基于马氏距离的数模联动滚动轴承寿命评估

Life Prediction of Digital Analog Linkage Rolling Bearings Based on Mahalanobis Distance

作为高速列车走行部系统的关键部件,滚动轴承的剩余寿命预测直接决定了系统的剩余寿命。因此,对滚动轴承进行退化状态监测和剩余使用寿命(RUL)预测具有重要的工程意义。将马氏距离作为滚动轴承振动信号及温度信号特征指标的度量,建立相应性能指标序列,从指标序列中观测滚动轴承的退化状态,结合以指数模型为退化模型的数模联动方法进行RUL预测,并引入基于3σ准则的首次预测时间(FPT)以触发RUL预测过程,可以实现滚动轴承RUL的可靠预测。将这种方法和传统神经网络的寿命预测方法进行对比实验,结果表明,本文基于马氏距离的偏离指标数模联动寿命评估方法具有较高的预测准确度。

As a key component of the high-speed train running gear system,the prediction of the remaining life of rolling bearings directly determines the remaining life of the system.Therefore,monitoring the degradation status and predicting the remaining useful life(RUL)of rolling bearings is of great engineering significance.Using Mahalanobis distance as a measure of the vibration signal and temperature signal characteristics of rolling bearings,a corresponding performance index sequence is established.The degradation state of rolling bearings is observed from the index sequence,and RUL prediction is carried out using the exponential model as the degradation model.A first predicting time(FPT)based on 3o criteria is introduced to trigger the RUL prediction process,which can achieve reliable prediction of RUL of rolling bearings.A comparative experiment was conducted between this method and the traditional neural network life prediction method,and the results showed that the deviation index based on Mahalanobis distance in this paper has a high prediction accuracy for the numerical model linkage life evaluation method.