摘要
针对测控天线传动链常见的螺栓松动故障,提出了一种子带功率均值(APSDS)的振动信号特征提取方法,通过搭建天线座架试验平台进行故障模拟和振动数据采集,并利用多层神经网络进行故障分类诊断及预测试验。试验结果表明,本文所提出的APSDS法能够有效增强故障状态的特征区分度,平均故障分类准确率达到95%以上,该方法的提出可在天线健康管理系统中具有较好的实用价值。
In view of the common bolt looseness fault in the antenna transmission chain,a vibration signal feature extraction technology based on APSDS is proposed in this research.Meanwhile,fault simulation and vibration data acquisition are carried out by building an antenna pedestal test platform,and then fault classification diagnosis and prediction tests are conducted by using multi-layer neural network.The experimental results show that the proposed APSDS method can effectively enhance the feature discrimination of fault status with the average fault classification accuracy of over 95%.The proposed method has good practical value in the antenna health management system.
作者
李蒙
周昊天
霍克强
刘泽成
王京城
李扬
LI Meng;ZHOU Haotian;HUO Keqiang;LIU Zecheng;WANG Jingcheng;LI Yang(The 54th Research Institute of CETC,Shijiazhuang Hebei 050081,China)
出处
《河北省科学院学报》
CAS
2023年第3期30-37,共8页
Journal of The Hebei Academy of Sciences
关键词
天线试验平台
故障诊断
子带功率均值
特征提取
神经网络
Antenna test platform
Fault diagnosis
Average of power spectral density sub-bands
Feature extraction
Neural network