A new global variance reduction technique based on pseudo flux method

In radiation shielding problem as well as nuclear reactor analysis applications, Monte Carlo (MC) method is considered as the most useful tool for its obvious advantage to deal with heterogeneous and complex system model. However, for the deep-penetration problem in shielding calculation, analog MC method is inefficient since particles have little contribution to the estimators in far-source region. Numerous variance reduction techniques have been proposed to improve the calculation efficiency such as the importance sampling and weight window (WW) technique. But these techniques still cannot realize global variance reduction. In this paper, a new pseudo flux method is proposed to realize global variance reduction. It based on forward MC calculation by MCNP and does not need secondary modeling and much implement to MC code. Forward particle flux and relative error (Re) are used to form WW thresholds. The empty mesh cells which unsampled in farsource region would lead to zero WW thresholds. The zero transport parameters have a significant influence on the calculation efficiency. The pseudo flux method based on direct contribution algorithm of point detector flux estimation is used to form pseudo flux as the unsampled tally value. This method would improve the calculation efficiency substantially and accelerate convergence. From the numerical test calculations, the Re distribution was flattened in whole phase region and empty mesh rate was reduced. The efficiency was considered to be better than analog simulation by comparing figure-of-merit (FOM) values. The goal of global variance reduction has been achieved. The proposed method based on forward MC estimates and pseudo flux could be useful variance reduction techniques in MC deep-penetration calculation.