基于多尺度振动传递率函数和灰度矩相对熵的损伤识别方法  

Damage Identification Based on Multi-scale Transmissibility Function and Grey Moment Relative Entropy

作　　者：杨斌[1] 刘吉彪[2] 程军圣[1] 

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室,长沙410082 [2]河南工业职业技术学院,南阳473000

出　　处：《中国机械工程》2015年第12期1639-1644,共6页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(51375152)

摘　　要：为了从结构振动响应中直接提取损伤敏感参数,对激励未知情况下的结构损伤模式进行识别,提出了基于多尺度振动传递率函数和灰度矩相对熵的损伤识别方法。对振动响应进行小波分解,通过计算不同尺度下的互相关函数,求解得到多尺度振动传递率函数。为了通过多尺度振动传递率函数提取损伤指标,计算了不同尺度范围内的多尺度振动传递率函数的灰度矩向量,该指标不受工况变化的影响。通过计算测试数据与样本数据灰度矩向量之间的相对熵,对结构的损伤模式进行了识别。实验结果表明,该方法能有效识别结构的损伤模式,适用于激励未知且工况变化情况下的损伤识别。To extract the damage sensitive features directly from the vibration response, a damage detec- tion method was proposed based on the multi-scale transmissibility function and the grey moment relative en- tropy, without requirments for the exciting forces. The multi-scale transmissibility function was obtained by calculating the cross correlation function between vibration responses under different scales after decomposed by the wavelet decomposition, and then the grey moment vector was acquired as the damage indicator by cal- culating the grey moment of the multi-scale transmissibility matrix under different scale ranges, which was independent of the different loading conditions. And the damage pattern could be identified by the relative en- tropy among the grey moment vectors of the test data and the sample data. Experimental tests undertaken on a experiment structure validate the proposed method.

关 键 词：结构损伤检测 小波分解 多尺度振动传递率函数 灰度矩 相对熵 

分 类 号：TH113.1[机械工程—机械设计及理论]