Evaluation of variance reduction techniques in BEAMnrc Monte Carlo simulation to improve the computing efficiency

Abstract

The application of Variance Reduction Techniques (VRT) in Monte Carlo (MC) codes brings significant improvement in efficiency and a significant profit in the time of simulation. Monte Carlo radiation transport codes have provided by these techniques to solve the time problematic. BEAMnrc Monte Carlo code has several techniques, that have enabled to be the best and fastest code, Including rang rejection, photon forcing, splitting of photon or electron, (Direction, Selective and Uniform) Bremsstrahlung Photon Splitting and Russian roulette (RR). This paper aims to analysis the variance reduction techniques available in BEAMnrc code, validation BEAMnrc model for simulation a Saturne43 accelerator of 12 MV photon beam, and comparing between MCNPX and BEAMnrc. The results obtained show that employing of Direction Bremsstrahlung Photon Splitting (DBS) technique alone or combined with others techniques lead to enhance the efficiency in BEAMnrc simulation. It improves by 16 times compared with Selective Bremsstrahlung Photon Splitting (SBS), 3.7 times (without RR) and it increases about 29 times (with RR) compared to Uniform Bremsstrahlung Photon Splitting (UBS), and also the efficiency increases by factor of 1130 compared to analogue simulation. A good agreement between measurement and calculation of PDD curves for 10×10 cm2 of 12 MV photon beam with local differences less than 2% has been obtained. BEAMnrc has an efficiency greater than MCNPX code in case of analogue simulation (without VRTs).

Keywords

• Variance reduction techniques (VRTs)
• Monte Carlo
• BEAMnrc
• Saturne43