changing cell dimension during simulation

[Hamid]

Hello Bernd,

I am trying to simulate coarsening of cementite particles in Fe-C system using MICRESS. For that I need to start with small particles first, thus I have to use smaller cell size. However, at the later stage of coarsening when the cementite particles have grown, I would like to change the cell size to larger one to speed up the process and reduce the simulation time.
I am wondering if that is possible in MICRESS?

As always, Thank you
Hamid

[Bernd]

Hi Hamid,

as you can see, I split your question from the other thread because the topic is a bit different...

Changing the resolution in the course of a simulation is a difficult task because it is not just enough to rescale the grid spacing - the already existing grains have to be remapped on the new grid, i.e. the number of grid cells need to be changed. This, of course, cannot be done on a running simulation, the only chance is to make a restart with the new resolution.
The way to try it would be to write out the actual grain structure as well as all the composition fields into ASCII or vtk files, to make a restart with the new reduced number of grid cells, and to read in again the corresponding grain and concentration information.
It is not possible to read in diffuse interfaces by this way, so there will be some numerical disturbations after restarting until you get back proper diffuse interfaces with the new resolution. Especially in case of TQ coupling there could be also numerical problems, but it should be possible to manage that.
You certainly have seen that there was a thread on how to read in concentration fields from file.

Bernd