Safety assessment for functionally graded structures with material nonlinearity

Yuan Feng,Di Wu,Lei Liu,Wei Gao,Francis Tin-Loi

DOI: https://doi.org/10.1016/j.strusafe.2020.101974

IF: 5.712

2020-09-01

Structural Safety

Abstract:<p>A machine learning aided reliability assessment framework is presented for functionally graded material (FGM) structures under plane strain/stress conditions with the consideration of elastoplasticity. The material nonlinearity of the FGM is modelled through the implementation of the Tamura-Tomota-Ozawa (TTO) model. For safety evaluation of FGM structures, the volume fraction of FGM has been modelled through spatially dependent uncertainty as random field for the concerned composite. In order to solve the complex stochastic elastoplastic problem, a further developed machine learning aided technique called the extended support vector regression (X-SVR) with a generalized Dirichlet feature mapping function has been introduced and then, the corresponding probabilistic features, including the statistical moments, probability density functions (PDFs), and cumulative distribution functions (CDFs), of the concerned structural responses can be effectively established for assessing the reliability of FGM structures. Moreover, the proposed approach is competent to deliver critical information regarding the uncertain system inputs which can be beneficial for subsequent safety assessment and structural designs for the FGM. Two test functions and two numerical examples have been adopted to visualise the accuracy, stability and capability of the proposed safety assessment framework for FGM structures.</p>

engineering, civil