基于多传感器融合的航空发动机剩余寿命预测  

The remaining life prediction of aero-engine based on multi-sensor fusion

作　　者：江鹏伟 许卓韬 段霞 巩瑞东 柴森浩 Jiang Pengwei;Xu Zhuotao;Duan Xia;Gong Ruidong;Chai Senhao(School of Mechanical Engineering,North University of China,Shanxi Taiyuan,030051,China)

机构地区：[1]中北大学机械工程学院,山西太原030051

出　　处：《机械设计与制造工程》2025年第2期52-56,共5页Machine Design and Manufacturing Engineering

基　　金：国家自然科学基金(51905496,52305140);山西省重点研发计划(202102020101015)。

摘　　要：针对单一传感器数据难以准确表征航空发动机退化状态,且在退化建模时存在初始退化状态异构等问题,提出了一种基于退化趋势一致性的多传感器数据融合方法和考虑初始退化状态异构性的维纳过程模型进行航空发动机剩余寿命预测。首先,提出了一种退化趋势一致性评价指标来衡量不同传感器数据的信息贡献率,确定最优的融合权重;其次,对发动机进行退化建模,考虑了初始退化状态和时间的随机性,并推广到考虑单位之间可变性的情况;再次,通过极大似然估计算法,估计出模型中的未知参数;最后,通过C-MAPSS实验数据对比,验证了该方法在提高数据质量和预测精度方面的优势。Aiming at the difficult to accurately characterize the degradation state of an aircraft engine with single sensor data and the heterogeneity of the initial degradation state during degradation modelling,a multi-sensor data fusion method based on the consistency of the degradation trend and a Wiener process model considering the heterogeneity of the initial degradation states is proposed for predicting the remaining life of an aircraft engine.Firstly,a degradation trend consistency evaluation index is proposed to measure the information contribution rate of different sensor data and determine the optimal fusion weights.Secondly,the engine is modelled for degradation,taking into account the stochastic nature of the initial degradation state and time,and extended to consider the case of inter-unit variability.Finally,the unknown parameters in the model are estimated in a great likelihood estimation algorithm.The advantages of the method in terms of improving data quality and prediction accuracy are verified by comparison with the C-MAPSS experimental data.

关 键 词：维纳过程 多传感器数据融合 航空发动机 剩余寿命预测 布朗运动 

分 类 号：TH17[机械工程—机械制造及自动化] TH1133.33