Suggestion on subroutine USDFLD

[skatemaverick86]

Hi there,
I have some issues for what concerns user subroutine USDFLD. I am trying to model a CREEP damage on a 2D P.Strain model (CPE4).
Basically I want to set the Young Modulus (E) of the elements to be equal to 1 [MPa] when the damage reaches 0.99 so that the element is no longer able to support stresses and that represents a crack propagation.
I have two issues:
1) is there a problem using USDFLD on 4 integration points elements? Let me explain: For example sometimes it happens that the damage reaches 0.99 in three nodes while the fourth node is still under 0.99. This means that stresses are close to zero in 3 integration points while are really high on the remaining point. Can this affect my solution and especially the damage calculation on the remaining integration point?
2) I would like to understand how abaqus works when calculating the material properties according to the field variable (the damage in this case).
for example this elastic properties definition:
*Elastic, dependencies=1
140000.,0.3, , 0
1. ,0.3, , 0.99
Does this implies that Abaqus is interpolating linearly the Young modulus from 140000 to 1 in the damage range that goes from 0 to 0.99?
In case yes, I do not want this because the elastic modulus should remain the same until damage=0.99 and then be setup to 1. I've tried this definition but I don't know if it's good enough.
*Elastic, dependencies=1
147016., 0.3, , 0.
147016., 0.3, , 0.98
1., 0.3, , 0.99
1., 0.3, , 1.
Any suggestions? Thanks in advance
Fede

 

[Mustaine3]

Afaik will Abaqus itself reduce the stiffness with Damage Evolution, so I don't see a reason for your plans. Read the manual regarding that topic and look at the variable SDEG.