Each Frame has 2 of each phase cable and 2 x neutral

So through each "hole" you have a complete set of L1/L2/L3 & N - so things should be reasonably balanced through individual hole.

If the cables are all in parallel and corresponding ones connect back together after the steel frames as well as at the generator, then you might get some imbalance due to the two paths not being exactly matching - but hopefully that's pretty small. Technically probably a non-conformity, but (unless we're looking at massive currents) unlikely to be a problem in practice. The next issue of BS 7671 might water down the requirement even more, if the DPC was anything to go by.

   - Andy.

Thanks AJ

Yes cables are in parallel and current is 1250A

Forgot to say cables are AWA

You may have far more of an eddy issue with the currents that will be induced in each of the aluminium armours. (*)

As above, so long as things are bunched in a way that flow and return currents are side by side and more or less balanced there will be no issues from the wiring frame or steel wall.
As an example if you wanted to sabotage things, you would have had to put phases L1 and L2 through one path, and phase L3 and neutral through the other path - and even that may work if there were only single phase loads on L3 and only 400V cross bar loads on L-L2, but of course that is not what has been done.

Mike
* each individual armour runs parallel to one current far more than it runs parallel to any other so a voltage will appear between one end of the armour and the other in proportion to the current in the inner core.
  Then either all 3 armours need to be so well connected at the ends, that  although a large current flows there is no heating, because the resistance of glands and gland plate is low, or the 3 armours need to not connect at one end or the other,  so that current up one armour and back down the others does  not flow.  One either finds massive conductive gland plates, or insulating ones, both approaches are valid.

That problem disappears again for multi core armoured cables, so long as the currents in the cores enclosed within one armour are more or less zero-sum as with 3 phase +N or single phase L+N -  they usually are.

All cables are made off in Ali Glands at both ends in to Ali gland plates 

Thanks for the response, most helpful. 

I also Read

2006 20 Autumn Wiring Matters

So I understand for an alternative application where low voltage service cables are single core AWA and the cables are suitably earthed at the load end but the DNO won't terminate the AWA at the supply end. Doesn't this present a safety concern for the DNO if the service cable is long and supply rating is high?

Armour can be single point bonded or solid bonded, there are pros and cons to each.

Long cables and/or high current would tend to have larger voltages between the armour ends if single point bonded - you'd need specific calculations to see if the resulting voltage was acceptable or not. The problem can be mitigated though - e.g.  by splitting the run into some multiple of 3 and 'rotating the phases' so the end-to-end voltages tend to cancel out.

Or you could go for solid bonding, but do the bonding yourself just after the DNO equipment.

   - Andy.