Numerical solution hybridized by machine-leaning-based algorithm to provide an efficient method for analyzing thermomechanical shock behavior of the circumferentially-graded graphene-plates reinforced composite sandwich panel

Abstract

The present research bridges the gap between a numerical solution with efficient machine-learning-based algorithms to provide a strong platform for predicting thermomechanical shock behavior of the sandwich cylindrical panels with polymer-made core and circumferentially-graded graphene-plates reinforced (CG-GPLR) nanocomposite face-sheets. The comprehensive declaration of the elasticity theory is considered to establish the governing equations of the system in three orthogonal directions. The well-known Lord-Shulman heat transfer theorem is implemented to present the time-variant nature of the system’s heat. The differential quadrature method (DQM) is used as a numerical approach to solve the spatial dependency of the governing differential equations. While the Laplace transform solves the time dependency of the governing differential equations. In order to decline the computational cost of predicting the time-variant analysis of the thermomechanical shock response of the system, an efficient machine-learning-based algorithm is employed in a way that it only requires the thermomechanical shock information of the selective set of points known as the training set. After completion of the training procedure, this machine-learning predictor can easily predict the thermoelastic response of all points included either as the train set or a new point. Validation of the applied solution is performed through a comparative process between the present results with those determined in the published literature. The results of this study are in excellent compatibility with those obtained in a solo numerical solution.