Improved constitutive model and processing map of Al-Cu-Mn-Mg alloy based on backpropagation artificial neural network

Hot compression tests were performed on Al-Cu-Mn-Mg alloy at various temperatures (300, 350, 400, and 450 degrees C) and strain rates (0.01, 0.1, 1, and 10 s- 1). To predict the flow stress of the alloy, an Arrhenius-type (AT) model and an improved model were developed. The improved model, which incorporates a compensation factor and a backpropagation artificial neural network, demonstrated superior predictive accuracy compared to the AT model. It achieved a higher correlation coefficient of 0.9999 and a lower average absolute relative error of 0.580 %. Based on data from the improved model, an activation energy-processing (AEP) map was constructed. The optimal hot working parameters for Al-Cu-Mn-Mg alloy were determined as 400-450 degrees C and 0.01-0.1 s- 1 using the AEP map. Additionally, continuous dynamic recrystallization (CDRX) and discontinuous dynamic recrystallization (DDRX) were observed in Al-Cu-Mn-Mg alloy. CDRX grains form through sub-grain rotation, while DDRX grains form via grain boundary bulging.