As the demand for high-performance lightweight materials in electric vehicle manufacturing increased, AA3003 alloy, due to its excellent hot deformation properties, became a key research subject for improving product quality and optimizing process performance, but few current material models can be used to simulate the synergistic evolution of deformation and performance. The hot deformation behavior of a homogenized annealed extruded AA3003 alloy was investigated through isothermal hot compression tests conducted at temperatures between 250 and 550 ℃ and at strain rates ranging from 1 to 10 s⁻1. The experimental results revealed that the true stress-true strain behavior of the AA3003 alloy exhibited the characteristics of dynamic recovery (DRV) and dynamic recrystallization (DRX). Subsequently, three constitutive models with different mechanisms were established: a strain-compensated Arrhenius model for deformation simulation, an internal state variable (ISV) model considering multiple internal variables for performance prediction, and a particle swarm optimization-artificial neural network (PSO-ANN) model for rapid design. The findings indicated that the PSO-ANN model exhibited superior prediction accuracy and more robust generalization ability in comparison to the modified Arrhenius and ISV models, with evaluation parameters <InlineEquation ID="IEq1"> <EquationSource Format="TEX">$R$</EquationSource> <EquationSource Format="MATHML"><math> <mi>R</mi> </math></EquationSource> </InlineEquation>,<InlineEquation ID="IEq2"> <EquationSource Format="TEX">$AARE$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="italic">AARE</mi> </mrow> </math></EquationSource> </InlineEquation>, and <InlineEquation ID="IEq3"> <EquationSource Format="TEX">$RMSE$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="italic">RMSE</mi> </mrow> </math></EquationSource> </InlineEquation> being 0.9998, 1.5%, and 1.48 MPa. A three-dimensional hot processing map was developed based on the PSO-ANN model, and the microstructural analysis revealed the presence of large undeformed grains and fine DRX grains in regions with low power dissipation coefficients. The optimal processing window for the deformation process was identified through the application of the developed 3D hot processing map (350–450 ℃, 1–10 s⁻1). The ISV model was integrated into the ABAQUS software for the simulation of material deformation and the associated properties, and the resulting predictions of microstructure and performance were validated. Graphic Abstract

Three Constitutive Equations and a Three-Dimensional Hot Processing Map of AA3003 Alloy at Elevated Temperatures for Process Design and Performance Prediction

期刊论文
发表日期：

2025年5月20日

查看全文

Xiaohui Fang,
Fuguo Li,
Miaomiao Wan,
Qian Zhao,
Xinru Jia

摘要

As the demand for high-performance lightweight materials in electric vehicle manufacturing increased, AA3003 alloy, due to its excellent hot deformation properties, became a key research subject for improving product quality and optimizing process performance, but few current material models can be used to simulate the synergistic evolution of deformation and performance. The hot deformation behavior of a homogenized annealed extruded AA3003 alloy was investigated through isothermal hot compression tests conducted at temperatures between 250 and 550 ℃ and at strain rates ranging from 1 to 10 s⁻¹. The experimental results revealed that the true stress-true strain behavior of the AA3003 alloy exhibited the characteristics of dynamic recovery (DRV) and dynamic recrystallization (DRX). Subsequently, three constitutive models with different mechanisms were established: a strain-compensated Arrhenius model for deformation simulation, an internal state variable (ISV) model considering multiple internal variables for performance prediction, and a particle swarm optimization-artificial neural network (PSO-ANN) model for rapid design. The findings indicated that the PSO-ANN model exhibited superior prediction accuracy and more robust generalization ability in comparison to the modified Arrhenius and ISV models, with evaluation parameters $R$ $R$ , $AARE$ $AARE$ , and $RMSE$ $RMSE$ being 0.9998, 1.5%, and 1.48 MPa. A three-dimensional hot processing map was developed based on the PSO-ANN model, and the microstructural analysis revealed the presence of large undeformed grains and fine DRX grains in regions with low power dissipation coefficients. The optimal processing window for the deformation process was identified through the application of the developed 3D hot processing map (350–450 ℃, 1–10 s⁻¹). The ISV model was integrated into the ABAQUS software for the simulation of material deformation and the associated properties, and the resulting predictions of microstructure and performance were validated.

Graphic Abstract