WEIGHT WINDOW GENERATION BASED ON PRE-CALCULATED RESPONSE MATRIX

In deep-penetration calculation, the variance reduction method is very important. However, the generation of variance reduction parameters is usually costly and the result is not always satisfying. In order to balance the accuracy and cost of generation of variance reduction, a lot of variance reduction methods have been proposed and the mesh-based weight window is widely accepted by academic research and industrial calculation. The generation of mesh-based weight window parameters is always a topic in variance reduction research and a lot of Monte Carlo program has implemented different types of weight window parameter generators. In this paper we proposed a new mesh-based weight window generator based on response matrix iteration prior to the Monte Carlo simulation. And we can show that this method can provide a quick estimation to the flux level of different mesh cells in the weight window mesh thus it can give out the estimated mesh-based weight window parameters for global variance reduction. Considering that Monte Carlo programs have used GVR (Global Variance Reduction) method to update weight window parameters during the Monte Carlo simulation, this method can provide a relatively acceptable iteration start point.