Error stress results with subroutine USDFLD

[ttavio]

Hello.

I am modeling an elastic cylinder with coupled damage, using an axisymmetric model and a USDFLD subroutine, which changes the modulus of elasticity, based on the value of the damage variable.

In this model, internal and external pressure is imposed on the cylinder, where over time the internal pressure is relieved while the external pressure is maintained.

For this model I developed the analytical solution, based on the Lame solution, so that the results obtained show that the stresses (S11, S22 and S33) do not change, despite the damage event.

However, when I simulate in Abaqus the stresses S22 and S33 vary, due to changes in the modulus of elasticity. I simulated two cylinders with different elasticity modulus and disregarded the damage. For these, the stress S11, S22 and S33 were the same.

Why when I use the USDFLD subroutine and consequently the modulus of elasticity varies during the simulation, the stresses also vary, knowing that in the Lamé solution the stresses depend only on the geometry and the load on the cylinder?

 

[FEA way]

What kind of damage criterion do you use ? How is it calculated in your subroutine ? It would be best if you shared the subroutine code here or provided more details about how it works.