What could be the problem? I heard various suggestions that I needed to change the mesh, infact I have a minimum orthogonal quality of 0.56 which is pretty acceptable. When I previewed the motion of the mesh, it was working well.once I go to the solver to calculate, I face this problem.
Dear Reza,
Thank you for the reply. I am using the pressure based solver for a 2D axis symmetric transient problem. I am using PISO scheme and first order for all the parameters except presure for which I am using PRESTO. Before the calculation could start, it flags something like multiple subdivision required on --- zone. Why does it need to do it when I have defined the interfaces and walls properly. I can post you with more data if your email id is made available. the dynamic mesh utility is working fine till 100000 iterations with 1e-7 time stepping. It is still on run. I have some contact regions or interfaces, some moving walls. I more doubt is that if we need to specify each and every edge as a wall or interface based on the utility - taking the subdivision problem into account!!.
Dear Durga,
Is your boundary under a generalized motion(transition + rotation) or it is constrained to have reciprocating displacement?
As you are using dynamic mesh you should consider the mirror artificial walls to be set as interfaces appropriately, while due to your boundary motion you should take care of adjacent stationary/moving zones(interiors) appropriate definition, as well.
To make sure of your physical nature consistency, I suggest you to do an steady state fixed boundary simulation to make sure whether the convergence/divergence issue is resolved.
Have you also tried the dynamic mesh tutorial of Fluent which simulates a 2-D axisymmetric problem having a circular arc boundary moving?
I can give you the tutorial files via e-mail if desired.
Thank you for the reply. Please find the attachments. The displacements is a translational displacement and defined by a simple profile file. The 'moving walls' have been defined as moving in the BCs, the stationary have been defined stationary.
In the dynamic zones, the moving wall have been defined 'rigid body', the stationary as ' stationary' and the deforming as 'deforming'. The fluid enclosed in the moving walls has been defined as deforming and layering and smoothing scheme are being used with a spring constant of 0.4. I could see the layering happening.
While doing named selection for the stationary walls, I named 2 edges for each of the stationary walls except the big cylinder section. Infact the real model doesn't bear any stationary walls except the big cyliinder section where the compression based flow field is generated. For simplicity I have considered them as walls to simulate 2 zones of compression and 1 zone of expansion. Is this alteration a problem. At this moment, I am not too very worried about the accuracy of the results. Did I intrude in to the physics of the system and make a basic manipulation?
- Badges
- Science topic