Simulating nuclear particle transport in stochastic media using Perlin noise functions

Abstract

Many nuclear particle transport problems require the simulation of stochastic media – that is, media whose composition varies randomly as a function of position. Examples include biological tissues, mineral slurries and crushed mineral ores. Such materials are usually represented in particle transport codes as being homogeneous with an appropriate average composition. This approximation can introduce significant errors. A new method is described for simulating the transport of neutral particles in such media using a three-dimensional Perlin noise function [Comput. Graph. 19 (3) (1985) 287]. The technique allows many properties of the stochastic medium to be specified, including the mean scale of the inhomogeneities, their shape and surface roughness and the proportions of the different media making up the stochastic medium. A key property of the new method is that the tracking speed is independent of the number density of the inhomogeneities. The tracking algorithm has been implemented into the EGSnrc [The EGSnrc code system: Monte Carlo simulation of electron and photon transport, NRCC Report PIRS-701, 2000] Monte Carlo code. An example application is described, modelling X-ray fluorescence from a particulate mineral sample. In contrast to the homogeneous approximation, the new algorithm accurately reproduces the observed particle size dependence seen experimentally. The increase in computational time is fairly modest, even for very small particle sizes.