material orientation in case of laminated tube - NX6 1

[RCMT2]

Does anybody know, how to define material orientation for laminated tube in NX6? I´ve already done similar analysis for plate and I was succesful. I know, how to change material orientation using a vector or a coordinate system, but it seems, that the solver has a problem with the two remaining directions, that should be normal and tangent to the circular cross-section. I tried to change to the cylindrical coordinate system while meshing, but it didn´t work. Or, is the problem somewhere else?..
Thanks

 

[potrero]

I assume that you're doing your FEM modeling with NX6 Advanced Simulation?

How do you want your material primary directions oriented? Along the tube axis?

 

[potrero]

Also, is it a shell model (2D elements)?

 

[potrero]

If you set your orientation to be along the tube axis (try creating a CSYS along the axis using Insert > Model Preparation > Coordinate System ... ) then the shell element orientations will be established. Then, you can create the laminate angular orientations by changing the layup in the laminate modeler. Make sense?

 

[RCMT2]

to potrero:
Thank you very much for your reply. You´re right, I´m doing this FEM with NX6 Advanced Simulation. The material directions should be parallel to the tube axis. I´m using shell elements CQUAD8. I tried the last tip you suggested, but without any succes. Maybe also wrong material definition of the lamina - there is a lot of options and I have to check it properly. But again, thanks a lot

 

[potrero]

Perhaps, try meshing with CQUAD4 instead (though this shouldn't matter). Are you sure you are manually setting the mesh orientation (ie: Right-click on the mesh in the Simulation Navigator > Edit Mesh Associated Data > Define Orientation > Select your user-defined CSYS ?)

 

[RCMT2]

to potrero:
Thanks again for your advice! I'm quite sure about setting the mesh orientation . As you said, I tried to mesh with CQUAD4. There were some warnings but the solution was done. But stil I had to think about putting on CQUAD8, which better represent the geometry. I'm not very sure because I haven't checked the results yet, but hopefully, I figured out the problem. After I defined the material orientation by selecting cylindrical CSYS I also assigned this CSYS to the nodes of the model as a nodal displacement CSYS. It removed error message 9137 from my simulation. I hope I'm not wrong.

 

[potrero]

Despite the fact that CQUAD8's have midpoint nodes and "look" like they "represent the geometry" better, I'm not sure they actually are better finite elements to model the curved surface, than CQUAD4's. NASA has written some papers on this topic, I seem to recall.

Good to hear you're up and running.