摘要
由于活塞敲缸响和活塞销响是两种常见的、却难以区分的柴油机异响故障,这里对EQ6BT柴油机这两种故障的缸体振动信号进行Morlet连续小波变换,作出小波变换系数的尺度-能量谱,并提取出尺度为3~20范围内的最大尺度能量作为BP神经网络的输入向量,实现了对该柴油机两种异响故障的诊断。结果表明,利用文中所设计的小波神经网络能非常准确地诊断出EQ6BT柴油机活塞敲缸响、活塞销响两种异响故障及其故障的严重程度。
Piston knocks and piston pin knocks are frequent,and hard to distinct in diesel engines.Through the use of maximum scale-power coefficients,which are extracted from the scale-power spectrum of the vibrant signal of EQ6BT diesel engine generated by the continuous wavelet transform,as the input vector of a BP neural network,a way is proposed to combine wavelet transform and BP neural networks to realize the exact diagnosis of those two faults.The analytical result proves that,by utilizing the wavelet neural networks in this paper,the two abnormal sound faults of diesel engine are distinguished efficiently,and the serious degrees of the faults are monitored as well.
出处
《振动.测试与诊断》
EI
CSCD
2006年第3期208-211,共4页
Journal of Vibration,Measurement & Diagnosis
关键词
小波变换
神经网络
柴油机
故障诊断
wavelet transform neural networks diesel engines fault diagnosis