Remove or fix tunnel distinction from physical volume info registry

[Laurie Nevay]

From coverage, we see that the physical volume info registry doesn't make use of flag of whether the physical volume info belongs to a tunnel segment or not. This is likely due to the factorisation of beam line construction and placement in detector construction. The flag was originally for a level of optimisation - beam line components would be hit more often than tunnel segments as the beam has to hit the accelerator before reaching the tunnel. Generally we should keep the map of physical volumes as small as possible for the most efficient search for curvilinear coordinates. This was used to keep tunnel segments in a secondary map that was searched afterwards.

I think now since we register the curvilinear world physical volumes this optimisation is no longer useful. It's also not currently used.

I will remove this for the next version.