In order to develop steels that have high strength and ductility, it is crucial to predict and control the microstructure formation during the austenite-to-ferrite transformation. A multi-phase field (MPF) model enables us to simulate the complicated microstructure formation during the austenite-to-ferrite transformation based on the chemical free energy of the systems. However, for the accurate prediction of austenite-to-ferrite transformation behavior, unknown parameters and physical properties involved in the MPF model have to be identified from experimental data. Furthermore, uncertainty quantification of parameters included in the MPF model is essential to validate the MPF model. In this study, we develop a parameter estimation framework using the Markov chain Monte Carlo (MCMC) method to estimate unknown parameters and their uncertainties involved in the MPF model from experimental data, along with experimental error. The developed parameter estimation method is applied the MPF model and estimate the interface mobility included in the MPF model using the transformation ratio obtained from a formaster test. The results show that the MPF simulation using the estimated interface mobility can successfully reproduce the experimental data. Furthermore, the estimated interface mobility had high reliability, as confirmed by the uncertainty quantification of the estimated interface mobility.
Suzuki et al. (Wed,) studied this question.