Hello, thank you very much for reading my question. Please, I need some help!

I am simulating a tensile test using the Standart Tensile Specimen (ASTM E399-17, figure 1) to obtain a prediction of the material anisotropic behavior.

Using Abaqus explicit, Hill48 anisotropic yield criterion, Johnson-Cook failure criterion (figure 3), and Elastic Modulus as input (119332 MPa) , I took the value of the elastic modulus from the experimental curve.

In figure 2 are the values of the hill48 yield stress ratios necessary for the simulation in Abaqus, the user also have to introduce the real stress x plastic strain experimental data (figure 4).

The problem is the resulting stress x strain curve, I don't know why the elastic modulus in the resulting curve is slower than the experimental one that I introduced as an input (experimental and simulation curves in figure 5). Does someone knows why?