摘要
对多转子系统中转子故障诊断进行了研究,以转子正常、偏心、不平衡和弯曲四种工作状态为例,采用径向基函数(RBF)神经网络对故障进行诊断。通过快速傅里叶变换和能量谱对转子振动信号进行特征提取,并将提取的特征向量作为神经网络的输入,实现多转子故障类型的识别。结果表明,利用能量谱和RBF神经网络能够有效地识别转子故障类型。
Aims at the research of rotor fault diagnosis in the muhiple rotors system. Four kinds of work states rotor normal, rotor eccentricity, rotor imbalance and rotor bending are diagnosed base on Radial Basis Function （RBF）neural network. The charac- teristics of the rotor vibration signals are extracted through the Fast Fourier Transform （FFT） and the energy spectrum. The extrac- ted characteristic vectors are inputted to the nerve network and the identification of multiple rotors diagnosis types are realized. The result shows that using the energy spectrum and RBF neural network can identify the rotor fault types effectively.
出处
《现代制造工程》
CSCD
北大核心
2013年第5期126-130,共5页
Modern Manufacturing Engineering
基金
江苏省科技支撑计划项目(BE2011046)
关键词
多转子
能量谱
RBF神经网络
故障类型诊断
multiple rotors
energy spectrum
RBF neural network
fault types diagnosis
