部分可观测信息条件下装备剩余寿命预测

 针对部分可观测信息条件下退化系统的剩余寿命(RUL)预测问题,综合利用装备的历史寿命信息和性能退化信息,采用隐马尔可夫模型(HMM)对系统进行状态评估,得到系统的转移概率矩阵和观测概率矩阵;采用Bayes方法不断更新系统状态空间的条件概率分布;利用比例故障率模型(PHM)对系统进行可靠性分析,得到系统的故障率和可靠度函数,进而得到装备的剩余寿命分布。研究表明,该方法可较准确地预测装备的剩余寿命,为保障人员提供科学的维修决策依据。

Prediction of Remaining Useful Life for Equipment with Partially Observed Information

In order to predict the remaining useful life(RUL) for a degraded system with partially observed information,the historical lifetime data and performance degradation data are fused together.Firstly,the hidden Markov model(HHM) is used for state evaluation to get the transition probability matrix and observation probability matrix of the system.Secondly,the Bayesian method is used to renew continually the conditional probability distribution of the equipment’s state.Then,a proportional hazards model(PHM) is used for reliability analysis to get the failure rate and reliability functions of the system.The remaining useful life distribution for the equipment is thus obtained.Case study indicates that the method can improve prediction precision effectively,which can help provide logistics personnel with a scientific basis for maintenance decision making.