Abstract: Aiming at the problem that the existing Remaining Useful Lifetime (RUL) prediction method fused with multi-source degradation data ignore the influence of the Random Failure Threshold (RFT),a multi-source degradation data fusion and RUL prediction method considering the RFT was proposed.A fusion coefficient determination criterion considering the RFT was established,and the multi-source degradation data were fused into a single Health Index (HI).The Wiener process with linear drift was used to establish the degradation model of the obtained HI,and the unknown parameters were estimated by Maximum Likelihood Estimation (MLE) method and updated Bayesian principle.Then,the analytical expression of RUL's Probability Distribution Function (PDF) under the influence of RFT was derived based on the full probability formula.Finally,the aero-engine degradation data was taken as an example to analyze,and the result proved that the proposed method could effectively improve the accuracy and precision of the RUL prediction and had engineering application value.

Key words: random failure threshold, data fusion, Wiener process, remaining useful lifetime prediction

摘要： 针对现有融合多源退化数据的剩余寿命预测方法忽略随机失效阈值影响的问题,提出一种考虑随机失效阈值的多源退化数据融合与剩余寿命预测方法。首先,建立考虑随机失效阈值的融合系数确定准则,将多源退化数据融合为单一健康指标;其次,采用带线性漂移的维纳过程建立所得健康指标的退化模型,利用极大似然估计法求解模型的未知参数,并基于贝叶斯原理对其进行更新;然后,基于全概率公式推导出随机失效阈值影响下剩余寿命概率分布的解析表达式;最后,以航空发动机退化数据为例进行分析,证明了所提方法能够有效提升剩余寿命预测的准确性与精度,具备工程应用价值。

关键词: 随机失效阈值, 数据融合, 维纳过程, 剩余寿命预测