Incident Energy with CL Fuse Equations 1

[SJBatTCE]

Does anyone else have a problem with the IEEE 1584 incident energy equations for current limiting fuses?

Incident energy is based on the arcing fault current, not the bolted fault current. Arcing fault current is determined from the bolted fault calculation but takes several other variables into consideration, like system voltage, gap, etc. The CL fuse incident energy equations only use bolted fault values.

Why are the CL fuse equations based on bolted fault rather than arcing fault values?

For now, I am not using the equations in my analysis. Are my concerns unfounded?

Thanks.

 

[burnt2x]

The bolted-fault current available at a certain location/ equipment is used to calculate the theoretical arcing fault current, that is why.

You must have mixed understanding on these things. Equipment fault withstand is different from arc flash protection. We use the bolted-fault current to determine whether our equipment can withstand a full, bolted-fault short-circuit. If our equipment is underrated, we will be forced to find a higher-rated equipment. If the rating we are looking for is not available, or acquiring those could render the project not feasible, we are going to use CL fuses so that our equipment rating can be up to code.

The other item you have trouble with is compliance with arc flash protection. You can either set your breaker trip setting to the fastest possible to minimize the incident energy level, hence a lower category arc flash protection requirement or reduce the available fault current by using limiters (reactors, CL fuses, etc.). The determination of the arcing fault current (theoretically) requires the use of the bolted-fault current. that is why you need to know the bolted-fault current available at the location to be able to know the arcing fault current.
Others who are very good at CL fuses could chime in, patience!

 

[SJBatTCE]

I don't care about equipment withstand here. I'm conducting an arc-flash hazard analysis. I've already determined in another part of this study that the equipment is suitable for an infinite source bolted fault. Where did I mention equipment fault withstand?

In IEEE 1584 Section 5.2, an equation is given to determine the arcing fault level from the determined bolted fault at that location. Several variables are used, such as system voltage, gap between conductors, enclosed or not, etc. Section 5.3 gives an equation to determine incident energy based on many of the same variables. Fine.

Now, if applicable, IEEE recommends using the incident energy equations for CL fuses in Section 5.6. These equations are based only on the bolted fault values; no mention of system voltage, gap or enclosure type. How have they verified that the arcing fault value is in the current limiting region of the fuse curve?

I've made available a chart based on a 50kA bolted fault illustrating the change in arcing fault current with system voltage and conductor gap. You can see a rather drastic difference at the 13mm gap between 600V and 208V. In order to consider the CL fuse to have a current limiting effect, the arcing fault must be in the current limiting region of that fuse's curve. (See attachment.) I've also uploaded a TCC of the KRP-C CL fuse showing the highest and lowest arcing faults for the 13mm gap (dotted lines), along with the current limiting flags for the fuse. (Diagram linked in next post.) Obviously, the arcing current for a 600V system is within the current limiting region but the 208V is not.

Why, then, is it recommended to use the CL fuse equation based on bolted fault only, without any consideration for the arcing fault variables?

I haven't investigated this before because, until recently, SKM's application of CL fuse equations has been rather buggy. Recent releases have mostly fixed the problem and I've considered using the CL fuse option.

 

[SJBatTCE]

Post to link to KRP-C TCC.

 

[jghrist]

The IEEE 1584 equations for incident energy of current limiting fuses that are in the current limiting range are empirical. They are based on tests with the fuses in place and use a regression analysis of measured energy and available bolted fault current. Trying to separate the current reducing effect of the arc and the current limiting effect of the fuse would be complicated and is unnecessary.

 

[dpc]

The reality is that if you have a current-limiting fuse that is operating in the current-limiting region at the arcing current level, the clearing time is so fast, that even if the standard TCCs are used with the "basic" IEEE arc-flash equations, the arc-flash levels are quite low.

EasyPower also has an option for using the IEEE equations for current limiting fuses, but we generally do not use them.

David Castor

http://www.cvoes.com

 

[PHovnanian]

Worst case incident energy for arcing faults occurs at fault currents less then those for bolted faults. So the CL fuses may not do anything to reduce incident energy.

Think about the difference between a (theoretical) bolted fault and an arcing fault. With a bolted fault, the energy is dissipated by the entire circuit resistance. For the 'ideal' zero resistance bolted fault, no energy is dissipated at the fault location itself. Now, start adding some real resistance to that fault (an arc, for example). While the power dissipated at that location goes up, the resistance has the effect of reducing the overall fault current. This reduction may move that point below that where a CL fuse has any effect.

 

[SJBatTCE]

Thanks for the responses.

dpc, what is your reason for not using the equations?

PHovnanian precisely states my concern with using the equations.

 

[dpc]

The equations are based on limited testing with a limited number of fuses. Trying to extrapolate that to cover all CLFs is questionable to us. Especially since we don't see a need to do so.



David Castor

http://www.cvoes.com

 

[DanDel]

In actuality the calculations, algorithms, and recommendations in use now for IEEE 1584 & NFPA 70E Arc Flash analysis are still in their infancy.

You can count on continuous changes over the coming years based on more testing (hopefully).

 

[RonShap]

Be sure that you are not using infinite primary for the incident energy calcs as it will offer lower than recommended PPE/incident energy due to the inverse characteristic of an OCPD.
Always try to determine the lwoest available from the utility or other source.

 

[alehman]

Ron, That brings up a good point. Therein lies a very big problem with arc flash calculations for utility-fed systems. Obtaining that number can be next to impossible unless your utility is unusually cooperative.

Alan
“The engineer's first problem in any design situation is to discover what the problem really is.” Unk.

 

[SJBatTCE]

I actually haven't had too difficult of a time getting realistic fault currents from utilities. I've done arc-flash analysis for companies across the U.S. and had pretty good cooperating with most utilities, once they understood why I needed an accurate value. Usually you just need to talk to another engineer rather than some key account rep or customer service person relaying a message to engineering.

I usually ask for the fault current on the primary side of the service transformer. It's easier for the utility engineer to determine. If you ask for the secondary side fault they'll assume you want to figure AIC for equipment and tell you the infinite source for the lowest impedance of the service transformer class.

 

[alehman]

All I've ever been able to get is secondary values, which they just look up in a table. I've not been too successful getting responses from engineering groups except when we have primary service. Maybe I'm just not persistent enough.

Alan
“The engineer's first problem in any design situation is to discover what the problem really is.” Unk.

 

[PHovnanian]

The problem with handing out available fault currents, from a utilities point of view, is their liability should they make a system change.

Handing out a distribution transformer secondary figure isn't that big a deal if they assume a zero impedance at its primary. They can always add language stating that this data is only valid for this unit.

But for primary services or networked secondary, most utilities don't want the responsibility of informing a bunch of customers should they change something a ways upstream that alters their available fault current.

And should they implement such a procedure, the issue of whom that customer might be changes over time. When its new construction, there's an engineering firm attached to the project. But years down the road, the customer is just the person who pays the monthly bill. If they get a letter informing them that their fault current is going from 100 kA to 150 kA, they've got no idea what the significance of this is.

 

[SJBatTCE]

From experience as a former utility engineer, the utility provides this information as a service, not as design information. NFPA 70E recommends updating a facility's arc-flash study every five years (if I'm wrong, correct me) ,if no major changes have occurred on the facility system, in order to account for utility system changes.

Not being a lawyer, I don't believe the utility is liable or required to notify a customer that source impedance to their facility has changed unless it is stated in the service agreement.

 

[burnt2x]

With new system interconnections/ change of power source providers, utilities can never be as up-to-date as they want to be. Having said that, the info you get will be true on that very instant. Expect systems to change every now and then, economics drives utility managers to make their bottom lines bigger (profit). I guess we will have to be content with data utilities give, hoping that their side of the grid experiences as little system change like what you expected.

 

[davidbeach]

It's worse than that, we can never have a completely accurate model. What the people maintaining the system model know about what exists in the field is always behind the times, and what is in the field is always changing. Nobody builds lines with exactly the same configuration at every pole or tower down the line. System modeling software assumes that all you need are positive and zero sequence impedances and the zero sequence mutual couplings between lines; completely ignoring all the other couplings. The system model then assumes (generally) that all sources and lines are in service. If you are far enough away from generation, which generation is in service may not matter much as your source impedance will be almost entirely lines and transformers. What lines and transformers are in service or not may make a significant difference though.

From the utility side we really don't care much about minimum fault levels at a given service location, just maximum so it will be difficult to ever find out what the lowest bolted fault current you might expect. But that is a very important number for arc flash analysis (often more important than the highest possible bolted fault value). What we can generally do a good job of is telling you what the highest possible fault current we can provide will be regardless of any change we might make on our system, and that is the important number for equipment sizing. As to getting more realistic numbers for other studies, your mileage may very.

 

[Maak]

At low voltage the arcing current is often less than 50% of the bolted fault current (can be less than 30%). No wonder an IEC standard puts it this way: "In the case of current-limiting fuses, the maximum arc energy may occur at current levels below the maximum interrupting rating." (IEC 62271-200)

An interesting solution might be the combination of CL fuses and an arc quenching device. The quenching device would practically guarantee high current and fast operation of CL fuses.

 

[SJBatTCE]

davidbeach, I agree concerning the realistic accuracy of system models, both utility and industrial. Sometimes I feel I am doing surgery in the dark with the arc-flash analysis. But at least the hazard is being addressed.

I think I've gotten a fairly good sampling of the community's feelings on the subject of CL fuse equations and arc-flash calculations in general. We do what we can with what we've got.

Thanks, everyone!