Invited Speaker

Abstract: Data-driven methods based on machine learning (ML) are increasingly being used for constitutive modeling of advanced aerospace materials, including metallic alloys and fiber-reinforced composites. However, their reliance on extensive datasets has hindered further development. To address this limitation, we propose an innovative mechanics-informed ML approach to predict the elastoplastic behavior and anisotropic features using small training datasets. For elastoplastic metallic materials, a novel strain reconfiguration strategy is proposed to improve the learning capability and predictability of the data-driven model, along with a two-step training method. A compatible numerical implementation algorithm is developed to incorporate the data-driven approach into a finite element calculation. This method is applied to learn and predict the mechanical response of Ti-6Al-4V titanium alloy under multiple loading conditions, including tension, compression, shear and impact loads. For orthotropic composites, we establish a decomposition and equivalence method for stress-strain tensors. Two independent artificial neural networks are employed to capture the deviatoric behavior and the volumetric-fiber coupling behavior, respectively. Additionally, an incremental algorithm is introduced to map the one-dimensional scalar stress back to the three-dimensional stress tensor. The proposed model is validated using a dataset generated by direct numerical simulation of a representative volume element (RVE) of composites, and further applied to the simulation of textile composites. The consistency between the constitutive model and the data highlights the advantages of the proposed approaches: integrating mechanics with ML significantly enhances predictive accuracy, even with limited data.

Biography: Dr. Zhang is Professor and Ph.D. supervisor for the School of Civil Aviation, Northwestern Polytechnical University. His research direction lies in the fields of multi-scale mechanics of composite materials, impact dynamics, and strength of aerospace engines structures. Dr. Zhang has awarded more than 20 scientific research projects (5 from NSFC). He is recipient of the National High-level Talent Youth Program, Shaanxi Youth Scientist Award and National Outstanding Young Scholar in Explosive Mechanics et al. Dr. Zhang has published more than 150 journal papers, with Google Citations over 5400. He serves as editorial board member for several scientific journals, e.g. Compos Struct, Acta Mechan Sin, Chin J Aeronaut, J Aero Eng et al., and active members of the Chinese Society for Composite Materials and The Chinese Society of Theoretical and Applied Mechanics.