Buckling response of CNT based hybrid FG plates using finite element method and machine learning method

Ravi Kumar,Ajay Kumar,Divesh Ranjan Kumar
DOI: https://doi.org/10.1016/j.compstruct.2023.117204
IF: 6.3
2023-06-03
Composite Structures
Abstract:In this study, a C 0 finite element model (FEM) based on modified third-order shear deformation (MTSDT) theory in conjunction with a deep neural network (DNN), extreme gradient boosting (XGBoost), and random forest (RF) models are implemented to predict the critical buckling load of a carbon nanotube (CNT)-reinforced hybrid functionally graded (FG) plate structure. Cubic and quadratic variations of the thickness terms are adopted in the longitudinal and transverse displacement fields, respectively. A zero transverse shear condition is assumed for the outermost surfaces of the plate. A C 0 FEM is developed and coded in a MATLAB environment. The MTSDT-based FEM is validated with previous literature and found suitable for finding new results. The DNN model is coded in Python and XGBoost and the RF models are coded in RStudio. Using the random value of the elastic properties of the material as an input parameter and the buckling strength as an output parameter, a regression problem is expressed using the given machine learning method to predict the buckling response of the CNT-reinforced hybrid FG plates.
mechanics,materials science, composites
What problem does this paper attempt to address?