Is there any reason to oversize power cables to motors? 4

[bdn2004]

We have a project to replace a bunch of large cables feeding a significant number of 250Hp motors. The cables are being run in cable tray single file. As far as I can tell the engineering firm did the calculation right per the NEC and calculates the ampacity of the 500kmil cables after all the various temperature deratings to be around 382 Amps. Per the Code the motor leads must be sized with an ampacity of 302 x 125% = 378 Amps.

Because this is close, they say to bump it to 750kcmil. Now in my mind the 125% requirement already covered the safety factor. Is there any reason or any data anyone can provide that would justify spending this extra cash? Like will we get twice the life out of the cables?

 

[jghrist]

Smaller longer cable would reduce the fault current at the load end. If the load end equipment was not rated for the full fault current, this could conceivable help, but I would not consider it a good design. A lower fault current could also increase clearing time and increase arc flash incident energy.

 

[stevenal]

The short circuit does not always occur at the end of the cable. Thats why NEC requires the interrupting device to be able to interrupt the available current at the interrupting device, not at at some distant remote end.

 

[bdn2004]

Thanks for the interesting discussion. It makes me wonder if decisions like this shouldn't have to go through an engineer instead of selecting it from a table in the NEC.

It reminds me of this somewhat toothless Board I sat on a few years ago in a fairly large city to make recommendations for the city's electrical code above or below the NEC. Whenever we had a meeting about once every 3 months the IBEW was right there demanding to be heard for something else to be done strictly by licensed electricians. Maybe at some point they wrestled this away from engineers.

 

[electricpete]

I vote burnt2x a LPS for bringing up an important aspect that was not mentioned in the early posts. (I certainly didn’t think of it when I read this thread.)

My comments applied to the cable damage scenario he described. (not to breaker interrupting capacity).

I agree with you your comment bdn2004 – I can’t find any table in the NEC that will help you meet I^2*t requirements for the cable. But there’s a lot I don’t know.

Again I would like to hear rbulsura explain what he meant by this:

It is well known (to those who consider short circuit ratings of the cables) that overcurrent protection provided per NEC, especially for LV installation, is more than adequate to protect against thermal damage to cables during a short circuit

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[rbulsara]

electricpete:

One can check this with a software all they want. Fuse/breakers sized for overcurrent protection per NEC would always turn out to be adequate for the thermal damage curve of the cables. The ampacity tables of NEC are not without any basis nor are the limits on OCPD rating there in based on conductor ampacity.

NEC does not preclude proper engineering however, as you already quoted per NEC 110, however. So when in doubt, check it.

NEC is for minimum safety code against fire and personnel hazard, not an assurance of complete engineering or meeting design objectives.

Rafiq Bulsara

http://www.srengineersct.com

 

[mayanees]

...and to reinforce what rafiq is stating, when you plot the time-current-characteristic of an overcurrent device, say a 100 amp circuit breaker, and include on the drawing the thermal damage (withstand) curve for the #3 CU cable that's connected to it, the situation is always such that the cable is completely protected up until it runs out to the 1000 second part of the curve. At that point the breaker protects the overload capacity of the cable and it's much closer to the melting curve for the cable.
That's to say that when you protect a cable with the NEC approved protection, there is always adequate protection for short-time withstand type faults.
As alluded to in his response - that's not always the case with medium voltage cables, particularly since NEC has much less stringent protection requirements over 600 volts.
John M

 

[electricpete]

Thanks Rafiq and John. I appreciate your comments. I have no doubt I can learn a lot about NEC and protection from you guys.

What about short circuits when a high fault current level is available? There is a finite delay for the instantaneous element to sense the short circuit and the breaker to open the circuit. During that time the cable is exposed to the fault current. I don't think any of the tables consider the available fault current as an input.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[mayanees]

Pete,
Here's a Time-Current-Characteristic curve showing a 4000 amp main on the secondary of a 3000 kVA transformer.
The instantaneous is off, which is often done to improve selectability with downstream distribution breakers.
But you can see the 4000 amp damage curve is far to the right of the maximum available fault current that the cable would see - denoted with a flag marking SQD NW, which is about 55kA.
You can see that the cable/4000A bus is totally protected - even with the Instantaneous off.
John M

 

[electricpete]

Thanks John. A large bus will have a large withstand which will be difficult to challenge without an unrealistic high fault current. How about a smaller conductor like your previous example with 3AWG cable.

Attached is I^2*t curve for 3AWG copper cable (26 mm^2 cross section) showing time to rise from 75C to 100C. The equation of the resulting curve is I^2*t = 7.5E6 Amp^2*sec.

The withstand time for various current levels is:
I t
10000 0.074919467
20000 0.018729867
30000 0.008324385
50000 0.002996779

If we have 50,000A fault current, we need to interrupt in 0.003 sec – doesn’t sound likely.
If we have 30,000A available, we need to interrupt in 0.0083 sec which is still less than half a cycle and probably not realistic. I read somewhere one half cycle is the time used by ICEA for rating cable short circuit. I would think it depends on the interrupting device and some surely will not be that fast. Let me know if I am looking at it wrong.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Correction in bold

Attached is I^2*t curve for 3AWG copper cable (26 mm^2 cross section) showing time to rise from 75C to 150C

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[mayanees]

.. hard to resist this topic line, but limited on my availability.
You're certainly looking at it correctly.
I've attached another hastily assembled TCC that shows a #2 cable with 100 and 110 amp SQ D breakers.
Note that the breaker with adjustable instantaneous protects - but only at the Low Inst setting. The thermal-mag breaker with no adjustments leaves the cable vulnerable in the short-time/inst region.
This pretty much says that it needs to be looked at on a case-by-case basis, which is a good thing because it keeps us engineers working!
But relaize that this is a very special case where a 100 amp cable is terminated at a point where there's 55kA of fault current, which usually wouldn't be the natural order of things. At a switchgear bus with that level of fault current the breakers are usually on the order of 400 amps or larger, and an 400 amp cable would tolerate that level of instantaneous fault current.
John M

 

[davidbeach]

But you only need to protect a conductor against damage from through faults, and smaller conductors tend to limit through fault currents. If the fault happens in the conductor it is damaged any way, no matter how fast you interrupt the fault. The catch 22 is that as fault current goes down the clearing time goes up.

 

[burnt2x]

Thanks for the heads-up, David.

 

[rbulsara]

If the cable damage itself is not the cause of the fault, cable itself seldom gets damaged ( beyond repairs) during a through fault.

Most cable and insulation damage (and fire) does not occur because of high fault current, but rather during the low fault current events such as arcing faults. Low fault currents do not clear the OCPD, building up localized heat, melting insulation, ionizing air and resulting in a much greater fault and damage. Another common cause for cable insulation damage is loose connections, followed probably by misapplication of cable insulation types.

Rafiq Bulsara

http://www.srengineersct.com

 

[mayanees]

To David's point, absolutely, and thanks for that clarification.
If that switchgear had 50 kA of available fault current, and 25 feet of #2 AWG cable is terminated on a 100 amp breaker, the fault current is reduced to about 30 kA at the termination of the cable.

John M

 

[rcw retired EE]

Most low voltage cable circuits are self protecting from damage to short circuits at the end of the run. The impedance of the cable reduces the short circuit level to where the typical MCB clearing time protects the cable.

I use a spreadsheet calculation that looks at the source short circuit MVA and calculates the length of cable needed to self protect the circuit from damage.

My spreadsheet says a #4 copper wire fed from a 100 KA bus by a 100 A MCB needs 35 feet of conductor length to be self- protected. At 50 kA the length drops to 27 feet.

A #2 awg fed by a 225A frame MCB needs 38 feet at 100 kA and 25 feet at 50 kA.

As long as the cable length is more than this critical length, a bolted fault at the motor, the panel or equipment fed by the wires will not permanently damage the wires.

 

[stevenal]

And the breaker at the switchgear location supplying the #2 must be capable of interrupting 50 kA per NEC 110.9 and not just 30 kA line end value.

Davidbeach,

If you protect based on the cable end fault, you will be replacing more cable than if you protect for the close-in fault, in the event of a mid-cable fault. Or do you always replace the entire cable and avoid splicing?

 

[burnt2x]

If we consider the specific short-circuit withstand of a #2 cable, is it good allowing 50kA on it? ICEA says #2 can safely be allowed to be subjected with around 36kA (assuming thermoplastic insulation)for 1 cycle (0.017 sec).
Am I missing something here?

 

[davidbeach]

When I was doing that type of calculation in my designs, the first point at which a splice might be possible was beyond the critical length (to use rcwilson's terms). Now working at much higher voltages the fault currents are generally low enough and the clearing fast enough that I'm not much worried about cable damage.

 

[bdn2004]

Getting back to my original question...I just talked to the engineer in charge of the area and he told me that these motors are started under a heavy load and the starting currents are over typically over 1000 amps for as long as 15 seconds. For one, is this really atypical? Would the NEC Tables still be applicable?

Can anyone refer me to an example problem where this is worked out so that I can get all the required information and make this a more scientific analysis? Thanks.