OpenMC In Situ Source Convergence Detection

Aldrich, Garrett Allen¹  Univ. of California, Davis, CA (United States)]  Dutta, Soumya¹  The Ohio State Univ., Columbus, OH (United States)]  Woodring, Jonathan Lee¹ 

[1] Los Alamos National Lab. (LANL), Los Alamos, NM (United States)

关键词: SPATIAL DISTRIBUTION;    CONVERGENCE;    NEUTRONS;    COMPUTERIZED SIMULATION;    ENTROPY;    MONTE CARLO METHOD;    RADIATION TRANSPORT;    STOCHASTIC PROCESSES;    CRITICALITY;    O CODES Comp;   

DOI  :  10.2172/1253484
RP-ID  :  LA-UR--16-23217
PID  :  OSTI ID: 1253484
Others  :  TRN: US1601375

美国|英语

来源: SciTech Connect

【 摘 要 】

We designed and implemented an in situ version of particle source convergence for the OpenMC particle transport simulator. OpenMC is a Monte Carlo based-particle simulator for neutron criticality calculations. For the transport simulation to be accurate, source particles must converge on a spatial distribution. Typically, convergence is obtained by iterating the simulation by a user-settable, fixed number of steps, and it is assumed that convergence is achieved. We instead implement a method to detect convergence, using the stochastic oscillator for identifying convergence of source particles based on their accumulated Shannon Entropy. Using our in situ convergence detection, we are able to detect and begin tallying results for the full simulation once the proper source distribution has been confirmed. Our method ensures that the simulation is not started too early, by a user setting too optimistic parameters, or too late, by setting too conservative a parameter.