Fix OpenMC lost particles
I determined the cause of the lost particles. Overlaps were generated because the pin lattices were not constrained by a prism/surface. This fixes that by either determining what the innermost radial surface is if there are radial regions defined in the subassembly, or by calculating the bound using the supplied pitch and number of rings provided.
Another bug I found is that the calculated radius for adding a wire wrap was incorrectly dividing the wire wrap diameter by 2. That is fixed and tests updated accordingly.