Suppression du régime transitoire initial des simulations Monte-Carlo de criticité

Abstract : Criticality Monte Carlo calculations aim at estimating the effective multiplication factor (k-effective) for a fissile system through iterations simulating neutrons propagation (making a Markov chain). Arbitrary initialization of the neutron population can deeply bias the k-effective estimation, defined as the mean of the k-effective computed at each iteration. A simplified model of this cycle k-effective sequence is built, based on characteristics of industrial criticality Monte Carlo calculations. Statistical tests, inspired by Brownian bridge properties, are designed to discriminate stationarity of the cycle k-effective sequence. The initial detected transient is, then, suppressed in order to improve the estimation of the system k-effective. The different versions of this methodology are detailed and compared, firstly on a plan of numerical tests fitted on criticality Monte Carlo calculations, and, secondly on real criticality calculations. Eventually, the best methodologies observed in these tests are selected and allow to improve industrial Monte Carlo criticality calculations.
Document type :
Theses
Complete list of metadatas

Cited literature [34 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00779772
Contributor : Florent Breuil <>
Submitted on : Tuesday, January 22, 2013 - 3:39:07 PM
Last modification on : Tuesday, October 23, 2018 - 2:36:04 PM
Long-term archiving on : Tuesday, April 23, 2013 - 3:54:57 AM

Identifiers

  • HAL Id : tel-00779772, version 1

Citation

Yann Richet. Suppression du régime transitoire initial des simulations Monte-Carlo de criticité. Mathématiques générales [math.GM]. Ecole Nationale Supérieure des Mines de Saint-Etienne, 2006. Français. ⟨NNT : 2006EMSE0036⟩. ⟨tel-00779772⟩

Share

Metrics

Record views

437

Files downloads

457