CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ICEM] 3d blocking - divergence in Fluent (http://www.cfd-online.com/Forums/ansys-meshing/113597-3d-blocking-divergence-fluent.html)

jyh3134 February 22, 2013 15:51

3d blocking - divergence in Fluent
 
3 Attachment(s)
Hi guys, I'm stuyding a fairing for Human Powered Aircraft.

I managde to make hexa grid using ICEM-CFD. (see picture)

I use O-grid around fairing, and set 1st cell distance to 0.04 for y+<=1.
I checked 2x2x2 determinant (above 0.7) , and angles(above 18 deg).

and I convert to unstructured mesh, and set BC.

After I ran the simulation in Fluent, the residuals of continuity diverged at 200~300 iterations. (check, scaling also done)

The flow condition is
U=8m/s
k-omega SST (even laminar model failed, too)
others : default for quick trying.

I'd like you to see ICEM files for detail, I linked it below.
https://www.dropbox.com/s/h3drel9swnyfb1l/diver.zip

(You should uncheck 'SOLID' part whick is inner block in fairing.)

I hope someone helps me for finding my fault.

best&regards.

jyh3134 February 22, 2013 20:18

solve
 
I solved the problem by regenerating new mesh by myself.;)

The aspect ratio caused this problem i think.

Thanks for your interest.:D

Best & Regards

Far February 23, 2013 06:13

Look on this blocking. Fairing3 has o-grid in the wake region.


https://dl.dropbox.com/u/68746918/Fairing2_Far.zip

https://dl.dropbox.com/u/68746918/Fairing3_Far.zip

jyh3134 February 23, 2013 08:20

Thanks Far!!
 
Above all, I'm very glad to receiving your help.:)

I like your first suggestion due to easy being easy to understand blocking.

You should split around SOLID part and wake region with o-grid
after split tail of fairing with o-grid(for inner circle), right?

and some other blocking strategy of yours are very helpful for me.

It's well worth enough to run simulation again with new mesh on your suggestion, and I'll let you know the results in Fluent.

-----------------------------------------------------------------
The simulation failed again.:eek:
So I'm certain that working in geometry had some error.(dividing parts or rapiaring geometry topology,etc.)

So I need to make new blocking again like yours with another geomety which was worked well.
It's very similar to previous one, but not identical, so can't match blk file.



Anyway, Thanks for your help again, and I have a question to you personally.

Are you working in ANSYS inc? I know SIMON is, and think that you should be.

I believe that the greatful help and support of yours(likes threads, YOUTUBE videos, etc.) is one of the strengths of ANSYS.;)
Best& regards

Far February 23, 2013 09:35

Quote:

You should split around SOLID part and wake region with o-grid
after split tail of fairing with o-grid(for inner circle), right?
After making o-grid around surface (fairing), I made split at some lengths downstream. Then I chose those downstream blocks (use index control) and made the o-grid. Blocking may look like something complex due to many edges.

Quote:

So I'm certain that working in geometry had some error.(dividing parts or rapiaring geometry topology,etc.)

So I need to make new blocking again like yours with another geomety which was worked well.
It's very similar to previous one, but not identical, so can't match blk file.
No need to re-create blocking. If your geometry is similar, you can use older blocking with minor adjustments.
http://www.cfd-online.com/Forums/ans...-geometry.html


Quote:

Are you working in ANSYS inc? I know SIMON is, and think that you should be. I believe that the greatful help and support of yours(likes threads, YOUTUBE videos, etc.) is one of the strengths of ANSYS.;)
;) I hope, one day I will be there...

PS : what are the units and boundary conditions you are using... Is there any thread in Fluent section?

jyh3134 February 23, 2013 11:59

I used BC with Velocity Inlet(upstream, top,bottom,side) , Pressuare Outlet and symmetry. Because the flow is incompressible. ( U=8m/s)

Re is about 1.5e6, so I used k-omega SST turbulent model.

Thanks for kind explain, the index control still isn't familiar with me.
I'll try it again tommorow.(now I'm busy to write onther report.:()

and also wish you get a reward for being so helpful.:) You'll be there I'm sure.

BTW is there more learing material or examples about 3d blocking? :confused:

I read tutorial manual and watched videos(ANSYS inc, yours(turboengineer, right?)), lots of threads in this forum and managed to try them.

But still I can't make structured mesh with the geometry(my final goal) of my study.
someone probably can tell me that the theoretical training already has done, that's problem about my experience and efforts.

But I want to know how you can improve your blocking skills. and i think many people want to know, too.

Best& Regards

Far February 23, 2013 12:24

Did you check the mesh quality in Fluent?


You made the model in mm or m?

Quote:

But I want to know how you can improve your blocking skills. and i think many people want to know, too.
You keep (and other) on giving me geometries ;)

Far February 23, 2013 12:42

I have planned to work on following cases (from geometry to post processing):

1. Wing body (different configurations)

2. Automotive (Ahmed body)

3. Centrifugal and axial compressor

4. Centrifugal pump and volute

5. Wing

6. Heat transfer (preferably heat exchanger and film cooling in gas turbine)

7. Vertical and horizontal axis wind turbine

8. Formula one

9. Any thing else in demand :cool:

We plan to provide high quality solution from scratch to final report. I am also planning to add T-grid, ICEM tetra surface meshing (Patch dependent and patch independent). I also want to attempt full hexa meshing on Helicopter and complex air-craft. We will put ready-made solution and also on call one-to-one sessions.

cfd seeker February 23, 2013 13:07

Quote:

Originally Posted by Far (Post 409693)
I have planned to work on following cases (from geometry to post processing):

1. Wing body (different configurations)

2. Automotive (Ahmed body)

3. Centrifugal and axial compressor

4. Centrifugal pump and volute

5. Wing

6. Heat transfer (preferably heat exchanger and film cooling in gas turbine)

7. Vertical and horizontal axis wind turbine

8. Formula one

9. Any thing else in demand :cool:

We plan to provide high quality solution from scratch to final report. I am also planning to add T-grid, ICEM tetra surface meshing (Patch dependent and patch independent). I also want to attempt full hexa meshing on Helicopter and complex air-craft. We will put ready-made solution and also on call one-to-one sessions.

Nice idea :). What is the name of your company/project so that we can advertise you :)

jyh3134 February 23, 2013 13:23

https://www.dropbox.com/s/fza18hdf6n...0for%20HPA.zip

my recent prj, tin file is here and unit is mm. It made no problem.

quality of old one in Fluent I checked,
Minimum Orthogonal Quality : 2.29615e-05 !!
Maxiimum Aspect Ratio : 1.24145e+06

This value is not equal to the value I check quality of converted uns mesh in ICEM-CFD (above 0.3)
How it can be???

By the way..

Oh, most of projects you listed are present in your Youtube channel.:eek:

I'll try more and more your projects even something seems to be not to associated with my study. Everything is helpful I think.:)

Far February 23, 2013 13:31

These tutorials are available, but I want to increase their quality 10 times with written material and voice.

Mesh is too fine. It should not exceed 0.5-0.6 million.


All times are GMT -4. The time now is 11:47.