A transport-based condensed history algorithm

Journal Article · Wed Sep 01 04:00:00 EDT 1999 · Transactions of the American Nuclear Society

OSTI ID:678131

Tolar, Jr, D R ^[1]; Larsen, E W ^[2]

---

1. Lawrence Livermore National Lab., CA (United States)
2. Univ. of Michigan, Ann Arbor, MI (United States)

Condensed history algorithms are approximate electron transport Monte Carlo methods in which the cumulative effects of multiple collisions are modeled in a single step of path length s{sub 0}. This path length is the distance each Monte Carlo electron travels between collisions. Current condensed history techniques utilize a splitting routine over the range 0 {le} s {le} s{sub 0}. For example, the PENELOPE method splits each step into two substeps: one with length {xi}s{sub 0} and one with length (1 {minus} {xi})s{sub 0}, where {xi} is a random number from 0 < {xi} < 1. Because s{sub 0} is fixed (not sampled from an exponential distribution), conventional condensed history schemes are not transport processes. Here the authors describe a new condensed history algorithm that is a transport process. The method simulates a transport equation that approximates the exact Boltzmann equation. The new transport equation has a larger mean free path (mfp) than, and preserves two angular moments of, the Boltzmann equation. Thus, the new process is solved more efficiently by Monte Carlo, and it conserves both particles and scattering power.

• OSTI ID: 678131
• Report Number(s): CONF-990605--
• Journal Information: Transactions of the American Nuclear Society, Journal Name: Transactions of the American Nuclear Society Vol. 80; ISSN 0003-018X; ISSN TANSAO
• Country of Publication: United States
• Language: English

Related Subjects

66 PHYSICS
99 GENERAL AND MISCELLANEOUS
ALGORITHMS
BOLTZMANN EQUATION
CHARGED-PARTICLE TRANSPORT
ELECTRONS
MONTE CARLO METHOD