The phenomenological and physically based models,using the true stress–true strain curve data obtained under various hot working conditions of 850–1200°C and 0.001–10 s−1,were developed and improved for AerMet...The phenomenological and physically based models,using the true stress–true strain curve data obtained under various hot working conditions of 850–1200°C and 0.001–10 s−1,were developed and improved for AerMet 100 ultra-high strength steel.The predictability of the developed constitutive models was verified and compared.The determination coefficient and average absolute relative error were 0.9988 and 3.72%for the improved version of the modified Zerilli–Armstrong model,0.9985 and 3.96%for the improved version of the modified Johnson–Cook model,0.9947 and 4.59%for the strain-compensated Arrhenius-type model and 0.9913 and 5.43%for the improved Khan–Huang–Liang model,respectively.The results showed that the improved versions of the modified Zerilli–Armstrong model have the best predictability among the studied constitutive models.Comparing the predictability before and after the improvement,the average absolute relative error was increased by 65.14%for the modified Zerilli–Armstrong model and 58.45%for the modified Johnson–Cook model.This indicates that the phenomenological improvement of physically based constitutive models allows us to develop effectively constitutive equations with high prediction accuracy.展开更多
基金support from the Central Research Fund of Kim Chaek University of Technology(MIRAE 2023779).
文摘The phenomenological and physically based models,using the true stress–true strain curve data obtained under various hot working conditions of 850–1200°C and 0.001–10 s−1,were developed and improved for AerMet 100 ultra-high strength steel.The predictability of the developed constitutive models was verified and compared.The determination coefficient and average absolute relative error were 0.9988 and 3.72%for the improved version of the modified Zerilli–Armstrong model,0.9985 and 3.96%for the improved version of the modified Johnson–Cook model,0.9947 and 4.59%for the strain-compensated Arrhenius-type model and 0.9913 and 5.43%for the improved Khan–Huang–Liang model,respectively.The results showed that the improved versions of the modified Zerilli–Armstrong model have the best predictability among the studied constitutive models.Comparing the predictability before and after the improvement,the average absolute relative error was increased by 65.14%for the modified Zerilli–Armstrong model and 58.45%for the modified Johnson–Cook model.This indicates that the phenomenological improvement of physically based constitutive models allows us to develop effectively constitutive equations with high prediction accuracy.
