何庆飞;王汉功;陈小虎;毋文峰

第二炮兵工程学院

    摘要：针对柴油机缸盖振动信号的非平稳时变特点,提出应用小波包能量法提取故障特征向量,并将提取的特征向量作为BP神经网络的输入向量进行学习训练。训练后的神经网络可以利用测量的振动信号判断柴油机的气阀机构故障状况。实践证明该方法在柴油机振动诊断中是有效可行的,对其他设备的故障诊断也具有借鉴意义。

    关键词： 小波包;;神经网络;;柴油机;;故障诊断

      Abstract: Aiming at the non-stationary and time-varying characteristics of vibration signals measured on cylinder head in diesel engine,a method of extracting fault characteristic vectors by the energy of wavelet packet over the signals is presented.The neural networks' input vectors are characteristic vectors,then learning and training the BPNN.After that training,the BPNN can be used to classify faults of valve train diesel engine using the input vibration signals.The practice proves that this method can efficiently diagnose and classify faults.In addition,the approach can also be applied to the fault diagnosis of other equipments.

      Keywords: Wavelet Packet;Neural Network;Diesel Engine;Fault Diagnosis

参考文献：

[1]黄强等.基于小波理论变换和神经网络的柴油机故障诊断方法的研究[J].柴油机设计与制造,2006,1(14):9~14.

[2]刘世元,杜润生,杨叔子.内燃机缸盖振动信号的特性与诊断应用研究[J].华中理工大学学报,1999,27(7):48~50.

[3]夏勇等.基于小波分析与神经网络的气阀机构故障诊断研究[J].内燃机学报,2001,19(3):235~240.

[4]焦李成.神经网络系统理论[M].西安:西安电子科技大学出版社,1996.

[5]Wichehauser M V.Lectures on Wavelet Packet Algorithms[J].Information Theory.1992,38:713~718.

[6]胡昌华,张军波,夏军,等.基于Matlab的系统分析与设计——小波分析[M].西安:西安电子科技大学出版社,1999.