Identification of Vivo Material Parameters of Arterial Wall Based on Improved Niching Genetic Algorithm and Neural Networks

Abstract

Cardiovascular diseases are seriously threatening human health and the incidence rate is high. Many scholars are devoted to studying arterial mechanical properties and material parameters. In this study, the bovine artery was selected as the experimental object and the uniaxial tensile test was carried out by cutting the specimens along its axial, circumferential and [Formula: see text] directions. The finite element software ABAQUS and hyperelastic Holzapfel Gasser Ogden (HGO) constitutive model were used to simulate the experimental process. Niche technology is introduced on the basis of genetic algorithm, and the program of Improved Niche Genetic Algorithm for material parameter identification is compiled based on Python language. In addition, BP Neural Network was constructed based on Tensorflow mathematical system. The material parameters of the constitutive model of bovine artery in different directions were identified by finite element method and experimental data. The results show that Improved Niche Genetic Algorithm and Neural Network, respectively, combined with finite element are both effective and accurate methods for predicting the parameters of arterial vascular hyperelastic materials, which can provide reference and help for the study of arterial vascular mechanical properties.