Morning all

Imagine, if you will, a customer main LV switchboard, form 4, with 2500A ACB incomer off the transformer tails, on the left, onto top busbars rated at 2500A. The ACB long time overcurrent is set to 1.0xIn, and while the load is usually some way below that, the factory engingeering staff aren't keen having settings below maximum to allow for future flexibility. As with many industrial settings, downtime for works is precious and available only sparingly. The switchboard has been in service for many years and contacting the original manufacturer for calculations is not likely to be possible.

Management would like an embedded generator (which type is immaterial) that would need a 100A OCPD off this switchboard. The two 200A switchfuses furthest from the incomer, bottom right, are used for existing loads (fused let's say at 63A and 125A), but the next way up is free. We could make it more tempting by suggesting it's already equipped with a 200A switch fuse needing only the right size BS88s, so an easy win.

It would, at first glance, appear sensible to use this spare way, requiring minimal works and distruption to the ongoing site processes. I would argue that it is not possible to overload the busbar below the generator connection because that's limited by the sum of the outgoing OCPDs = 63+125 = 188 < 2500A. Indeed the ways are only rated 200A each so even if circuits change, you can only get 400A without major surgery. Likewise it is not possible to overload the busbar upstream because even if you had 2600A of total load, because it's now fed from either end, no section of busbar could be called on to carry that much. An exception might be if it was a switchboard with multiple vertical droppers and the sum of each OCPD coming off one of those exceeded 2500A* since in that case the application of diversity relies on the incomer ACB being present.

But then we refer to BS7671 551.7.2, which says we can't do this, because 2500 + 100 > 2500A and instead we must either demand that the ACB is dialled down, and if that is not possible for any reason, we have to inform management that sorry, you can't connect the generator on the outgoing side, so instead we need to spur off the transformer tails into a switchfuse or something similar. Assuming there is room, and aside from the ugly equipment configuration it would necessitate, this means a shutdown to carry out relatively high risk work (commercially: If something goes wrong and it's not safe/possible to re-energise on schedule, the whole factory can't operate). Understandably, no-one is keen.

So... Is this time for a departure? On the grounds that we've assessed the risk in accordance with the underlying principle and the proposed design is no less safe than simply following that regulation. Or has something been missed?

Is the fact that it's form 4, so the fusegear is well spaced, of import, or could one apply similar logic to form 2 boards with adjacent MCCBs, as long as nothing is running too warm?

And/or is there some protection one could reach for to sum the incoming supplies?

Finally, if this generator was, in fact, connected via an existing submains, 551.7.2 strictly only says one needs to consider where the connection is made. But surely the fact that "it was there before" isn't much of a defence so one should consider everything upstream in the same fashion. Which makes connecting generators to anything other than the main switchboard somewhat fraught, unless whoever did the original install was kind enough to size equipment larger than the upstream protection.

For the avoidance of doubt, I do absolutely get, in principle, that it does need consideration. For example, if one had a domestic CU with an 100A service fuse, rooftop PV, batteries in the garage able to feed internal loads via G100 ELS, an EV charger, and electric heating (etc) the risk of cooking the CU with busbar currents well over rated for protected periods is quite high.

I know some here have been involved and gave talks on this regulation when it came out, so would very much appreciate your thoughts. Apologies if this all made clear somewhere (I didn't spot it in an IET COP last night); I've not been involved with this type of project since the regulation was changed.

Thanks
Jam

NB: This is an amalgamation of a number of sites I've seen lately. I've added some details to help give a picture but it is entirely fictional.

* If the vertical bars are not fully rated, due to diversity assumptions, the problem, if there is one, is then more to do with the design of the switchboard than the connection of the generator, so I'd rather not confuse things here.