how should I get centroidal values in Quadratic shell elements ? 1

[Znjmech]

Hi All

having Quadratic shell elements, I see that I can obtain centroidal results in Ansys ?

 

[chris456s]

I had the same problem and have a solution.
I have discussed this problem to death with Ansys and with Bureau Veritas.
Bureau Veritas as well as ABS, DNV and a few other standards base their stress acceptance criteria on centroidal values and their in-house software gives those values by default.
In some software the center value is the basic value stored in the results file and all other values are extrapolate from that value.
Unfortunately, in Ansys elements there is not always a Gaussian point directly at the center. Even when there is, center values are not readily exposed in the Workbench mechanical environment and are not by default stored in the results file. This poses a significant problem when trying to compare results to the acceptance criteria of certain standards.

There are two solutions that have worked for me.
Option 1) In a stress object set the "Position" parameter to "middle". For surface bodies this gives the membrane stress rather than the "top and/or bottom".
In the same stress object set the "Display Option" parameter to "Elemental Mean". This will result in the average stress value of the nodes involved in each element.
Of course this is not the exact value at the center, but for Von-Mises Stress it is acceptably close and Classification design reviewer have accepted this for me.

Option 2) Use the "ETABLE" command (see the help text) in a command object. Using the "AVG" ETABLE option results in the "Store averaged element centroid value". This is the default setting for ETABLEs. By default the results are even labeled "centroid value" and (if the "SHELL,mid" command is given prior to the generation of the ETABLE) they are nearly identical to the results seen in option 1.

I hope this helps.
If you find any better solutions, please let me know.

 

[Znjmech]

Thank you so much, sure I will