Problem with the reaction forces in a 3-point bending model

[T.Gmai]

Hi everyone,

I am modelling a 3 point bending test on a SENB specimen with a fatigue precrack using Abaqus/standard. For the FEM model I've used 3D solid to model the specimen and discrete rigid body to model the punch and roller, as for the crack I've used contour integral. I applied a displacement in the RP of the punch.

I need the load-displacement curve from the model but when I compared the results to the experimental test I have obtained higher reaction forces, 2 times higher . I tried to refine the mesh but it didn't work also I've checked my material, it's fine.
I have two different steels, the results of one material are good, I have this problem only with the second material.
any suggestion on how or where could be the problem?

Thank you in advance

 

[FEA way]

Are both models exactly the same, apart from material data ? It would be best to copy one of them and then change only the material constants to make sure that there are no accidental differences between these models.

 

[T.Gmai]

The geometry is exactly the same, the only difference is the crack length. I tried to change the material in one of the model with the closes crack length that I need but still have the same issue. one of the material is highly ductile, the other is more brittle I am afraid that the model is not capturing well the material behavior.

 

[FEA way]

Which material model do you use ? Linear elastic or another one ?

 

[T.Gmai]

Elastic-plastic for both, to capture the plastic behavior I used true stress strain curve from the tensile test.