Grouping Factor Alternatives

Question

I have a situation in a Theatre environment where the requirement for application of grouping factors has been picked up rather late in the process and the consequent cable size increases (which are significant if the figures from the table in BS7671 are applied) mean the containment is too small and we can't get the cables in the breaker terminals or the connector terminals. 
 Scenario is a series of panels with 400A supplies, with protective devices of various sizes and a number of outlets ranging from 32A single phase up to 3-phase Powerlock (for 250A supplies). The original design was that this was all in one panel, so the panel building standards applied, to the internals, rather than BS7671. Then somebody decided to split the outlets away from the panel by varying distances (from around 2m to 20m) and connect them with trunking. Now BS7671 needs to be applied to the installation of the link cables. Sadly at the time they just used the internal panel cable sizes to size the containment routes and didn't allow for grouping. The panels are used for connection of temporary equipment that a touring show or a given production may need, so not very predictable really. 
 There are typically 6 or 7 circuits running in each trunking link. They are generally of different ratings/capacities and are overall limited by the 400A supply to the breaker panel. This does however mean that they could all be loaded to more than 30% (the figure below which they could be ignored for grouping - generally this works out at 50-70% depending on the panel) so without imposing an operational limitation below the panel maximum which the client may not be very happy about, I'm struggling to see a way round using the tabulated grouping factors. These take the 95mm2 cable for the Powerlocks up to 400mm2 (based on a 70 degree operating temp). We're generally looking at 2 sizes up from the nominal for most cables with this method. 
 The grouping factors say they are for identical cables equally loaded. We have different sized circuits and they may be loaded evenly or unevenly. 
 Whilst we can come up with a number of scenarios of circuits loaded at different levels, and it is is unlikely that they would all be used at once and proportionately evenly loaded, it is entirely possible that all circuits from a given panel may be used at once to more than 30% each. On this basis what options do we have but to use the tabulated grouping factors (which seem to be generally regarded as being pretty conservative)? 
 There are a lot of practical limitations on site, as were considering options around changing to run some (or all) of the circuits in SWA and keep them spaced apart but we likely can't do this everywhere. 
 Anyone have any good engineering suggestions for an alternative approach to the grouping factors/cable calculations? This all starts to get particularly ridiculous where the breaker panel and outlet box are either side of a wall and we may need to put additional termination boxes on both sides so we can upsize about a metre of cable run! 
 Jason

mapj1 · Accepted Answer

I do not think it is quite that strong. You may or may not be able to comply with '7671; you will need to show that the cables will not be damaged by the extra heat due to grouping for cases where the simple table is not adequate. That would need some rigorous mathematical analysis of the thermal situation, and possibly some test jig measurements to back it up. 
 In terms of the paper trail, as a non-standard design it would need to indicate how that is to be accomplished- " see attached analysis doc XYZ for derivation of project specific grouped cable ratings" 
 Rather like using the ERA tables for current rating instead of the ones in BS7671, or makers data for some exotic materials but you would be providing the intellectual back up yourself . And that attached doc needs to be able to stand full scrutiny - neck on the block and all that. 
 But this is a lot of work, likely to raise eyebrows, causes trouble at future inspections, and unless there is a really good reason not to do so, I'd recommend you to be putting some extra trunking in to augment that already purchased and spacing things out so the grouping effect is less onerous and it then meets the more normal way of looking at it. 
 That said, if you like to be able to refer back to a 'standards approved' approach for cables that are neither loaded to 100% nor to less than 30% one could adapt the method of BS IEC 60287 for calculating the current rating of groups of non bundled cables sharing a closed channel . This method calculates an effective increase in ambient temperature from the total power dissipation of all the cables in the channel and and the perimeter length through which heat is to be lost. The effective increase in ambient temperature is added to the external ambient, and the cable size is then selected on the basis of the de-rating factor for this higher temperature . The semi-empirical equation for the effective increase in ambient temperature, dt in degrees C given in BS IEC60287-2-1 is 
 dt = 3*Ptot /perimeter 
 where: Ptot is the total heat dissipation from all of the cables in watts per metre of length P is perimeter length channel, 2 &times; width + 2 &times; depth, in metres the heat dissipation can be calculated from the voltage drop values given in BS 7671. The heat dissipation, per core, is given by I 2 R, where I is the load current and R is the conductor resistance in ohms per metre length. Take acre with the voltage drop values given in BS 7671 for a single-phase circuit - it equates to 2 &times; R. as it is for a 'there and back' path. 
 As it is usually used for mains supply cables in big ducts, I suspect this method is better for very large cables and very large channels, and underestimates the cooling ability of thinner structures - much as a constant current per copper area under estimates cable current carrying capacity. as an aside pre-war MOD designs based on a constant 1000A per square inch oversized cables less than 50A and ran far too hot on larger ones. 
 
 Mike.

Grouping Factor Alternatives

Join us to get the best from IET EngX.

Community Rules & Guidelines