时变系统基于KDE的剩余寿命预测研究

Research on Residual Life Prediction of Time-varying Systems Based on Kernel Density Estimation

考虑到机械系统实际运行中参数随时间不断演化,以及自身监测样本数据不够,同类系统故障样本缺少等,导致剩余寿命预测时模型结构假设和参数估计不够准确的问题,推导建立了时变系统基于KDE的剩余寿命预测模型。与传统数据驱动方法相比,该方法既可避免预测模型假设不够准确的问题,同时也可以动态地预测未来任意时刻的剩余寿命,从而为系统的未来退化趋势提供先期预警,为设备维护维修提供指导。

Considering that the parameters in the actual operation of the mechanical system evolve with time,the insufficient data of self-monitoring samples,the lack of failure samples of the same kind of systems,and so on,resulting in inaccurate assumptions of the model structure and parameter estimation in the residual life prediction,this paper deduces and establishes the residual life prediction model of the time-varying system based on KDE.Compared with traditional data-driven methods,this method can not only avoid the problem of inaccurate prediction model assumptions,but also dynamically predict the remaining life at any time in the future,so as to provide early warning for the future degradation trend of the system and provide guidance for equipment maintenance and repair.