Theoretical and Applied Mechanics Letters 12 (2022) 100363.
doi: 10.1016/j.taml.2022.100363
Abstract:
Engineering structures often are subjected to a variety of load types.Some parts of them may experience a complex deformation history,for example,plastic deformation under cyclic loading.The materials of such structures often include nonlinearity and path (history) dependence.Appropriate nonlinear constitutive models considering loading path (history) can assist in structural design,help monitor structural health and assess the redundancy life of the engineering structures.Many constitutive models have been proposed to describe the behavior of elastoplastic materials[1-4]based on an explicit-functions-based paradigm with basic elements such as the plastic flow,loading and unloading judgment etc..However,the formulation of function-based constitutive model along with the traditional approach is a challenging and time-consuming job.It requires complex mathematical derivation and intuition into the plastic deformation mechanisms of the materials.In recent years,with the development of the data-driven approach and the machine learning by the deep neural networks[5,6],building the constitutive models through the combination of data science and solid mechanics attracted the research attention of the researchers around the world.
Zefeng Yu, Chenghang Han, Hang Yang, Yu Wang, Shan Tang and Xu Guo. Elastoplastic constitutive modeling under the complex loading driven by GRU and small-amount data[J]. Theoretical & Applied Mechanics Letters. doi: 10.1016/j.taml.2022.100363.