Experimental Prediction of the Fracture of 6XXX Aluminum Alloys

Abstract

The fracture behavior of Al-6DR1 sheets during the stamping process is of specific importance in the automotive industry. Efforts were made to reduce the costs associated with fracture prediction by using numerical simulations instead of experimental testing. The motivation for developing the fracture surface is to improve the prediction of fractures in simulation that then can be used to guide the stamping/forming tool design process. The theoretical framework employed in this thesis is based on two fracture models to predict the material behavior: The Modified Mohr-Coulomb (MMC) and the Hosford-Coulomb (HC). Furthermore, a hybrid and a direct calibration method are used to get the models parameters. The hybrid method is based on a numerical-experimental approach to get the variation of triaxiality and lode angle during deformation and it is also coupled with a damage accumulation rule. While the direct calibration method is based on a pure experimental approach, where the triaxiality on lode angle were assumed to be constant for the suggested experiments all the way to the fracture initiation stage. The specific tests used in the direct calibration are hemispherical punch stretching tests to induce equi- biaxial strain, pure shear tests, 3- point bend tests and Marciniak tests to induce plane strain, and hole-expansion tests to induce fracture under uniaxial tension strain. To capture the effects of stress triaxiality and lode angle experienced in the material fabrication process, a range of stress states including was needed including pure shear, uniaxial tension, plane strain tension and equi-biaxial tension. The generated fracture surface is then validated and incorporated in numerical models that simulate the deformation process and allow for prediction of critical locations part locations that are likely to fracture during forming. Such predictive capabilities are important in the tool design stage.