摘要
基于局部均值分解的自适应时频分析特性和样本熵的非线性量化能力,结合径向基(RBF)函数神经网络,提出一种基于LMD样本熵和径向基函数神经网络的结构损伤识别方法。首先,应用局部均值分解方法将结构振动原始信号自适应分解为若干乘积函数分量(PF分量);然后提取前3个PF分量的样本熵,实现对PF分量的特征量化;最后将分量的样本熵作为损伤特征向量,利用径向基神经网络对高速列车比例车体下底板进行识别。实验结果证明,采用该方法识别结构损伤时,结构损伤位置和损伤程度的识别精度分别为96.97%和96.25%,证明了此方法在结构损伤诊断方面的有效性和准确性。
Adaptive time frequency analysis based on local mean decomposition and nonlinear quantization ability of sample entropy,combined with radial basis function（ RBF） neural network. A method of structural damage identification based on local mean decomposition（ LMD） sample entropy and radial basis function neural network is proposed. Firstly,the original signal is decomposed into a number of product function components（ PF component） by LMD to the original signal of structure vibration.Then extract the sample entropy of the first 3 PF components to realize the feature quantization of the PF component. Finally,the sample entropy of the component is used as the damage characteristic vector. The radial basis function neural network is used to identify the bottom plate of scaled carbody for high-speed train. The experimental results show that while this method is used to identify structural damage,the damage identification errors of location and degree are 96. 97% and 96. 25% respectively. The validity and accuracy of this method in structural damage diagnosis are proved.
作者
王名月
缪炳荣
李旭娟
杨忠坤
WANG MingYue;MIAO BingRong;LI XuJuan;YANG ZhongKun(Traction Power State Key Laboratory, Southwest Jiaotong University, Chengdu 610031, China)
出处
《机械强度》
CAS
CSCD
北大核心
2018年第3期522-527,共6页
Journal of Mechanical Strength
基金
国家自然科学基金项目(51375405)
牵引动力国家重点实验室自主项目(2016TPL_T10)资助~~
关键词
损伤识别
局部均值分解
样本熵
径向基函数神经网络
Damage identification
Local mean decomposition
Sample entropy
Radial basis function neural network
