Assume the arc-flash boundary ( AFB ) for an arcing fault on the line-side of an Incoming Air Circuit Breaker ( ACB ) of a 415V MCC is 10m.
Can the AFB be defined as the locus of a semi-circle of radius 10m with the line-side of the ACB as the centre of the semicircle ?
If the Incomer ACB is in the centre of a 12m long MCC and the AFB for an arcing fault on the main horizontal bus is 2m, then is calculating AFB for other areas other than the line-side of the MCC Incomer ACB meaningless ?
ie. at each end of the MCC, the AFB for an arcing fault on the line-side of the Incomer ACB exceeds the AFB for an arcing fault on the main horizontal bus at the end of the MCC.

It is more of a three dimensional spherical type of boundary rather than a simi-circle. Literally, the distance from the arc where the Ei falls to 1.2 cal/cm2. Depending on the design of the MCC i.e. individual cubicles with branches isolated from mains, many will treat the main and branches separate. One set of calculations for the branches with the main defining the duration and another set of calculations for the main.

It sounds like your case however, the incomer boundary might override any other boundaries. The easiest thing to do is remember the AFB only affects people that don't belong there (no PPE) so just go with the big boundary and keep everyone out of the area when the Arc Flash hazard is present.