If the installation is already complete and in use, then I would suggest measuring the temperature of the steel gland plate under full load. You may find that it is a non issue.

Still a non compliance, but if the temperature rise is modest, then arguably an acceptable non compliance. My own observations suggest hat eddy current heating is a non issue at 100 amps or less, and usually anon issue at up to a few hundred amps. There are, I suspect too many variables to calculate the degree of heating, hence my suggestion to measure rather than calculate.

If the installation is already complete and in use, then I would suggest measuring the temperature of the steel gland plate under full load. You may find that it is a non issue.

Still a non compliance, but if the temperature rise is modest, then arguably an acceptable non compliance. My own observations suggest hat eddy current heating is a non issue at 100 amps or less, and usually anon issue at up to a few hundred amps. There are, I suspect too many variables to calculate the degree of heating, hence my suggestion to measure rather than calculate.

broadgage said:

Still a non compliance, but if the temperature rise is modest, then arguably an acceptable non compliance.

I get what is meant by the above and agree. However, if the work is new then by virtue of the BS7671 definition of “non-compliance”, it would need to be attended to before certification was issued or, alternatively, identified as a “departure”. 

So by “departure” it would be a deliberate decision by the designer of an electrical installation to not fully comply with the requirements of BS 7671. For me it seems more like a mistake.

it may be a mistake, or it may have been decided early on that there is negligible effect to worry about.

The regs on this are framed assuming that all steel enclosures have the worst possible magnetic loss, and that all single cable are in intimate contact with the steel and and are already carrying a current very close to maximum.

Calculating the heating effect is quite tricky as you usually don't have an analysis of the steel, or the full info on how it is cooled. It may  be OK to assume that any field less than 100milli-teslas p-p at 50Hz is most unlikely to cause serious heating but this will still flag up many cases that are in practice not an issue at all.

(more info on estimating  losses like that here )

You don't need much of an air gap around the wire, or especially good steel to make the problem negligible for sensible currents. (i.e ones not pushing hard at the cable rating.)

Before reaching for the dremel and vacuum cleaner, personally I'd be fitting temperature stickers to the plate and seeing if there is any evidence of overheating..  If not I'd leave it.

Mike.