A theoretical approach for thermomechanical shock analysis of doubly curved panel with respect to geometrical and physical parameters using machine-leaning-based algorithm

In this work, for the first time, bending information of the doubly curved panel under thermo-mechanical shock loading (MSL) with respect to geometrical and physical parameters via both numerically, and machine-leaning-based algorithms is presented. The graphene nanoplatelets composites are used to reinforce the current composite curved panel along with latitudinal direction. It should be note that the current curved panel is reinforced in latitudinal direction with GPLs due to thermo-MSL in the latitudinal direction. The material properties of the current doubly curved structure are obtained with the aid of role of mixture and Halpin-Tsai micromechanics model. As well as this, GPL distribution's thermal conductivity for each pattern of GPL is considered for high accuracy of the current work's model. Three-dimension (3D) elasticity theory, linear thermo-elasticity principle, constitutive heat conduction equation, and relaxation time equation are used to present the mathematical modeling of the current work. After obtaining the governing equations, the equations are solved via numerical solution using Chebyshev-Gauss-Lobatto function, and spatial and Laplace latitudinal methods. Deep neural networks (DNNs) as a machine learning method are used to train and test the obtained results from mathematical modeling. After predicting the results, the results are compared with the outcomes of mathematical modeling and show that with the low computational cost of DNNs algorithm, the results can be predicted instead of other high computational cost methods. Finally, some suggestions for improving the thermo-mechanical behavior of the current doubly curved panel under external shock loading.