Arc Flash Calculation - MCC with integral LV xfmr and panelboard

Status
Not open for further replies.
N

nhee2

Senior Member
Location
NH
We have a 480 VAC MCC with integral 3-ph step down transformer and 208V 3-ph panelboard. The incident energies calculated for the panelboard are greater than those calculated for the MCC. I believe this would require the entire MCC (it is small, only 5 or 6 sections) to be labelled with the higher incident energy of the LV panelboard. Is this correct? Under the older IEEE 1584 this would have been disregarded, but
 
the transformer and panelboard are typically in separate compartments from the rest of the gear. I don't see how under "normal" conditions the IE rating would extend outside of those compartments. I think they would if someone were working live there though. It might need to be labeled that way.
 
There is going to be two schools of thought here. Go the most conservative route and label the entire MCC the higher arc flash rating off of the secondary of the small transformer which likely can't even carry arc flash at 208V. (still under discussion by the "experts")

The other way is like petersonra says, "only label the sections with the transformer and the 208 panelboard at the higher arc flash label." assuming you can prove that the 208V arc flash will not leak around.

I think most engineers just go the conservative approach because why lose sleep at night? Then there is the guys who will yell at you or make fun of you for being so conservative.

Only question is this, "what would you do?"
 
thanks for the responses. In this case the transformer and panel board are actually in different vertical sections. while I don't see a practical way for someone working in one of the other sections to initiate a fault on the secondary in those two sections, we'll probably label the whole thing with the higher IE value.
 
powerpete69 said:
The other way is like petersonra says, "only label the sections with the transformer and the 208 panelboard at the higher arc flash label." assuming you can prove that the 208V arc flash will not leak around.
Click to expand...

IMHO
This is kind of like the issue with labeling a non-compartmented transformer terminals.
If an arc is created within the 208V panelboard would it begin to envelop the 480V MCC bus? If it eventually grew large enough to involve all 3-phases of the MCC buses, then wouldn't the arc flash energy be based on the 480V rating? As soon as the 480V MCC fault is cleared the 208v transformer would also lose power, unless it is fed from an external source.

If you are going to assume the 208V fault clearing time is accurate and will not be shortened, you almost have to assume the 208V fault stays contained.

Again, IMHO,
Stop putting panelboards into MCC buckets.
 
Thanks jim dungar.

I was not thinking of a 208V fault initiating a subsequent 480 V fault (if that was what you were suggesting), rather the 208V flash event impacting someone in an adjacent section/bucket. again i'm not sure there is a practical way for that to occur, but my understanding is that generally the MCC will be labelled for the higher 208 V incident energy
 
nhee2 said:
We have a 480 VAC MCC with integral 3-ph step down transformer and 208V 3-ph panelboard. The incident energies calculated for the panelboard are greater than those calculated for the MCC. I believe this would require the entire MCC (it is small, only 5 or 6 sections) to be labelled with the higher incident energy of the LV panelboard. Is this correct? Under the older IEEE 1584 this would have been disregarded, but
Click to expand...
I'm having a hard time figuring out how that could be. ALL of the energy for a fault on the 208V side has to come FROM the MCC bus. How could you possibly be adding energy to the circuit? If you were doing a SEPARATE IE calc on the 208V system and starting over with an "infinite bus" feeding the transformer, then that would be your mistake. The AFC on the MCC bus will be the starting point for determining the AFC on the 208V side.
 
Jraef said:
I'm having a hard time figuring out how that could be. ALL of the energy for a fault on the 208V side has to come FROM the MCC bus. How could you possibly be adding energy to the circuit? If you were doing a SEPARATE IE calc on the 208V system and starting over with an "infinite bus" feeding the transformer, then that would be your mistake. The AFC on the MCC bus will be the starting point for determining the AFC on the 208V side.
Click to expand...
Maybe not. Anytime you get on the secondary of a transformer, your arc flash calories go way up per what I've seen in EasyPower.
 
Jraef said:
I'm having a hard time figuring out how that could be. ALL of the energy for a fault on the 208V side has to come FROM the MCC bus. How could you possibly be adding energy to the circuit? If you were doing a SEPARATE IE calc on the 208V system and starting over with an "infinite bus" feeding the transformer, then that would be your mistake. The AFC on the MCC bus will be the starting point for determining the AFC on the 208V side.
Click to expand...
The duration of the fault is longer - takes longer for the xfmr primary protection to open on the low-level fault then it does for the main on a L-L fault on the 480 side. (I think)
 
nhee2 said:
The duration of the fault is longer - takes longer for the xfmr primary protection to open on the low-level fault then it does for the main on a L-L fault on the 480 side. (I think)
Click to expand...
Right. Its the longer duration in breaker trip time that builds the arc flash event.
 
Wanted to add my 2 cent.
You can increase the working distance to the max allowable number. That reduces the incident energy.
If you’re running the calc based on IEEE1584:2002, then apply the exception from 9.3.2 allowing to default everything fed by a xfmr smaller than 125kVA and less than 240V to 1.2cal.

In your case, I would look at the AF boundary. How many buckets would be affected in the worst case. And as the others stated, if it’s compartmentalized, I’d rate this panel with a higher incident energy only. It’s not even connect to the mcc bus bars but fed by a transformer.

Just imagine you have a service entry rated mcc with the main breaker in its own vertical section. The line side may be 68cal but the load side is 2cal. There is no way, I’d apply 68cal on each bucket. Why? If I short anything on the load side of the service disconnect, I’m exposed to 2cal only. I’d apply the 68cal label only on the section carrying the service disconnect.


Sent from my iPad using Tapatalk
 
Also, which it doesn’t look like it was mentioned, the danger in an arc flash event is based on the current in the event (in addition to time). So it’s not extremely strange for the IE to be higher at the secondary than it was for the primary, but this usually doesn’t happen unless you have a large KVA transformer. It’s the available fault kVA that’s transferred through a transformer, not kA, so if you have a zero impedance transformer (not real), you can see how you’ll have higher IE on the secondary.
 
I see most of the time that the current on the secondary time is too low to pick up the INST of the main principal. The PD arcing fault current overlaps with the STD leading to an arcing time of +2s. So I agree that current is a driver but it’s most of the time that the current is too low leading to a long tripping time. On the other hand, at the service entrance the fault current is so high that you still end up with a high incident energy even though you initiate a trip after 20ms.


Sent from my iPad using Tapatalk
 
Status
Not open for further replies.
Top