Multiscale fatigue-prediction method to assess life of A356-T6 alloy wheel under biaxial loads

• Published in: Engineering failure analysis Vol. 130; p. 105752
• Main Authors: Duan, Yong-chuan, Zhang, Fang-fang, Yao, Dan, Hu, Jin-hua, Dong, Rui, Zhao, Xu, Guan, Ying-ping
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2021
• Subjects: Biaxial wheel fatigue test; Data-mapping algorithm; Mean stress; SDAS; Shrinkage cavity; Mean stress; Data-mapping algorithm; Shrinkage cavity; Biaxial wheel fatigue test; SDAS
• ISSN: 1350-6307; 1873-1961
• DOI: 10.1016/j.engfailanal.2021.105752

•An efficient tetrahedral mesh data-mapping algorithm is developed for defects and SDAS data.•A mesoscopic fatigue-strength prediction method considering the effects of the defects and SDAS is developed based on the meso-cell model and stress-gradient theory.•Combined with the BP neural-network model, Two S-N data prediction methods driven by the coupling model and data and by solely data are established.•According to the biaxial physical experiment, a biaxial virtual experiment is established for the wheel. The real stress of a car wheel can be reproduced by biaxial tests. However, in the cases of complicated loads, these tests are expensive and time-consuming. Additionally, shrinkage cavities and uneven microstructures can be introduced in A356 cast aluminum alloys, which has a certain influence on fatigue life. Therefore, a new simulation method for the biaxial wheel fatigue test, including the effects of the shrinkage cavity and secondary dendrite arm spacing (SDAS), is urgently needed to avoid the dispersion of the simulation results and reduce costs. In this paper, a new multiscale biaxial fatigue simulation method is proposed. An efficient tetrahedral mesh data-mapping algorithm is developed, in which the natural coordinates are introduced, and transfer of the SDAS and porosity between the cast wheel and finished wheel are realized. Based on the meso-cell model and stress-gradient theory, a mesoscopic fatigue-strength prediction method with defects and SDAS effects is developed. The two pieces style fatigue strength surface is determined. S-N data prediction methods driven by the coupling model and data and solely by the data are developed respectively. The generalization accuracy is within 4%, the structure of the pure data model is simple. The prediction accuracy is verified by performing a uniaxial tensile experiment. A wheel biaxial simulation model is established, and a continuous biaxial load is realized using the sequence amplitude curve family. Finally, a new multiscale biaxial fatigue simulation method with multiple load spectra is created using Fe-safe. The prediction result for the minimum life starting point (considering casting defects) is consistent with the test results, and the overall prediction results are significantly improved. The proposed method lays a solid foundation for optimization design and Big Data fatigue prediction of aluminum alloy wheels.