Learning Forum

[satri]

If a model gives consistent results that converge to a particular value when the edge size is split with ONLY MULTIPLES of 4 (not to any other edge size) can you consider that as grid-independent?

 

 

[Nanda Veralla]

If your mesh is fine enough, you’ll get mesh-independent solution and you’ll see a minor difference in averaged and unaveraged results.

Mesh size affects the accuracy of the results. High gradients of quantities such as stresses and strains, through the element, tells us that element size is too large. Display options such as averaged, unaveraged and node difference methods can be used to check if the element size in your model is appropriate.  Have a look at this Ansys innovation course on numerically accurate results, in this course you’ll get to learn about appropriate element size and obtaining accurate results.

Numerically Accurate Results - ANSYS Innovation Courses

Regards,

Nanda.

Guidelines for Posting on Ansys Learning Forum

How to access ANSYS help links

 

 

[satri]

No, I don't see a minor difference. I have a circle with a number of divisions on it as (30,32,45,48,60,75,76) among these I get converged results for 32,48,76 but not for 30,45,75. that's why I am not sure if this solution of fluent can be called grid independent.

please let me know

[peteroznewman]

I suggest you post questions about convergence of Fluent models in the Fluids section of this site. When you do, please show the mesh, describe the boundary conditions and the flow model (laminar, turbulent) and show the results so you can get some advice targeted to a CFD model.