摘要
根据混沌信号处理理论,研究分析了关联维数、最大Lyapunov指数和Kolmogorov熵等转子碰摩声发射源的非线性动力学特征量的特性。并将这些特征量作为神经网络的输入,研究了碰摩声发射源的定位问题。实验结果表明,该方法能较好地解决碰摩声发射源定位问题,为转子碰摩故障诊断提供依据。
A nonlinear dynamic analysis of the acoustic emission(AE) signals in rotor rubbing was conducted based on the chaotic signal processing technique and such new features of nonlinear dynamics as correlation dimension,largest Lyapunov index and the Kolmogorov entropy were proposed herein.Simulation results reveal the neural network model employing these new features improves the localization efficiency of the AE source and contributes to the diagnosis of rotor rub-impact faults.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2011年第9期1013-1017,共5页
China Mechanical Engineering
基金
国家自然科学基金资助项目(60472058
60975017
51075068)
江苏省自然科学基金资助项目(BK2008291)
关键词
声发射
定位
碰摩
混沌
神经网络
acoustic emission
localization
rub-impact
chaos
neural network