Elmer Discussion Forum

Hi I looked at the code. To me it seems that the 'spatial 2' activates the PhaseSpatial flag that is used to evalute C_p from Enthalphy inside the DiffuseConvective* assembly routine. The problem is that this is that it requires the gradient operator and hence cannot be done outside the routine wher...

Hi, Yes, you can provide the effective heat capacity yourself. However, it should be smeared over such a wide temperature interval that you get accurate enough numerical integration by the default rules. So over a few elements perhaps. -Peter

Hi Stan Probably it is Elmer. I would guess that the problem is in deallocating the created variables. The Velocity uses for certain time integration models an existing column of the PrevValues table. My guess is that this is not properly tested for and deallocation is attempted twice. Deallocating ...

Hi

You need a phase change model since the melting heat is a delta peak for specific heat that cannot be accurately integrated with the default integration rules. See
viewtopic.php?f=7&t=971&p=3151&hilit=spatial+2#p3151

-Peter

Hi, Solvers are executed in order of numbering if not otherwise determined by the "Exec Solver" keyword. So I think you save results before computing and get the right results for inlet because Elmer enforces by default Dirichlet conditions also for the initail guess. Now, either give Save...

Yes, all will be set by DefaultDirichletBCs. Assuming that the default variable is "VarName" it operates on it, or if it is a vector, then on components "VarName i" where i=1,2,...,dofs. -Peter

Hi

You're right. There was a bug and this testing is also rathe stupid bacause it hardly saves any time. So I removed it on the current svn version.

Thank you for reporting.

-Peter

Hi I've also hard time figuring what you want to do. You could solve the system hierarchically by first using StatCurrentSolver for conductors and then StatElecSolver for insulators. Then you should give them different names (both cannot be "Potential") and assign them different Equation d...

Hi The OS workflows are not that well streamlined. Therefore when starting with simple problems I would suggest using a rather short workflow. If you look what people usually use for mesh generation you'll see that gmsh is rather popular. See: http://www.elmerfem.org/forum/viewtopic.php?f=9&t=20...

Hi Elasticity equation exhibits a phenomenon called locking for linear elements. Therefore it is advisable to use quadratic elements. They will give much better accuracy with the same computational cost. The easiest way to increase the accuracy is to set "Element = p:2" in the Solver (or i...