[Other] Mesh refinement technique

ArathoN · February 6, 2014, 15:24

Hi foamers,

I'm studying the backward facing step case and i want to find the mesh where the solution (i want the wallstressshear) is independent from the discretisation error.

Do any of you have advice on how i can do the refinement? I imposed the first grid point to be in the viscous region and i fixed its valor, then i increased the number of the grid point in the y-direction keeping the coordinate of the first one the same. Is this correct?

Another question how do you decide the grid size along the x-direction? I know that this direction isn't important for RAS models (except in my case for the step region) because generally it is the parallel to the walls; but is there any useful rule? Until now i've decided the grid size by guaranteeing a smooth transition in the key region (in my case in the region near the step the mesh size in x-direction is almost the same of the one in the y-direction). Also i try to guarantee that the line intercepting the center-cells stay inside the geometry (in my case was critical the region near the step).
Are there any more useful rules?

Is there any limit to the expansion ratio? I read that the ER for the normal-direction in the SSTKOmega model should be less than 1.15; are there any other limits for the other models? Does The ER in the parallel-direction pose any problem like the ER in the normal-direction?

Any useful publication is appreciated.

ArathoN · February 8, 2014, 10:45

Please can someone help me here or pinpoint me to any good report/docs?

ArathoN · May 5, 2014, 08:28

I'm beginning the simulation with an uniform mesh, my objective is to refine the mesh until i found the finest mesh with invariant flow properties (until they became independent from mesh size). But is there any criterion to choose how to grade the mesh? I'm thinking to use the info from the uniform mesh to better define the grid points on the boundary layer, in particular to better define the viscous layer (i need to study the reattachment of the flow).

Do you know any paper or book useful? or You can leave your opinion.

February 6, 2014, 15:24	Mesh refinement technique	#1
ArathoN Senior Member ArathoN Join Date: Jul 2011 Posts: 137 Rep Power: 16	Hi foamers, I'm studying the backward facing step case and i want to find the mesh where the solution (i want the wallstressshear) is independent from the discretisation error. Do any of you have advice on how i can do the refinement? I imposed the first grid point to be in the viscous region and i fixed its valor, then i increased the number of the grid point in the y-direction keeping the coordinate of the first one the same. Is this correct? Another question how do you decide the grid size along the x-direction? I know that this direction isn't important for RAS models (except in my case for the step region) because generally it is the parallel to the walls; but is there any useful rule? Until now i've decided the grid size by guaranteeing a smooth transition in the key region (in my case in the region near the step the mesh size in x-direction is almost the same of the one in the y-direction). Also i try to guarantee that the line intercepting the center-cells stay inside the geometry (in my case was critical the region near the step). Are there any more useful rules? Is there any limit to the expansion ratio? I read that the ER for the normal-direction in the SSTKOmega model should be less than 1.15; are there any other limits for the other models? Does The ER in the parallel-direction pose any problem like the ER in the normal-direction? Any useful publication is appreciated.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[snappyHexMesh] Creating very fine, high accuracy mesh with snappyHexMesh	JD_Welch	OpenFOAM Meshing & Mesh Conversion	10	July 5, 2023 12:50
y+ = 1 boundary layer mesh with snappyHexMesh	Arzed23	OpenFOAM Running, Solving & CFD	6	November 23, 2022 16:15
[snappyHexMesh] SnappyHexMesh running killed!	Mark JIN	OpenFOAM Meshing & Mesh Conversion	7	June 14, 2022 02:37
decomposePar problem: Cell 0contains face labels out of range	vaina74	OpenFOAM Pre-Processing	37	July 20, 2020 06:38
[snappyHexMesh] snappyHexMesh & Mesh around a missile	sasanghomi	OpenFOAM Meshing & Mesh Conversion	2	October 15, 2014 20:54

February 8, 2014, 10:45		#2
ArathoN Senior Member ArathoN Join Date: Jul 2011 Posts: 137 Rep Power: 16	Please can someone help me here or pinpoint me to any good report/docs?

May 5, 2014, 08:28		#3
ArathoN Senior Member ArathoN Join Date: Jul 2011 Posts: 137 Rep Power: 16	I'm beginning the simulation with an uniform mesh, my objective is to refine the mesh until i found the finest mesh with invariant flow properties (until they became independent from mesh size). But is there any criterion to choose how to grade the mesh? I'm thinking to use the info from the uniform mesh to better define the grid points on the boundary layer, in particular to better define the viscous layer (i need to study the reattachment of the flow). Do you know any paper or book useful? or You can leave your opinion.