Switchboard Short Circuit Rating

[sibeen]

I'm doing a study on a relatively old building and have run across something that I'm not sure how to handle.

The building has a main switchboard which has a design kA rating of 63 kA for 1 second. The main ACB circuit breakers have a Icu / Ics / Icw rating of 75 kA. The board is fed by three emergency generators, each of 2250 kVA at 400 volts, 50 hZ.

Te generators have a subtransient reactance of 12.6%, transient reactance of 17.6% and a zero sequence reactance of 3.16%. The feed from the generators to the switchboard is approximately 30 metres. It is slightly more complex than that as the third generator has lower subtransient and zero sequence impedances.

When I do a calculation I find that the symmetrical RMS short circuit current for a three phase fault is approximately 78 kA, which is above the circuit breaker and board rating. This current drops very rapidly, so that after approximately 1 cycle is is down around 55 kA.

When I do the calculation for the single phase short circuit current (L - E)I get a figure of approximately 88 kA. Again this current falling off rapidly.

I'm not a switchboard person, so my question basically is, what should the fault current rating of the switchboard and the circuit breakers be in this situation?

 

[dpc]

Since you don't give us a location or local code reference, it's hard to answer. In the USA, this would be an NEC violation. Below 1000 V, the 1/2 cycle momentary fault current is used. The rate of fault current decay is not a factor. Low voltage devices have simply a "short circuit rating" or "AIC". The one second rating sounds like a withstand rating for the bus, not the circuit breaker?

 

[sibeen]

dpc, thanks for the quick response. I'm in IEC land for code purposes. The thing that has thrown me is the single phase fault. I've tried looking around for references, but they all seem to concern themselves with a bolted three phase situation; and whenever I've seen a reference to single phase it assumes that it is going to be less than the three phase.

I'm 99% sure that this is an issue no matter what code you are using, but I just want some form of assurance that there isn't a 'caveat' hidden somewhere before I raise this as a major defect with the client.

 

[dpc]

From what little I know of IEC ratings for low voltage devices, it is an issue there as well. I generally advise the client that the first step is to verify that the equipment rating data is correct. That gives them a little time to come to grips with the situation.

 

[FreddyNurk]

If the board has a type test certificate available, there were provisions in the standards relevant for testing (might be the predecessor to IEC60439) to allow for a higher current for a shorter time. I've seen various considerations on it, including attempting to extrapolate a higher current for a shorter time period without any testing data, blindly approving the board for service on the basis of short fault duration and others, though the approach that I'd want to take is to verify busbar bracing and withstand for the short duration (its obviously not a thermal issue for 1-2 cycles noting the decay, but there's still the question of how much bracing is required for that very short time period).

Obviously if the test certificate turns up and they did exactly that (say, 80kA for 1/2 second) then you've got no issues at all, but I suspect that the original certificate won't be able to be found. Beyond that it becomes an exercise in how much risk is held against the consideration of whether a fault occurs when all generators are connected at once, which is often a similar argument to grid connected generators synchronising onto a board and exceeding the board rating for the duration of load transfer, an associated argument I'm not particularly comfortable with either.

 

[ScottyUK]

You need to look at the assymetric peak current, especially when there is local generation, because that will determine the peak mechanical force on the bracing. The answer 'might' be in the IEC 60947 series of standards but I'm away from my desk. I'll have a look later if I remember.

I don't know which legislation you're working to but under British law there is no dispensation whatsoever for exceeding the capability of the equipment, even during short-term paralleling.

 

[LDKGR]

Hi sibeen,

you didn't write anything about the distribution philosophy. All these generators work in parallel ?

 

[sibeen]

Scotty, I'm in Australia so I assume the rules would be very similar. I've had a look through AS3439.1, which is equivalent to IEC60439.1. Truly my eyes did glaze over The peak current is mentioned, but only given by the short circuit power factor. Doesn't really say much else.

LDKGR, yes the generators operate in parallel in an N+1 arrangement.

I'll probably be able to get around the problem, and not cause the client too much grief, just by changing the generator control scheme and ensuring that only two generators are connected to the bus at any time.

 

[LDKGR]

sibeen, in my opinion you should install couplers to split the switchboard (I know that it's not cheap or easy). In this way you can reduce the SC and maximize the flexibility of the installation. If my suggestion is impossible, pay attention that in your way you need a second level of insurance (locking).

 

[FreddyNurk]

Bear in mind that often in an N+1 arrangement all generators may need to be connected to the bus for changeover, if its an islanded system, thus it gets somewhat harder to implement a control system modification to prevent exceeding the rating.

The predecessor standard to AS3439 used to detail the tests required, at least it did if my memory serves correctly, and it was possible to test for greater current at shorter duration. Every time I've been up against this, though, the test certificate has been nowhere to be found.

ScottyUK's correct, its the peak current and bracing, most of the smaller generation I've seen won't maintain a significant fault level for more than about 10 cycles without a significant drop, though I'm aware the scale of equipment ScottyUK deals with does (well, something has to, otherwise there'd be no perception of stiff grid...).

If its a backup system how do they transfer back to grid?

 

[sibeen]

Freddy, I'm on site tomorrow, so I'll have a look around, but can't say I'm hopeful of finding anything like a test certificate. The board was built in '91.

Saying that, I assume I'm stymied anyway, as the single phase fault current certainly exceeds the circuit breaker fault rating by quite a margin.

As to the current drop off. I did the calculation for the three phase fault. At the generator terminals a fault of 28.6 kA at t = 0. At t = 40mS (2 cycles) it has already fallen to 16.5 kA. So there is going to be a large force generated on the busbars with the initial fault, but it falls very rapidly off in strength. The 16 kA fault falls off reasonably slowly after that as we enter the transient reactance period. At t = 1S I still have around 9 kA.

The board is set up with six ATSs, each consisting of a 2 x 4000 amp ACB for the mains and the generator bus. So once mains is available they just do an open transition change on each ATS.

 

[ScottyUK]

Sibeen,

BS EN 60947-2 requires that for circuit breakers of breaking capacity greater than 50kA the test conditions be as follows: short-circuit power factor = 0.2; time constant = 15ms.

For breakers with a breaking capacity between 20kA and 50kA the power factor increases to 0.25 with a time constant of 15ms.

 

[FreddyNurk]

sibeen, you've obviously got the nameplate for the board, if you can find the manufacturer, some of them keep test certificates and records, or can point you in the direction of the relevant documentation. Last time I went through this the board was made in '76 and the manufacturer still had the original drawings (but no test certificate...). Granted, in recent times I've seen more recent boards with no details and no manufacturer support. Since you're likely looking for a type test certificate rather than one for the specific board, there's still a chance.

 

[waross]

When I have been faced with an issue of the Available Short Circuit Current of the transformer bank being greater than the ASCC of the service entrance equipment, I have considered the impedance of the connecting conductors. Fortunately the feed was fairly long and served to reduce the ASCC at the service entrance below the rating of the equipment. This was accepted by the AHJ.
A cheap fix may be to lengthen the cables from the generators.
I have also seen this solution employed by other designers.

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

 

[sibeen]

BS EN 60947-2 requires that for circuit breakers of breaking capacity greater than 50kA the test conditions be as follows: short-circuit power factor = 0.2; time constant = 15ms.

Scotty, yep, that seems to be the standard in the IEC world, AS3439.1 gives the same figure, it's equivalent to an X/R ratio of 5. Plug that into the appropriate formula and it means that the peak asymmetrical current is 2.2 times the RMS rating of the circuit breaker. So my 75 kA circuit breakers have an asymmetrical peak rating of 165 kA, of course assuming that 25 year old Masterpact breakers were tested to the same limits. The new one's certainly are...but. I'd better call Schneider on that one.

Now I need assurance that the same thing can be applied to the switchboard busbar and bracing system. I talked to one switchboard manufacturer who told me the type tests for his 100 kA boards state that they have a peak withstand of 220 kA. So that seems to be exactly the same. Now if I could be assured that the 25 year old 63 kA switchboard that I have had the same test certificates I'd sort of know where I stand.

Saying that, with the three generators supplying the main switchboard I'm going to exceed the symmetrical RMS rating and the peak asymmetrical ratings anyway, albeit for a very short period.

The client actually wants to put another generator onto this system, one of the reasons I've been kicking tyres on the site. I suspect I may have to suggest an intermediary generator switchboard, with higher kA rated circuit breakers, higher bus rating, and using the extra cable length (as per waross), and use the current limiting aspect of the new circuit breakers to sort out the issues.

 

[ppedUK]

Sibeen,

I this board split into busbar sections? If not, how are you managing to run 3 x 2250kVA generators on the same bar? In IEC land, the largest thermal rating of LV switchboard is 6,300A, but your 3 x 2250kVA gens = 9,740A. You could only run 2 gens at a push on a 6,300A rated board, so why would you want to add another gen? If the bar is split then you shouldn't have the fault level problem.

BTW, 4000A ACBs are not normally current limiting.

 

[sibeen]

ppedUK, the current configuration of the generators is as an N+1 arrangement onto a 6000 amp generator bus. Although the generators are 2250 kVA, this is the standby rating, and for the purposes of loading the prime rating is used, so are effectively 2000 kVA sets.

In IEC land, the largest thermal rating of LV switchboard is 6,300A

I've had a read through IEC60439 part 1 and cannot find anywhere that states that the maximum bus rating is 6300 amps. Can you please expand on that.

Cheers.

 

[ScottyUK]

I don't think there's a limit in the standard, it's just that manufacturers don't generally make ACBs bigger than 6300A because the copperwork becomes difficult and thermal management becomes challenging.

There are MV breakers specifically for generator applications which thermal ratings in the 20,000A range but they are a completely different type of device and are both massive and expensive.

 

[LDKGR]

Sibeen, I am experienced in "IEC land" and I know that switchboard FATs in the area of 6300A isn't an easy stuff even for "famous" manufactures. As ScottyUK wrote thermal management is the big problem. Who is the manufacturer of your switchboard ?

 

[healyx]

Hi Sibeen,

The breakers in this board have to satisfy 2 different ratings. The breakers from individual generator feeds will not see the 88kA calculated total, they will only see their share of that (1/3) or 30kA rms. Breakers feeding other loads from this combined bus do need that 88kA+ rating.
If it is a open transition n+1 scheme, doesn't that mean the bus will only ever see 2/3 of 88 (approx 60kA). If that is the case, would that come under the 63kA board rating?
In answer to your other question regarding 3 phase faults being larger that L-N faults. It is common for generator peak currents to be L-N due to the relative magnitude of the zero sequence impedance compared to transient quantities. Zero sequence impedance is a component of L-N faults, but not L-L-L faults. I suspect you could consider a neutral earthing resistance/reactance on the generator to increase zero sequence impedance and bring down peak fault current, although I am not an expert on this.
I had a quick calc with assumed cables and got about 92kA peak if 3 generators can parallel and 62kA if an open transition interlock with only 2 generators in parallel used. If it is the later, it looks ok to me.
One thing I noticed is that if you are going to try to retrospectively satisfy the arc fault requirements as per AS3000 (which would not have been in effect in 1991), or even just to ensure you pick up the minimum fault current (about 13kA assuming 4x500mm2 cables) you will need to wind the short time setting on the breaker back a long way and check the knock on effects from this downstream.