Prediction of Natural Frequency of Composite Plates with Non-canonical Shape Using Convolutional Neural Networks

This work is focused on the application of convolutional neural networks (CNN) to predicting the natural frequency of composite plates with non-canonical form of plan under clamped boundary condition. For training and testing of the CNN model, a number of numerical experiments have been carried out using R-functions (RFM) and Ritz methods by varying the geometry of a multi-layered plate of symmetric layout. The plate is modeled using the refined theory of the first order, taking shear deformations in account. The input features to the CNN are chosen to represent the geometrical form of the plate. The results predicted by CNN are in a good agreement with the RFM modeling results. The convergence study of the CNN architecture is also provided.