2D-GDQM and adaptively tuned deep neural network for frequency analysis of the sandwich disk with honeycomb resting on elastic foundation

Abstract

The 2D-GDQM (Two-Dimensional Generalized Differential Quadrature Method) and adaptively tuned deep neural network are two computational methods that have been proposed for the frequency analysis of sandwich disks with honeycomb cores. Sandwich disks are commonly used in aerospace and automotive industries due to their lightweight and high strength properties. However, accurately predicting their natural frequencies is crucial for ensuring their structural integrity. The 2D-GDQM is a numerical method that discretizes the equations of motion of the sandwich disk using a grid-based approach. The method has been used to obtain accurate and efficient frequency solutions for various types of sandwich structures. In this study, the 2D-GDQM was applied to analyze the frequency response of the sandwich disk with honeycomb core with first and higher-order deformation theories to model displacement field. Additionally, an adaptively tuned deep neural network was used to predict the natural frequencies of the sandwich disk. This method involves training a deep neural network with a dataset of frequency solutions obtained from the 2D-GDQM simulations. The neural network is then optimized to provide accurate predictions for new cases. The results of this study showed that both the 2D-GDQM and the adaptively tuned deep neural network can provide accurate predictions of the natural frequencies of the sandwich disk with honeycomb core. The 2D-GDQM was found to be more computationally efficient, while the neural network approach can be more flexible and adaptable to new geometries or material properties.