| Summary: | Isothermal constant-strain-rate compression experiments of S280 ultrahigh-strength stainless steel were conducted at 800–1150 °C, 0.001–10 s<sup>−1</sup>, and 70% height reduction. The flow stress behaviors were analyzed based on the compression data. The strain compensation Arrhenius constitutive relation, multiple linear regression constitutive relation, and back-propagation (BP) neural network constitutive relation of this alloy were established for the first time. The S280 ultrahigh-strength stainless steel is characterized by a positive strain rate and negative temperature sensitivity. Its flow stress at high temperature (1000–1150 °C) and low temperature (800–950 °C) is generally at the steady state and the softening state, respectively. The three new flow stress constitutive relations all meet the requirements for engineering applications in terms of predictive precision. The BP neural network constitutive relation shows the highest predictive precision, with correlation coefficient R of 0.999 and average absolute relative error <i>AARE</i> of 1.04%. The strain compensation Arrhenius constitutive relation shows the lowest predictive precision, with R of 0.994 and <i>AARE</i> of 14.748%. The multiple linear regression constitutive relation shows the modest predictive precision, with R of 0.994 and <i>AARE</i> of 6.24%.
|
|---|
