Advanced FE model validation of cold-forming process using DIC: Air bending of high strength steel

Abstract

Recent advances in mechanical and civil engineering are noticed in many innovative designs that frequently employ cold-formed High Strength Steels (HSS). Typical mobile cranes benefit from the advanced properties of these steel grades in a bent configuration. Here, the majority of load-carrying members are produced through cold-bending and subsequent welding procedures. These cold-foring processes induce residual stresses and strains that must be considered when assessing the structural integrity and service life of bent sections in an assembly. Finite Element Analysis (FEA) offers a unique solution here to reproduce the bending process accurately. However, this analysis must be verified using representative validation methods. If these methods remain scarce, basic or incomplete, the credibility of a sensitive FE model may be compromised. In the present paper, a series of model validations are proposed that rely on the global and local response of the material during or after bending. A benchmark specimen and an air bending set-up have been developed from a numerical design of concepts and fine-tuning of tool dimensions, ensuring the appropriate bending conditions. Local validation is pursued using stereo Digital Image Correlation (DIC) to capture the strain fields, generated during plastic bending of a 12 mm thick S690QL plate. The crux of the problem is twofold: firstly, strain calculation methods used in DIC and FEA are fundamentally different, hampering a correct and honest comparison. Secondly, consistent point-to-point comparisons of experimentally acquired (DIC) and numerically computed (FEA) strains are more susceptible to uncertainties related to processing settings and differences in coordinate frame. Moreover, the main advantage of the introduced ground truth validation is the ability to level the FEA data through identical filters as the DIC experiment. Unlike a direct comparison, this levelling approach auto-adopts an unconditionally equal strain calculation, based on nodal displacement fields, independently of a local (shell) or global (solid) element formulation. This paper aims at clarifying the need for this ground thruth validation in pursuance of higher fidelity FE-models for metal forming simulations.