基于改进小波包能量的梁式结构损伤识别  

Damage detection method for beam structures based on improved Wavelet Packet energy

作　　者：刘习军[1] 商开然[1,2] 张素侠[1] 霍冰[1] 孙良[1] 

机构地区：[1]天津大学机械工程学院,天津300072 [2]中建交通建设集团有限公司,北京100000

出　　处：《振动与冲击》2016年第13期179-185,200,共8页Journal of Vibration and Shock

基　　金：国家自然科学基金(51009107);天津市基础重点基金(13JCZDJC27100);天津市青年基金(13JCQNJC04200)

摘　　要：利用小波包变换技术将结构的节点曲率响应分解到不同模态,选择能量较大的结构单阶模态响应构造损伤指标,使得损伤指标具有较强的噪音鲁棒性和损伤敏感性。通过统计损伤指标超过损伤预警值的次数来识别损伤,进一步降低了环境噪音对识别结果的影响。通过多种不同损伤工况的简支梁数值算例和实验验证了该方法的有效性。结果表明,该方法能够用于工程实际,当响应信号中的噪音在10%以下时,此方法能够准确的识别出较小程度的结构损伤。Structural nodal curvature responses were firstly decomposed into different modes with the wavelet packet transformation. Then,the modal response with larger energy was selected to construct the damage index with better noise robustness and damage sensitivity. To further reduce the effect of noise on detection results,the number of the damage index exceeding the damage warning value was counted to detect the damage of a structure. The effectiveness of the proposed approach was verified with numerical simulations and tests for damage detection of a simply supported beam under several different damage conditions. The results indicated that the method can be used to identify small structural damages accurately when the noise in response signals is less than 10%.

关 键 词：节点曲率 损伤识别 小波包变换 噪音鲁棒性 

分 类 号：U447[建筑科学—桥梁与隧道工程] O327[交通运输工程—道路与铁道工程]