87 Differential tripping main breaker??? 5

[rockman7892]

I have a problem where a SEL 387 differential is tripping our main 5kV breaker. The differential zone is between the secondary bus of our substation transformer and on the load side of our main breaker right after the breaker. The cable that is between the differential zone is 9 runs of 1000MCM for a distance of 1000ft.

The SEL 387 keeps tripping the main and indicating an 87 trip for phases B and C only. The only thing that we did different was added an additional small load in the plant to this breaker however the breaker was closed when we added this load and did not trip. It was not until we opened the main breaker ( unrelated reason) and then tried to close it again that it tripped on differential. After it tripped we tried to close this main again an it tripped on the B and C phase again.

I'm confused b/c nothing that we do on the load side of the breaker (in the plant) should effect the differential zone since what goes in must come out. I'm thinking that there is a problem with the cable run between the substation transformer, or something wrong within the main breaker.

Any ideas???

 

[davidbeach]

I don't see the half cycle shift looking at the L currents (local) and the X currents (remote end). The IA/IB/IC currents in stevenal's image would be the actual measured currents, no time shift, and the L currents would be what is used in the differential calculation, time shifted by half the round trip channel delay. But it is clearly the case that the CTs at one end, probably the 387L end, are connected backwards.

Not an inrush problem, but a CT connection problem. You will trip for load with this CT connection.

 

[davidbeach]

Also, in SEL's free (registration required) AcSELerator program, there is a fairly complete subset of the 5601 software that will allow anyone to look at the event file as posted. The L and X currents are perfectly in phase and plot directly on top of each other. The relay is calculating a perfect operate current, both ends of equal magnitude and phase angle, both into the line. The only thing that kept it from tripping sooner was that the minimum trip threshold had never been reached before.

This system was put into service without ANY commissioning. Somebody did not do their job.

 

[rockman7892]

david beach

I will go out and check the CT connections. If it is the CT connections why is it tripping with a transient inrush current but not with a steady state current?

I agree that both currents appear to be in phase on top of each other. Shouldn't the relay trip with them both being in phase and not 180deg out of phase. I'm assuming they get 180deg out of phase by the arrangment/placement of the CT at the 387 end.

Are you saying that the IAX, IBX, and ICX currents are the currents from the 387L remote relay, and the time shift is due to a data transmission time between relays?

 

[davidbeach]

Yep. There is a minimum trip setting in the relays, below which it won't trip on differential. Your load is still below that minimum setting, but the inrush is above it. Steady state load equal to your inrush current would cause a trip.

Differential schemes work best when currents can be compared at the same instant. With data being communicated between the relays at both ends of the line there is some channel delay time. To minimize the errors associated with that delay, the relays at both ends delay their own current data by half the round trip channel delay, so that they are comparing data gathered at the time the remote data was gathered. If the channel delay is symmetrical this works well; part of the angular spread of the "C" shaped restraint region is to account for channel asymmetry.

 

[stevenal]

Per David's suggestion, here is a replot. Read IAL as local A phase current, and IAX as remote end A phase current (over channel X) and IAT as the summation of the two (or A phase differential current).

Agree with David. Techs and or engineers involved may need some training.

 

[stevenal]

Comtrade files.

 

[stevenal]

again

 

[stevenal]

last one

 

[slavag]

David, Stevanal some Q to you.
Normal situation ( I don't know this relay, is depend on the relay algorit.)
local and remote current in same phase angle or 180deg between them.

 

[davidbeach]

Read my post of 23 Jun 08 0:13. Ideal restraint is same magnitude, 180[°] phase angle difference. Both ends should be measuring current into the line.

In this case it appears one end is measuring current in and the other is measuring current out. No settings available for phase angle compensation.

 

[stevenal]

slavag,

While the SEL alpha plane algorithm may be unique, the connection requirement is not. Differential relays should more properly be called summation relays, since they will trip on a non-zero sum of the two currents. Through current will always show as 180 degrees out of phase and will sum to zero.

 

[edison123]

Great discussion especially stevenal and davidbeach (lps for you both).

So, it was the CT's after all ?

 

[slavag]

OK.
David , Stevenal, thanks a lot again.
Yes, it's standard solution : summation.
First suggestion of Edison was right, start to you too.
As usually 99% of this faults, CT's and non stability test by load.
Best Regards.
Slava
Now, Q's, what Rochman have do with utilities

 

[edison123]

slavag

you are too generous.

Mine was just a drive-by shooting, theirs was a planned assassination (of nailing the right root-cause for this problem).

 

[rockman7892]

Thanks for all of the responses. I am going to confront the utility with the findings of this thread.

Just so I can explain it to the utility correctly I am going to present them with the following: The reason that the relay is tripping is because there are a set of CT's that are installed incorrectly most likely at the 387L relay. Because they are installed correctly, both sets of currents are in phase with each other instead of being 180deg out of phase. There is a minimum trip setting on the relay, and during steady state we are below this trip setting so the relay does not trip. During inrush the current is above this setting (500A + 500A =1000A) and is causing the relay to trip. The CT's will need to be looked at. Am I explaining this correctly?

I find it hard to believe that the utility technicians and engineers did not notice that the phase currents were in phase. I agree that they may not have adequate training.

To answer Slavag's question I am an engineer at a cement plant and the utility services our substation which we own and is on-site. They are in charge of all the relaying and settings in the substation as well as relay's and settings for the differential relay on our main cables. This is a new system and we are in the commisioning phases.

 

[davidbeach]

You are correct, except for one typo.

Because they are installed correctly,

should be

Because they are not installed correctly,

 

[stevenal]

Another small clarification.

The way the CTs are physically installed is fine, it is just the CT secondary wiring that needs to be corrected. Even if it is the primary polarity that doesn't match the drawings, it can be adjusted for on the secondary side. Usually easier to change the drawings than to change the primary polarity.

 

[rockman7892]

One last question regarding cable differential relaying in this application.

Should the remote set of CTs be placed on the line side (before) or load side (after) of the main breaker on the switchgear that is being fed?

 

[jghrist]

The CTs should be placed on the load side so that the breaker is in the differential zone. If there is bus differential on the 4 kV switchgear, it should use CTs on the source side so that the zones overlap and the breaker is in both the line and bus differential zones.

 

[rockman7892]

I was finally able to download the SEL software necessary to view the event files that I had. I noticed when looking at these files that although the the waveform plot appeared as if the two sets of currents were in phase, the phasor plot showed the two sets of currents (IA & IAX etc..)were 180deg out of phase. Any thoughts on why this is?

I'm not sure if these are from the same event file as I posted before, however they are still from the same events that was tripping our main.