Evaluation of constitutive equations for modeling and characterization of microstructure during hot deformation of sintered Al–Zn–Mg alloy

Abstract

Commercial simulation software like QFORM and DEFORM do not possess adequate material data for alloys produced through powder metallurgy. Therefore, in the present study, hot compression tests of Al-5.6%Zn–2%Mg aluminum alloy were conducted at temperature ranges (300 0C-500 °C) and strain rates (0.1 s−1-0.0001 s−1). The impact of compression temperature and strain rate on the flow curve behaviour and microstructure evolution-associated mechanisms were investigated through experimentally acquired flow curves and EBSD analysis. Four constitutive models were constructed; namely the Arrhenius-type, modified Johnson Cook (MJC), modified Zerilli-Armstrong (MZA), and an artificial neural network (ANN). The results demonstrated that among the models considered, the ANN and Arrhenius-type models exhibited the lowest AARE values of 0.486 % and 3.36 %, respectively. Conversely, the MZA and MJC models displayed higher AARE values of 8.84 % and 3.93 %, respectively. Notably, the Arrhenius-type model emerged as the most suitable prediction model due to its capability to handle nonlinear relationships between factors. However, in scenarios where material properties are unknown or experimental data is limited, the MJC model can serve as a simpler alternative. The MZA model was deemed unsuitable for accurately estimating flow stress in hot compression. Remarkably, the best-trained ANN model exhibited the highest predictive performance with an AARE of 0.486 % and an R-value of 0.99. This study offers fundamental insights to improve the accuracy of simulating hot compression procedures. EBSD analysis demonstrated that higher upsetting temperatures and lower strain rates promoted recrystallization, resulting in a more uniform microstructure.