Time delay from fault to breaker opening 3

[prc]

I have a query on protection. Request guidance from learned protection Gurus.

Let us take a 400 or 220kV transformer. Imagine a flash over from the bottom tail of bushing to tank. The system fault current will go into the arc in oil. In such a case, how much time will it take for the incoming breaker to trip? Which relay will trip it? Differential or over current? Please give typical breakup time for the below steps:
Arc start> CT sense > Relay start-finish to give signal to breaker > Breaker take over> clear the fault.

Will the CT get saturated due to this extremely high current and further delay the actuation of relay?

 

[davidbeach]

Unrestrained differential, relay time of a half cycle or so, lockout relay will add a few milliseconds, then breaker time. The relay can have it all figured out before the CT goes into saturation, but even if it doesn’t the saturation only increases the likelihood of a differential operating.

 

[sushilksk]

Differential and REF relays are faster for such type of faults. Pressure relief devices also operate withing 5-10 ms of operation of Differential / REF relays. Over currents relays have intentional time delays in settings for avoiding trip during through fault.

Differetial or REF Relay senses fault and contacts make in less than 20ms, add 10 ms for master trip relay to operate, after this CB fault clearing time of less than 40ms. Typical fault time will be 60-65ms.

CT saturation depends on bus fault level. In most of the cases fault is cleared without CT saturation.

 

[dpc]

Breaker opening time could be as long as 5 cycles, depending on the breaker. With differential protection, I'd assume a total fault duration of 150 ms a very conservative maximum duration. Assuming everything works, of course.

 

[prc]

Thank you gentlemen for your response. My interest is if we can clear low impedance faults quickly, we can avoid transformer tank explosion and fire. With today's technology, whether such fast acting digital relays available ?
1) Davidbeach- My research says 2cycles for CT sensing+ 2 cycles for relay sensing+ 2 cycles for breaker= 6 cycles min
Does this tally with your experience for 220 &400 kV system? Some engineers claim there are modern digital relays that act within 0.5 cycles instead of 2 cycles. Is it true?
2) Sushilksk- Your post mentions CT + 1.5 cycles for relay+ 2 cycles for Breaker = 3.5 cycles. Is it not too short ? What about CT sensing time from the start of arc?
3) dpc-You have mentioned a total of 5-7.5 cycles (all for 50 cycles) My understanding is also same. But if we can reduce it to say within 3 cycles, it will help in reducing transformer fires. I read, in US, critical old EHV transformers are retrofitted with fast acting breakers to reduce fault time so that tank explosion can be avoided with internal faults. Is it true?

 

[davidbeach]

There's no separate CT time, the CT secondary should be following the CT primary at all times, except for saturation. Saturation does not begin instantaneously and the unrestrained differential can make its decision prior to saturation. The relay manual says that the Raw Unrestrained element (threshold is 2.83 times the filtered unrestrained threshold) has a minimum operate time of 0.25 cycles and a maximum of 1.0 cycles. The Filtered Unrestrained element has a minimum operate time of 0.7 cycles and a maximum of 1.2 cycles.

Two cycle breakers exist. It should be possible to clear a bottom of bushing fault in 3 cycles or less. If absolute minimum trip time is essential, trip the breaker directly in addition to tripping the lockout relay to eliminate the lockout relay time.

 

[dpc]

The main variable will be the maximum breaker clearing time. The differential relay operate time can't get too much faster for digital relays - some sampling time is required. But replacing old breakers is expensive. Old breakers imply old transformers so the investment in a new breaker to replace an old transformer might be a tough sell. The G&W explosive device might be an option to improve on the 3-5 cycle breaker opening time.

 

[prc]

Thanks davidbeach for the great info. Apart from claims as per manuals, what is your experience from actual disturbance recordings from site? Should we add any extra margin over the theoretical time of arc to clearance as calculated?

What is the disadvantage of raw unrestrained differential relay over standard relay? Or this should be provided in addition to standard relay to protect tank?

dpc, what is the G&W explosive device? We cannot depend on pressure relief device for prompt pressure release esp in case of low impedance faults. If we use an arc detecting device(in parallel to conventional relays) will it be beneficial in clearing the arc quickly?

 

[sushilksk]

In your first post you have mentioned "for 400 or 220kV transformer flash over from the bottom tail of bushing to tank" In this case fault current will be very high as there is no trnasformer impedance. Fault current will be same as bus fault level of substation. It may be 20kA to 40kA. If fault is at bottom tail, this level of current will create enough pressure to damage the tank, np matter how fast you clear the fault. At present most of the circuit breakers at 400 and 220kV level has fault clearing time of 40ms. If you are lucky fault may be on air portion of bushing, tank will be saved.

 

[stevenal]

C37.04 allows for an extra half cycle for <= 3 cycle breakers for a close-open operation.

What do you mean by standard relay? Differential is standard for the transformers in question. Modern microprocessor relays will include restrained (% and harmonic) differential, unrestrained differential, over-current, restricted earth fault, and frequency and voltage based elements. I'm not sure anything will beat unrestrained 87.

To protect the tank, use pressure relief devices.

I found a device on the G&W web site that might be described as "explosive," but it is a type of current limiting fuse that only responds to over-current. The voltage rating is much too low also.

If your current transformers are in the bushings, the air fault sushilksk described will need to be cleared by other protection.

 

[cranky108]

If you really want to protect a transformer, it might be better to protect it from things that can cause damage. Like faster clearing of through faults, and oil monitoring for water and other things.
At least here we try to replace transformers at about 30 to 35 years, and we have fewer failures because of that.

Also reducing down line faults from transformers is very helpful in extending the life, or at least reducing the duration of the faults.

If you can, you should look at current limiting fuses to reduce fault magnitudes and not have such large faults currents that would cause explosive conditions.

 

[davidbeach]

The raw and filtered are both part of 87U. The filtered makes sure you don't get an 87U trip below the 87U trip point while the raw requires sufficiently more current that it can operate without waiting for the filtering to occur and still not trip below the 87U threshold. As stevenal says, it's all one relay with all of the functions.