摘要
随机退化设备在实际运行中会产生非线性、隐含性等问题,对其剩余寿命预测会产生不确定性影响。现有剩余寿命预测方法尚未系统研究隐含非线性退化建模及相应的剩余寿命分布。因此,采用Wiener过程,建立了隐含双重非线性退化模型;利用设备现场监测数据,更新了隐含状态的后验分布;利用全概率公式,基于首次达到失效阈值的时间分布推导出设备剩余寿命分布;基于激光器实测退化数据设定仿真参数,对所提方法的正确性和合理性进行了对比验证。
Stochastic degradation equipment will produce the problems of nonlinearity and implicitness in actual operation,which will cause uncertainty on the remaining useful lifetime(RUL)prediction.Existing RUL prediction methods have not systematically studied implicit nonlinear degradation modeling and corresponding RUL distribution.The Wiener process is used to build the implicit dual-nonlinear degradation model.The online monitoring data of equipment is used to update the posterior distribution of implicit states.By using the full probability formula,the RUL distribution is derived based on the first hitting time(FHT)distribution.The simulation parameters are set based on the degradation data of the laser,which verifies the accuracy and rationality of the proposed method.
作者
蔡忠义
王泽洲
张晓丰
李岩
CAI Zhongyi;WANG Zezhou;ZHANG Xiaofeng;LI Yan(Equipment Management&UAV Engineering College,Air Force Engineering University,Xi’an 710051)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2020年第6期1410-1416,共7页
Systems Engineering and Electronics
基金
国家自然科学基金(71901216)
中国博士后科学基金(2017M623415)资助课题。
关键词
剩余寿命预测
WIENER过程
退化建模
测量误差
双重非线性
remaining useful lifetime(RUL)prediction
Wiener process
degradation modeling
measurement error
dual-nonlinearity
