SUBSTEPS_IMPL

number of implicit substeps with constant time step size in EULERIMPL setting

Default: SUBSTEPS_IMPL = 1.0 In order to save more computation time within the EULERIMPL scheme the MESHFREE time step size \( \Delta T_{k}\) can be increased by a factor of SUBSTEPS_IMPL. If the automatic time step size control method based on local errors (see TOL_T, TOL_v, TOL_keps) determines a time step size \( \Delta t_{k}\), then the size of the next MESHFREE time step will be \( \Delta T_{k} =\) SUBSTEPS_IMPL \( \cdot \; \Delta t_{k}\). In order to conserve the given error tolerance the physical entities must be computed with \( \Delta t_{k}\), so that SUBSTEPS_IMPL steps are needed to compute one MESHFREE time step. Furthermore during one MESHFREE step a constant \( \Delta t_{k}\) is used in order to reuse the same matrix, what saves a lot of computation time. Remark: Take care that SUBSTEPS_IMPL is not bigger than 10! Otherwise the automatic time step size control method based on local errors will reject the result and repeat the substep with a smaller \( \Delta t_{k}\). Repetitions of substeps can be suppressed by setting NB_OF_ACCEPTED_REPETITIONS = 0, but then a big SUBSTEPS_IMPL will lead to very inaccurate results. Therefore we recommend SUBSTEPS_IMPL \( \in [2,6]\) to avoid repetitions.
This item is referenced in:
SUBSTEPS_EXPL number of explicit substeps for solving TRANSPORT part in EULEREXPL setting
SUBSTEPS_IMPL number of implicit substeps with constant time step size in EULERIMPL setting
EULERIMPL Higher order implicit Eulerian or ALE motion (recommended among the Euler implementations)