Nonpositive Jacobian Error in Mixed Shell-Solid Model Imported via Gmsh

General Questions
1

Hello everyone,

I have been working on an analysis that I originally ran in ANSYS. I exported the mesh to Gmsh in BDF format and then imported it into PrePoMax using the INP format. I set up all the connections similarly to what I had in ANSYS, and the mesh is identical, yet I am getting a nonpositive Jacobian error. My model includes both shell and solid geometries. I am investigating a grounding (bottom impact) scenario for a sailing yacht the keel is modeled as a solid and the hull is modeled as shell elements. The upper part of the keel is connected to the hull, while the lower part is free. I defined nodesets at the bolt locations and applied rigid body constraints, and I also added a compression only contact between the two sections of the keel. The analysis runs successfully in ANSYS, but when I replicate the same setup in PrePoMax I get this error. Any help on identifying the cause and resolving it would be greatly appreciated. Thank you.

image
image1222×635 116 KB

2

image
image1555×670 551 KB
Here is my mesh.

3

Do you have adjustment enabled in any tie constraints or contact pairs ? If yes, disable it.

1 Like
5

Can RigidBody be used in shell elements ?

6

In linear analyses it works without issues (unless CalculiX 2.23 is used). In nonlinear ones it may not converge, but there are workarounds such as making individual parts/regions solid: Known CalculiX limitations

7

“Zenggkk” describes a shell model with solids in one-sided contact. In this case, are linear calculations feasible?

8

Contact indeed introduces nonlinearity. So if convergence issues are encountered, it would be good to apply rigid body constraints to solid regions (or use tie constraints if contact is not required). But first things first - the negative jacobian issue needs to be fixed to let the analysis run. And let’s wait for the OPs feedback on that.

1 Like