limiting short circuit current at chiller

[malachi constant]

I am the engineer on a project currently under construction. We have a chiller being installed in the vicinity of a 208V utility transformer. The chiller is fed directly from the transformer, with a fused switch at the chiller. The electrical inspector noted that the nameplate data shows the chiller (which is on-site) has a nameplate SCCR of 5kA, and he stated that he expects the available short circuit current to be less than this at the enclosure.

The available current at the utility transformer is on the order of 80kA, by the time it reaches the disconnect it is on the order of 25kA. Using the over-up-down method on the 500A fuses yields a let-through current on the order of 10kA. As near I can tell he is correct that this does not meet code, is therefore unsafe and will not pass inspection.

I have done some preliminary research and believe my best options are to install an impedance transformer or see if modifications can be made to the control panel to increase the SCCR. Both of these would be new territory for me. Does anyone hear have any insight as to which is more feasible / cost effective? Or have other viable solutions?

On a related note, how does one go about preventing this? I typically either perform the short circuit study prior to issuing the drawings, but in this case the Owner wanted a third party to perform them. In this case I still ball-parked the panel ratings. Do I now need to be coordinating known or estimated SCCR ratings with the mechanical engineer prior to him writing his spec/schedule? Should mechanical start specifying minimum SCCRs at chillers and other HVAC equipment? (Or whatever the code says needs rated - I don't have it in front of me.) Previously just getting panels and switchboards coordinated was sufficient, but it looks like I need to step up to a different level. Makes sense, and is probably not that big of a deal...but I doubt it is standard practice for 90% of the engineering world, and the coordination involved is a little daunting.

Anyway. The question at hand is what to do now, with this situation. Thoughts on those options? Thanks!

 

[GoldDigger]

Can you find current limiting fuses that are rated to limit to 5kA?

 

[malachi constant]

Excellent question. I looked at two Bussman fuses - forget which off the top of my head. Both were stated to be excellent for (or best at, or something like that) limiting short circuit current. That's what I was hoping would do it, but it's not looking promising. If there are specific fuses I should look into I gladly will, but I don't know where to start and am not holding my breath.

 

[augie47]

I had a similar install a few years back and then only solution the engineers came UP with was installing current limiting reactors.
What size load are we addressing ?
ON small loads I have seen the problem solved by increasing wire lengths considerably.

 

[iwire]

Can you find current limiting fuses that are rated to limit to 5kA?

Do they really limit current? The first fault cycle would still be full current would it not?

I don't know that is why I am asking.

 

[malachi constant]

I had a similar install a few years back and then only solution the engineers came UP with was installing current limiting reactors.
What size load are we addressing ?
ON small loads I have seen the problem solved by increasing wire lengths considerably.

Have this at a couple chillers. One about 450A, one a hair over 800A, both 208V. I considered increasing wire lengths but seems very impractical for loads of this size. Think it would cost a bazillion dollars and cause more harm than good.

 

[jim dungar]

The requirements that all equipment be rated for the available fault current has been part of the NEC (article 110.10) for more than 35 years. Our Short Circuit studies often identify VFD and ATS equipment that is not sufficient.

The default SCCR value is most often 5kA, so that tells me your chiller manufacturer was 'too lazy' to determine the actual SCCR.

Change your mechanical specs.

 

[kwired]

The default SCCR value is most often 5kA, so that tells me your chiller manufacturer was 'too lazy' to determine the actual SCCR.

I can buy into that. How often does this kind of equipment end up near the supply and has a high available fault current? Probably over 90% of the time. Same manufacturer will cry that you need a more stout supply if you have voltage drop issues - but if you have voltage drop issues chances are available fault current is not all that high:roll:

 

[dkidd]

Here is a link to a document, and an upcoming webinar:

http://www1.cooperbussmann.com/pdf/b748ff0b-1cc6-406b-b1e7-5264a02a5324..pdf
http://www.graybar.com/company/even...ngs-electrical-equipment-industrial-machinery

Also, there have been some discussions on the forum that you might find helpful.

 

[ron]

Here is a link to a document, and an upcoming webinar:

http://www1.cooperbussmann.com/pdf/b748ff0b-1cc6-406b-b1e7-5264a02a5324..pdf
http://www.graybar.com/company/even...ngs-electrical-equipment-industrial-machinery

Also, there have been some discussions on the forum that you might find helpful.

Bussman has always been a proponent of current limiting fuses and the up-over-down method, although I think they changed the name / format in some publications.

Since it is not a tested combination, I do not use fuses to reduce fault current for equipment ratings. I often run feeders long. You can also use an isolation transformer or reactor. In all three cases, be careful of voltage drop.

 

[petersonra]

My boss decided at some point in the distant past that if the customer does not otherwise specify a desired SCCR that we will design for 10kA.

This works pretty OK for most panels, but when you have a 1000A or 1600A CB it makes little sense to make such an assumption.

If the chiller people used fuses in their design there is a pretty good chance that they could legitimately make some kind of calculation that states what the real SCCR is, and it is probably well in excess of 5kA.

The real problem is that people buy stuff and don't tell the supplier what they need so the supplier does whatever is the most cost effective. HVAC suppliers are notorious for this, but I cannot blame them. It is a very competitive market segment, and the reality is that going above and beyond the bare minimum adds cost. If the customer is unwilling to pay for that added cost, IMO that is the customer's fault and not the supplier.

 

[winnie]

The 'up-over-down' method is not appropriate when applied to systems that place multiple overcurrent protective devices in series.

A current limiting fuse will respond so quickly to a bolted fault that you never see a _full_ cycle of current. The I^2*T value is reduced so much that the stresses on downstream passive devices will be reduced as if there were less available short circuit current. My understanding is that current limiting fuses (and the 'up-over-down' calculating technique) are perfectly acceptable for protecting passive things such as bus bars or elements that are not actively trying to respond to the overcurrent.

But dynamic impedances; devices that are explicitly trying to change state in response to the overcurrent will see the _entire_ fault waveform until such time as one of the devices breaks the current flow. A device rated for 5KA might not be able to survive the _initial_ part of a 50KA fault, or in attempting to open against that 50KA fault, that 5KA device might prevent the 'current limiting' device from doing its job. Whatever the reason, without testing the series combination you don't know that the 'current limiting' device will properly protect the downstream OCPD.

Back to the original question, with a question for other forum members: would it be appropriate to use a 'line reactor' (such as are often used with variable speed drives) as a method for reducing available short circuit current? Would it be appropriate to make inductors with 'air coils' of feeder wire? What about running the feed as 'isolated phases' to get a larger loop area between the phases and introduce some current limiting inductance?

-Jon

 

[petersonra]

I think making your own inductors is a bad idea.

If you need to add impedance, just do it.

I think I would start by looking at where the SCC came from. A lot of times the numbers come from a so called infinite bus calculation and result in much higher numbers than are really available.

 

[ron]

would it be appropriate to use a 'line reactor' (such as are often used with variable speed drives) as a method for reducing available short circuit current?

I think a line reactor is a valid way to reduce short circuit current, however an air core reactor will have lots of magnetic fields "flying around" and may cause problems in the immediate vicinity of where it is installed if there electronics / controls.

I have used solid core line reactors, but often it is double checking calculations for data errors, and running the feeders long.

 

[pfalcon]

...
The real problem is that people buy stuff and don't tell the supplier what they need so the supplier does whatever is the most cost effective. HVAC suppliers are notorious for this, but I cannot blame them. It is a very competitive market segment, and the reality is that going above and beyond the bare minimum adds cost. If the customer is unwilling to pay for that added cost, IMO that is the customer's fault and not the supplier.

30kA for our heavy industrial machining site. Specified in our Annex B to the NFPA79. Our buss fuse is also listed there.
Panel SCCR of 5kA is definitely the lazy method.
Bussman publishes tables for tested combination fuses with their products.

 

[petersonra]

30kA for our heavy industrial machining site. Specified in our Annex B to the NFPA79. Our buss fuse is also listed there.
Panel SCCR of 5kA is definitely the lazy method.
Bussman publishes tables for tested combination fuses with their products.

You can certainly make up any spec you want but if your purchasing department does not enforce it, it really does not do much good.

The HVAC market, especially on the commercial side, is notoriously cheap. If they can save $5 on a $100k order, they will do it. I have seen HVAC projects where several different brands of the same type of fuse are used because a 60A class J fuse from one company is 50 cents cheaper than the 40A class J fuse also used so there is one brand of class J fuses for the 40A fuses and a different one for the 60A fuses.

It is not really a matter of laziness. It takes engineering time to take a close look at a project to make sure it really meets whatever SCCR is agreed on. It may not seem like a bunch but an extra few hours of engineering time is a couple hundred bucks and on projects that can be lost for that kind of money, I can understand why someone would not want to add cost unless they have to.

Why do you think a lot of HVAC motors run closer to their service factor limit a lot of times instead of their nominal ratings? It is solely about cost.

 

[Jraef]

... The default SCCR value is most often 5kA, so that tells me your chiller manufacturer was 'too lazy' to determine the actual SCCR.

Change your mechanical specs.

I can buy into that. How often does this kind of equipment end up near the supply and has a high available fault current? Probably over 90% of the time. ...

... The real problem is that people buy stuff and don't tell the supplier what they need so the supplier does whatever is the most cost effective.

And in that last statement is the lesson to be learned. When you leave out that detail, the low bidder will invariably be the one who pays no attention to that aspect and takes the "lazy" way out by slapping a 5kA number on it and kicking the can down the road. The solution (for future projects) is to add wording to the effect of; "All electrical panels for equipment will be properly rated for installation in the available short circuit fault current environment that exists at the location." The low ball bottom feeders will cry foul, but the responsible ones will do their homework, or ask what the AFC is at the switchgear feeding their equipment and design to that (if they can) rather than try to guess about final values at the terminals.

 

[augie47]

and, unfortunately, way too many inspectors won't even address the issue.
My boss tells me "you are the only inspector where I ever hear 'fault current' " as if it's a negative issue to him. :happysad:

 

[don_resqcapt19]

... The solution (for future projects) is to add wording to the effect of; "All electrical panels for equipment will be properly rated for installation in the available short circuit fault current environment that exists at the location." The low ball bottom feeders will cry foul, but the responsible ones will do their homework, or ask what the AFC is at the switchgear feeding their equipment and design to that (if they can) rather than try to guess about final values at the terminals.

I don't see that as the real solution. It is not the equipment suppliers job to dig up that type of information. It is the designers job to provide that information to the equipment supplier.

 

[texie]

and, unfortunately, way too many inspectors won't even address the issue.
My boss tells me "you are the only inspector where I ever hear 'fault current' " as if it's a negative issue to him. :happysad:

I agree. On the properly engineered jobs most EEs do a good job of detailing the ratings required at every point in the system. One issue that I see too often on well engineered jobs is the EC/electricians not realizing that all that preassembled gear and panelboards they have on the job, not all breakers are the same AIC rating. So when they find they need a breaker for an extra circuit in a given panel they just go find a spare breaker from another panel that they don't realize has a lower rating. Had job recently where we discovered a lot of this. When I asked the journeyman about it I got the standard answer of " they look the same to me". I gave him 2 QO breakers, 1 a 10 K and the other a 65 K and asked if he could tell the difference. He swore they were the same breaker. We need more education out there.