Hello,
I work on a casting model with a phase change in a metallurgy process.
The 2D-axy model works fine and i begin the 3D model.
I work on a 3D coupled problem with different physics: mf for excitation, ht and spf. ht and spf are strongly coupled.
To achieve the solution of my model i have to compute something like this:
stationary solver
step1:mf
step2:ht1
step3:spf1
step4:ht2
step5:spf2
...
step10:spf10
After this sequence i reach the solution of the 2D-axy model.
The problem is that my mf.Qh source term for electromagnetic heating (ht) and Lorentz force (spf) is increasing after each step! I think that problem is linked to mesh density.
I use a specific mesh for mf and ht (dense) and spf (light). spf can not be meshed as fine as ht and mf because of memory issue.
What can i do to avoided the increase of source term? What kind of solver should i use preferentialy?
Thank you for your help
Julien
I work on a casting model with a phase change in a metallurgy process.
The 2D-axy model works fine and i begin the 3D model.
I work on a 3D coupled problem with different physics: mf for excitation, ht and spf. ht and spf are strongly coupled.
To achieve the solution of my model i have to compute something like this:
stationary solver
step1:mf
step2:ht1
step3:spf1
step4:ht2
step5:spf2
...
step10:spf10
After this sequence i reach the solution of the 2D-axy model.
The problem is that my mf.Qh source term for electromagnetic heating (ht) and Lorentz force (spf) is increasing after each step! I think that problem is linked to mesh density.
I use a specific mesh for mf and ht (dense) and spf (light). spf can not be meshed as fine as ht and mf because of memory issue.
What can i do to avoided the increase of source term? What kind of solver should i use preferentialy?
Thank you for your help
Julien