Computational geometry for reactor applications

Conference · Fri Jan 01 04:00:00 EST 1988 · Transactions of the American Nuclear Society; (USA)

OSTI ID:5530248

Brown, F B; Bischoff, F G

Monte Carlo codes for simulating particle transport involve three basic computational sections: a geometry package for locating particles and computing distances to regional boundaries, a physics package for analyzing interactions between particles and problem materials, and an editing package for determining event statistics and overall results. This paper describes the computational geometry methods in RACER, a vectorized Monte Carlo code used for reactor physics analysis, so that comparisons may be made with techniques used in other codes. The principal applications for RACER are eigenvalue calculations and power distributions associated with reactor core physics analysis. Successive batches of neutrons are run until convergence and acceptable confidence intervals are obtained, with typical problems involving >10{sup 6} histories. As such, the development of computational geometry methods has emphasized two basic needs: a flexible but compact geometric representation that permits accurate modeling of reactor core details and efficient geometric computation to permit very large numbers of histories to be run. The current geometric capabilities meet these needs effectively, supporting a variety of very large and demanding applications.

• OSTI ID: 5530248
• Report Number(s): CONF-881011--
• Conference Information: Journal Name: Transactions of the American Nuclear Society; (USA) Journal Volume: 57
• Country of Publication: United States
• Language: English

Related Subjects

22 GENERAL STUDIES OF NUCLEAR REACTORS
220100* -- Nuclear Reactor Technology-- Theory & Calculation
ACCURACY
COMPUTER CODES
COMPUTERS
CRAY COMPUTERS
EFFICIENCY
EIGENVALUES
MONTE CARLO METHOD
PHYSICS
POWER DENSITY
PROGRAMMING
R CODES
REACTOR COMPONENTS
REACTOR CORES
REACTOR LATTICES
REACTOR PHYSICS
THREE-DIMENSIONAL CALCULATIONS
VECTOR PROCESSING