You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For true-SDC, the sponge does not work since construct_old_sponge_source is a no-op. This is because we treat the sponge implicitly with the CTU evolution, but SDC always needs the source at the time of the quadrature node, and therefore only ever calls the old time sources.
The text was updated successfully, but these errors were encountered:
the way we implement the Sponge in Castro is essentially that of Almgren et al. 2008 (Maestro paper III), which is an implicit update to the velocity:
U^new = U^old - dt kappa f U^new
this is equivalent to doing:
U^new = U^old / (1 + dt kappa f)
this is also referenced in the original Castro paper (Almgren et al. 2010) and the wdmerger paper (Katz et al. 2016).
For a MOL integration, we can't really do the implicit update, since we are just computing the advective terms at a particular instance in time. Instead, we should just implement a sponge source in these cases that has the form:
For true-SDC, the sponge does not work since
construct_old_sponge_source
is a no-op. This is because we treat the sponge implicitly with the CTU evolution, but SDC always needs the source at the time of the quadrature node, and therefore only ever calls the old time sources.The text was updated successfully, but these errors were encountered: