欢迎光临《计算力学学报》官方网站！

Inversion of elastic modulus of stiffened double modulus rectangular thin plates based on PSO-BP model

DOI：10.7511/jslx20210913001

 作者 单位 E-mail 刘韡 西安建筑科技大学 理学院, 西安 710055 张宇鹏 西安建筑科技大学 理学院, 西安 710055 1633649163@qq.com 宋永超 西安建筑科技大学 理学院, 西安 710055

为了利用少量观测点高效反演材料的弹性模量,本文利用基于粒子群算法优化PSO(Particle Swarm optimization)的误差逆向传播网络BP(back propagation)建立反演模型,利用粒子群算法对误差逆向传播网络的参数进行寻优。以解析法求解双模量矩形薄板在不同弹性模量时的挠度作为训练样本,输入四个观测点的挠度值,利用PSO-BP模型对板的弹性模量进行反演。结果表明,PSO-BP模型可以建立挠度与弹性模量的联系,PSO算法可以提高BP模型的精度,加筋双模量矩形薄板的三个弹性模量的最大残差分别为39.052 kPa,73.513 kPa和64.207 kPa,最大相对误差分别为1.722%,3.681%和3.637%。本模型可为工程实践提供参考和指导。

In order to retrieve the Young’s moduli of materials efficiently by using a small number of observation points,the inversion model of error Back Propagation Network (BP) based on Particle Swarm Optimization (PSO) is established in this paper,and the parameters of error back propagation network are optimized by particle swarm optimization.The deflection of double modulus rectangular thin plate with different elastic moduli is solved by an analytical method as the training sample.The deflection values of four observation points are taken as the input,and the Young’s moduli of the plate are determined through inverse analysis from the PSO-BP model.The results show that the PSO-BP model can establish the relationship between deflection and elastic modulus,and PSO algorithm can improve the accuracy of the BP model.The maximum residuals of three values of the Young’s moduli of a reinforced double modulus rectangular thin plate are 39.052 kPa,73.513 kPa and 64.207 kPa respectively,and the maximum relative errors are 1.722%,3.681% and 3.637% respectively.The model can provide reference and guidance for engineering practice.