Abstract

The Random Ray Method is a recently developed multigroup method for performing 3D neutron transport. The new method uses randomly sampled rays that are ray-traced through a geometry to iteratively solve for the flux distribution and eigenvalue—a process which shares some common roots with the deterministic Method of Characteristics. Beyond being very simple and easy to implement, the method’s stochastic nature provides surprising advantages compared to existing transport methods. The Advanced Random Ray Code (ARRC) implements the random ray method and can converge full core 3D light-water reactor problems in high-fidelity on both CPU - and GPU -based supercomputers. Random ray has also recently been implemented in the OpenMC and SCONE Monte Carlo codes. In this seminar, I will explain how (and why) random ray transport works, demonstrate the method on a variety of problems, compare its performance against other transport methods, and present a brief roadmap for future research topics in the field.

For further information on this talks please email Dr Paul Cosgrove: pmc55@cam.ac.uk or Jo Boyle jb780@cam.ac.uk