基于改进粒子群算法的直升机振动信号盲分离  

Blind source separation of helicopter vibration signal based on improved particle swarm algorithm

作　　者：申江江 高峰[1] 曲建岭[1] 袁涛[1] 刘锦涛[1] 

机构地区：[1]海军航空工程学院青岛校区,青岛266041 [2]海军航空工程学院航空训练基地,青岛266109

出　　处：《电子测量与仪器学报》2016年第9期1372-1378,共7页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金(51505491)资助项目

摘　　要：振动信号分析是直升机状态监控的重要研究内容之一。针对某型直升机加装的振动信号采集系统,提出一种基于改进粒子群优化算法(PSO)的盲源分离方法。介绍了盲源分离(BSS)基本理论及改进的基于动态因子的粒子群优化算法,将二者结合应用于某型直升机振动信号盲分离研究。仿真实验及噪声中实测直升机振动信号的盲分离都取得了良好的分离效果,结果表明相比于直接使用FastICA方法,所提出的方法具有分离精度高、迭代次数少等优点。Vibration signal analysis is one of significant research contents for helicopter condition monitoring. Aiming at the vibration signal acquisition system installed a helicopter, a blind source separation （BSS） method based on improved particle swam optimization （PSO） is proposed. BSS basic theory and improved PSO based on dynamic factor are introduced, and they are combined and put into BSS research of helicopter vibration signal. The blind source separation of simulation experiment and real-acquainted helicopter vibration signals in noise environment obtains good separation result. And compared with using FastlCA directly, the result shows that this algorithm proposed in this paper has advantages of high precision and less iteration times.

关 键 词：直升机振动监控 独立分量分析 粒子群优化算法 动态因子 小波降噪 

分 类 号：TN911.7[电子电信—通信与信息系统]