摘要
应用小波包分解技术提取齿轮箱振动信号中的故障特征向量,并以此作为改进BP神经网络的输入,对神经网络进行训练,建立了齿轮箱运行状态分类器,用以识别齿轮箱的运行状态。试验结果表明,小波包分解与神经网络相结合的齿轮箱齿轮故障识别方法是可靠的,可以准确识别齿轮箱的故障。
The wavelet package was applied to decompose vibration signals of a gearbox to get the fault feature vectors. The feature vectors were employed as the input samples to train an improved BP neural network,and then the running state classifier of the gearbox fault was set up. The experimental results show that the proposed method is effective for gearbox fault diagnosis.
出处
《振动．测试与诊断》
EI
CSCD
北大核心
2009年第3期321-324,共4页
Journal of Vibration,Measurement & Diagnosis
基金
山西省科技攻关项目(编号:2007032054)
关键词
小波包
BP神经网络
齿轮箱
故障识别
wavelet packet BP neural network gearbox fault recognition
