> well you probably should clarify that - you are certainly "unwinding" at the cell interfaces to get an upwind bias in the scheme, right? So that could be alternatively looked at as a central + >diffusive discretrization… So I would contend the artificial viscosity is (i) less direct and (ii) physically based, but could be thought of as viscosity nonetheless. >Certainly if you computed the interface cell flux as the average of the neighbors things would go to hell in a hurry?
The funny thing is that I do not unwind or upwind. The formulation I was talking about is based on centered fluxes, that is they are computed as the average of the neighbors things. Not all the centered fluxes works. The formulation based on the skew-symmetric form of convective term and centered fluxes introduced by Di Pietro and Ern http://www.ams.org/journals/mcom/2010-79-271/S0025-5718-10-02333-1/ actually works in convection-dominated flows. With convection-dominated I mean the following http://youtu.be/G45eHT-hnUQ, not Re=10^6 RANS. By the way that was done with Gnuid, which is based on libMesh ;-). I also use Riemann solvers things for incompressible flows. Artificial compressibility is introduced only at the flux level and there is no \frac{\partial pressure}{\partial time}. What I can say is that the difference is appreciable at low order when the jumps at interfaces are significants. Riemann solver based fluxes helps stability but when the order is increased the differences tend to disappear, as expected. For compressible flows and RANS we use exact Riemann solvers as a first guess. Inexact Riemann solvers introducing additional dissipation might help in some (hard to desperate) cases. Lorenzo On Fri, Apr 19, 2013 at 1:02 AM, Kirk, Benjamin (JSC-EG311) < [email protected]> wrote: > > On Apr 18, 2013, at 5:29 PM, Jed Brown <[email protected]> wrote: > > > lorenzo alessio botti <[email protected]> writes: > > > >> In my experience dG works without stabilization and additional > artificial > >> viscosity > > well you probably should clarify that - you are certainly "unwinding" at > the cell interfaces to get an upwind bias in the scheme, right? So that > could be alternatively looked at as a central + diffusive discretrization… > So I would contend the artificial viscosity is (i) less direct and (ii) > physically based, but could be thought of as viscosity nonetheless. > > Certainly if you computed the interface cell flux as the average of the > neighbors things would go to hell in a hurry? > > > ------------------------------------------------------------------------------ Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis & visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter _______________________________________________ Libmesh-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
