[reespanau]

Hello all!

I'm working on a CFD simulation applied to the study of turbulence in river reaches. I'm using pisoFoam solver. I run my 3D turbulence model in 2 domains. The first one with a coarse grid and the second one with a fine grid.
I want to quantify the loss/gain of information when we are moving from a fine grid resolution (validated base case) to a coarse grid resolution.

I want to use the velocity field as the variable to quantify this loss or gain of information.

There is a technique to compare these two domains and quantify these differences in the results?

I appreciate your help.


L Espana.

[FMDenaro]

Originally Posted by reespanau

Hello all!

I'm working on a CFD simulation applied to the study of turbulence in river reaches. I'm using pisoFoam solver. I run my 3D turbulence model in 2 domains. The first one with a coarse grid and the second one with a fine grid.
I want to quantify the loss/gain of information when we are moving from a fine grid resolution (validated base case) to a coarse grid resolution.

I want to use the velocity field as the variable to quantify this loss or gain of information.

There is a technique to compare these two domains and quantify these differences in the results?

I appreciate your help.


L Espana.

I don't know what formulation you are using but for a generic numerical solution what you need to assess is the grid convergence, that is how the solution change for refined grids. But what you should consider is the h+ step size as parameter.

Only two grids are not sufficient. You can use some averaged velocity profile at different location. But the things depend also on what you are doing (RANS/URANS/LES).

[reespanau]

Thank you FMDenaro, I'm using a 3D DES model. I agree with your advice. Do you know a tool or command in OpenFOAM or python code to study the mesh convergence?

I appreciate it.

L