remeshing due to negative volume error
I'm trying to simulate a blade moving with 3 DOF in space (translate along a 2D plane and rotate within the plane). I am using a moving mesh method to do this.
I have a blade with a cylindrical subdomain around it. The outer domain is a large square. I am attempting to translate and rotate the subdomain within the larger domain. This is so the mesh within the subdomain can remain constant while the mesh in the large domain absorbs all deformation as the flow in the outer domain is very minimal.
When i try to move my subdomain at different velocities i run into negative volume errors. I understand this is due to my mesh calapsing on itself. I can avoid the error using a mesh stiffness based on mesh volume size however this causes the mesh to stretch greatly on 1 side of the inner domain and squish on the other (this makes sense as when the mesh increases slightly in size each time step the stiffness will decrease causing it to exponentially stretch).
When i run my simulation using a mesh stiffness based on boundary distance the mesh retains a much better structure however, when i increase the velocity (from 0.5 to 2m/s and up) i cannot get it to finish a run without coming across a negative volume error.
My first question is does anyone have any suggestions as to fix the negative volume error while still using a boundary distance based mesh stiffness. I've played a bit with the time step size but that doesn't seem to fix it.
My second question is if i have to set a remeshing critiria than is it possible to set it to only remesh the outer domain (keep in mind the inside domain is not a different domain but a subdomain) in order to reduce computational time and increase accuracy through not remeshing where the main part of the flow is.
My last question is just about 1 setting. When setting a moving mesh, for my plate wall boundary condition, the "Wall Velocity Relative To" option. Should that be checked? and if so which should be used boundary frame or mesh motion. I'm not quite sure what that means.
After examining my mesh further, i can see that the negative volume is occurring further away from the subdomain walls. It seems as though the mesh is folding in on itself to the side of the subdomain (so if it moves in x direction, elements fold in on its self along the y side away from the interface).
Any help is appreciated. Even if anyone can point me towards remeshing tutorials. I've tried to find some stuff but nothing really takes you through it step by step.
|All times are GMT -4. The time now is 17:56.|