Parallel Multicore cables - current sharing

[healyx]

Does anyone have a source/reference that shows that current sharing between parallel multicore cables is equal?

I am talking about 3 phase, unshielded multicore cables where the 3 phase cables are arranged in a triangle (triplex/trefoil) and have a consistent, constant twist rate (transpostion) during manufacture.

I have been modelling this mathematically using the impedance matrix method and it appears that provided the parallel cables run together with the twist in the same direction - current sharing is equal (based on inductance - unequal heating due to layout may effect resistance but that isn't the focus of this question).

Doesn't seem to matter if you parallel 2,3 or 4 sets together and doesn't seem to be affected by spacing between sets or starting orientation of cable phases.

I was hoping for a source to confirm my analysis.

 

[dpc]

Thousands upon thousands of installed systems seem to imply that the current sharing must be relatively equal, provided the lengths are very close to the same. If there was a major issue, it would have surfaced decades ago. The NEC does not require that the three phases be triplexed or lay in any particular geometry. If all three phases are in the same conduit, the flux cancellation is quite good.

 

[healyx]

Thanks for the response dpc. The question is specific the multicore cable as described. I agree anecdotal evidence would suggest it should be OK, but I'm looking for a reference where it has been proven.
You seem to be making reference to parallel single core cables. There is alot of material available for parallel single core cables that show that arrangement is important. Spacing between sets seems to reduce the incidence of current imbalance for single core cables.

 

[7anoter4]

If I1 and I2 are perfect equal and symmetrical arranged in space Magnetic Flux 1=Magnetic Flux 2 and the result total it is zero.
Magnetic Flux 3 has no compensation then a remaining flux could be. If a magnetic flux is zero the inductance is zero and also the reactance
between these conductors.
In the case using 3 cores cable as a single-phase the entire cable it is as a single core cable and a reactance between this cable
and other phase cable has to be calculated according to distance and arrangement type.
A problem could be if the cable would run through metallic conduit or if the cable would be armored. The EMF induced could produce
a build-up voltage at the not grounded end or a supplementary losses-then heating-if both ends are grounded.
So, I agree with dpc, of course.

 

[dpc]

I'd think that actual installed systems would be a better proof than someone's academic research and theoretical calculations. But that's just me.

I agree that triplex should have better flux cancellation than random lay of single conductors.

In my experience, the problems with paralleled conductors come on extremely short runs where it is the slight differences in resistance due to difference in length and conductor temperature that cause the imbalance, not the inductive reactance.

It is common to see large 480 V services with up to 12 or more large 600 V conductors per phase, installed in the cheapest, sloppiest ways imaginable. I still cringe when I see this, but I'm not aware of any problems related to current imbalance related to inductive reactance. Of course, probably no one bothers to measure the current split after installation unless one of the conductors fails, so I'd expect some imbalance to exist.