Choosing Transport Events for Initiating Splitting and Rouletting

A study was performed to determine which transport events should be used to initiate a weight window lookup to achieve the best variance reduction performance. A weight window lookup potentially triggers particle splitting (in important regions of phase space) or rouletting (in unimportant regions),...

Full description

Bibliographic Details
Main Authors: Evan S. Gonzalez, Gregory G. Davidson
Format: Article
Language:English
Published: MDPI AG 2021-03-01
Series:Journal of Nuclear Engineering
Subjects:
Online Access:https://www.mdpi.com/2673-4362/2/2/10
id doaj-950faba8946340f1997dd9daad0ae033
record_format Article
spelling doaj-950faba8946340f1997dd9daad0ae0332021-09-09T13:50:13ZengMDPI AGJournal of Nuclear Engineering2673-43622021-03-012109710410.3390/jne2020010Choosing Transport Events for Initiating Splitting and RoulettingEvan S. Gonzalez0Gregory G. Davidson1Nuclear Engineering and Radiological Sciences, University of Michigan, 2355 Bonisteel Blvd., Ann Arbor, MI 48109, USAOak Ridge National Laboratory, 1 Bethel Valley Rd., Oak Ridge, TN 37830, USAA study was performed to determine which transport events should be used to initiate a weight window lookup to achieve the best variance reduction performance. A weight window lookup potentially triggers particle splitting (in important regions of phase space) or rouletting (in unimportant regions), thereby optimizing computational effort. Potential initiating transport events include collisions (both pre- and post-collision), geometry surface crossings, traversing a mean-free path, and streaming across a weight window boundary. Permutations of these initiating events were tested on an urban model with background radiation sources and a spent fuel cask with a neutron dose mesh tally. Generally, all methods perform better with finer weight window meshes. Tracking on weight windows performs well for coarse weight window meshes, while a combination of splitting each mean-free path, geometric surface crossing, and before collisions performs well for fine weight window meshes.https://www.mdpi.com/2673-4362/2/2/10Monte Carlovariance reductionFW-CADISshift
collection DOAJ
language English
format Article
sources DOAJ
author Evan S. Gonzalez
Gregory G. Davidson
spellingShingle Evan S. Gonzalez
Gregory G. Davidson
Choosing Transport Events for Initiating Splitting and Rouletting
Journal of Nuclear Engineering
Monte Carlo
variance reduction
FW-CADIS
shift
author_facet Evan S. Gonzalez
Gregory G. Davidson
author_sort Evan S. Gonzalez
title Choosing Transport Events for Initiating Splitting and Rouletting
title_short Choosing Transport Events for Initiating Splitting and Rouletting
title_full Choosing Transport Events for Initiating Splitting and Rouletting
title_fullStr Choosing Transport Events for Initiating Splitting and Rouletting
title_full_unstemmed Choosing Transport Events for Initiating Splitting and Rouletting
title_sort choosing transport events for initiating splitting and rouletting
publisher MDPI AG
series Journal of Nuclear Engineering
issn 2673-4362
publishDate 2021-03-01
description A study was performed to determine which transport events should be used to initiate a weight window lookup to achieve the best variance reduction performance. A weight window lookup potentially triggers particle splitting (in important regions of phase space) or rouletting (in unimportant regions), thereby optimizing computational effort. Potential initiating transport events include collisions (both pre- and post-collision), geometry surface crossings, traversing a mean-free path, and streaming across a weight window boundary. Permutations of these initiating events were tested on an urban model with background radiation sources and a spent fuel cask with a neutron dose mesh tally. Generally, all methods perform better with finer weight window meshes. Tracking on weight windows performs well for coarse weight window meshes, while a combination of splitting each mean-free path, geometric surface crossing, and before collisions performs well for fine weight window meshes.
topic Monte Carlo
variance reduction
FW-CADIS
shift
url https://www.mdpi.com/2673-4362/2/2/10
work_keys_str_mv AT evansgonzalez choosingtransporteventsforinitiatingsplittingandrouletting
AT gregorygdavidson choosingtransporteventsforinitiatingsplittingandrouletting
_version_ 1717759955207979008