Paralleled gensets and fault current

L

lielec11

Senior Member
Location
Charlotte, NC
Occupation
Electrical Engineer (PE)
We have a system where three (3) 2000kVA gensets are paralleled onto a common switchgear. When we run our fault calculations in SKM we assumed all generators are running at the time of a fault. Therefore at the time of a fault, as a rotating machine, each generator will contribute fault to the system for a short period of time like a motor would. We submitted our report as such showing the three (3) generator mounted breakers (NOT on the paralleling gear) failing because the combined fault current from all three gensets is greater than the AIC of the breakers (100k). The gear itself is fine as it is rated 200kA.

Here is where we disagree with the EOR on their response back. They are trying to direct us to run the fault calculations for the generator breakers as if the fault itself is isolating one generator from the system. They're claiming that the genset breakers will only see contribution from the OTHER two gensets at the time of a fault.

I disagree because if there is a fault when all three gens are running, all three will contribute fault current to that device. I understand that if there is a fault at one of these breakers it will isolate one generator from the rest of the system, but it will still contribute fault current to the main CB in question. Therefore, the fault current MAY be to high for the breaker to actually isolate the fault as intended.

Does this make sense? Am I a moron?
 
It would depend on how the generators are paralleled. I see both of your points.

If the 100kAIC rated breakers are used for the individual breakers to each generator than I agree with the EOR. If the paralleling breaker, or transfer of that main bus is rated 100kAIC than I would be inclined to agree with you and would take a closer look at the way the protection is put in place. I don't know if series ratings would apply.
 
I'm not sure how this snippet is going to translate here but I hope you can see the general idea. The screenshot shows the results with all 3 running. Based on your analysis and this image I think you're leaning towards their side of the argument. The device/bus in question is PD-GEN-G1 MAIN and BUS-0029 (the same can be applied to G2 and G2 main as well though).

Where am I wrong in my thinking that BUS-0029/PD-GEN-G1 MAIN would not be exposed to the fault from all 3 gensets? This is all theory and I'm clearly missing something simple.

1749568368703.png
 
lielec11 said:
Where am I wrong in my thinking that BUS-0029/PD-GEN-G1 MAIN would not be exposed to the fault from all 3 gensets? This is all theory and I'm clearly missing something simple.
Click to expand...
In your diagram, if the fault occurs between GEN-G1 and PD-GEN-G1 MAIN, the device PD-GEN-G1-MAIN will only see fault current from the other two generators. The fault current from GEN-G1 does not pass through PD-GEN-G1-MAIN for that fault location.

And if the fault occurs anywhere else, then PD-GEN-G1 MAIN will only see fault current from GEN-G1.

Cheers, Wayne
 
wwhitney said:
...if the fault occurs between GEN-G1 and PD-GEN-G1 MAIN, the device PD-GEN-G1-MAIN will only see fault current from the other two generators. The fault current from GEN-G1 does not pass through PD-GEN-G1-MAIN for that fault location...


Cheers, Wayne
Click to expand...
The fault doesn't PASS THROUGH the device, agreed, but shouldn't the fault current be included AT the device? Without looking up a specific definition, the AIC rating is the amount of fault current a device can handle while still operating as intended (i.e. close on a fault). So, shouldn't we count all the potential fault current the device may see when evaluating it?
 
wwhitney said:
The device only "sees" current flowing through it.

Cheers, Wayne
Click to expand...
So why does SKM not run their evaluation this way (I know, probably a question for them)?

I've shared another snippet below, this time showing the fault currents when we fault BUS-0029 directly. You can see it aggregates both the current from the other two gens (~72kA) with the fault current from GEN-G1 (~62kA). Based on our theoretical conversation this is incorrect.

EDIT: I also moved BUS-0029 between GEN-G1 and the MAIN to better match your description, and the results are the same either way FYI.
1749570203790.png
 
lielec11 said:
EDIT: I also moved BUS-0029 between GEN-G1 and the MAIN to better match your description, and the results are the same either way FYI.
Click to expand...
But for this scenario, if the SKM fault analysis showed the current through PD-GEN-G1-MAIN, it would only be from the other two generators and would not include the current from GEN-G1.

As to the fault current shown of 134598A at BUS-0029, that is the current that would feed the fault itself. But the bus any distance away from the fault would not see that current, it would either see 62650A or 71961A, depending on which side of the fault we consider.

Cheers, Wayne
 
wwhitney said:
But for this scenario, if the SKM fault analysis showed the current through PD-GEN-G1-MAIN, it would only be from the other two generators and would not include the current from GEN-G1.
Click to expand...
But I don't think it does, see the image below. Clearly the ~72kA from the other gens goes "through" PD-GEN-G1 MAIN because it shows the total value of 134kA at BUS-0029 in this scenario too.

I think I've overanalyzing this, but I'm not sure how to show the results/explaining them when the software shows levels near 130 as opposed the 70 or 62 only. I am leaning towards agreeing with everyone else here but I want to be able to explain it succinctly, as well as have the numbers back me up.
1749570985174.png
 
Draw it with a pen with an arrow. I think you will get a better idea of what Wayne is saying if you do.

The current at the fault point is correct but that is not the current seen by PD-GEN-G1 MAIN. That 134kA is going to "ground". The current seen by PD-Gen-G1 MAIN is the fault current is CBL-GEN-G1.
 
Elect117 said:
Draw it with a pen with an arrow. I think you will get a better idea of what Wayne is saying if you do.

The current at the fault point is correct but that is not the current seen by PD-GEN-G1 MAIN. That 134kA is going to "ground". The current seen by PD-Gen-G1 MAIN is the fault current is CBL-GEN-G1.
Click to expand...
The arrows are on the diagram, what other arrows would be missing? Regardless, SKM still shows the devices in the report as "failing", which based on our discussion, we (me now included) do not agree with. So, either SKM has their software setup wrong, or we need to clarify in our report that the results as indicated by the software in this instance, are incorrect.
 
Elect117 said:
The arrow from G1 to ground and the arrow from CBL-GEN-G1 to ground trying to get back to their respective sources.
Click to expand...
I see what you mean now. The software still shows all 130kA at the device, so I guess we will just have to explain why in this case, that is incorrect.
 
You could get it to show the fault contribution from G1 by inserting a PI equivalent with negligible impedance between G1 and BUS-0029. Pick the symbol for the PI that looks like a cable. Then edit the datablock so your PI shows the same parameters as the cable branches.

BUS-0029 will show the higher fault levels as before, but the PI branch between there and the generator only shows the contribution from G2 and G3 going toward G1.
 
lielec11 said:
I see what you mean now. The software still shows all 130kA at the device, so I guess we will just have to explain why in this case, that is incorrect.
Click to expand...
As long as you are playing with your model; try adding a bus and conductor between breaker PD-GEN-G1MAIN and the generator, use a minimum amount of very large conductor. I am looking to see how SKM handles this when there is a clear separation between the line and load sides of the breaker.

gen -> trial bus -> trial conductor -> breaker -> Bus 29
and/or
Gen-> Bus 29 -> breaker -> trial conductor -> trial bus.

Looks like Joe beat me with the idea.
I remember have to use PI equivalents to fix SKM quirks.
 
I agree that the contacts of the set mounted breaker will not see more than 2 generator's maximum worth of fault current, but if the fault happens on the line or load side of one of the set mounted breaker, the breaker terminations will see 3 generators worth of fault current and needs to be braced for such. The standards for breakers generally do not differentiate the AIC rating of a breaker and the withstand rating of a breaker mounting / terminations, so you would still fail at 100kA.
 
ron said:
if the fault happens on the line or load side of one of the set mounted breaker, the breaker terminations will see 3 generators worth of fault current and needs to be braced for such.
Click to expand...
You mean if the fault is breaker terminal to breaker terminal (on the same side of the breaker, either line or load)? Are the magnetic force effects from the currents going into the terminal from the two different sides additive or subtractive?

Cheers, Wayne
 
JoeStillman said:
You could get it to show the fault contribution from G1 by inserting a PI equivalent with negligible impedance between G1 and BUS-0029. Pick the symbol for the PI that looks like a cable. Then edit the datablock so your PI shows the same parameters as the cable branches.

BUS-0029 will show the higher fault levels as before, but the PI branch between there and the generator only shows the contribution from G2 and G3 going toward G1.
Click to expand...
Can you elaborate on this for me? I've never used a PI equivalent before, let along for a hack like this. Also, wouldn't I need to do this on ALL the branches? Otherwise the other two breakers will show the full fault current over 100kA? What is the thought process behind the PI equivalent?
 
You must log in or register to reply here.
Top