Arc Flash Fed by Generators-What's The FCT

[ThePunisher]

Hi all,

I have this dilemma of a system for a 575 V, 3 phase pump motor with starter fed directly from a 600 V, 3 phase generator.

The generator is rated 56 kW, 3 phase, 60 Hz but the generator circuit breaker is 200 AT MCCB, Square D type JD. The power cable between the generator and the starter is only 3C-4AWG, RECK, 1000V.

1. Circuit breaker is oversized and would require 70AT MCCB to protect the 4AWG cable. Since the generator is only rental and we have schedule issues, we decided to simply add the 70 AT MCCB between the generator 200AT breaker and the starter. We will install the 70AT MCCB 3 meters away from the generator.

2. When we ran arc flash study, there is no issue between the 70AT MCCB and the power panel.

3. However, the 200AT generator circuit breaker's fault clearing time (FCT)for generator contribution of only 560A is resulting to a large arc flash energy at the line side of the 70AT MCCB.

Someone informed me that generators normally are not able to sustain the arc current and should not be an issue. However, my dilemma is that HOW would we simulate that in ETAP? How much FCT we would consider (I was thinking to assert 2 seconds), but I don't have any basis.

I believe, the AVR would simply give up during fault conditions and shutdown the generator...but I do not have any data on how long would the generator shifts from substransient to synchronous and shut down. If I have that duration, I may consider that time as FCT and assume sustained subtransient current to be conservative. I was thinking of checking the ETAP typical decrement curve but still the FCT total must be assumed for the generator.

I appreciate any help or guidance. Thank you very much.

 

[mparenteau]

Can you get the decrement curve from the manufacturer? I'm not familiar with ETAP, but SKM allows you to "hand" input custom curves such as this.

Mike

 

[ThePunisher]

I do not have the generator manufacturer decrement curve but ETAP is able to generate one using typical data of Impedances and time constants.

 

[ThePunisher]

Thanks mparenteau.

I do not have the actual decrement curve but I have an ETAP generated curve based on typical impedances and time constants generated by ETAP.

see above post.

My concern with the decrement curve is it shows up to synchronous long time....but can't estimate how much time the AVR/Exciter would shut down the generator

 

[mparenteau]

I could be wrong here, but I would think the decrement curve would hold over the time period (1000s), as in theory, the genset needs to provide enough short circuit current for OCPDs to clear, i.e. no AVR/exciter shutdown. Even if that means a sustained fault. Now you have me curious...

Mike

 

[ThePunisher]

mparenteau, I totally agree with you.

The decrement curve would indicate the resulting short circuit current based on subtransient, transient and synchronous reactances over time. However, the curve does not consider any time to shutdown by the AVR/Exciter/Governor.

In our case, since the generator circuit breaker is oversized, or even, the resulting short circuit current is too low to hit the magnetic element of the MCCB, it will take some time to trip it. An current overload relay at the generator can do the job of faster clearing, but we do not have that.

Line side of the 200AT, that's generator's controls now. But my concern is the line side of the 70AT where the 200AT will only trip around 50 + seconds on bolted fault and worst on arcing current (85% of bolted).

Hence, I would rely on generator controls for the fault clearing.

 

[mparenteau]

Maybe the AVR manual prescribes a field current limit...which would give a time to shut down the AVR?

 

[waross]

It is not safe to AssUMe anything about the AVR.
Self excited AVR:

The AVR power is derived from the output terminals of the generator. A fault causes the terminal voltage to drop which in turn cuases the AVR output to drop. Voltage collapse is rapid.​

Self excited AVR with field forcing:

In order to force a generator to develop enough fault current to operate protective devices, A CT(s) may be used to generate a voltage across a resistor. This voltage is rectified and fed to the field to mitigate voltage collapse and hold up the fault current so as to facilitate the operation of protective devices.​

PMG powered AVR:

A small permanent magnet generator is added to the end of the generator shaft to power the AVR. These AVRs are much less sensitive to voltage collapse.​

Iron AVR:

the output of a constant voltage transformer is fed to a tapped resistor. The generator no load voltage is set by selecting the appropriate resistor tap. CTs monitor the load current and the CT outputs are dropped across a suitable tapped resistor. A voltage is tapped off and rectified and added to the field voltage to compensate for loading. A feature odf this system is that, by the selection of the proper tap, The output may be set to droop with increasing load, remain constant with increasing load or increase with increasing load. In some applications the voltage may be set 8 or 10 percent low and a conventional AVR added for more precise voltage control. An advantage of adding electronic AVR is that in the event of failure of the electronic AVR the generator will still produce usable power at a slightly lower voltage.​

Capacitor excitation, mostly smaller machines, less than about 10 or 20 KVA:

These generators are self excited by transformer action of the load current. The voltage is set by the value of an external capacitor. (I have used a lot of these. One experience with a short circuit when the generator breaker tripped instantly.)​

You need to know what EMF will be induced by the AVR action under fault conditions. The fault current may be driven by less thyan rated voltage.
A work around: Flag the area between the generator breaker and the second breaker with arc flash warnings.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[aussiejohn2]

Waross has it right: it depends. Decrement curve is often (always?) drawn for constant excitation. If you have PMG driven excitation system that can sustain high levels of field forcing, output will be sustained until (/unless) over excitation limiter or rotor thermal protection kicks in.

 

[FreddyNurk]

AVRs, unless specifically referenced in the list provided by waross, won't drop all excitation for at least a set time period (the units I've familiar with state 10 second capability).
I've not used ETAP for a while, but it can definitely produce a decrement curve (which can include the requirement to sustain excitation under fault conditions).

Where it gets complicated, if my experience with other software packages is anything to go by, is whether or not the decrement curve is taken into account when calculating arc flash hazards.
It does sound like part of your issue is that the SC trip on the generator breaker is a general duty one rather than a lower level unit specifically for generators, in which case there's no much you can do about it for a rental set.

If there is a generator controller onboard the set that also does overcurrent you may be able to use those settings if they're better suited to tripping under such conditions and incorporate them into the model. Otherwise waross' suggestion of restricting the area is a good one.

EDMS Australia

 

[dpc]

I believe, the AVR would simply give up during fault conditions and shutdown the generator.

You can't assume this. Many exciters have an excitation support or boost system that can increase fault current substantially. As noted, the generator decrement curves assume a constant excitation.