Learning Forum

[abdullahhassan]

Hi, 

Below is the image of my geometry. i am trying to mesh it via fluent meshing. 

I have use multiple variation of body sizing, max and min cell size in surface mesh and volume mesh but the minimum cells i am obtaining is 3.2M. 

The cells count can increase with more small size meshing but i am unable to reduce the cell count and simulation got very very slow with this number of cells. 

can anyone guide me on how to reduce the cell count. 

[Keyur Kanade]

You will need to define correct sizing. Please use some local sizing such as face sizing. 

In global settings please use only curvature sizing. If you use proximity then it will increase cell count. 

Please go through help manual for more details 

https://ansyshelp.ansys.com/account/secured?returnurl=/Views/Secured/corp/v221/en/flu_ug/tgd_user_sf_types.html

Regards,

Keyur

How to access Ansys Online Help Document

Guidelines on the Student Community

Fluids Engineering Courses | Ansys Innovation Courses

 

[abdullahhassan]

Hi, thank you for your guidance. do surface meshing max and minimum size has an effect on overall volume meshing as volume meshing has its own maximum and minimum sizing. 

[merveala01]

Hi, from my personal experience using poly mesh is better to reduce both surface and volume mesh cell count. Also I think the most important part is small circle sides for you. Please use local mesh and other sides can more bigger. 

[Rob]

Are you meshing the metal casing too? 

[abdullahhassan]

Dear Merva & Rob,

Thankyou for your guidance.

Yeah the whole geometry is being meshed. I have put local sizing to internal domain with inflation layers and after playing with some sizing, the total cell count reduced to 2,159,730 (2.15M). I have used the poly hexcore mesh and minimum orthogonal quality is around 0.19. 

But the calculation time is still very slow. I am using eulerian multiphase model. 

How can i improve the computational speed of my simulation. 

I have done another simulation with different geometry of same nature & cell count was 1.6M at that time and polyhedra cells was used. the computational speed of that simulation is much better than today's simulation.

[Rob]

Not meshing the casing may well help. Have a look at Fluent's thin wall options. 

Depending on how much RAM you have you may be causing the hardware to page to disc if the RAM needed is higher than what is available. Then there are model specific effects.