摘要
在某导弹支撑座模型宽带随机振动实验的基础上,针对其连接螺栓松动所产生的支撑座结构响应的非平稳特性,采用小波包分析的方法得到缩减的信号特征;然后利用BP神经网络的模式分类功能,进行了螺栓松动程度的损伤识别研究。实验结果表明,小波包结合神经网络的方法可以有效地识别该支撑座连接螺栓的松动程度。
On the basis of broad-band random vibration test on a clamping support in a missile, wavelet packet analysis method was used to obtain the reduced features of structural response signals, which have non-stationary characteristics due to the attachment bolt looseness. Then damage identification of severity of bolt looseness was studied by utilizing the pattern classification function of the BP neural network. Test results show that the combination of wavelet packets and neural networks can effectively identify the severity of bolt looseness in the clamping support.
出处
《机械科学与技术》
CSCD
北大核心
2006年第1期102-106,共5页
Mechanical Science and Technology for Aerospace Engineering
基金
国家自然科学基金项目(10176014)资助
