摘要
提出了一种改进的基于 ICA 法的航空发动机故障信号分析方法,通过将观察信号分块并利用各块来修正分解矩阵,使其能更准确、完整地利用信号所含信息。最后,将该方法应用于发动机振动信号的分离,仿真结果与理论基本相符。该方法对于航空发动机转子振动故障诊断具有一定的参考价值和理论指导意义。
A fault-diagnose method based on mended ICA was presented.It makes use of information involved in observed signals more precisely than common by dividing signals into many blocks which can be used to modify separated matrix.Applied results in aeroengine vibration signals proved that the method is feasible and effective. This method is contributive and instructional to aeroengine vibration fault-diagnose.
出处
《弹箭与制导学报》
CSCD
北大核心
2005年第S1期220-221,224,共3页
Journal of Projectiles,Rockets,Missiles and Guidance
关键词
航空发动机
振动
盲源分离
ICA
aeroengine
vibration
blind source separation
ICA
