Elmer Discussion Forum

Hi Well, basically they refer to the case directory and the relative mesh directory to it. ElmerGUI assumes that both are "." but which means that all the mesh.* files are also located in the case directory. Often its convenient to have the mesh in its own directory i.e. to use Mesh DB &qu...

Well, as the coupling is only conditionally stable you should look at the courant number. I.e. when you get the velocity field the timestep should be so small that v*dt is in same order as the size of the elements. Typically natural convection rolls result to timesteps that are a fraction of a secon...

Hi The outlook does indeed look fishy. What ElementType are you using? There are two versions that vtk can read: the legacy vtk format and the xml vtk format for unstrcutured meshes i.e. the vtu format. I think the latter is used more frequently and hasn't to my knowledge had any serious problems la...

The files related to the solution of the GUI based tutorials may be found at
http://www.nic.funet.fi/pub/sci/physics ... GUI.tar.gz

The cases were run with 64-bit Windows 7 computer running the 32-bit binary revision 4844 of Nov. 25, 2010.

-Peter

Hi To activate heat transfer to fixed external temperature with heat transfer coefficients (h) doesn't require that the object in constant temperature is meshed. So for me it seems that there is no reason to use the thermal gap model. Simply give External Temperature = 307 Heat Transfer Coefficient ...

Hi I agree with you that introducing the Navier-Stokes into the greenhouse simulation would be an overkill. However, in Elmer if you turn on "convection" it is assumed that there really is a flow field computed. There is no effective heat conduction models that would take into account the ...

Hi

The weak coupling between N-S and Heat Solvers in only conditionally stable. If your simulations otherwise looks ok this might be the reason. You could try to reduce the timestep, add some relaxation (~0.5--0.7), or try with Implicit Euler.

-Peter

Hi,

For example, So list the dependent variables separated by commas, and refer to them in the components of tx following C-convention i.e. numering starts from zero.

-Peter

Hi "SaveLine" solver of module "SaveData" should do roughly want you want. It saves all the variables and all the nodes at given boundary. It also tries to optimize the order so that if the BC happens the be a line the points would come in order. Of course if the "line"...

Sure, you could use Eulerian and Lagrangian approach. In the Eulerian you simply let "Heat Source" be a function of position, temperature etc. Using MATC script this can be done without coding. You might also want to use a Lagrangian approach where the moving heat source deforms also the g...