Neural network-based ductile fracture model for 5182-O aluminum alloy considering electroplastic effect in electrically-assisted processing

Abstract

Complex components made of 5182-O aluminum alloy are usually formed at high temperatures due to their low ductility at room temperature, and the advanced current assisted processing can achieve energy-saving, high-efficiency, and green production compared with traditional hot forming. In this study, the coupled effects of electrical pulse, temperature, strain rate, and strain on flow behavior and ductile fracture are comprehensively investigated by experiments and constitutive modeling. In order to investigate the influence of electroplastic effects in different stress states, shear, hole and notched specimens are tested by isothermal tensile test and electrically-assisted isothermal tensile test in temperature ranges from 300 to 423 K and strain rates from 0.001 to 0.1/s. The evolution of the microstructure is characterized by electron backscatter diffraction to reveal the mechanism of the experimental phenomenon observed. An artificial neural network model is developed to characterize the dynamic hardening behavior and ductile fracture, and further embedded in ABAQUS/Explicit for numerical simulation. The results show that the experimental results are highly non-linear and coupled for the flow curves. At the same temperature, electric pulses have an independent effect on suppressing the Portevin–Le Chatelier effect and can reduce the deformation resistance, thereby forming products at low energy. The non-monotonicity of the fracture loading paths obtained from the hybrid experimental–numerical analysis is affected by strain rate hardening, thermal softening and electroplasticity. The neural network-based plasticity model is calibrated and validated to describe fracture initiation considering stress state, current density, temperature and strain rate.