Model your specimen and apply the velocity calculated from Hopkinson bar calculation at both ends of the specimen (or find the relative velocity and apply at one end with other end fixed) to reproduce the stress-strain curve. Use an appropriate constitutive model for the material, which can capture strain rate dependence, as most of the material exhibit strain-rate dependence.
I collected numerical strains from the skins of the bars. After that, I calculated velocities using SHPB governing équations and then strain in the specimen. The problem is that the strain in the specimen is sensitive to the litle differences of the numerical velocities results, which affects the Stress-strain behavior law. Stresses were determined from interaction domains using CFN3.