Parallel sets and OCPD

[ggunn]

OCPD is there to protect conductors, so the conditions of use adjusted ampacity must be greater than the next size down OCPD from the one being used. So far, so good, but what about parallel sets? A fault on a single conductor in a two parallel set feeder gets all the current from the OCPD, not just half of it. How is that OCPD protecting the conductors?

 

[dkidd]

Use cable limiters to provide parallel cables with protection.

http://www.cooperindustries.com/con...opper_aluminiumcablelimiters600vac250vac.html

 

[iwire]

OCPD is there to protect conductors, so the conditions of use adjusted ampacity must be greater than the next size down OCPD from the one being used. So far, so good, but what about parallel sets? A fault on a single conductor in a two parallel set feeder gets all the current from the OCPD, not just half of it. How is that OCPD protecting the conductors?

The fault goes both directions so unless / until it burns open they remain in parallel.

The only time I have seen cable limiters used was per power company requirement to protect the power companies network.

 

[ggunn]

The fault goes both directions so unless / until it burns open they remain in parallel.

The only time I have seen cable limiters used was per power company requirement to protect the power companies network.

Well, duh. Friday morning brain fart. The wires are connected at both ends.

The faulted wire still gets a bit more current than the other one because the current path to the fault is longer through the other wire to the end and back around, but yes, I see it now. Move along, nothing to see here...

 

[pv_n00b]

The OCPD on parallel conductors will not protect a single conductor from overload, that's why we make sure the design will encourage current to divide equally. But it will still protect a single conductor from a bolted fault, the single conductor will carry enough fault current to trip the OCPD by itself, even if it might take longer, and some fault current will take the long way around to the fault through the unfaulted conductor(s).

Cable limiters can actually cause problems because they will take the faulted cable out and then the full load will be carried by the remaining cables, possibly overloading them in the process. They are good in supervised installations where the fault can be indicated and repaired quickly without the fault isolating a critical load.

 

[jaggedben]

Edit: response was intended for the other thread. Got my EGC threads mixed up.

 

[GoldDigger]

The fault goes both directions so unless / until it burns open they remain in parallel.

When you have two parallel wires, the fact that the two sets are connected at both ends does allow roughly comparable fault current to flow in both wires.
The closer to the load end the fault occurs the closer the current in the two wires will be. If the fault is near the supply end AND if the wire resistance is the major resistance in the fault current calculation you could have as much as a 2:1 ration between the two fault currents.
Where it gets interesting is when you have three or more parallel sets. In those cases you can end up putting a much higher fraction of the fault current through the single faulted wire.
If you have a limited current fault, whatever the mechanism of that might be, then you theoretically could fail to reach even the thermal trip of the breaker while overloading the faulted conductor by a factor of 3 or more.

 

[wwhitney]

When you have two parallel wires, the fact that the two sets are connected at both ends does allow roughly comparable fault current to flow in both wires.
The closer to the load end the fault occurs the closer the current in the two wires will be. If the fault is near the supply end AND if the wire resistance is the major resistance in the fault current calculation you could have as much as a 2:1 ration between the two fault currents.
Where it gets interesting is when you have three or more parallel sets. In those cases you can end up putting a much higher fraction of the fault current through the single faulted wire.
If you have a limited current fault, whatever the mechanism of that might be, then you theoretically could fail to reach even the thermal trip of the breaker while overloading the faulted conductor by a factor of 3 or more.

I'm having a little trouble with the above, but I'm not 100% clear on how to calculate fault currents. Is the following correct?

Suppose we have a 100V source with a 10 milliohm source impedance. Then a bolted fault at the source would allow 10,000 amps to flow. (Is that right?)

Now we want a 400A feeder with at most a 2% VD, so we need a feeder of resistance at most 5 milliohms. We provide two parallel conductors of resistance 10 milliohms each. If we fault at the far end of the feeder, the total impedance from the source to the fault is now 15 milliohms, so the fault current is 2/3 of 10,000 amps.

Suppose one of the conductors faults 1/5 of the way from the source. Then there are two parallel paths, one with a resistance of 2 milliohms, the other with a resistance of 18 milliohms. The combined resistance is 1.8 milliohms, so the total impedance from the source to the fault is 11.8 milliohms, and the fault current is 100/118 of 10,000 amps. 1/10 of this current takes the "long path" of 18 milliohms, and 9/10 of this current takes the "short path" of 2 milliohms.

Suppose instead we had chosen to provide 3 conductors of 15 milliohms each, and one conductor faults 1/5 of the way from the source. Then there are still two parallel paths (treating the other two conductors in parallel as one conductor), one with a resistance of 3 milliohms, the other with a resistance of 12 milliohms + 15/2 milliohms = 19.5 milliohms. The combined resistance is 2.6 milliohms, so the total impedance from the source to the fault is 12.6 milliohms, and the fault current is 100/126 of 10,000 amps. 2/15 of the current takes the "long path" of 19.5 milliohms, and 13/15 of the current takes the "short path" of 3 milliohms.

Cheers, Wayne

 

[GoldDigger]

Looks right to me.
Many times the actual numbers give a better picture than "gut feeling".
Thank you.

To complete the picture, the one thing to add onto your analysis (which was intended only to calculate fault current) was that in the first case you have each conductor capable of handling X/2 A where X is the nominal circuit current.
In the second case you have each conductor capable of handling X/3, so an equal current in the short path end of the faulted conductor would be more of an overload on that conductor.
As long as the total fault current is above the instantaneous trip value in both cases, it is unlikely that the faulted conductor will be damaged.
But if the combination of fault resistance and circuit resistance takes us into the thermal zone of the breaker, there may be a problem.

 

[wwhitney]

Looks right to me.

OK, then the example I worked out above shows that the current in the faulted conductor (short path) will approach the total fault current as the fault point moves towards the point of supply. Also that if the location of the fault is fixed, then the disparity in current between the two fault paths is worst with two conductors. With multiple conductors, all the unfaulted conductors are in parallel and serve to reduce resistance of the "long path" and hence the ratio of the resistance of the two paths.

Cheers, Wayne

 

[wwhitney]

The example I worked out shows that a fault of just one conductor in a parallel can carry nearly all the fault current, when the fault is located near the point of supply.

So suppose we have an 800A feeder protected by an 800A circuit breaker, consisting of (4) parallel sets of 200A conductors. In what sense is the individual 200A conductor protected by that 800A circuit breaker? One conductor could fault to ground with enough resistance so that the fault current is only 700A, with 600A of that carried by a short section of the faulted conductor. The circuit breaker won't do a thing unless the fault progresses to a lower resistance fault with a higher fault current.

Is this just considered an acceptable trade off for the benefits of using parallel conductors?

Cheers, Wayne

 

[iwire]

Is this just considered an acceptable trade off for the benefits of using parallel conductors?

I would say the large number of parallel installations being installed vs the limited number of them burning down tends to show that the trade off is a fine one.

 

[jaggedben]

I agree with iwire, and I'll venture to try to explain why what he says seems to be the case...

...So suppose we have an 800A feeder protected by an 800A circuit breaker, consisting of (4) parallel sets of 200A conductors. In what sense is the individual 200A conductor protected by that 800A circuit breaker? One conductor could fault to ground with enough resistance so that the fault current is only 700A, with 600A of that carried by a short section of the faulted conductor. The circuit breaker won't do a thing unless the fault progresses to a lower resistance fault with a higher fault current. ...

It seems to me the parts in red represent a problem regardless of any other information. 700A worth of power (multiply by voltage but it hardly affects the point) is being thermally dissipated via a resistive path that was not designed to thermally dissipate 700A. And that's your fire starting event right there! It hardly matters whether you've got a single conductor or parallel conductors that are rated less than the fault amps, because the conductors will probably last considerably longer and take considerably less damage than the material conducting the fault. Either you get lucky and the fault quickly burns itself open, or you get unlucky and the building is on fire, at which point the ampacity of the conductors is essentially moot. And that is true regardless of whether you have a single conductor or a paralleled conductor that is overloaded.

An OCPD just isn't designed to protect against a 'resistive fault'. Requiring some kind of fancy overcurrent protection for each conductor in a paralleled circuit would actually be a more stringent requirement than for non-paralled circuits, because in either case a 'resistive fault' is a major hazard that simply isn't addressed by an OCPD. Moreover, it seems to me that we are really only worried about an operating overcurrent when we do ampacity calculations, because almost all actual faults trip the OCPD. By operating overcurrent I mean an overcurrent due to a bad load calculation or misuse of the circuit, as compared a fault between conductors and/or ground. The chances are really low of having a 'resistive fault' whose resistance falls right in the window where you get a sustained current without tripping the OCPD. And the chances are even lower that a fault conducts an amount of current that happens to fall in the window between the ampacity of one paralleled conductor and the whole circuit. If there's a worry about that kind of fault, then you'll need something other than an OCPD anyway, like GFCI, GFP, etc.. The code does require GFP for a lot of high ampacity feeders, so that hopefully should handle some of the concern, and it should work regardless of paralleled conductors or not.

 

[wwhitney]

I agree with iwire, and I'll venture to try to explain why what he says seems to be the case...

Thanks for your thoughts, definitely helpful.

Cheers, Wayne

 

[Zach in WA]

parallel conductors

parallel conductors

This is slightly off topic.
So, parallel conductors, say I have (2) sets of 250kcmil copper. In the 75°C column if (1) 250kcmil = 255A. So multiply 255 by 2= 510A. So, what if instead I added the (2) 250kcmil together to get 500kcmil that =380A in the 75°C column. Where in the NEC does it say to do one instead of the other?
go easy on me.
thanks

 

[Dennis Alwon]

I am not sure the nec states what you want to know but parallel conductors have more surface area then the equivalent kcm cable. More surface area means the conductors can carry more current. Also if one conductor carries x amps then why would 2 conductors carry less than 2x amps.

 

[GoldDigger]

The resistance of the wire is inversely proportional to the cross sectional area. But in addition to the resistance, the current carrying capacity is increased by better ability to shed heat. And heat shedding depends on surface area.
Although the two wires and the single wire represent the same cross section, the two have a greater surface area than the single.

Sent from my XT1585 using Tapatalk

 

[Carultch]

I am not sure the nec states what you want to know but parallel conductors have more surface area then the equivalent kcm cable. More surface area means the conductors can carry more current. Also if one conductor carries x amps then why would 2 conductors carry less than 2x amps.

If you put both sets in the same raceway, derate factors apply. Which in my general experience, may mean with 90C wire & 75C terminations you need to go one size above the size that has sufficient 75C amps for half the ampacity required.

In general, ability to curtail voltage drop is proportional to the KCMIL of the conductor. 1x500 does about as well as 2x250, provided that you need 380A or less for your application. 2x250 does slightly better in AC, because the inductive effects are slightly less in smaller wire.

However, for local factors alone, ampacity of an individual conductor is not proportional to KCMIL. If you want to understand the trend, there is a formula for it that can be used under engineering supervision, but it usually is easier to just use the table. Like others have said, ampacity is a thermal problem. The conductor surface area plays an important role, and that surface area is proportional to the square root of the cross sectional area. The amps per kcmil is greater for smaller conductors. At #10 Cu wire, you get about 3 amps per kcmil. #4 Cu wire, about 2 amps per kcmil. At 250 kcmil Cu, you get about 1 amp per kcmil. So it is strategic to parallel (provided 1/0 and larger), instead of running single conductors.

 

[ggunn]

If I am running parallel sets to carry a given amount of current, I divide the current by the number of sets and use the tables and derates as usual.

 

[pv_n00b]

When do you start to consider parallel conductors over single conductor runs? Personally, I start to look at paralleling conductors if 500kcmil Cu or 750kcmil Al are not filling the bill. Bigger becomes much more difficult to deal with and increasing size does not provide a proportional increase in ampacity when the conductors get larger.