Dear all,

I'm wondering which tangent modulus is used by Abaqus in the linearization of hyperelastic behavior in linear perturbation steps (e.g. direct solution steady-state dynamics). It seems to differ from the commonly used "DDSDDE". I discovered this when extending the verification example "User subroutine for hyperelastic model" for a Neo-Hookean solid by a linear-perturbatin step. The results differ significantly from the results obtained from the built-in Neo-Hookean material model, although the documentation claims that both implementations should yield identical results. Any thoughts on that? Could it be that in leaner-perturbtion steps Abaqus uses a different tangent modulus to preserve work conjugacy when calculatiing the variations? Any help is appreciated!

Kind regards