three point bending experimental data vs simulation

[Aminpd]

I am working on a simulation of three point bending using ABAQUS. I used Johnson-Cook hardening model. In the experiment i rectangular steel bar was bent and afterwards unbent and the Force-Displacement(F-D) data was extracted. There were no strain gauges on the specimens therefore the strains should be calculated analytically. The goal is to compare the experimental data with the simulation to modify the hardening model.
The flexural stress does not match the stress calculated by abaqus!
I am interested in the converting the Force-Displacement(F-D) curve into Stress-Strain(S-S) curves in Three-Point-Bending tests. My problem is that the Stress equations mentioned in Some papers (one is attached in this Post) do not work in the plastic regime but somehow the S-S curves were calculated and it looks like that they are simply are a scaled form of F-D curves. The other problem is to calculate the residual stress for the unbending part, which I could not find useful resources to deal with it.
I am looking forward to hearing from you.

Best regards

 

[FEA way]

In Abaqus you can create a stress-strain curve by taking some representative points, probing stress and strain at these points and then creating combined XY plot of these values.