基于有限元与改进SVM的飞行器结构无损检测模型设计

Design of aircraft structure nondestructive testing model based on finite element and improved SVM

针对传统飞行器结构无损检测中存在的准确度低且易造成二次破坏等问题,以有限元仿真为数据基础,提出一种基于改进支持向量机的飞行器结构无损检测模型。该模型使用主元分析法对数据主特征进行分析,解决了有限元仿真数据维度高的问题;利用二叉树的思想改进了传统支持向量机,使其具备多特征分类能力,并对多数据特征加以分类,提高了模型的收敛准确度;还通过引入粒子群算法优化多分类向量机的惩罚因子及核函数参数。实验测试结果表明,所提模型可实现分类器参数的性能优化,平均分类准确率较对比算法提升了约1.4%。

In allusion to the problems of low accuracy and easy to cause secondary damage in the traditional nondestructive testing of aircraft structure,a aircraft structure nondestructive testing model based on improved support vector machine(SVM)is proposed based on finite element simulation data.In the model,the principal component analysis(PCA)is used to analyze the main characteristics of the data,so as to solve the high dimension of finite element simulation data.The traditional SVM is improved by means of the idea of binary tree to make it have the ability of multi-feature classification and classification of multi-data features,and the convergence accuracy of the model is improved.The particle swarm optimization(PSO)is also introduced to optimize the penalty factor and kernel function parameters of the multi-classification vector machine.The experimental testing results show that the proposed model can realize the performance optimization of classifier parameters,with an average classification accuracy improvement of about 1.4%compared with the comparative algorithm.