2D and 3D geometry influence to the CFD result

keda xu · August 24, 2021, 01:54

The sonicFoam turbulence model is used to solve the ejector CFD problem. When using the 2D ejector geometry and mesh, the result is good, but when using the 3D geometry and mesh (generated by the ansys fluent), there is only 4 time step before stopped without any error.

The boundary conditions are not changed. Dose anyone know what the solution of this problem is? because of the 3D mesh quality?

geth03 · August 30, 2021, 04:24

the mesh quality for the 3D Geometry is really good.

however, as you can see from your output your simulation runs without converging. either change the maxIter for the smoothsolver, like 2000 or more, to see if it does lead to convergence, or change the solver for U completely.

Upczju · March 10, 2023, 03:43

Quote:

Originally Posted by keda xu

The sonicFoam turbulence model is used to solve the ejector CFD problem. When using the 2D ejector geometry and mesh, the result is good, but when using the 3D geometry and mesh (generated by the ansys fluent), there is only 4 time step before stopped without any error.

The boundary conditions are not changed. Dose anyone know what the solution of this problem is? because of the 3D mesh quality?

hello，do you solve this problem, now I have same error. Please give some advice. best wishes for you!

August 30, 2021, 04:24		#2
geth03 Senior Member Join Date: Dec 2019 Location: Cologne, Germany Posts: 355 Rep Power: 8	the mesh quality for the 3D Geometry is really good. however, as you can see from your output your simulation runs without converging. either change the maxIter for the smoothsolver, like 2000 or more, to see if it does lead to convergence, or change the solver for U completely.