基于状态监测数据的航空发动机剩余寿命在线预测  

Online Remaining Useful Lifetime Prediction for Aero-Engine Based on Condition Monitoring Data

作　　者：李航[1] 张洋铭 LI Hang;ZHANG Yangming(Equipment Management&UAV Engineering College,Air Force Engineering University,Xi’an,710051,China;Beijing Institute of System Engineering,Beijing,100020,China)

机构地区：[1]空军工程大学装备管理与无人机工程学院,西安710051 [2]北京系统工程研究所,北京100020

出　　处：《南京航空航天大学学报》2020年第4期572-579,共8页Journal of Nanjing University of Aeronautics & Astronautics

摘　　要：针对现有基于状态监测数据的航空发动机剩余寿命预测研究未能综合考虑隐含退化建模和同步更新漂移/扩散系数的问题,提出一种基于状态监测数据的航空发动机剩余寿命在线预测方法。首先,基于非线性Wiener过程构建带比例关系的航空发动机隐含退化模型;其次,基于多台同类发动机的历史状态监测数据,对退化模型参数进行离线估计;然后,基于目标发动机的实时状态检测数据,利用贝叶斯原理同步更新退化模型漂移/扩散系数;最后,推导出航空发动机的剩余寿命概率密度函数。结合实例分析,验证了本文所提方法较传统方法具有更高的预测准确性与精度,具备潜在工程应用前景。For the problem of remaining useful lifetime(RUL)prediction of aero-engine,the present methods have not comprehensively considered the hidden degradation modeling and drift/diffusion coefficient synchronous updating.An online RUL prediction for aero-engine based on the condition monitoring(CM)data is presented in this paper.Firstly,the proportional degradation model of aero-engine is established based on the nonlinear Wiener process.Secondly,based on the historical condition monitoring data of similar engines,the degradation model parameters are estimated offline by using the maximum likelihood estimation(MLE)method.And then,based on the real-time condition monitoring data of the target engine,the drift/diffusion coefficient are synchronously update by using the Bayesian principle.Finally,the RUL probability density function of aero-engine is derived.The example analysis shows that the proposed method has higher prediction accuracy and precision than the traditional one,and has potential engineering application prospects.

关 键 词：状态监测 航空发动机 剩余寿命预测 WIENER过程 隐含退化建模 

分 类 号：TB114.3[理学—概率论与数理统计]