Automatic optimized Wielandt's acceleration method of source convergence for Monte Carlo criticality calculation

Slow source convergence of Monte Carlo criticality calculation is a significant problem. Several acceleration techniques have been proposed, whereas Wielandt's method is an efficient method. However a bad estimated eigenvalue (k_e) which is selected often according experience makes the calculation time consuming. In this study, automatic optimized Wielandt's acceleration method has been developed to get optimized estimated eigenvalue based on dominance ratio (DR) for improving the efficiency of source convergence and implemented in Super Monte Carlo Simulation Program for Nuclear and Radiation Process (SuperMC). Numerical calculated results of 5x5x5 array of metal spheres and Hoogenboom benchmark showed that automatic optimized Wielandt's acceleration method can enhance the convergence efficiency. (author)