CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   moving mesh (

Eugene June 26, 2008 19:04

moving mesh
Does anyone get some experience in moving mesh?

Vuik June 26, 2008 22:18

Re: moving mesh
what do you want to know about the moving mesh? Which kind of equations are you solving, and with which kind of grid?

eugene June 26, 2008 22:23

Re: moving mesh
I want to study gas flow in a vibration bed. Need to solve N-S equation. All the mesh will be moved up and down in a similar speed. There is no mesh deformation. Can you provide some resource on this area? Thanks

Vuik June 27, 2008 04:30

Re: moving mesh
you can use a cartesian grid for your simulation. It is cheaper than curvelinear girds. In cartesian grid, the boundary doesn't fit the cells. There are two known techniques to solve this problem, namely cut-cell, or ghost-cell immersed boundary method. In cut-cell method, the cell will be cut, thus you will have arbitrary shaped cells. In 2D is possible to handel with these arbitrary cells, but in 3D case, I don't advice you to use cut-cell. The geometry of the cells become very complex in 3D. Ghost-cell method is based on forcing the velocity by considering a force term in the momentum equations. This term forces the velocities to be zero on the solid boundaries. They are also applicable for moving boundaries, by applying the interpolation (forcing) on every time step. For cut cell: A CARTESIAN GRID METHOD FOR VISCOUS INCOMPRESSIBLE FLOWS WITH COMPLEX IMMERSED BOUNDARIES , T. Ye1, R. Mittal , AIAA"99"3312

For Ghost-cell: A ghost-cell immersed boundary method for flow in complex geometry, Yu-Heng Tseng *, Joel H. Ferziger Journal of Computational Physics 192 (2003) 593"623

wi June 27, 2008 06:32

Re: moving mesh
what is the simplest case to start with i.e. 1-D euler equations ???

eugene June 27, 2008 07:21

Re: moving mesh
Thanks Vuik and Wi for your valuable information. Is there any demo code for me to have a reference. My case is indeed very simple. A 2D geometry and simply vibrating in one direction. Cheers

Harish June 27, 2008 13:34

Re: moving mesh
If you do not have any mesh deformation, you can use rigid body motion of the grid over deforming mesh or immersed boundary method. In this way you will not need to calculate the grid metrics at every timestep and would also not be required to reconstruct the grid at every timestep and it would speed up your computations. Look for papers by Steger around 1980s in AIAA journal.

eugene June 27, 2008 14:02

Re: moving mesh
Thanks, Harish. I will search these papers. By the way, I am still wanting to know whether there is any open-source code for a reference?

All times are GMT -4. The time now is 01:25.