Learning Forum

[shahria]

Hello,

 

I am trying to implement periodic boundary conditions using the CE command in my APDL script. My model involves non-linear materials, transient analysis and large deflection analysis. I have been able to run the model without any errors. However, I can see some unusual deformation in my model. It appears that the boundary conditions are not properly implemented and I receive the warning message in the picture. I tried different versions of the model with Nlgeom off and applied the same set of constraint equations without any issue. So, I wonder if there are any limitations to using constraint equations in large deformation analysis(Nlgeom on) in ANSYS APDL. I would appreciate any insights and suggestions on this matter.

 

Regards,

Shahriar.

 

 

[dlooman]

You're right.  Constraint equations don't update with changes in the position of the model (large deflection.)  I don't see any workaround.  Is the periodic condition translational or cylindrical (like on the low and high sides of a 45 degree sector of a cylinder)?

[shahria]

Hello dlooman,

 

I appreciate the clarification. I am modeling rectangular 3D RVEs of composites consisting of 3 pairs of opposite faces, and my mesh is non-periodic. I was trying to apply periodic boundary conditions in my model so that all the opposite face pairs would have similar deformation(mirror image). There are some automated techniques to achieve that in LS-Dyna and Abaqus. But I couldn't find any in APDL. That's why I tried implementing the periodic boundary conditions from scratch with constraint equations. Please let me know if there are other approaches to implement periodic boundary conditions on RVEs in APDL rather than constraint equations.

 

Best regards,

Shahriar.

[dlooman]

I don't know of any.  It sounds like you may not have matching meshes on the periodic faces.  That's usually a requirement.  Could  you describe the deformation that requires large deflection?  It can't be a large rigid body rotation of the model.  That would definitely invalidate the periodic assumption.

[shahria]

Hello dlooman,

 

I am modeling the matrix face with the Three Network Model(TNM) which requires nlgeom on (large deflection). I am currently doing a shear test which might have a rigid body rotation component. You can see a typical scaled deformation plot in the picture.

 

 

[dlooman]

The displacements seem very small.  Is the periodic condition between the right and left side of the model in the image?  How much rotation would you estimate there is?

[shahria]

Hello,

 

I have provided that figure to give an idea about the desired deformation and rigid body rotation. The periodic boundary condition is not applied here. My model reaches an average shear strain of 0.1, which corresponds to a maximum rotation of around 5 degrees. Is it large enough to invalidate the constraint equations?

[dlooman]

It just occurred to me that you might be able to use generalized plane strain instead of coupling the two sides.

[shahria]

Hello,

 

Can you please elaborate on why you recommended generalized plane strain? Just so you know, my model is not uniform along the out-of-plane direction.