From f9c6c53431e6173a1a50bcccfb52edcb01f3ed3e Mon Sep 17 00:00:00 2001 From: bangerth Date: Wed, 29 Feb 2012 16:46:22 +0000 Subject: [PATCH] A couple more notes. git-svn-id: https://svn.dealii.org/trunk@25194 0785d39b-7218-0410-832d-ea1e28bc413d --- deal.II/examples/step-43/doc/results.dox | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/deal.II/examples/step-43/doc/results.dox b/deal.II/examples/step-43/doc/results.dox index 4392c1c48a..cbc8d7dac4 100644 --- a/deal.II/examples/step-43/doc/results.dox +++ b/deal.II/examples/step-43/doc/results.dox @@ -1,12 +1,6 @@

Results

-Still to do: -- play a bit with beta, c_R, CFL -- switch back to residual-based viscosity -- always recompute the Darcy solution whenever we refine the mesh - - The output of this program is not really much different from that of step-21: it solves the same problem, after all. Of more importance are quantitative metrics such as the accuracy of the solution as well as @@ -96,3 +90,12 @@ we solved the system. Consequently, the criterion was pessimistically stated: what we should really compare is the solution we would get at the current time step with the extrapolated one. Re-stating the theorem in this regard is left as an exercise. + +There are also other ways to extend the mathematical foundation of +this program; for example, one may say that it isn't the velocity we +care about, but in fact the saturation. Thus, one may ask whether the +criterion we use here to decide whether $\mathbf u$ needs to be +recomputed is appropriate; one may, for example, suggest that it is +also important to decide whether (and by how much) a wrong velocity +field in fact affects the solution of the saturation equation. This +would then naturally lead to a sensitivity analysis. -- 2.39.5