Advanced virtual modelling aided stochastic nonlinear dynamic stability analysis of the GPLR-FGP plate in thermal environments

Abstract

This paper explores the crucial yet challenging task of stochastic nonlinear dynamic buckling investigations of the GPLR-FGP plate under biaxial impacts in thermal environments with multi-dimensional uncertainties for the first time. An advanced virtual modelling technique, namely the Extended Support Vector Regression (XSVR) with a fresh S-spline polynomial kernel, is newly developed to alternatively depict the underpinned and sophisticated relationship between the inevitable system uncertainties and concerned nonlinear buckling response in an efficient fashion. Adequate statistical information of stability performance is furnished for the safety assessment and reliability-based design of the GPLR-FGP composite. The superior competence, advisable exactness, and appealing efficiency of the proposed uncertainty quantification framework are vividly demonstrated through benchmark tests and comprehensive numerical experiments against established machine learning algorithms. Moreover, the user-friendly modularity feature, and unique capabilities of expeditious information update and responsiveness highlight the versatile applicability of the proposed technique in realistic structural implementation under swiftly changing environments.