geometry repair
 

Hi！ I have a problem of my geometry. I attached a picture. Have you found some holes of in this surface? Does it affect the mesh quality? How could i deal with it? Thanks!

ICEM is very much tolerant of small problems in the geometry. Give it a try and come here back.

But I can't generate prim mesh. And the mesh quality is only 0.4.

0.4 is ok for fluent !! give it a try...

But when i import the mesh into fluent, the convergence curve is bad, like this picture.So I think it's the problem of the mesh. I am not sure where the problem it is.

i think it's more because of your geometry than your mesh.
You outlet is way too close to your inlet... what are you trying to model ??

Oh, maybe. I want to analyse the pressure drop of air flow through the widest path unit (it's made of four grain kernels). So I can't change the model. I am not sure if i use the structed mesh the mesh quality will change. Because I guess the Y-GRID may be better. So do you have some idea how to use structed mesh on my model? Thank you so much!

Blocking can be done. You can refer to this thread, where discussion on blocking is under way.

Thank you so much for your help! I will read it then think about how i can do.

:)Hi, I used Y-grid to my geometry. But there is a problem of my mesh. I am not sure if it is the problem of block or the surface of the geometry. I think there is no problem of my block. Could you check it for me? Many thanks!

The only way to learn ICEM is to share the .tin file ;)

thanks a lot!

oh, complex model. Trying my best and will comeup with solution very soon.

:pit is really complex. I have been confused with it for a month. I tried unstructed mesh but the the convergence is not good in fluent. So I am trying to use structed mesh now. Hope you can sovle it. And I will keep trying. Thank you so much for your help!

The prisms in boundary layer and tetra in the rest of the domain is good choice for this model. And convergence should be good with refined boundary layer. It is some sort of intake?

Sorry, I just started learning ICEM and don't know how to use different meshes for different areas. This is a model for airflow. The inlet is at the bottom surface. And the outlet is one of the side surfaces of this model. Somebody said that the convergence is bad becuase the inlet is too close with the outlet. I am not sure. I think i can't change the model. So, just improve the mesh quality.

You can extend the inlet and outlet domain and apply slip BOCO on them and record the data at the original inlet and outlet planes.

Sorry, i really don't have an idea about your advice. I am a newcomer. I think if the structed mesh can't change the convergence, then I will try another method. So, firsly, try the structed mesh,maybe it will work.:p

I am also sending you the link of training material through pm

Thank you so much for your tutorial! You are really so nice!

I also have a question. If I checked the model, for example the surface after import the igs file into ICEM and it is said that one surface is checked to have a problem. What should i do to repair the surface?

Hi, I solved the problem of the mess mesh. But the mesh quality is still bad. So far, I don't know what the problem it is.

attach blocking

attach blocking?

Yes. .blk file....

I am so sorry. It is 3 am in my region. So, just a little tired. I attached a new file now.Thank you!

Hi, Far. I think I found what the problem exsits. I tried to change the outlet to the upper surface to make the inlet not too close to the outlet. And the convergence is better than before. And I guess the uncrtucted mesh I generated is good enough for the fluent calculation. So, as you said, it is not the problem of the tetro mesh, it is the problem of the geometry. But the structed mesh is also bad. But anyway, could you give me a detailed explanation about the idea you said " extend the inlet and outlet domain and apply slip BOCO on them and record the data at the original inlet and outlet planes. " Thank you so so much for your help. I am looking forward to your reply.

Best Wishes!

keep the inlet and outlet in your geometry, in the meshing tab, declare them as "internal wall". Now, extend your outlet. In fluent, the extended outlet will be the real outlet, and the outlet that you declared internal wall will be interior.
The extension is for a stability in the calculation, you are more intererested in knowing what's going wrong in your outlet (interior one). you kept it interior so you can visualize it easily in the post processing. am i clear enough ??? let me know...

Hey Ali
Lets try our skills of hexa meshing on this geometry. What do you say?


@yuhehuan

If you apply the no slip boundary condition on the extended domain then you will have the extra effects of the boundary layer growth and which are not present in the original geometry.

@far i will , i'm a little bit busy right now... we keep in touch

could you get a good blocking so far ??

Hopefully, I will get one by tomorrow. Any update on multi-zone?

it worked for one geometry but not the other. trying to investigate the problem more... i realized that the one in which it didn't work had a bad geometry... I'll try everything once again this week end.

Hi, Ali. I am sorry I didn't completely understand your idea. I tried to extend the outlet surface using "extend surface" command. But I am not sure which sub-command shoud be used. Need I extend both inlet and outlet in ICEM or in FLUENT? Thank you so much!

Regards!

Rong

I tried to extend the inlet and outlet of my geometry. But I am not sure it's correct. Is is like this after extended?

yes that's was i was reffering to, may be more if you can handle large large element. keep your original outlet. declare them as internal wall, and export them as interior...

Hi, Ali. I tried to defined the inlet and outlets as internal walls. But I am not sure how to define. Do I just name it? I divided mesh on the whole geometry after extend the inlet and outlets. The result is as the following picture. Do I just need to divide mesh on the original mesh and split mesh on the internal walls or divide mesh on the geometry including the extension part? Many thanks!