Ding She

Asymptotic Wielandt Method and Superhistory Method for Source Convergence in Monte Carlo Criticality Calculation

Abstract In loosely coupled systems and large-scale systems, Monte Carlo criticality calculation suffers from slow fission source convergence because of the high dominance ratio (DR). In previous work, the Wielandt method and the superhistory method have been separately proposed to accelerate source convergence. However, although both methods decrease the DR, they are found not able to sufficiently accelerate fission source convergence. In this paper, the effective DR is defined and used to analyze the effectiveness of the Wielandt method and the superhistory method and to theoretically prove that they cannot reduce the computational time to converge the fission source. Accordingly, both methods are modified by adjusting the source population of inactive cycles, and their efficiency after adjustment is also compared. Moreover, the asymptotic Wielandt method (AWM) and the asymptotic superhistory method (ASM) are proposed, and the rules of deciding asymptotic parameters are also discussed. The new methods are implemented into the RMC code and validated by calculating loosely coupled problems and large-scale problems. Numerical calculation results show that AWM and ASM are practical and efficient for source convergence acceleration, which can save 75% to 90% of the computational time to reach a converged fission source.

An Equivalent Homogenization Method for Treating the Stochastic Media

Dispersion fuel is used in high-temperature reactors (HTRs) and some other advanced reactors. It contains a stochastic mixture of microsphere fuel grains or burnable poison grains embedded in a matrix material, which leads to the so-called double heterogeneity problem in the neutron transport calculation. This work investigates an equivalent homogenization method to deal with the stochastic media. In this method, the stochastic media are transformed to a homogenized material by introducing spatial self-shielding factors and preserving first-collision probabilities. A transmission model is proposed to calculate the first-collision probabilities and the self-shielding factors. In addition, the method is extended to treat the stochastic media with multitype grains. The applicability and correction techniques for the proposed method are discussed. The proposed method has been implemented in a lattice physics code named XPZ for HTRs. Numerical results are presented for typical HTR fuel pebbles and are validated against Monte Carlo solutions. It is concluded that the proposed equivalent homogenization method is promising for treating the double-heterogeneity problem and can be conveniently implemented in existing lattice physics codes.

RMC1.0: Development of Monte Carlo Code for Reactor Analysis

This paper describes a newly developed Monte Carlo code used for reactor analysis called RMC1.0, which is based on ACE format library. RMC1.0 is able to estimate criticality eigenvalue, and tally flux/spectrum with collision estimation method or tracking length method. Series of benchmarks and other examples are calculated for validation, which prove that RMC1.0 gives a good performance in both accuracy and efficiency compared with mcnp5. Despite its limitation in geometry processing, RMC1.0 has made a profitable attempt in self-development of Monte Carlo code for reactor analysis.Copyright