An Overview of Geometry Representation in Monte Carlo Codes

Geometry representations in production Monte Carlo radiation transport codes used for linear-transport simulations are traditionally limited to combinatorial geometry (CG) topologies. While CG representations of input geometries are efficient to query, they are difficult to construct. In the Integrated-TIGER-Series (ITS) Monte Carlo code suite, a new approach for radiation transport geometry engines has been implemented that allows for Computer Aided Design (CAD), facetted approximations, and other geometry types to simultaneously define an input geometry. These techniques effectively allow the user to trade off problem setup time and computational time. For instance, radiation transport on CAD geometries can be 100 times slower than transport on comparable CG geometries. However, since linear-transport Monte Carlo can be solved with high parallel efficiency on massively-parallel computers, CAD representations of complex systems can be analyzed.