Well, as I said deal.II is integrated into another software (DAE Tools, mostly chemical engineering applications i.e. the fluid flow coupled with several scalar transport equations). Now, it works well. Basically, the system assembly is done only once (no adaptive meshes used). The system matrices/load vector are used to generate a set of diff. equations + additional equations for surface integrals. That way, the surface integrals are evaluated and available at every time step. All these equations are a part of the larger DAE system solved using Sundials IDA. The boundary conditions and generation terms (typically very non-linear) can be set using the variables from the "outer" software and the surface integrals are stored in the variables from the outer software. This typically produces a coupled system with a non-linear FE problem in it. The Newton type non-linear solver in IDA is used to solve the system (during the time step taken by the DAE solver; the analytical Jacobian is available).
Typical results from a FE system are quantities that flow in/out of the system (thus the need for surface integrals). I have a couple of working examples at the moment: a) Heat transfer (surface integrals are the total heat transferred) b) Cahn-Hilliard equation (i.e. used in electrochemistry - modelling of phase separation). This is particularly interesting since it produces excellent animations. Something like this on Wikipedia: Cahn-Hilliard equation <https://en.wikipedia.org/wiki/Cahn%E2%80%93Hilliard_equation>. Many FE/FV software has this as an example (i.e. FiPy, libMesh also somewhere). I can generate some very cool videos for the gallery. It needs a more detailed explanation, but this is the basically it. I am planning to write some longer description soon Dragan On Friday, September 23, 2016 at 10:13:55 PM UTC+1, Wolfgang Bangerth wrote: > > On 09/23/2016 11:30 AM, dragn...@gmail.com <javascript:> wrote: > > Oh yes, that' exactly what I wanted (in the post-processing case). > Thanks. > > And if I want to do it during integration? > > What do you mean by "during integration"? What are you integrating at > that time? > > Best > W. > > -- > ------------------------------------------------------------------------ > Wolfgang Bangerth email: bang...@colostate.edu > <javascript:> > www: http://www.math.colostate.edu/~bangerth/ > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to dealii+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
