In general burial is a funny thing, as the thermal properties of the backfill sand soil or whatever, are very dependent on local moisture levels and the proximity to the surface, and real performance can actually be very variable. Assuming that cable depth is ~ half metre or so then any neighbour cable that is further than the surface is not so significant as the surface is the dominant cooling surface (but of course the cables may be deeper...).

The model of use a constant grouping factor and up to some distance and then switching it off suddenly is a bit of a simplification. It also does not consider the long term drying out of the ground around the cables altering the behaviour.

I would look at the factors for 0.125m, 0.25m and 0.5m and use a bit of extrapolation to determine where the grouping distance is sufficiently big enough to be ignored. At the end of the day, you have an  internally heated cylinder (ie the cable) - so a bit of 3 dimensional heat transfer analysis should show you that the heat flux is being lost to an equivalent cylinder of increasing diameter as you move away from the cable


For BS 7671 compliance, you can assume no grouping factors beyond 0.5m however


That said, unless you want to pay a fortune for trenching costs (in your example, you already have a trench approx. 1.8m wide) then it's often far more cost effective to dig a narrower trench, accept some grouping factors at say 0.25m spacing and use selected backfill to get the ground thermal resistance down to get you back to the cable size you intended to use.


Narrower trench, less excavation disposal and a smaller amount of imported sand or stonedust bedding is cheaper from a civil point of view, and still shouldn't result in larger cables due to narrower spacing as you've compensated by reducing the soil thermal resistivity


Regards


OMS