Premium
Enforcing conservation at Monte Carlo level in a coupled finite volume—Monte Carlo simulation
Author(s) -
Mortier Bert,
Baelmans Martine,
Samaey Giovanni
Publication year - 2018
Publication title -
contributions to plasma physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.531
H-Index - 47
eISSN - 1521-3986
pISSN - 0863-1042
DOI - 10.1002/ctpp.201700180
Subject(s) - monte carlo method , physics , finite volume method , conservation of mass , momentum (technical analysis) , estimator , energy conservation , coupling (piping) , statistical physics , kinetic energy , discretization , computational physics , mechanics , classical mechanics , mathematics , mechanical engineering , ecology , mathematical analysis , statistics , finance , engineering , economics , biology
Plasma edge simulations generally require coupling a finite‐volume (FV) fluid plasma discretization with a kinetic neutral Monte Carlo (MC) model. From the MC neutral simulation, source terms are estimated and inserted into the FV plasma simulation, transferring MC stochasticity to the FV simulation. While the FV simulation is based on the physical principle of conservation, most common MC source term estimators do not strictly conserve mass, momentum, and energy. Through the coupling, this also breaks the conservation in the FV simulation. Losing conservation can lead to numerical instability on the one hand and the inability to fulfil global mass, momentum, and energy balances on the other. Presently, the SOLPS‐ITER code rescales the neutral flux launched from the target to mitigate the instability. In this paper, we propose to instead independently rescale each of the source term estimations. We show that rescaling introduces a bias that depends on the number of MC particles and is negligible compared to the statistical error in the simulation. The resulting method restores all conservation properties in the FV/MC code, which will allow monitoring of the convergence through stabilization of the global quantities.