基于CNN-BiLSTM的航空发动机滑油流量故障诊断预测方法研究  

Research on Aircraft Engine Lubricating Oil Flow Fault Diagnosis and Prediction Method Based on CNN-BiLSTM

作　　者：张青[1] 赵洪利[1] 杨佳强 Zhang Qing;Zhao Hong-li;Yang Jia-qiang(College of Aeronautical Engineering,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学航空工程学院,天津300300

出　　处：《内燃机与配件》2024年第8期84-87,共4页Internal Combustion Engine & Parts

基　　金：中国交通教育研究会2022-2024年度教育科学研究课题(JT2022YB326)《基于智慧民航建设的“新工科”运维类课程教学模式改革与实践》。

摘　　要：航空发动机滑油系统为整个发动机的传动系统、轴承齿轮等部件提供滑油,是保证航空发动机正常运行的重要系统,因此准确对航空发动机滑油量进行预测是对保证飞机飞行的安全有重要意义的。为了提高预测准确性,提出了一种基于CNN-BiLSTM的航空发动机滑油流量预测模型,可以同时捕捉数据中的空间特征以及时序关系。以某航QAR数据进行验证,结果与CNN和LSTM模型进行对比,左发预测准确率提升了2.43%和7.85,右发预测准确率提升了7.97%和10.82%,证明了本文所提方法的有效性,为航空发动机滑油流量故障诊断的预测方法提供了新的解决方案。The aircraft engine lubrication system provides lubricating oil for the entire engine's transmission system,bearing gears,and other components.It is an important system that ensures the normal operation of the aircraft engine.Therefore,accurate prediction of the aircraft engine lubricating oil quantity is of significant importance for ensuring flight safety.In order to improve prediction accuracy,a prediction model for aircraft engine lubricating oil flow based on CNN-BiLSTM is proposed,which can capture both spatial features and temporal relationships in the data.Verified with QAR data from a certain airline,the results are compared with CNN and LSTM models.The accuracy of left engine prediction has increased by 2.43%and 7.85%,and the accuracy of right engine prediction has increased by 7.97%and 10.82%.This proves the effectiveness of the proposed method in this study and provides a new solution for predicting aircraft engine lubricating oil flow.

关 键 词：航空发动机 CNN-BiLSTM 滑油流量预测 深度神经网络 快速存取(QAR)数据 

分 类 号：V263.6[航空宇航科学与技术—航空宇航制造工程]