I'd certainly agree that 551.7.2 is in need a of bit of work still - if only because as written it only covers final circuits and switchgear but omits distribution circuits.

I agree that it's entirely possible to take "layout" into account to show that no conductor can carry more than its rated current - and 551.7 should allow such approaches to be taken into account in general.

The switchgear aspect it a little interesting though. From other discussions here it seem that the thinking behind that bit of the reg isn't simply down to overload protection of the bus-bars, but also tries to consider the heating effect of all the protective devices. With your layout the heat generated would be equivalent to 2700A - 2500A from the supply, plus 100A as is comes into the DB and another 100A as the load supplied by the generator goes out again. Personally I'm less than convinced it's a real problem in reality. At a CU level we'd effectively being htold we need to worry about 116A equivalent, yet the Irish typically use a DP MCB as their incoming device - so even if that was as low as 63A rated we'd potentially have 189A worth of heat dissipation (63Ax3) in a box that's substantially similar to what we'd have in the UK. Whether the same considerations should apply to a form 4 DB, others will know better than me.

   - Andy.

The OP talks about ratings of certain components, but what Regulation 551.7.2 actually requires is that the I_nA rating applied by the manufacturer not to individual components, but to the entire switchgear assembly, i.e. in this case switchboard is selected so that

I_nA ≥ I_n + I_g(s)

This is not a wholly brand new issue - it's been highlighted in the second edition of the IET Code of Practice for Electrical Energy Storage Systems since 2020 as it was introduced into IEC 61439.

Jam said:

So... Is this time for a departure?

Well, if when you look at the whole assembly, that doesn't comply, perhaps not.

My understanding of this is that it originates from investigation into fires in switchgear and controlgear assemblies in another country.

Therefore, I wouldn't suggest a departure is appropriate to address what could be considered a breach of EAWR, Regulation 5 (as well as non-conformity to fundamental principles of BS 7671).

Well - if you modify the current distribution along the bus bars by reversing some or all of the current flow, you are changing it from the designer's original intention - that needs a few moments thought.  In effect you are assuming the designer's mantle - which may sound bad, but actually folk do that all the time. It's only bad if you do so and make a change without realizing the consequences.

There are a few danger cases.

If it had been connected at the end that meant in effect there was some part of the bus now sweating at more than the design rating, that might be bad.  But if we can be sure that the load tapping sequence is such that is not an issue, then we are OK. This gets harder to be sure as the genset power in creases - in extreme if it was say capable of doubling the available current, then clearly it needs every item revisiting. If as the example it adds 200A to a 2000A system, then we are probably within the tolerance of the original. What must not happen is the addition of several lots of only  10% - 'mission creep' is dangerous.
We need to be sure of the ADS, as we are adding rotating inertia to the system, so things will stay a bit live while things spin down. Hopefully the change in PSSC is not that great, but again if it was a bigger genset, it could be and in any case it needs to be double checked.
I'd be wanting to know how near the knuckle it really was with the original design first, and how much the specs had already crept over previous years. The knowledge it is a form 4 thing in sets the scale but the questions to ask and things to think about do not change much. If the thing already smells hot and the copper bars have that blued by heat  look then it may not be adequate already.
Connecting 'unseen' does not feel good.
Mike.

Thanks all, interesting thoughts.

AJJewsbury said:

From other discussions here it seem that the thinking behind that bit of the reg isn't simply down to overload protection of the bus-bars, but also tries to consider the heating effect of all the protective devices.

Indeed, and I think that is my main concern. When I've asked switchboard designers for their heat load calcs (for determining room ventilation requirements) they've quoted the sum of all devices at max load; I presume that's standard? Even if not I would expect that there's tolerance due to the difference between worst case (I presume some combination of maximally loaded ways) and in-practice loads (with Ib < In and hence heating reduced considerably due to I2r).

And what about simple busbar assemblies, with external switchgear tapping off?

gkenyon said:

The OP talks about ratings of certain components, but what Regulation 551.7.2 actually requires is that the I_nA rating applied by the manufacturer not to individual components, but to the entire switchgear assembly, i.e. in this case switchboard is selected so that

I_nA ≥ I_n + I_g(s)

That's kind of my point: There's a plate on the switchgear that says xxA so for full conformity to BS7671 we have to carry out some serious work to the incoming supply, when in practice it's not likely to be a problem for at least some specific cases.

gkenyon said:

This is not a wholly brand new issue - it's been highlighted in the second edition of the IET Code of Practice for Electrical Energy Storage Systems since 2020 as it was introduced into IEC 61439.

...Which is fair enough but only if you're installing energy storage systems, and it was "only" a Code of Practice. Now it's in BS7671 and applies to any generating set, so more likely to be contractually binding, insurers will be watching etc.

gkenyon said:

Therefore, I wouldn't suggest a departure is appropriate to address what could be considered a breach of EAWR, Regulation 5 (as well as non-conformity to fundamental principles of BS 7671).

Absolutely agree that it would not be appropriate to use a departure where danger may arise. The question is whether there are scenarios where danger will / can not arise beyond the simple check set out in the Regs.

mapj1 said:

If it had been connected at the end that meant in effect there was some part of the bus now sweating at more than the design rating, that might be bad.  But if we can be sure that the load tapping sequence is such that is not an issue, then we are OK.

Agreed, hence elaborating on this hypothetical scenario re the switchgear. In this case the frame sizes limit the maximum OCPDs that could be applied such that this can't occur.

mapj1 said:

We need to be sure of the ADS, as we are adding rotating inertia to the system, so things will stay a bit live while things spin down.

Agreed, though this is dealt with elsewhere in BS7671.

mapj1 said:

I'd be wanting to know how near the knuckle it really was with the original design first, and how much the specs had already crept over previous years. The knowledge it is a form 4 thing in sets the scale but the questions to ask and things to think about do not change much. If the thing already smells hot and the copper bars have that blued by heat  look then it may not be adequate already.
Connecting 'unseen' does not feel good.

Fair comment; visual inspection and/or review of historic data to be taken into account.

Please don't get me wrong; I'm not trying to cut corners. Indeed as a designer it is no skin off my teeth to simply say "Tough luck, them's the rules", particularly where safety is concerned. And I agree with the underlying principal. But I am also mindful of the implications, some of which are not simply financial. For example, the increased complexity of modified arrangements may themselves give rise to issues and in some cases it is likely that such modifications would be carried out under time pressure, or possibly even requiring new HV substations. These are forseeable risks, so need to be justified.

[Edit: Deleted duplicate post]

PS - And where does this leave an option, for example, to monitor the total current into the switchboard (let's say via a set of CTs on each supply and a summation transformer to an overload (overcurrent) relay)?

Where did the first (longer) reply go?

[Edit: Deleted duplicate post]

"The switchboard has an "xx A" plate on it and if, as is almost always the case, the protective device is set to that rating, no generation can be connected downstream."

True - the question is who is responsible for up-issuing that plate in effect, and saying it can be changed - hence my comment about someone coming in and assuming the mantle of design authority.

What is needed is enough info to re-plate it to say

'xxA via input no 1 and

zz A via input no 2,

total outgoing load not exceeding XX+ zz

(Domestic consumer units could/ should say not  more than 100A via main switch, no more than 16A of  embedded generation via any MCB and total load not to exceed 116A for example - but only if someone knew enough about the design to be confident to say to say it was OK - in effect this is the  assumption made but it is not made explicitly, which may not be wise. )

In an ideal world  your factory panel would also say XX amps, Joe Bloggs 01483 xxxyyy  and you could call them and say 'Joe- can we chat about re-purposing that design you did in 1973? ..' Of course that is total fantasy. We usually have no idea who or what assumptions about heat flow, de-rating etc were made by the original makers back at some time and place far away.

It is also a fair bet that if generation was not part of the original remit , then it was never considered that some of the current could ever be going the wrong way, and coming in one of the outputs, so the rating plate will be for single input multiple outputs only - so it may well be OK, it was just never considered.  And to be safe, the full gamut of things that may go wrong need to be considered before doing it.

Again, there will be loads of cases that are just fine, but they need some analysis before leaping.

Mike.

Jam Your post was caught in the automatic spam prevention queue due to 'text repetition' i.e. because you'd repeatedly quoted Mapj (Mapj said) etc. 

It's now been manually approved