核独立分量分析的随机滤波剩余寿命预测模型

A Stochastic Filtering Prediction Model for Residual Useful Life Based on Kernel-Independent Component Analysis

基于随机滤波的预测模型是剩余寿命预测方法的一个重要分支,当前制约滤波模型的一个重要问题就是如何对大量高维非线性状态监测数据进行特征降维,以易于模型参数求解。通过线性回归处理了非定期换油保养对油液数据的影响;运用核独立分量分析进行特征降维,消除了各维数据之间相关性对模型预测精度的影响;建立了基于油液增量的滤波模型,并设计了极大似然估计方法求解模型参数;最后实例验证了模型的有效性和实用性。

The stochastic filtering prediction model is an important branch of residual useful life methods.At present,a difficult problem that restricts the filtering model developing is how to make a dimensional reduction when with a great deal of non-linear condition monitoring data,for it would make the parameters estimating be more easily.Firstly,the metal concentrations data with nonscheduled oil changes activities is analyzed by linear regression method.Secondly,Kernel Independent Component Analysis（KICA） algorithm is used to accomplish the dimensional reduction process,which would avoid the influence of correlations between data dimensions to model prediction precision.Thirdly,a stochastic filtering prediction model is established based on the increment features of oil data.Fourthly,it designs a maximum likelihood estimation method to estimate the unknown parameters.Finally,the validity and practicability of the model are validated by an example.