Reference points created by center of gravity

[Gunnar]

I don’t know if there is a solution for this, if not then this is just for information:
I noticed that reference points, which are created in the center of gravity of a node set, sometimes lead to very large deviations, if the node set is referenced by asymmetrical geometries. In the example, the actual center of gravity is at Z = -11.43 (calculated by CAD). For the mesh with regular hexahedral elements the value of -8.785 is calculated which corresponds to a deviation of 23%, while the center of gravity for the tetrahedral mesh only deviates by 0.37%.

For the second example, the correct value would be Z = -5.753

[synt]

even it’s allowed in PrePoMax, center of gravity should be using surface set instead of node.

example in similar case

[Gunnar]

Thanks that is way better Then all outer surfaces needs to be selected.

[FEAnalyst]

Btw. Abaqus has an option to automatically move the reference point to the COM of the rigid body. Too bad CalculiX can’t do it.

[synt]

center of gravity is depending on object type

• node group set for edge line or curve
• surface group set for 2D plane face
• element group set for 3D part (unavailable)

usng node or surface group set for 3D part is actually inappropriate.

[Matej]

@Gunnar

I noticed that reference points, which are created in the center of gravity of a node set, sometimes lead to very large deviations

Well, as the name suggests, the center of gravity is computed based on the selected nodes. It was a feature used in the past, where you could select two nodes and then use this feature to find their midpoint (the feature between two points was not yet implemented).

@Gunnar

Thanks that is way better Then all outer surfaces needs to be selected.

No, the volumetric center of gravity might not be the same as the combined center of gravity for the volume faces.