A Strain-Compensated Arrhenius Model for TC18 Alloy Optimized via Genetic Algorithm and Response Surface Methodology

Abstract:

Accurate prediction of hot deformation behavior is critical for numerical simulation and intelligent manufacturing of titanium alloy forgings. The high-temperature deformation characteristics of TC18 titanium alloy were investigated using isothermal compression tests on a Gleeble-3800 thermal simulator over the temperature range of 720–840 °C and strain rate range of 0.001–1 s^–1. To precisely describe the flow behavior, three strain-compensated Arrhenius (SCA) constitutive models were developed and evaluated: a linear regression-fitted SCA (LR-SCA) model, a genetic algorithm-optimized SCA (GA-SCA) model, and a response surface-modified SCA (RS-SCA) model. The results demonstrate that genetic algorithm optimization effectively identifies global optimal parameters, while the response surface modification successfully accounts for the coupling effects of temperature and strain rate. Statistical evaluation reveals that the RS-SCA model achieves superior predictive capability, with a correlation coefficient (R) of 0.9980 and a mean absolute relative error of 2.16%, a significant improvement over the LR-SCA model's 7.01%. The reliability of the established RS-SCA model is further validated through its secondary development in DEFORM-3D, where finite-element simulations under stable processing conditions show excellent agreement with experimental load-displacement curves. This robust constitutive model provides a solid foundation for finite element simulation and process optimization of TC18 alloy components in thermomechanical processing applications.