Transformer differential protection - Need your wisdom 2

[BINGMD]

Transformer: 230kV(Yg)-115kV(Yg)-34.5kV (Delta with a zig-zag grounding transformer).

At 34.5kV side, there are two wind power generation feeders and a capacitor bank feeder. CT is 1200:5 C400.

Available 3 phase fault is about 8kA and 1LG fault is about 5kA.

Relay: SEL-387-5

Question: Is there any porblem to combine CTs of the two wind power generation feeders to one CT inputs of SEL-387-5, i.e., one restrain element? The CT burden is low and no satuaration is expected. Combining source feeders seems no problem.

Thank you for your response.
Regards,
Bing

 

[davidbeach]

If you have a fault on one of the 34.5kV circuits outside the zone of protection, the current from the other circuit will not show up as restraint current in the relay and you could have a differential operation for an out of zone fault. It is always dangerous paralleling CTs on sources into a differential relay (does not apply to high Z bus diff). You can get away with paralleling loads, but all sources should have their own restraint input.

 

[dpc]

I agree with David. Back in olden times, we might have combined the CTs into one input and lived with the risk, but with easy availability of multiple input digital relays, I would always try to have a separate input for each set of CTs, especially for sources.

 

[BINGMD]

Davidbeach,

Nice point, thank you.
In our case, the wind power contribution is about 300 Amp per feeder for a three phase fault. By comparing to 8kA total 3 phase fault level, can the removal of the small portion of restraint currents be allowed and still have correct selectivity for an out-of-zone feeder fault?

Thanks again.
Regards,
Bing

 

[BINGMD]

Sorry for my miss.

One thing needs to clarify is

SEL-387 only has four sets CT inputs.

The above case has five sets CT required for 230kV, 115kV, 34.5kV feeder 1, 34.5kV feeder 2, and 34.5kV capacitor bank.

So, that's why a paralleling CTs is considered.

Thank you,
Regards,
Bing

 

[davidbeach]

You need to spring for one more set of CTs, on the 34.5kV between the transformer and the first tap. That would certainly provide the best protection. On the other hand, with your weak sources, you might (but I'd still recommend not) get away with paralleling the two 34.5kV sources if you set the relay such that the missing restraint current won't cause a trip for an out of zone fault. Changing the settings could degrade your transformer protection. Another approach would be to not include the caps in the differential scheme, put a separate relay on the cap bank circuit and use a directional instantaneous element to detect faults in the cap bank and block the differential for cap bank faults. This may require slowing down the differential as well. All in all a single set of CTs on the 34.5kV where it leaves the transformer would be the best solution.

 

[stevenal]

Agree with Davidbeach, you need CTs at the transformer. Do you need to spring for them, though? No bushing CTs already present on the transformer? Even if they are in use, the low extra burden of the 387 might not be a problem.

 

[JensenDrive]

I think the fault current you mentioned was for the 34.5kV side. Gee: 1200:5 CTs on a system 5kA (SLG) implies about 20.83A in the CT secondary. Compare this to the 300A from the wind farm giving only 1.25A fault current. Will the 1.25A cause any effect on relay restraint? I do not thing so. The concern about summing CTs is best seen in an example: Suppose you had 50A coming in one feed to a bus, and going back out on a second feed. Suppose one of the CTs partially failed and gave only 40A. If you paralleled the CTs into one input, the relay would see 10A coming in on an input, that is all. It would see operate = 100% of the restraint, and would trip. If you brought the two CTs into different relay inputs, you would see 50A restraint and 10A operate, so operate = 20% of restraint. The relay does not trip.

I agree that you should not parallel CTs on sources in general, but when a source is 5% of the total fault current, as in your case, I think you can easily ignore the common recommendation and go ahead and parallel the CTs.