基于SOM神经网络的三维编织复合材料制件损伤诊断模型研究

Research on Damage Diagnosis Model of 3D Braided Composite Parts Based on SOM Neural Network

三维编织复合材料制件主要应用于承载器件,在实际应用过程中经常会受到高温、高压、高空辐射等环境因素的影响,容易出现制件疲劳甚至损伤情况.笔者建立了基于自组织特征映射神经网络(SOM)的损伤诊断模型,实现了三维编织复合材料制件的损伤监测和数据分析处理;设计了基于MATLAB的仿真实验平台,实现了制件损伤监测与分析的可视化.实验获得的损伤监测数据与制件实际测试的数据相同,证明了建立的基于SOM的损伤诊断模型在三维编织复合材料损伤监测和数据分析处理方面的可行性,对提高我国复合材料制件的健康监测应用水平具有应用价值.

The 3D braided composite parts are mainly used in load-bearing devices.In the practical application process,they are often affected by environmental factors such as high temperature,high pressure,high altitude radiation and so on,and are prone to fatigue and even damage.A damage diagnosis model based on self-organizing feature map neural network(SOM)is proposed to realize the damage monitoring and data analysis of 3D braided composite parts.A simulation experiment platform based on MATLAB is designed to realize the visualization of damage monitoring and analysis.The damage monitoring data obtained by the experiment is the same with the actual test data of the parts,which proves the feasibility of the proposed damage diagnosis model based on SOM in the three-dimensional braided composite damage monitoring as well as data analysis and processing.The feasibility of damage monitoring as well as data analysis and processing has important application value for improving the application level of health monitoring of composite parts in China.