How to improve mesh to get a more accurate result?

Aarib · November 23, 2020, 03:58

Hello everyone!
I'm designing a 2D nozzle with exit flow of mach 5. By reducing the element size I get an output of 4.784. If I decrease the cell and face size then the mesh becomes finer but the solutions starts diverging. Can anyone tell me how can I get an accurate mesh? The image of the mesh attached gives 4.78 mach. If i try to increase its quality or size of face cells or elements then the solution diverges.

EmmTheof84 · November 23, 2020, 08:07

You haven't attached an image so I can't comment on your mesh. However, assuming your y+ is within the proper range for the turbulence model you are using, I wouldn't worry about refining the mesh more manually. I would try to use mesh adaptation as this approach will add more cells where needed (big pressure gradients) and not in the whole domain.

Also bear in mind that your mesh might be ok as it is. Try using second order upwind for solution formulation, or even one of the third order methods, and also try playing with solution relaxation. Reduce the values for density and pressure, this might solve your convergence issue but will add computational time.

November 23, 2020, 03:58	How to improve mesh to get a more accurate result?	#1
Aarib New Member Aarib Irfan Join Date: Mar 2020 Posts: 18 Rep Power: 6	Hello everyone! I'm designing a 2D nozzle with exit flow of mach 5. By reducing the element size I get an output of 4.784. If I decrease the cell and face size then the mesh becomes finer but the solutions starts diverging. Can anyone tell me how can I get an accurate mesh? The image of the mesh attached gives 4.78 mach. If i try to increase its quality or size of face cells or elements then the solution diverges.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
sliding mesh problem in CFX	Saima	CFX	46	September 11, 2021 08:38
[snappyHexMesh] Creating multiple multiple cell zones with snappyHexMesh - a newbie in deep water!	divergence	OpenFOAM Meshing & Mesh Conversion	0	January 23, 2019 05:17
Writing a dynamic mesh motion result to fixed mesh	albcem	OpenFOAM Running, Solving & CFD	0	December 23, 2015 03:24
[snappyHexMesh] snappyHexMesh won't work - zeros everywhere!	sc298	OpenFOAM Meshing & Mesh Conversion	2	March 27, 2011 22:11
Icemcfd 11: Loss of mesh from surface mesh option?	Joe	CFX	2	March 26, 2007 19:10

November 23, 2020, 08:07		#2
EmmTheof84 New Member Manolis Theofilos Join Date: Nov 2020 Location: London Posts: 11 Rep Power: 6	You haven't attached an image so I can't comment on your mesh. However, assuming your y+ is within the proper range for the turbulence model you are using, I wouldn't worry about refining the mesh more manually. I would try to use mesh adaptation as this approach will add more cells where needed (big pressure gradients) and not in the whole domain. Also bear in mind that your mesh might be ok as it is. Try using second order upwind for solution formulation, or even one of the third order methods, and also try playing with solution relaxation. Reduce the values for density and pressure, this might solve your convergence issue but will add computational time.