Sensitivity Analysis of the Finite Difference 2-D Cellular Automata Model for Phase Transformation during Heating

Abstract

Sensitivity analysis of the Finite Difference Cellular Automata model for Dual Phase steel phase transformation during heating was performed in the present work. The main goal of the work was to determine the process parameters that are most important throughout transformation and should be particularly considered during the identification of model parameters: deformation, coefficients related with grains nucleation, activation energy, pre-exponential factor, and curvature parameters. The Morris OAT is a screening method capable of recognising the important factors of model and global sensitivity analysis was computed using this method. Different responses of the model outcomes were obtained by changing subsequent model input parameters. Results from Morris OAT design showed that deformation and activation energy have the most significant impact on the kinetics of phase transformation whereas average grain size strongly depends on all of the model parameters. Next, local sensitivity analysis was considered to check the behavior of each parameter locally. Finally both global and local sensitivities were compared and it was found that local sensitivity analysis in the case of such complex models can lead to inaccurate results.