Inversion Of Orthogonal Anisotropic Lamina Parameters Based On Lamb Wave

摘要

Using guided wave characteristics of dispersion and multi-mode, two inverse methods based on Genetic Algorithm (GA) and Deep Neural Network (DNN) is presented to determine the elastic constants and initial stresses of carbon fiber-reinforced composite (CFRC) lamina. The input of GA is the phase velocity of A0 mode at specific frequency in each direction, and the output is the elastic constants of CFRC lamina. The input of DNN is the phase velocity of A0 mode at several low frequencies and the output is the prestress value of CFRC lamina. This article will also discuss the differences between Evolutionary Computation and Artificial Neural Network in inversion of material parameters.