I have done aorta simulations in Fluent for three meshes. How can we say which one has the better quality for my further tests? The results are nearly the same.
To check the quality of your mesh, you can use the Report Quality button in the General task page.
http://aerojet.engr.ucdavis.edu/fluenthelp/html/ug/node155.htm
An important factor that show mesh quality is skewness. if you use Gambit in mesh examine select equisize skew and choose range. if your skewness is below 0.75-0.8 for 3D geometry your mesh is acceptable and if it is great than 0.85 your mesh is worst.
age mayel bodid mitonid ba man tamas begirid ke kamel baraton tozih bedam.
For structured meshes you can do a Richardson extrapolation to determine the discretization error of your mesh (grid independency test).
Skewness is an important factor for mesh quality. Furthermore, actually you can try to do the calculation first, and identify whether the mesh is well enough for the calculation. For example, you can calculate it for 100 steps, then check the convergence.
All answers are right, and Abhilash has given the fluent tool to measure if the mesh is OK or not. Nevertheless, a good mesh is determined also for the models that you are using. This metrics are only geometric information about the mesh, no if the mes is good or not for a set of EDP's e.g. SST, k-e. or LES models. Each physics have its own needs when you generate a mesh. Some problems require a maximum element size or a minimun, or both to capturate the physics properly - what I think that is your doubt. "Is my mesh suitable to my problem?" A good metric is not enough to assurance that you have a good mesh or a poor mesh. The physic nature of the problem must be accounted for in this process also.
The good metric will assurance to you that the convergence problems will be avoided or minimized. Residues oscillations, slow convergence, divergence, can arise from this poor mesh or the inadequate use of one mesh that dos not capture properly the physics.
With the assumption that are all OK with the metrics and the physics. And if you have three grids with solution very next, measure that difference, and based on criteria of change of solution, you can determine if you have a grid independence and how these 3 meshes you can work inside of your criteria.
Computational effort can be a issue and a coarse mesh can be used to work, but with the knowledge that you does not know if you problems is mesh independence.
Good look Patricia!
you can able to do this by grid independent test through this test you can able to judge the best mesh for your problem.
- Badges
- Science topic