A Numerical Evaluation of Forming Failure of an Aluminum Sheet due to Splitting Damage in Hole Expansion Process

Abstract

The reduction of car body weight is the principal issue of car manufacturers for reducing fuel consumption. Aluminum alloys are attractive materials for the automotive industry because they have low density and adequate strength, but they may exhibit crack formation during manufacturing processes. Generally, crack formations emerge because of tool geometry and material anisotropy. Accordingly, determination of the forming limits of aluminum alloys is essential. The hole expansion test (HET) is a significant formability process used in the automotive industry because it gives information about the stretch-flangeability limits of the material. Edge splitting (edge fracture) is a failure type seen in HET, and it limits the stretch-flangeability of the material. Therefore, the prediction of edge splitting is an essential issue for engineers in the automotive industry. In this work, HET of AA6016-O aluminum alloy was simulated with the finite element (FE) method to assess the influence of yield functions on failure prediction in HET. To this end, Hill48, Yld91, and a homogeneous fourth-order polynomial type yield criteria (HomPol4) were selected to identify the anisotropic behavior of the sheet. Analyses were carried out in Marc commercial FE software, and the Hypela2 user subroutine was incorporated into FE code. Thickness distributions in the rolling direction (RD), diagonal direction, and transverse direction (TD) of the part and around the hole edge were also predicted, and it was observed that only the HomPol4 criterion predicted excessive thinning at two locations near the RD and TD, separately. On the other hand, Yld91 and Hill48 predicted lower strain levels when compared with HomPol4. Finally, plastic work distribution around the hole edge was considered, and the predictions were compared with the experimental damaged sample. This comparison showed that HomPol4 predicted a robust plastic work localization in RD, which is consistent with the damaged sample.