ANSYS SHELL181 element equivalent

guillegiraldo

Hi,

I have been asked to select the closest element to ANSYS' SHELL181 from Code_Aster.

The main characteristics of the element are:

- 4 nodes

- 6 DoF per node

- Transverse Shear-Flexible with k=5/6

- Logarithmic strain - true stress

- Finite strain

- Optional multi-layer

- In-plane-rotation stiffness through penalty method

- Optional reduced integration with hourglass control

By research from the documentation, the choice is narrowed to DSQ or Q4G, but I am not able to find a clear difference between those two, or a performance indicator.

I wonder if someone more experienced could give some insight to help me make a choice.

Greetings,

pierre_j

Hi,

To my knowledge, there is no finite strain shell element in Code_Aster.

If you really intend to model "large" strains in Code_Aster (let say up to 20%), your best bet will be 3D HEXA20 element with reduced integration (3D_SI modelization).

Bests,

Pierre

guillegiraldo

Thanks for your reply Pierre,

The analysis is going to be linear elastic, so I don´t think there is any problem with large strains.

Now, I think that in this case, the real question is about performance of DSQ vs Q4G models, compared to the ANSYS element.

pierre_j

Hi,

Well I believe the best way to know is to have a look with your own tests.

Note, you might find interesting chapter 5 ("Conclusions") of documentation u2.02.01.

There is a table that gathers application fields for the different shell elements.

Bests,

Pierre

guillegiraldo

What I can see from the document is:

- The optimal element according to tests is MEC3QU9H, which corresponds to model COQUE_3D (curved shell) with quadratic quadrilaterals (9 nodes).

- If there is no curved shells, between DST and Q4G, DST is more precise for in-plane bending, requiring less elements.

What do you think?

pierre_j

Hi,

There is also a sketch indicating you which shell element might be more suitable according to the thickness you are considering (is it really a "thin" structure or not?).

Bests,

Pierre

guillegiraldo

Yes, I considered that, but both elements are in the same category in h/L ratios and in type of possible analysis,

As the analysis is going to be a ship structure, h/L ratios are small, I mean, less than the 1/10 and even less than 1/20.

At the moment the only difference I have found between the elements is the shape function for the shear deformation (linear for DST, quadratic for Q4G, I might be mistaken, please correct me if I'm wrong), and the note about the in-plane bending.

jeanpierreaubry

hello

guillegiraldo wrote:
As the analysis is going to be a ship structure, h/L ratios are small, I mean, less than the 1/10 and even less than 1/20.

and DKT or DST (DSQ) are not enough????????????

jean pierre aubry

guillegiraldo

jeanpierreaubry wrote:
and DKT or DST (DSQ) are not enough????????????

I do think they are. In fact, I completely trust Code_Aster elements, as to offer it as a reliable alternative to ANSYS (in a professional environment).

The discussion so far has been interesting, and helped a lot, and any contribution is valuable, for the sake of the community!

Greetings to all, and again, thanks for your help.