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
张青 1赵洪利 1杨佳强1
作者信息
- 1. 中国民航大学 航空工程学院,天津 300300
- 折叠
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.
Abstract
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 im-portance 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 re-lationships 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 meth-od 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Key words
Aircraft engine/CNN-BiLSTM/Lubricating oil flow prediction/Deep neural network/Quick Access Recorder(QAR)data引用本文复制引用
基金项目
中国交通教育研究会2022-2024年度教育科学研究课题(JT2022YB326)
出版年
2024
NETL
NSTL
维普
万方数据