Constitutive models are considered a key prerequisite to investigate the thermal deformation behavior of materials. Accurate identification of their parameters is crucial for enhancing the predictive precision of the models. A novel optimization method for accurate inverse identification of parameters using the genetic algorithm (GA) in the Modified Zerilli-Armstrong (MZA) model was proposed in this work. Under conditions of temperatures ranging from 1223 K to 1473 K at intervals of 50 K and strain rates of 0.01, 0.1, 1, and 5 s−1, uniaxial isothermal hot compression tests were performed on 2Cr13 martensitic stainless steel (MSS) employing a Gleeble-1500D thermal simulation test machine. Based on the experimental data, the conventional linear regression method was utilized to solve the MZA model of 2Cr13 MSS. Initial values for the material constants I1, S1, C5, and C6 to be optimized were assigned drawing from the calculated results. The GA-based iteratively optimized MZA (G-ZA) model was established by minimizing the mean squared error as an objective function between the experimental and predicted flow stress values. Compared to the MZA model, a significant enhancement in predictive performance was achieved with the G-ZA model, while good generalization was also demonstrated. Both models were successfully integrated into the Forge® finite element analysis software through the secondary development of user subroutines. Numerical simulation results indicated that the G-ZA model demonstrated a better agreement with the experimental data in predicting the load-displacement curve. This validated that a more accurate constitutive model for the hot deformation of 2Cr13 MSS can be effectively constructed using the GA-based parameter inverse identification strategy.
Read full abstract