基于时变自回归模型与神经网络的滚动轴承故障智能诊断

摘要

使用时变自回归建模分析方法建立滚动轴承振动信号特征提取模型，基于基函数算法求解该模型的时变参数，并采用AIC准则确定模型阶数。在利用上述参数化模型对轴承振动信号进行特征提取的基础上，构建BP神经网络，有效地实现了轴承故障的智能诊断。%The feature extraction model for vibration signal of rolling bearings is established by using time varying au-toregressive modeling method.The time varying parameters for the model is solved based on basis function arithmetic, and the model order is determined by using AIC rule.On the basis of above-mentioned parameterized model for fea-ture extraction,a BP neural network is built,and the intelligent diagnosis for fault of rolling bearings is effectively real-ized.