Question about generating grids?

OpenFoam12 · July 21, 2017, 00:53

Hi All,

I am currently working on performing some grid dependence study, and I am facing an issue about knowing how to estimate the number of nodes required for an object as well as the progressions. I would be greatly thankful for any clarifications. Say that I have an object in the following picture;

Object and numerical domain;
CFDonline.jpg

My target is to test this object using epsilon turbulence model. Therefore, according to the law of wall as in the picture attached below, I need to be in the Log-law layer. y+ in this case ranges from 30< y+<300 as in the picture below.

Law of the wall;
500px-Law_of_the_wall_(English).svg.png
Ref: https://en.wikipedia.org/wiki/Law_of_the_wall

To preform grid convergence study, I need to create three or four grids and the finest grid has to be say approximately at y+=35. In this case, for free stream velocity 8m/s (Reynolds number: 5.4e+5) and boundary length 1.02m (D=1.02), estimated first cell height using a y+ Calculator https://www.cfd-online.com/Tools/yplus.php would be 1.4e-3 m. What is the rule of thumb to estimate the number of nodes in this case? Is estimating number of nodes based on first cell height?

About calculating progression on the mesh, I have come to the following tool, which seems handy, however, I don't know the total expansion ratio or Bias factor in Gmsh tool? Is there any tool to calculate growth if we have a multi - progression?
Mesh grading;
https://openfoamwiki.net/index.php/S...ng_calculation

Another Excel Tool;
http://www.mediafire.com/file/a915bz...as+factor.xlsx

I would be so thankful if anyone could provide some support.

Thanks you for your time.

Regards,

OpenFoam12 · July 22, 2017, 19:29

Any help!

clapointe · August 7, 2017, 23:35

I've been able to reproduce the figure you included before with LES fairly well. I just made sure to match the estimated size with the first cell, and then used grading to slowly coarsen the grid and reduce the overall cell count. It might take some tweaking, but trying with grading first will save a bit of time.

Caelan

July 21, 2017, 00:53	Question about generating grids?	#1
OpenFoam12 New Member Flyer Join Date: Apr 2017 Posts: 13 Rep Power: 9	Hi All, I am currently working on performing some grid dependence study, and I am facing an issue about knowing how to estimate the number of nodes required for an object as well as the progressions. I would be greatly thankful for any clarifications. Say that I have an object in the following picture; Object and numerical domain; CFDonline.jpg My target is to test this object using epsilon turbulence model. Therefore, according to the law of wall as in the picture attached below, I need to be in the Log-law layer. y+ in this case ranges from 30< y+<300 as in the picture below. Law of the wall; 500px-Law_of_the_wall_(English).svg.png Ref: https://en.wikipedia.org/wiki/Law_of_the_wall To preform grid convergence study, I need to create three or four grids and the finest grid has to be say approximately at y+=35. In this case, for free stream velocity 8m/s (Reynolds number: 5.4e+5) and boundary length 1.02m (D=1.02), estimated first cell height using a y+ Calculator https://www.cfd-online.com/Tools/yplus.php would be 1.4e-3 m. What is the rule of thumb to estimate the number of nodes in this case? Is estimating number of nodes based on first cell height? About calculating progression on the mesh, I have come to the following tool, which seems handy, however, I don't know the total expansion ratio or Bias factor in Gmsh tool? Is there any tool to calculate growth if we have a multi - progression? Mesh grading; https://openfoamwiki.net/index.php/S...ng_calculation Another Excel Tool; http://www.mediafire.com/file/a915bz...as+factor.xlsx I would be so thankful if anyone could provide some support. Thanks you for your time. Regards, Last edited by OpenFoam12; July 21, 2017 at 16:14.

July 22, 2017, 19:29	Any help!	#2
OpenFoam12 New Member Flyer Join Date: Apr 2017 Posts: 13 Rep Power: 9	Any help!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
a question about generating grid used by icem	sweet0575	CFX	1	April 13, 2009 23:31
Reloading grids in Pre	Pascale Fonteijn	CFX	1	July 7, 2003 13:33
Overset Grids	Anton Lyaskin	Main CFD Forum	5	April 30, 2003 09:17
UDF Face Fluxes & Partitioned Grids	Greg Perkins	FLUENT	0	April 6, 2001 01:50
Can I define grids motion as a time functions	Darcy	CFX	4	March 29, 2001 21:14

August 7, 2017, 23:35		#3
clapointe Senior Member Join Date: Aug 2015 Posts: 494 Rep Power: 14	I've been able to reproduce the figure you included before with LES fairly well. I just made sure to match the estimated size with the first cell, and then used grading to slowly coarsen the grid and reduce the overall cell count. It might take some tweaking, but trying with grading first will save a bit of time. Caelan