基于BP神经网络的圆锥滚子轴承故障诊断

摘要

圆锥滚子轴承在旋转机械中应用较为广泛，其故障对整机的运行状态将造成极大的影响，因此对其进行故障诊断十分必要。文中提出了基于神经网络的圆锥滚子轴承故障诊断方法。利用小波包分解对轴承的振动信号进行分析，将分解后得到的小波包能量矩归一化处理后作为特征向量，用标准数据的特征向量构成的训练样本对BP神经网络进行训练和测试，达到误差要求后，用该网络对圆锥滚子轴承的故障仿真实验数据进行故障诊断，诊断结果在误差范围内，达到故障诊断目的，验证了该方法在圆锥滚子轴承故障诊断中的有效性。%Tapered roller bearings is widely used in rotating machinery, the failure of the bearing will cause a great impact on the complete machine’ s running status. ,so its fault diagnosis is necessary. In this paper, we put forward a method, which is based on BP neural network ,to diagnose the fault of tapered roller bear-ing. Wavelet packet decomposition is used to analyzing vibration signals of the bearings, after the analysis, we get the wavelet packet energy,which can be used as a feature vector after the normalization process. U-sing the feature vectors that contains standard data as the training samples to train and test the BP neural net-work, after reaching the error requirements, this network is used to diagnose the fault of tapered roller bear-ing, diagnose results is in the error range, so we achieve the purpose of fault diagnosis, then can also verify the effectiveness of this method that is used to fault diagnosis of the tapered roller bearing.