Modeling an air duct with shell elements, applying pressure.

[Txoleski]

Hi,

It is possible to simulate an air duct with shell elements? I have doubts when i apply an inner pressure.

If I choose an intermediate surface to model it, the surface is bigger than the inner surface of the real pipe. So I suppose that the applying pressure has a bigger effect in the intermediate shell surface...

Ansys has also the possibility of using top and bottom surfaces in the last versions. Has anyone any experience with them?

What could be the best option?

Thanks in advance

 

[TGS4]

The only time that you should be using shell elements (read: thin-shell elements) is when your geometry is such that the thin-shell approximations are appropriate. If you have high D/t ratios, then you should be fine (>10). Otherwise, shells are not for you.

 

[CESG]

When you use a shell element it is wise to choose the elements to represent the mid-surface. Doing that, you guarantee that the results related to "mid" will correctly represent the displacements and stresses on that surface.

For instance, when you model a tube (or duct) by its midsize surface you know that the stresses results for "mid" of shell elements will represent the membrane stresses and that "top" and "bottom" stresses will represent the membrane plus bending stresses.

The other question that sometimes people have is when you need to model a multi-thickness part. In that case you may choose a midsize thickness surface (I like to chose the largest or more rigid component) and keep it to the entire model. The model will look weird but the results are usually good enough for a stress analysis.


MSc Carlos Simoes
Mechanical Engineer

 

[Txoleski]

Many thanks for your answers.

The d/t ratio is bigger than 10.

And what about pressure application? I suppose Ansys calculate the forces related to the pressure directly in each element area, but those elements represent mid-surface.So i suppose that Ansys is applying more force than in a real case. Does it make sense to do a correction in the appled pressure?

Thanks in advance.

 

[TGS4]

Txoleski - if your D/t ratio is greater than 10, then the effects that you write about are insignificant. Don't worry about it. You're into thin-shell theory - see any of the Timoshenko Thin-Shell Theory books.