Deep penetration radiation shielding problems often use Monte Carlo augmented with weight windows to obtain sufficiently low variances in acceptable run times. We propose using the random ray method as an approach to generating weight windows. Compared to deterministic alternatives, it allows for straightforward spatial discretisation and continuously samples angular phase space. We compare its performance with an alternative, fully Monte Carlo approach to weight window generation. Effect of factors like high order sources and anisotropy in RR, NJOY vs MC generated multi group cross sections are investigated. Finally, variance reduction for singe- and multiple-detectors will be implemented and tested, possibly including augmented angle-dependent weight windows.

Project lead: Valeria Raffuzzi