Transformer Inrush earth fault trip 4

[AEMB]

In the case of energisation of multiple transformers fed from a single board the main circuit breaker tripped on earth fault. The earth fault measurement is via residual current measurement. I assume that the CTs saturated and that caused the residual current to appear. The main breaker energises 3 x 3000 KVA transformers. I am trying to calculate the inrush curev for this scenario to see whether it would be sufficient to saturate the CTs. Any ideas?

 

[alehman]

It is possible one or more CT's saturate when energizing the transformer. You would have to know the burden and magnetization curve for the CT's to determine if that is the case. You may need to add or increase time delay on your earth fault relay.

 

[davidbeach]

The CTs undoubtedly saturated, but even if they didn't, transformer inrush can easily produce residual current that the relay will not be able to distinguish from a ground fault. As alehman suggests, lengthen your time delay to allow subsidence of the inrush current. To go beyond that, you could consider using the presence of second harmonic current to restrain operation of the ground element.

You could also put in transformer differential relays with the zone extending from the point the three circuits separate to the secondary breaker of each transformer. With the conductors in the zone of protection, the ground unit on the main breaker would only need to cover ground faults at the bus and could be set much higher.

More details on your system configuration would make it easier to provide additional suggestions.

 

[RalphChristie]

Can you confirm if it is instantaneous earth fault trips or overcurrent (IDMT) earth fault trips? Inrush currents may reach instantaneous peaks of 8 - 30 times those for full load, and what I have seen many times (on this size transformers) is instantaneous earth fault trips due to the unbalance in the system during energizing - due to the inrush current. (current in one phase will always be higher than the other two during inrush) I doubt if it is an IDMT earth fault trip.

If my assumption is correct, you can either incorporate an instantaneous time-delay (as suggested by above posters) or increase the instantaneous current setting.

 

[DiscoP]

Another option worth consideration is a stabilising resistor in series with the earth fault CT input, this would allow you to keep the protection instantaneous and keep the same pickup value.

 

[AEMB]

Thank you for the valuable information. There is no instantaneous setting on the relay. The relay is set to definite time with a 0.1 s time delay. Interestingly one of the downstream breakers also tripped on earth fault at the same time. I noticed that the time delay was set the same on both breakers. Is it possible that both sets of CTs saturated at the same time initiating trips on both breakers?

 

[RalphChristie]

AEMB

What type of relays do you use? Older electromechanical relays or electronic relays?

0.1 seconds is to short, (5 cycles on 50hz) it is almost instantaneous. Remember that inrush current is very high for the first few cycles, and then it decays. When you use a DTL-relay (definite time lag) you'll have to increase you time setting.

Can't you rather use an IDMT-relay? Normally, on a solid grounded system, I use a current setting of 10% (of full load current) and a very low time-multiplier (say 0.05 or 0.1 on a normal inverse curve) - this is very conservative. You also have to coordinate your up-stream relay with your down-stream relay. 0.3 seconds (electronic) or 0.4 seconds (electromechanical) grading between relays are normally sufficient.

Regards
Ralph

 

[Reactorman]

Inrush current will be at a maximium if the breaker closes on a voltage zero. Of course this can only occur on one of the phases since the other two phases cannot be also on a voltage zero. The book "Transformer Engineering" by Blume includes a formula for this calculation, however you would need to know a lot about the transformer design to make the calculation. Maybe you could ask the manufacturer if they have calculated this inrush current.

Good Luck

 

[Alex68]

AEMB

My opinion is that maybe CTs didn't saturated.
The inrush current of a transformer is like the current required to energise a R-L circuit. That is characterised by a strong unidirectional component which is realised like a 1ph fault by the protection.

The residual current configuration is normally affected by this error. I suggest to use a unique thoroidal CT involving the 3ph.

I prefer not to add a stabilising resistor to avoid over-voltages.

If it is not possible to delay the DT function, because the fault current could be very high, then I suggest to restrain the function with the 2nd harmonic.

 

[AEMB]

Apologies the actual time delay is 0.5s and it is also the same for the downstream breaker. Is it possible that the inrush can last that long. My thinking is that it should have substantially decayed after 0.1 s.

I have the curve from the manufacturers for one of the transformers. It indicates current decays from 1300A to 350 A in 0.5s. (Rated current 173A). There are four transformers in total. That means the current in the incomer goes from 5200A to 1400A in 0.5s). The CT ratio is 300/5 so you would expect them not to saturate.

Some interesting points regarding the use of core balance CTs and harmonic restraints, hoever I'm hoping I can solve the problem without additional kit.

 

[davidbeach]

The CT ratio is 300/5 so you would expect them not to saturate.

The ratio is not the only factor in determining whether or not the CTs would saturate. The type of CT and the amount of burden are important factors. If they are C800 and the burden is a fraction of an ohm it is most likely they didn't saturate; if they are C50 with a 10 ohm burden they probably did saturate. IEC designations would differ, but provide similar results.

I don't believe that you've ever said what current your ground fault element is set for. If set too low, the 0.5 sec time may not be any where near long enough.

The loading of the transformer is also a critical factor here. If you are energizing unloaded transformers, secondary breaker is open, the inrush will take much longer to decay than if you are energizing transformers which have full load connected. The more resistance on the secondary of the transformer, the faster the inrush will decay.

So, unloaded transformers, with a low set ground (earth) element set at 0.5 seconds, could very well still be producing enough inrush to trip at that time, even if the CTs are not saturated.

I've no idea what your installation looks like, but if you have numeric relays, you should be able to get the oscillography for the event and see what happened.

 

[daviandgo]

"I have the curve from the manufacturers for one of the transformers. It indicates current decays from 1300A to 350 A in 0.5s. (Rated current 173A). There are four transformers in total. That means the current in the incomer goes from 5200A to 1400A in 0.5s)"

I believe the simultaneous energization of four transformers can cause the total inrush to be higher than four times the individual inrushes. And as davidbeach said the inrush current measured in individual phases may differ considerably (higher residual).

Daviandgo