Reply To: Mesh convergence more mesh get lower stress

[sgkcompany]

Thank you for explainatinon, I never thought about high element aspect ratios can cause this problem.

I have to try how different shape of upper roller affect plate's deformation, upper roller doesn't ratate but I give it revolute joint so only rotate in z axis.

Yes all contact use normal lagrange, strange enough when I use augmented lagrange it seems harder to convergence even use 0.01 stiffness factor than normal one.
And plate usually move opposite direction (instead roll in, it move back).

It is true scale, thank you to indicate that, I will improve my element.

I use symmetry to cut the model xy plane and xz plane.
The plate touch the roller in begining.
No other condition applied.
auto substep off use 500, large deflection on.