Fracture prediction of 5052 aluminum alloy sheet forming using machine learning-calibrated DF-2013 uncoupled ductile fracture criterion

Ductile fracture is the primary failure mode in sheet metal stamping, directly affecting material forming limits and product quality. Accurate prediction of ductile fracture is critical for optimizing stamping process parameters, avoiding crack defects and cutting production costs. It also provides reliable theoretical support for the application of lightweight materials like aluminum alloys in complex forming processes. In this study, a machine learning-calibrated DF-2013 uncoupled ductile fracture model was applied to simulate the stamping process numerically. We conducted cup-forming tests and numerical simulations on 5052 aluminum alloy to develop a method for accurately extracting the equivalent fracture strain point. For 5052 aluminum alloy, the equivalent fracture strain should be extracted at the point where load-bearing capacity drops rapidly. At room temperature and a stamping speed of 5 mm/min, the experimental IE value of 5052 aluminum alloy is 7.887 (accompanied by arc-shaped cracks), while the simulated IE value is 8.101—yielding a prediction error of only 2.71%—and the simulated crack morphology is also arc-shaped. This result further verifies that the DF-2013 uncoupled ductile fracture criterion, calibrated via machine learning, is reliable for predicting ductile fracture in sheet metal forming. This calibrated criterion can thus be further used to predict fracture in the stamping of automotive heat shields. Through systematic analysis of stress-strain distribution contours and thickness variation characteristics during stamping, the fracture positions of heat shields were accurately predicted, providing theoretical support for effectively preventing potential stamping defects. Results indicate that stresses, strains, and thickness reduction rates are generally lower on smooth plate surfaces. In contrast, higher stresses, strains, and thickness reduction rates are observed around the protruding edges of the thermal shield and surrounding components. The primary fracture locations in automotive heat shields are concentrated along the protruding edges and around the periphery of the components.