Abstract
With state-of-the-art plasma edge codes, which consist of coupled finite-volume (FV) Monte Carlo (MC) codes, it is challenging to obtain accurate results for time-dependent simulations in a feasible computational time. For steady-state simulations, it has been recently demonstrated that the speed and accuracy can be drastically improved by choosing more suitable numerical parameters and including post-processing averaging. This article extends the methodology for accuracy assessment to time-dependent simulations. For a simplified one-dimensional (1D) plasma edge model, we compare the numerical accuracy of solutions obtained with an implicit approach, where the code system is fully solved in each time step, to those with a mixed implicit/explicit approach, where only the FV code is converged in each time step. We demonstrate the importance of choosing the numerical parameters adequately in both approaches.
Original language | English |
---|---|
Article number | e201900126 |
Journal | Contributions to Plasma Physics |
Volume | 60 |
Issue number | 5-6 |
DOIs | |
State | Published - Jun 1 2020 |
Externally published | Yes |
Keywords
- accuracy
- convergence
- errors
- Monte Carlo