Nagative Jacobian Issue

[sungho91]

Hi all,

I've encountered issues with negative Jacobian elements after enforcing specific node movements.
In my model, there are considerably large deformed elements present on the top and bottom boundaries. To retain the original shape of the domain during remeshing, I opted for a two-step remeshing approach (#3955).

This approach seems to be functioning as anticipated. However, I'm facing issues with negative Jacobian values when attempting to flatten the nodes along the top and bottom boundaries. I initially thought that the flattening process would resemble stretching, but I'm uncertain why this is resulting in negative Jacobian values. Also, I have a gut feeling that the issue could be resolved by adjusting the middle nodes together. I used submesh of top and bottome boundary to do flattening, however, I'm currently unsure how to control the middle nodes of elements on the top and bottom boundaries in the super mesh.

Any insights or comments would be greatly appreciated.

Thanks

[vladotomov]

This operation is changing both the local aspect ratio and skew angles (as evident on the flattened face). If by stretching you mean that it changes only the aspect ratio, this is not the case. Tangling is expected.

See miniapps/meshing/fit-node-position.cpp - it will move the top nodes where you want them while keeping the elements valid (by moving all internal nodes as well).