Sphere at Re=1e+7

danny9019 · June 13, 2018, 07:06

Hi,
i'm trying to simulate a 3D-Sphere in water at Reynolds Number 1e+7 (i.e. 10'000'000) in Star-CCM+, but at the moment i'm not achieving the correct CD result of about 0.19. My physics options are:

Velocity Inlet V = 10.014 m/s
Density rho= 998.6 kg/m3
Sphere Diameter D = 1m
Turbulence Model = K-w
Time = Steady
Fluid Type = Liquid.

The Mesh options on the body are:
Number of Prism Layer = 10
Stretching Factor = 1.5

Can someone please help me?
I really appreciate much.

-Danny-

FMDenaro · June 13, 2018, 07:25

Quote:

Originally Posted by danny9019

Hi,
i'm trying to simulate a 3D-Sphere in water at Reynolds Number 1e+7 (i.e. 10'000'000) in Star-CCM+, but at the moment i'm not achieving the correct CD result of about 0.19. My physics options are:

Velocity Inlet V = 10.014 m/s
Density rho= 998.6 kg/m3
Sphere Diameter D = 1m
Turbulence Model = K-w
Time = Steady
Fluid Type = Liquid.

The Mesh options on the body are:
Number of Prism Layer = 10
Stretching Factor = 1.5

Can someone please help me?
I really appreciate much.

-Danny-

Computing viscous drag at such Re number require a very fine grid. You cannot model the wall BC if your goal is computing the viscous forces

danny9019 · June 13, 2018, 07:30

So? What i can do?
Thanks very much.

FMDenaro · June 13, 2018, 07:42

Quote:

Originally Posted by danny9019

So? What i can do?
Thanks very much.

What about your present grid resolution at the wall?

danny9019 · June 13, 2018, 08:03

The first element near to the body is 0.0004m lenght. The Total prism layer thickness is 0.05m.
Thanks again

FMDenaro · June 13, 2018, 08:08

Quote:

Originally Posted by danny9019

The first element near to the body is 0.0004m lenght. The Total prism layer thickness is 0.05m.
Thanks again

you need to compare this height to the BL thickness, a fact expressed by the y+ value. To compute accurately the viscous stress at the wall you need at least 3-4 nodes at y+<1.

danny9019 · June 13, 2018, 08:11

So i have to obtain a y+ value less than 1?

FMDenaro · June 13, 2018, 08:15

Quote:

Originally Posted by danny9019

So i have to obtain a y+ value less than 1?

at least 3-4 nodes at y+<1

danny9019 · June 14, 2018, 05:52

Hi FMDenaro, i've setted the grid in this way:

Number of prism layer = 22
Total prism layer lenght = 5 cm

Achieving an y+ maximum of 0.92 like in the attached picture.

Now my Cd is 0.12
What can i improve to rise the Cd value to the right 0.19?
Thanks so much.

FMDenaro · June 14, 2018, 05:58

Quote:

Originally Posted by danny9019

Hi FMDenaro, i've setted the grid in this way:

Number of prism layer = 22
Total prism layer lenght = 5 cm

Achieving an y+ maximum of 0.92 like in the attached picture.

Now my Cd is 0.12
What can i improve to rise the Cd value to the right 0.19?
Thanks so much.

Still refining the grid close to the wall. Try also different model

danny9019 · June 14, 2018, 06:02

I'm running the simulation with the "All y+ treatment". Do you think it is better use "Low y+ model?

How can i refine the grid close to the wall? Increasing the number of layer? Because with 40 the Cd is still 0.12.

Many thanks.

FMDenaro · June 14, 2018, 07:51

Quote:

Originally Posted by danny9019

I'm running the simulation with the "All y+ treatment". Do you think it is better use "Low y+ model?

How can i refine the grid close to the wall? Increasing the number of layer? Because with 40 the Cd is still 0.12.

Many thanks.

More layers in the range y+<1 so that you can better describe the BL. Try to compare the results obtained with other models.

danny9019 · June 14, 2018, 08:20

In order to do this, i have to split the sphere in the regions where y+ is less than 1 and more than 1, right? In order to construct two different boundary layer?

Thanks again

FMDenaro · June 14, 2018, 08:25

That depends on the grid generation code you are using. Using RANS you can exploit the symmetry of the geometry to solve only a half sphere, using structured grid in the BL and unstructured grid in the outer region.

danny9019 · June 14, 2018, 08:48

I'm currently using RANSE and analyzing a quarter sphere exploiting the double symmetry. If i have get you i've to create two different boundary layer where: one more refined with more layer And the other less refined with less layers.
I have to obtain a value of y+ costant along the body right?
Many thanks FMDenaro

FMDenaro · June 14, 2018, 08:52

Quote:

Originally Posted by danny9019

I'm currently using RANSE and analyzing a quarter sphere exploiting the double symmetry. If i have get you i've to create two different boundary layer where: one more refined with more layer And the other less refined with less layers.
I have to obtain a value of y+ costant along the body right?
Many thanks FMDenaro

yes, you need to retain 3-4 grid points below y+=1 along the whole body

danny9019 · June 14, 2018, 09:01

Quote:

Originally Posted by FMDenaro

Still refining the grid close to the wall. Try also different model

So in the regions where is Yellow/ red i have to create a refined mesh and where is blue a coarse mesh?
Many thanks.

FMDenaro · June 14, 2018, 09:04

Unfortunately, I don't see any figure...

danny9019 · June 14, 2018, 09:07

In The picture attached before this morning

danny9019 · June 14, 2018, 10:32

Quote:

Originally Posted by danny9019

Hi FMDenaro, i've setted the grid in this way:

Number of prism layer = 22
Total prism layer lenght = 5 cm

Achieving an y+ maximum of 0.92 like in the attached picture.

Now my Cd is 0.12
What can i improve to rise the Cd value to the right 0.19?
Thanks so much.

This picture FMDenaro

June 13, 2018, 07:06	Sphere at Re=1e+7	#1
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	Hi, i'm trying to simulate a 3D-Sphere in water at Reynolds Number 1e+7 (i.e. 10'000'000) in Star-CCM+, but at the moment i'm not achieving the correct CD result of about 0.19. My physics options are: Velocity Inlet V = 10.014 m/s Density rho= 998.6 kg/m3 Sphere Diameter D = 1m Turbulence Model = K-w Time = Steady Fluid Type = Liquid. The Mesh options on the body are: Number of Prism Layer = 10 Stretching Factor = 1.5 Can someone please help me? I really appreciate much. -Danny-

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Possible to select Cells next to a SPHERE surface?	keepfit	OpenFOAM	2	November 30, 2016 15:47
[snappyHexMesh] Sphere in a channel by snappyHexMesh	arsalan.dryi	OpenFOAM Meshing & Mesh Conversion	0	November 14, 2016 16:01
[ICEM] meshing a sphere - large deviation from perfect sphere	murx	ANSYS Meshing & Geometry	25	August 15, 2012 12:37
CFX analysis of a sphere - drag too low	Rob Findlay	CFX	6	March 26, 2007 10:11
meshing F1 front wing	Steve	FLUENT	0	April 17, 2003 12:37

June 13, 2018, 07:30		#3
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	So? What i can do? Thanks very much.

June 13, 2018, 08:03		#5
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	The first element near to the body is 0.0004m lenght. The Total prism layer thickness is 0.05m. Thanks again

June 13, 2018, 08:11		#7
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	So i have to obtain a y+ value less than 1?

June 14, 2018, 05:52		#9
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	Hi FMDenaro, i've setted the grid in this way: Number of prism layer = 22 Total prism layer lenght = 5 cm Achieving an y+ maximum of 0.92 like in the attached picture. Now my Cd is 0.12 What can i improve to rise the Cd value to the right 0.19? Thanks so much.

June 14, 2018, 06:02		#11
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	I'm running the simulation with the "All y+ treatment". Do you think it is better use "Low y+ model? How can i refine the grid close to the wall? Increasing the number of layer? Because with 40 the Cd is still 0.12. Many thanks.

June 14, 2018, 08:20		#13
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	In order to do this, i have to split the sphere in the regions where y+ is less than 1 and more than 1, right? In order to construct two different boundary layer? Thanks again

June 14, 2018, 08:25		#14
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,981 Rep Power: 73	That depends on the grid generation code you are using. Using RANS you can exploit the symmetry of the geometry to solve only a half sphere, using structured grid in the BL and unstructured grid in the outer region.

June 14, 2018, 08:48		#15
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	I'm currently using RANSE and analyzing a quarter sphere exploiting the double symmetry. If i have get you i've to create two different boundary layer where: one more refined with more layer And the other less refined with less layers. I have to obtain a value of y+ costant along the body right? Many thanks FMDenaro

June 14, 2018, 09:04		#18
FMDenaro Senior Member Filippo Maria Denaro Join Date: Jul 2010 Posts: 6,981 Rep Power: 73	Unfortunately, I don't see any figure...

June 14, 2018, 09:07		#19
danny9019 Member Danny Join Date: Jun 2018 Location: Rome, Italy Posts: 36 Rep Power: 9	In The picture attached before this morning