【作者】 李凌均； 段晨东；

【Author】 Li Lingjun School of Mechanical Engineering,Zhengzhou University Zhengzhou 450001 Duan Chendong School of Electronic & Control Engineering,Chang’an University,Xi’an 710061

【机构】 郑州大学机械工程学院； 长安大学电子与控制工程学院；

【摘要】 针对结构损伤识别中缺少实际损伤样本的问题,提出基于小波包特征提取的支持向量机结构损伤诊断方法。该方法将结构振动信号小波包分解后的频带能量,经过多传感器数据融合后作为特征向量,输入到多分类的支持向量机中,实现了结构多损伤的识别和定位。应用该方法对IASC-ASCE模型进行了分析,试验结果表明,小波包分解频带能量能够较好地反映结构的损伤特征;多传感器数据融合能够使不同传感器的信息相互补充,减小了损伤检测信息的不确定性,提高了损伤诊断准确率。更多还原

【Abstract】 Due to the lack of actual damage samples in the structure damage diagnosis,a new method of support vector machines(SVMs)based on feature extraction with wavelet packet decomposition is proposed.According to the method,the energy sequences of different frequency bands decomposed by wavelet packet transformation are investigated.Then these energy sequences are fused as extracted feature vectors,which are input to a multi-classified support vector machines to implement multi-damage recognition and damage localization.Analyzed with the IASC-ASCE model,experiment results proved that wavelet packet energy sequence could reflect the damage condition perfectly.On the other hand,data fusion enriched diagnosis information and reduced the uncertainty of damage detection information,so the diagnosis accuracy was improved.更多还原