Modelling flat plate having more than two thickness using shell elements

[MR.VIJITH2802]

Hi

I am analyzing a uniformly loaded rectangular plate supported along the long edges using shell elements.

The plate is 10mm thick, except at center where a circular portion is 12mm.

I have modelled 10mm and 12mm plates as separate surfaces and used surface to surface contact as bonded. However, the results show that 12mm circular plate at center does not bend along with rest of the plate. The results show that there is a kink at the boundary of 12mm plate.

How does one model flat plate with two thickness using shell elements

[John_Holtz · Accepted solution]

Hi @MR.VIJITH2802 

Here's your post with the images attached! You should delete your post on the Inventor Nastran forum.

What result is shown in the image of the deflected shape. Is the result the displacement showing that the maximum in the rectangular plate and circular section is 2.4E-6? Everything along the center has a displacement in the red, so the two surfaces are at least bonded together. Does a different view show the thicker, circular plate is not bending?

[MR.VIJITH2802 · Accepted solution]

Hi 

Thanks for the solution. Modelling plate as solid, splitting the solid and then extracting surface has given good results.

Earlier I had used sketch command to draw rectangle and a circle inside the rectangle. Used surface patch command to generate 1) circular face and 2) Rectangular face minus the circle (screen shot attached). The deflections is continuous at the intersection of circular face and rectangular face.

Is the above procedure acceptable? I am I missing some steps.

Regards

Vijith V

[John_Holtz · Accepted solution]

Hi Vijith

Your steps for creating the patches are correct.

In many cases, it does not matter how the model is created, meaning solid model to midsurfaces versus surface model. By the time the models gets to the solver (through the Nastran file that Inventor creates), the solver has no idea how the model was created. What matters is whether the results are correct: the deformed shape is expected, the reaction forces and moments equal the applied loads. the magnitude of the displacements and other results is expected, and so on.

[MR.VIJITH2802]

Hi John

Thanks for the reply.

Regards

Vijith V