phase on the overall strength of material. A 2D RVE was generated on the basis of real micrographs. 3D cubeshaped
RVEs were defined and the phase fractions measured by the LOM were taken into account. A statistical
algorithm was implemented to describe the random distribution of phases within the RVE. The 3D RVE for a TRIP
microstructure consisting of 55% ferrite, 35% bainite, 5% austenite, and 5% martensite is illustrated in Fig. 2 (a).
Multiphase steels usually exhibit approximately isotropic mechanical properties. In this work, effective isotropic
formulations for the flow behaviors of each individual phase were assumed using the von Mises elastic plastic
material law. The constitutive models for the flow curves are the function of chemical composition of material, and
microstructural characteristics [8-10]. In general, the local carbon content is an important factor to incorporate.
During the annealing process, carbon partitioning takes place, which influences the volume fraction of different
phases and their mechanical properties. The true stress-strain curves used for the different individual phases in the
RVE simulations and their estimated percentage of carbon contents are depicted for steels DP600 and TRIP600 in
Fig. 2 (b). FE simulations of the 3D RVE applying uniform tensile boundary condition were carried out for the
investigated microstructures to verify the effective mechanical properties of each constituent phase. Averaged stressstrain
responses calculated by the RVE simulation were compared with the experimentally determined true stressstrain
curves obtained using both tensile and hydraulic bulge tests. The comparisons in Fig. 2 (b) show acceptable
agreements between calculations and experiments.