Quasi-axisymmetric analysis with CalculiX

Hello everyone,

I would like to perform an FE analysis with CalculiX for a geometry consisting of two rings (upper and lower) and between them a series of beams placed every 18°. So, one option would be to model a 3D sector (18°) with cyclic boundary condition. However, due to project specifications, I have to create the model in 2D. Then I would like to model the rings with CAX8 elements and the beam with B32 element.

quasi_axisym

My question is the following. Is it possible to perform such a quasi-axisymmetric analysis and specify that there is a cycle symmetry (the beam repeats every 18°)? I read in the documentation (CAX8 section) that it was possible to model quasi-axisymmetric structures with combination of axisymmetric elements with plane stress elements, and for example model a circular disk with holes along the circumference. However, I don’t think I could apply the method in my case. It is also mentioned the case of springs connected to axisymmetric elements (which seems close enough to my case), but no example or additional information that could help me.

If you have any examples or indications that could help, it would be very appreciated.

Thank you.

Regards,

Luca

That’s an interesting problem but I don’t think you can model it using the axisymmetric approach. You would have to create a 3D model and use cyclic symmetry if you want to simplify it. Otherwise, the only way that comes to my mind is to model this as truly 2D axisymmetric and properly reduce the stiffness of the middle part to account for the fact that in real life it consists of multiple separated beams instead of a continuous sleeve around the whole circumference. Of course, this will require additional analyses to determine the stiffness of this part so it might come down to 3D modeling anyway except that the final model will be two-dimensional.

It also depends on what you want to get from this analysis - if the connection between the rings is a region of interest or not.

1 Like

Anyways the analysis isn’t 2D in ccx. Actually, axisymmetric elements are inflated into a 2° wedge and the results are averaged back into a midplane if requested so on output.

Yet, the requirement for surrogate structure which is uniform in circular direction remains valid. That’s how bolted flanges under axial loads have been analyzed in early days of FEA.

The cyclic symmetry feature of ccx should be perfect for your application, it was designed with multiblade fans in mind.

Thank you for your answers.

Regards