Variance reduction methods applied to deep-penetration Monte Carlo problems
A review of standard variance reduction methods for deep-penetration Monte Carlo calculations is presented. Comparisons and contrasts are made with methods for nonpenetration and reactor core problems. Difficulties and limitations of the Monte Carlo method for deep-penetration calculations are discussed in terms of transport theory, statistical uncertainty and computing technology. Each aspect of a Monte Carlo code calculation is detailed, including the natural and biased forms of (1) the source description, (2) the transport process, (3) the collision process, and (4) the estimation process. General aspects of cross-section data use and geometry specification are also discussed. Adjoint calculations are examined in the context of both complete calculations and approximate calculations for use as importance functions for forward calculations. The idea of importance and the realization of the importance function are coverd in both general and mathematical terms. Various methods of adjoint importance generation and its implementation are covered, including the simultaneous generation of both forward and adjoint fluxes in one calculation. A review of the current literature on mathematical aspects of variance reduction and statistical uncertainty is given. Three widely used Monte Carlo codes MCNP, MORSE, and TRIPOLI - are compared and contrasted in connection with many of the specific items discussed throughout the presentation. 75 refs., 16 figs.
- Research Organization:
- Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
- DOE Contract Number:
- AC05-84OR21400
- OSTI ID:
- 5970446
- Report Number(s):
- ORNL/TM-9643; ON: DE86007080
- Resource Relation:
- Other Information: Portions of this document are illegible in microfiche products. Original copy available until stock is exhausted
- Country of Publication:
- United States
- Language:
- English
Similar Records
Monte Carlo techniques for analyzing deep penetration problems
Monte Carlo techniques for analyzing deep-penetration problems