An automotive steel wheel digital twin for failure identification under accelerated fatigue tests

Abstract

In this paper, a digital twin to predict failure in automotive steel wheels during experimental fatigue tests is proposed. The methodology involves a wheel finite element model that incorporates assembling and mounting processes, and CDTire/3D simulation package, employed to calculate tyre to rim generalised internal forces in several loading conditions. The two models are decoupled under small deformation assumption, which allows to reduce computational effort for batch simulations. The Digital Twin (DT) is designed to meet the standard requirements and the working principles of test benches for dynamic cornering, radial, and biaxial fatigue tests. The overall DT architecture is described with emphasis on different test bench logic and the introduction of simplifications to efficiently estimate failure under mean stress variation. Then, a simplified formulation of the critical plane McDiarmid multi-axial failure criterion is presented to face the non-proportionality of stress path under loading conditions. Finally, the methodology is validated against an industrial case study: the predicted failures are compared to experimental test result database for wheel topologies, material properties, test benches and loading conditions.