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.