structured mesh for external aerodynamics

Ryan0531 · January 31, 2014, 22:29

hi.. im doing my final year project related to aerodynamic analysis of a solar car.

I'm facing a problem in preparing a structured mesh in 3D with certain y+ value. im using ANSYS meshing in Workbench... i failed to obtain a good quality mesh.. can anyone share some technique with me or some tutorial? Thanks

lovecraft22 · February 1, 2014, 04:13

What is your car shape? Are you sure you want to do a structured mesh?

Ryan0531 · February 1, 2014, 08:01

the car body has airfoil shape..

my supervisor asked me to do a structured mesh since want to examine the boundary layer near the body..

below shows the overall shape

lovecraft22 · February 1, 2014, 08:58

Are you sure that eh wants you to make a "structured" mesh? It'll probably be impossible with your geometry IMHO...

Ryan0531 · February 1, 2014, 09:28

may i knw why it is impossible to do structured mesh?

if can't do structured mesh, is there any other method can be used to define the mesh which are around the boundary layer? thx..

lovecraft22 · February 1, 2014, 10:07

Have a look here for the definition of a structured mesh:
http://www.cfd-online.com/Forums/cfx...ured-mesh.html

As long as your geometry is simple, it is fairly easy to build one.
In your case you may want to do a regular unstructured mesh with different refinement regions around important parts of the car such as a spoiler for instance. Besides, as your professor suggested, you may also want to refine the region of the boundary layer

Ryan0531 · February 2, 2014, 00:25

thanks for reply...

so by refining the mesh near boundary layer is good enough to catch the flow pattern??

can we detect the y+ value for an unstructured mesh?

when i check the y+ value, what is the procedure actually? do i need to check surface by surface or just select the all domain?

sorry cos got too much problem to ask here... since im new to cfd..thx again

lovecraft22 · February 2, 2014, 05:20

Yes, you need to refine in your boundary layer regions. Usually, with NS codes what is done is to create "layers" there, as you can see in this picture (purple region of the mesh):

Then you can use the height of the layer closest to your surface to calculate the y+ value. The y+ is a local characteristic of your mesh and resulting flow, which mean that you need to check it in the regions of interest, so, yes, surface by surface but again, only where you think your boundary layer height prediction is important.

Ryan0531 · February 2, 2014, 22:05

Thanks for clear explanation.

I had done the unstructured mesh + refinement around the car body + inflation layer.. is this more that enough for my case?

i checked my mesh quality using skewness.. it shows that the max skewness is 0.999995, which is more than 0.95.. is this means that i need to repair my mesh? or i can just carry on to solving?

Again, thx for yr reply

joshm · February 3, 2014, 02:50

I'm not sure whether your thesis will require a lengthy literature review, but these solar cars are a fairly popular at universities around the world so there should be a bit in the literature about validation of CFD around similar geometries. These should get you headed down the right path, and with a bit of luck they might explain the logic behind their mesh generation.

Ryan0531 · February 3, 2014, 03:41

Since it is popular, almost all university keep it confidential.. i failed to find any papers that related to its aerodynamic analysis..

somemore, i need to finish this cfd analysis cos still need to conduct wind tunnel test.. lack of time actually..

So.. need opinions from experienced cfd guys..

joshm · February 3, 2014, 04:35

Well I'm probably not what you'd classify as experienced and you've provided no details on your case, but that near-wall boundary looks too coarse to "analyse the boundary layer..".

Ryan0531 · February 3, 2014, 07:49

okay.. then what should i do to refine those mesh again??

lovecraft22 · February 3, 2014, 13:26

Also, you don't have any refinement in the wake…

Ryan0531 · February 4, 2014, 05:15

current meshing at wake region still not fine enough???

lovecraft22 · February 4, 2014, 06:50

Quote:

Originally Posted by Ryan0531

current meshing at wake region still not fine enough???

Nope... But that is only one of the problems... Making the mesh setup usually takes some time, it is not just a matter of 5 minutes...

Ryan0531 · February 4, 2014, 21:04

ok.. Thanks for the advices..

sameben · February 6, 2014, 01:55

Hi
This is a mesh from GridPro.

https://www.dropbox.com/sc/abqj04fi9ntn2ht/gQMtQ9HLxX

Best
Sam

February 1, 2014, 10:07		#6
lovecraft22 Senior Member lore Join Date: Mar 2010 Location: Italy Posts: 460 Rep Power: 18	Have a look here for the definition of a structured mesh: http://www.cfd-online.com/Forums/cfx...ured-mesh.html As long as your geometry is simple, it is fairly easy to build one. In your case you may want to do a regular unstructured mesh with different refinement regions around important parts of the car such as a spoiler for instance. Besides, as your professor suggested, you may also want to refine the region of the boundary layer Ryan0531 likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[snappyHexMesh] SnappyHexMesh for internal Flow	vishwa	OpenFOAM Meshing & Mesh Conversion	24	June 27, 2016 08:54
Exporting structured mesh from ICEMCFD to Fluent?	jeevan kumar	FLUENT	1	January 23, 2012 11:21
[ICEM] Unstructure Meshing Around Imported Plot3D Structured Mesh ICEM	kawamatt2	ANSYS Meshing & Geometry	17	December 20, 2011 11:45
Block structured hexagonal mesh vs automated tetra mesh inside Workbench for CFD	Chander	CFX	3	November 27, 2011 16:24
How to control Minximum mesh space?	hung	FLUENT	7	April 18, 2005 09:38

February 1, 2014, 04:13		#2
lovecraft22 Senior Member lore Join Date: Mar 2010 Location: Italy Posts: 460 Rep Power: 18	What is your car shape? Are you sure you want to do a structured mesh?

February 1, 2014, 08:58		#4
lovecraft22 Senior Member lore Join Date: Mar 2010 Location: Italy Posts: 460 Rep Power: 18	Are you sure that eh wants you to make a "structured" mesh? It'll probably be impossible with your geometry IMHO...

February 1, 2014, 09:28		#5
Ryan0531 New Member Ryan Lai Join Date: Jan 2014 Posts: 10 Rep Power: 12	may i knw why it is impossible to do structured mesh? if can't do structured mesh, is there any other method can be used to define the mesh which are around the boundary layer? thx..

February 2, 2014, 00:25		#7
Ryan0531 New Member Ryan Lai Join Date: Jan 2014 Posts: 10 Rep Power: 12	thanks for reply... so by refining the mesh near boundary layer is good enough to catch the flow pattern?? can we detect the y+ value for an unstructured mesh? when i check the y+ value, what is the procedure actually? do i need to check surface by surface or just select the all domain? sorry cos got too much problem to ask here... since im new to cfd..thx again

February 2, 2014, 05:20		#8
lovecraft22 Senior Member lore Join Date: Mar 2010 Location: Italy Posts: 460 Rep Power: 18	Yes, you need to refine in your boundary layer regions. Usually, with NS codes what is done is to create "layers" there, as you can see in this picture (purple region of the mesh): Then you can use the height of the layer closest to your surface to calculate the y+ value. The y+ is a local characteristic of your mesh and resulting flow, which mean that you need to check it in the regions of interest, so, yes, surface by surface but again, only where you think your boundary layer height prediction is important.

February 3, 2014, 02:50		#10
joshm New Member Josh Join Date: Jun 2013 Posts: 19 Rep Power: 12	I'm not sure whether your thesis will require a lengthy literature review, but these solar cars are a fairly popular at universities around the world so there should be a bit in the literature about validation of CFD around similar geometries. These should get you headed down the right path, and with a bit of luck they might explain the logic behind their mesh generation.

February 3, 2014, 03:41		#11
Ryan0531 New Member Ryan Lai Join Date: Jan 2014 Posts: 10 Rep Power: 12	Since it is popular, almost all university keep it confidential.. i failed to find any papers that related to its aerodynamic analysis.. somemore, i need to finish this cfd analysis cos still need to conduct wind tunnel test.. lack of time actually.. So.. need opinions from experienced cfd guys..

February 3, 2014, 04:35		#12
joshm New Member Josh Join Date: Jun 2013 Posts: 19 Rep Power: 12	Well I'm probably not what you'd classify as experienced and you've provided no details on your case, but that near-wall boundary looks too coarse to "analyse the boundary layer..".

February 3, 2014, 07:49		#13
Ryan0531 New Member Ryan Lai Join Date: Jan 2014 Posts: 10 Rep Power: 12	okay.. then what should i do to refine those mesh again??

February 3, 2014, 13:26		#14
lovecraft22 Senior Member lore Join Date: Mar 2010 Location: Italy Posts: 460 Rep Power: 18	Also, you don't have any refinement in the wake…

February 4, 2014, 05:15		#15
Ryan0531 New Member Ryan Lai Join Date: Jan 2014 Posts: 10 Rep Power: 12	current meshing at wake region still not fine enough???

February 4, 2014, 21:04		#17
Ryan0531 New Member Ryan Lai Join Date: Jan 2014 Posts: 10 Rep Power: 12	ok.. Thanks for the advices..

February 6, 2014, 01:55		#18
sameben Member Sam Join Date: Mar 2009 Posts: 35 Rep Power: 17	Hi This is a mesh from GridPro. https://www.dropbox.com/sc/abqj04fi9ntn2ht/gQMtQ9HLxX Best Sam