基于优化信噪比算法的航空发动机振动信号分离  

Application of optimized signal noise ratio algorithm in aero-engine vibration signal separation

作　　者：滕光蓉[1] 李舜酩[2] 梁恩波[1] 王艳丰[1] 

机构地区：[1]中国燃气涡轮研究院,四川江油621700 [2]南京航空航天大学,南京210016

出　　处：《燃气涡轮试验与研究》2016年第2期26-31,共6页Gas Turbine Experiment and Research

基　　金：国家自然科学基金(50675099)

摘　　要：为克服噪声对航空发动机振动信号的影响,提出一种优化信噪比的航空发动机振动信号分离方法,其通过建立时延自相关目标函数,对最大信噪比盲源分离算法进行优化。利用该方法,对航空发动机仿真混叠振动信号进行分析,分离后的信号频谱与源信号一致;对具有故障的实测航空发动机振动信号进行分离,成功分离出不同故障的振动特征。本文提出的方法为航空发动机振动信号的监测和故障诊断,提供了一种新的思路。In order to overcome the influence from noise on the vibration signals of aero-engine, the opti-mized Signal Noise Ratio （SNR） algorithm was introduced to analyze aero-engine mixed vibration signals. Based on time delay auto correction function, SNR was used to optimize the maximum signal noise ratio Blind Source Separation algorithm. The simulated mixed vibration signals were analyzed and the separated frequency spectrum was found to agree with that of the sourced signals. The faulted vibration signals from the real aero-engine tests were separated, and different fault frequencies have been found. It is proved that the optimized SNR algorithm is largely effective on mixed vibration signals separation, and it can be widely used in aero-engine tests.

关 键 词：航空发动机 信噪比 时延自相关 盲源分离 混叠振动信号 故障诊断 

分 类 号：TN911.3[电子电信—通信与信息系统] V263.3[电子电信—信息与通信工程]