摘要
针对风力机叶片蒙皮多裂纹难以状态识别的问题,根据裂纹扩展释放能量的过程,推导主裂纹扩展AE信号的表达式,从而明晰了主裂纹扩展的AE信号特性及其与应力变化之间的关联。由于多裂纹扩展AE信号为卷积混合模型,提出一种对具有非平稳、非线性特性的卷积混合AE信号特征提取的方法,以输出信号的广义能量作为目标函数得到盲解卷的滤波器迭代式,采用Godard算法通过输出信号与估计值的误差调整滤波器系数,并根据相似系数选择适当的非线性函数以减少采集设备对AE信号的影响。最后在裂纹扩展试验中,预制不同尺寸的多缺陷,对叶片试件同时施加激振载荷和循环载荷,每间隔一定的循环次数采集不同状态的AE信号,同时采用具有非全局性的瞬时频率和特征尺度来识别多裂纹在不同扩展状态下的特征,从而明晰了信号特征与多裂纹生存状态的关联,形成了识别多裂纹复合材料损伤的评价机制。
Aiming at the difficult problem of multi-crack condition identification of wind turbine blade skin,the main crack propagation expression of AE signal is derived according to the energy release process of crack propagation.The AE signal characteristics of the main crack propagation and its correlation with the stress changes are ascertained. A feature extraction method for the convolution mixing AE signal with non-stationary,non-linear characteristics is proposed since the AE signals of multi-crack propagation is adapted to the convolution mixing model. Taking the generalized energy of the output signals as the objective function,the filter iterative expression of blind deconvolution is obtained. And the filter coefficient values are adjusted through the error between the output signals and the estimation signals using Godard algorithm. And the appropriate nonlinear function is selected according to the similarity coefficient to reduce the effect of acquisition device on AE signal. Finally,the blade samples with multiple defects with different sizes were prefabricated in the crack propagation tests; the excitation load and cyclic load were exerted on the blade samples; and the AE signals were acquired at the intervals of a certain number of cycles from the blade samples. The instantaneous frequency and characteristic scale were used to identify the features of the multi-cracks under different propagation conditions; and the correlation of the signal characteristics and survival status of the multiple cracks were ascertained,which forms the evaluation mechanism of recognizing the multi-crackdamage of composite materials.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2015年第1期110-117,共8页
Chinese Journal of Scientific Instrument
基金
国家科学支撑计划(2013BAF07B04)
中国博士后科学基金(2014M560220)资助项目
关键词
风力机
叶片
扩展裂纹
声发射信号
卷积混合
特征识别
wind turbine
blade
propagating crack
acoustic emission signal
convolution mixing
feature identification