Unraveling thermoelastic performance of the multi-directional functionally graded composite structure by data-driven deep-learning approach

Abstract

Multi-directional functionally graded materials (MD-FGMs) are a brand-new family of sophisticated composites with continually changing structural and mechanical properties but no internal boundaries and interfacial stress concentrations. This study is the first to examine the effects of an instantaneous thermal shock on the thermo-elastodynamic response of a multi-directional functionally graded (MD-FG) doubly curved panel resting on an elastic substrate due to the remarkable properties of these advanced materials. The peripheral discontinuity at the boundaries between the loaded and unloaded surfaces as well as the shear coupling between the various Winkler-spring components are both taken into account in this proposed mode of the Kerr-type elastic substrate. Through the use of 3D-DQA, the system’s equations may be solved for both the simply supported and the completely clamped boundary conditions. For the first time, it is examined how several thermal shock models, such as cosinusoidal, sinusoidal, and Heaviside functions, impact how the system’s stress and deflection parameters fluctuate. The authors of this paper provide a unique strategy for enhancing the computing efficiency of the approach by constructing a deep-learning-based method on the data produced by 3D-DQT and Laplace transform at specified locations due to the dynamic nature of the issue. In order to decrease the mechanical excitation of the existing curved structure brought on by thermal shock, some advice for relevant industries is offered at the conclusion.