Ground wire sizing above NEC recommendation

[skeshesh]

Hope all is well with everyone.

I was looking over considerations for EGC sizing and I can't establish a gut feeling for when it's appropriate to actually worry about upsizing the wire size. I understand that the let-through current may exceed the ground wire insulation rating depending on the OCPD, but can't imagine this possibility warrants short circuit calculations for every point in a system. Is this mainly a problem in very large power systems where the bigger KVA rating allows for significant available fault currents? Would much appreciate anyone who's had experience with either a failure of a wire due to this condition or who's had to consider this conditions to shine some light on what constitutes a case where these calculations should be conducted as a good engineering practice.

 

[kwired]

JMO, i would not be too concerned about a fault current damaging an EGC that is properly sized. That conductor likely can handle much more current for a longer time than any devices that it is connected to, including wire nuts or other connecting devices.

Insulation damage is not too likely, plus the conductor does not need to be insulated anyway (with a few exceptions).

Systems with a higher fault current available usually results in faster trip time. Systems with lower fault current usually results in slower trip time. Any issues will be more about how much energy is used during the fault and how much energy the conductor and any connected components can handle without damage.

 

[bphgravity]

It appears the code is pushing ever more towards this type of consideration. With the addition of Section 110.24 in the 2011 NEC, there is no doubt the CMP's are growing more and more concerned about this issue. Other examples would be Section 430.8 and 440.4(B). These two section relate back to Section 110.10 which relies much on A250 grounding and bonding practices.

Copper can withstand 1 ampere for 5 seconds for every 42.25 cm. The withstand rating can be expressed as IxIxt. So what's that mean?

Take a #8 EGC. Per Table 8 of Chapter 9, #8 has 16,510 cm. So, 16,510 / 42.25 = 391 amperes.

Using the expression above, 391A x 391A x 5 seconds = 764,405 A^2(t).

Now, figure the trip curve time of your overcurrent device. If the OCD will trip time is say 1/2 cycle, take the square root of 764,405 / .0083 (.5 cycles / 60 hz) = 9,597A.

This means that a #8 EGC will withstand nearly 10kA for 1/2 cycle. If the trip time is longer on the OCD, that withstand rating drops. If the SCCR or Available Fault Current exceeds the withstand rating, the conductor could fail.

Hence the note to Table 250.122 .

So think of it like this. If you have a 60A branch circuit supplying a motor with 7,000A of available fault current, what size EGC would be necessary to withstand the fault, assuming a 1/2 cycle trip OCD?

I think you will find the Table allowance of #10 AWG would potentially not be sufficient...

 

[don_resqcapt19]

If you don't want to do all of the math look at Table 2 in this document. Also these values are not the point where the EGC fails. It is the point where the copper reaches 150?C. Table 1 in the document gives the I?t values for the conductor to reach 250?C and for it to melt. 150?C is considered the maximum temperature that the insulation of common building wire can withstand without permanent damage.

 

[skeshesh]

Thanks for the responses everyone.

Don, I had read that document before and I don't have a hard time doing the calculation. Really what my concern was if any of you guys had done these calcs on regular basis for a particular scenario. I really don't see many engineers consider this issue, but at the same time an important part of good engineering practice is to have a gut feeling and do a calculation when necessary.

 

[kingpb]

What I find interesting is that the international community has been taking the short current rating of wire into consideration for some time.

Designing circuits in the IEC world often requires an increase in wire size to withstand SC currents. Got burned on that many years ago, it was a costly lesson to be learned.

Cable damage curves are easily come by and provide this information. They are built into ETAP and SKM.

 

[skeshesh]

Cable damage curves are easily come by and provide this information. They are built into ETAP and SKM.

You're commets regarding IEC are spot on. I talked to ETAP about this and they said they are incorporating EGC sizing recommendation based on let-through current in their next release. I haven't spoke to SKM since this client's spec call for ETAP but both software packages have been competative for some time and I'm sure SKM would be adding this feature sooner or later.