Investigation on strain dependence of dynamic recrystallization behavior using an inverse analysis method

Since it is difficult to accurately describe the nucleation behavior through analytical calculation or experimental observation, an inverse analysis method is used to evaluate the reasonability of two different nucleation models, i.e., a conventional constant nucleation model where the nucleation rate is assumed to be independent of strain and a dislocation-related nucleation model where it is strain dependent. In order to accurately estimate the parameters in the above two models, a flow stress-based inverse analysis method is developed. This method is implemented by coupling a cellular automata (CA) model with an adaptive response surface method (ARSM). Then the reasonability of each nucleation model is evaluated by comparing the level of agreement between the simulations and the experiments. Taking a low carbon steel deformed at 1173 K and 0.1 s−1 as an example, it is shown that the flow stress-based inverse analysis method is able to estimate the nucleation parameters accurately and the introduction of the dislocation-related nucleation model makes the simulation even closer to the actual physical process.