuit to ground. With the ONE ground lifted, there should be no continuity to ground. I don't know how many times I've found incorrect grounding on CT circuits, usually on transformer differentials.

old field guy

 

[Marmite]

Sanchezjl.
Cutting through the wood to get to the trees, the fundamental problem here is that the differential protection tripped on a through fault. The reason behind that is the lack of any zero sequence filtering. The relay configured as a Dy11 would provide the appropriate phase correction for the Dy1 transformer, however the transformer
has an earth connection on the secondary winding, so it
can deliver zero sequence current to the fault. Use of
star connected main CT’s provides a path for the zero sequence current to reach
the protection relay. On the primary side of the
transformer, the delta connected main primary winding
causes zero-sequence current to circulate around the
delta and hence will not be seen by the primary side
main CT’s. The protection relay will therefore not see any
zero-sequence current on the primary side, and hence
detects the secondary side zero sequence current
incorrectly as an in-zone fault.
The solution is to provide the interposing CT's (software generated in this case) on the secondary side
of the transformer with a delta winding, so that the
zero-sequence current circulates round the delta and is
not seen by the relay. Therefore, the rule of thumb is that a transformer winding with a connection to earth
must have a delta-connected main or ICT for unit
protection to operate correctly. For this particular scenario the relay should be configured as Yd11 to avoid maloperation for through faults. Hopefully you will see this before you waste too much time dismantling and testing.
Regards
Marmite

 

[stevenal]

Wrong path. When you set an SEL transformer differential to compensate for phase shift your software generated delta interposing CTs are included automatically.

 

[Marmite]

I confess to having no knowledge of this particular relay, however similar UK relays would have a number of vector groups available to choose from for the software primary and secondary interposing CT's which would apply the same appropriate phase correction overall. It was Sanchezjl's second post in which he states "Therefore, for the relay is a Dy11 because the low side is leading the high side by 60 degrees".which led me to believe that an inappropriate vector group for the interposing CT's may have been chosen.

 

[sanchezjl]

Stevenal, you are right. On Dy1 or Dy11 conections there is only a 30 degrees shift between high & low side. I was trying to say that when you feed a Dy1 transformer with CBA conections, the low side vector rotates 60 degrees, and now is leading the high side by 30 degrees. Thank for your observation.

So far, we tested the power transformer and is Ok. The CT ratio & excitation test are good. In the CT AC circuits it was only one ground point, in this one was in the relay cabinet inside the relay house. It was no grounded points in some other place.

Some additional information. When we put the transformer in service, it had 7.926 KVA. The transformer max. capacity is 33.000 KVA. Therefore, the restrain (RST) on the relay should be 7926/33000= 0.24. The relay reading on the RST were A=0.24, B=0.25, C=0.24. The operational reading were A=0.02, B=0.00, C=0.02.

What do I am missing?

Thanks a lot for your support!

 

[Marmite]

Sanchezjl
From what I understand from the earlier posts your transformer has phases A&C crossed on both the primary and secondary sides to turn it from a Dy1 to a Dyll. Where is the cross physically? ie is it outside the protected zone formed by the biased diff CT's or inside the zone? This could affect whether the relay "thinks" it is looking at a Dy1 or a Dy11 transformer.
Regards
Marmite

 

[sanchezjl]

Marmite:

Is outside the protected zone. The CT's that I'm using for the differential protection, is the primary & secondary CT's located on the transformer bushing. Therefore, nothing else is in the protected zone, but the transformer.

Like I said before, our entire transformer are Dy1, and are been fed with negative sequence (ACB), and therefore, what we really got is a "Dy11 system". We have the same relay in other places with the same settings with no problem at all.

Now, does somebody think that I'm missing something or now that we have settings with more sensitivity than the existing electromechanical ones; there is a possibility that we may have a problem inside the transformer or in the non load tap changer that tested Ok on the TTR?

From the relay events, I saw that the first element that called the trip was "A" phase and 0.25 cycles later the "C" phase also activates the trip order. Taken into account that the trough-fault current in the low side was on "C" phase; do somebody thinks that that may be a possibility, (a problem in the tap-changer when is above certain current magnitude) or I still have a relay problem?

Again, thanks a lot!

 

[Electic]

Could it be that one side CT's went into saturation thereby upsetting the differential balance?

 

[Marmite]

Sanchezjl:
With the CT's positioned as you describe, the relay is looking at a Dy1, not a Dy11, transformer and the software ICT's would need to be chosen to apply +30 degree correction overall, plus the zero sequence filtering.
It's extremely unlikely that a through fault current barely twice the normal full load current of the transformer would cause any coincident damage to either the transformer or the tap changer. The transformer should be able to stand 15-20% normal current for around 2 seconds without sustaining damage.
I'm convinced that you have a stability problem with the differential protection, either due to faulty CT's, faulty or incorrect secondary wiring, faulty or more likely incorrectly set/configured relay. I presume there is other protection on the transformer, such as restricted earth fault and buchholz which has also failed to see the transformer "fault". To put your mind at rest you could take oil samples from the main tank and tap changer if seperate for DGA testing. Also a gas sample from the Buchholz relay, if there is one.
Just because you have the same relay in other places with the same settings doesn't mean it's been set up correctly. More likely the same guy commissioned them all and made the same error on them all.
Regards
Marmite

 

[jghrist]

I would assume that the SEL-387 WnCTC settings are correct to compensate for the CT connections and zero-sequence blocking. Otherwise, the pre-fault CØ currents would not be in phase and around the same magnitude as noted in the OP.

The question is, why the 50 degree phase shift during the fault?

 

[stevenal]

sanchezjl,

I'm stuck. Prechecks confirm that settings are okay, and CTs also check out. Only the relay and transformer are left and the transformer continues to work. A self repairing internal fault in response to the low magnitude through fault seems unlikely, but DGA would help rule it out. I will be most interested in what SEL reports. Please post back when you know something. We also put C/c phase on H/x1, and the proper entry for the low side wye winding is 11, confirmed with security for through faults.

 

[sanchezjl]

What we are going to do is change the CT's set for the differential, keep the diff blocked, monitor twice a week the relay looking for other event to see the behavior of the differential function.

Meanwhile, the transformer (20 MVA base) will be protected with overcurrent (the SEL-387 and the SEL-501 backup).

We also are going to take a DGA the transformer just in case.

I will keep you informed.

Thanks a lot!

By the way, how can I paste a picture here.. I was trying to do it with the SEL report and I did not find the way.