Showing results for 
Search instead for 
Do you mean 
Reply

Flow simulation with sheet obstacle

Hi,
I need some help - hint.
I want to simulate flow of medium (air) with sheet obstacle inside. (fig 1)
In FEM model I generate 2D mesh on sheet body and 3D mesh on solid body independently (sheet is not embedded into solid).
I run the analysis (fig 2), but the result of flow shows that the 2D elements are not considered (fig 3).
I get back to FEM model and erase the 3D mesh. In simulation I create Fluid Domain (Fluid Mesh) object.
Now the results make sense (fig 4)
Question is: is this the only way how to take into consideration sheet obstacle in flow? Or is there any chatch in setting of the solution of the first try?
Note: In Femap I can embed the sheet into solid (non-manifold solid) so that 3D mesh match 2D, but in NX I can not do this (topic of another post "how to embed face into solid in NX").

Juraj

6 REPLIES

Re: Flow simulation with sheet obstacle

Maybe I understood you wrongly, but why can't you simply model the fluid domain (solid mesh) with cavity inside ? - meaning solid block wit subtracted smaller solid block from greater one? Is't it the thing you want to simulate?

 

Second thing is how it looks in reality? - mentioned obstacle can not be "suspended in the fluid". It must be fixed somehow right?

Re: Flow simulation with sheet obstacle

Hi Tomek,

if the obstacle would be solid shape - yes, it is easy-going with subtract functionality. But I have to consider thin sheet bodies (e.g. FR - PCB) in complex geometry. The shape of the medium is quite complex and to split it by these sheet bodies would be very unconfortable. And also there is another step - thermal issue, so the solution will be thermal+flow and I need resulst of temperature on these thin walled sheets.

 

Of course that in reality is the obstacle attached somehow to main body, but this is theory example.

Re: Flow simulation with sheet obstacle

See the following in the online help: Meshing for flow surfaces

Re: Flow simulation with sheet obstacle


juraj_zahradnik wrote:

Hi Tomek,

if the obstacle would be solid shape - yes, it is easy-going with subtract functionality. But I have to consider thin sheet bodies (e.g. FR - PCB) in complex geometry. The shape of the medium is quite complex and to split it by these sheet bodies would be very unconfortable. And also there is another step - thermal issue, so the solution will be thermal+flow and I need resulst of temperature on these thin walled sheets.

 

Of course that in reality is the obstacle attached somehow to main body, but this is theory example.


Tomek is correct.  It is trivial to do this, even for sheets.  I'd work at the i_prt level to do this, using split body for instance...

Re: Flow simulation with sheet obstacle

Hi all,

I accept this way by simpler models and if You have strong machine.

Accepted solution for me is to use Fluid Domain functionality which geneates mesh that considers 2D mesh inside without any spliting model. My opinion is that sheet metal parts and PCBs is the best way to model by 2D (midsurfacing). As regards 3D geometry inside the medium, it depends on the complexity and shape. E.g. thin plastic part (case) of constant thickness comes to me smarter to model by 2D elements as to use 3D which generates bigger model (file) what is more time consuming.

Have a good day,

Juraj

Re: Flow simulation with sheet obstacle


juraj_zahradnik wrote:

Hi all,

I accept this way by simpler models and if You have strong machine.

Accepted solution for me is to use Fluid Domain functionality which geneates mesh that considers 2D mesh inside without any spliting model. My opinion is that sheet metal parts and PCBs is the best way to model by 2D (midsurfacing). As regards 3D geometry inside the medium, it depends on the complexity and shape. E.g. thin plastic part (case) of constant thickness comes to me smarter to model by 2D elements as to use 3D which generates bigger model (file) what is more time consuming.

Have a good day,

Juraj


Modeling with 2D entities is exactly how it's been done since I-DEAS ESC, 20+ years ago.  While fully capable of doing eveything in 3D, 2D modeling is the faster way to go for sure.  PCB, components, fans.  The MAYA solver is capable of calculating the drag via the definition of obstacles on the flow surface properties...  In general, for thermal models of electronics enclosures, 2D is the way to go for sure.