基于神经网络与特征融合的损伤诊断方法  

Diagnosis method of structure damage using neural network and feature fusion

作　　者：刘义艳[1] 段晨东[1] 巨永锋[1] 赵学风[2] 韩旻[1] 

机构地区：[1]长安大学电子与控制工程学院,陕西西安710064 [2]西安交通大学电气工程学院,陕西西安710049

出　　处：《长安大学学报（自然科学版）》2008年第6期106-110,共5页Journal of Chang’an University(Natural Science Edition)

基　　金：陕西省自然科学基金项目(2005E205)

摘　　要：为了能准确地诊断复杂结构损伤是否产生以及产生的位置和程度,提出了一种小波包分解、多传感器特征融合和神经网络模式分类相结合的结构损伤诊断方法。首先,用正交小波包对多个传感器采集的振动信号进行小波包分解,并计算每个频带上的相对能量;然后把这些传感器信号的小波包相对能量融合,构成神经网络分类器的输入特征向量,从而实现损伤的诊断和评价。研究结果表明:正交小波包分解的频带能量分布能够较好地反映结构的损伤特征;特征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,使诊断信息具有更高的精度和可靠性,提高了诊断准确率。In order to diagnose the occurrence, position and degree of the damage of complex structures accurately, a method was presented by means of wavelet packet decomposition, multisensor feature fusion theory and neural network pattern classification. Firstly, taking the orthogonal wavelet as a basis function, vibration signals gathered from sensors were decomposed. Secondly, the relative energy of decomposed frequency band was calculated. Thirdly, the input feature vectors of neural network classifier were built by fusing wavelet packet relative energy distribution of these sensors. Finally, with the trained classifier, damage diagnosis and assessment were realized. The results indicate that： the frequency band energy decornposited by orthogonal wavelet packet could perfectly reflect the damage condition; the fused feature can make different information complementary, and reduce the uncertainty of damage detection information, so the diagnosis information has much more precision and reliability, and the diagnosis accuracy can be improved. 2 tabs, 3 figs, 10 refs.

关 键 词：小波包分解 频带能量 神经网络 特征融合 损伤诊断 

分 类 号：TU973.2[建筑科学—结构工程] TU311.3