Computer Modeling For Frequency Performance Of Viscoelastic Magneto-Electro-Elastic Annular Micro/Nanosystem Via Adaptive Tuned Deep Learning Neural Network Optimization

The presented paper is the first attempt to apply deep learning for predicting the frequency characteristics of a magneto-electro-elastic (MEE) annular nano/microdisk (MEEAD). The optimum amount of the factors participating in the mechanism of the fully connected neural network are achieved through the optimizer based on the momentum. The positive side of the mentioned approach employed in this investigation would be due to its high accuracy along with lower epochs required for training the multi-layered network. This scrutinization would be semi-computational research that estimates the vibrational behavior of a MEEAD employing a non-classical continuum model known as the modified couple stress (MCS) model. First order shear deformation theory (FSDT) and shell model would be provided for presenting their displacement fields. Then, Kelvin-Voight theory has been applied to model the viscoelastic foundation. Its non-classical governing equations, as well as related boundary conditions (BCs) of small-scaled MEEAD, would be achieved by considering the symmetric spinning gradient along with higher-order stress tensors for the strain energy. The provided non-classical theory would be able to capture the small scale in the MEEAD employing just one length scale of a material factor, then, the mathematical modeling of MEEAD according to the classical theory would be able to be recovered from the provided model by eliminating the material length scale factor. Ultimately, the non-classical governing equations would be solved by applying the generalized differential quadrature (GDQ) approach for multifarious BCs. Moreover, parametric research has been conducted to analyze the influences of the viscoelastic foundation, length scale factor, geometry of MEE, radial and circumferential mode number, radius ratio, and BCs on the frequency behavior of the MEEAD by applying MCST. The outcomes reveal that there would be a crucial radius ratio in which the relationship between these elements and crucial inserted voltage alters from direct to indirect relation.