| 基于MSCNN-LSTM编解码器的飞机辅助动力装置EGT预测模型 |
| |
| 引用本文： | 白春垣,孙有朝.基于MSCNN-LSTM编解码器的飞机辅助动力装置EGT预测模型[J].组合机床与自动化加工技术,2022(2). |
| |
| 作者姓名： | 白春垣 孙有朝 |
| |
| 作者单位： | 南京航空航天大学民航学院 |
| |
| 基金项目： | 国家自然科学基金(U1333119);国家重大专项基础研究项目(2017-Ⅷ-0003-0114,2017-VIII-0002-0113);中央高校基本科研业务费(56XBC20018);装备预研领域基金(61400020402);“十三五”装发共用技术项目(41402040410);南京航空航天大学研究生创新基地(实验室)开放基金项目资助(kfjj20190712)。 |
| |
| 摘 要： | 针对提高飞机辅助动力装置(auxiliary power unit,APU)排气温度(exhaust gas temperature,EGT)参数的预测精度问题,提出了一种基于特征选择和多尺度卷积-长短期记忆网络编码器-解码器的EGT预测模型。首先,利用随机森林方法确定重要度较高的APU监测参数;其次,利用多尺度卷积神经网络能够提取信号深度特征和LSTM网络能够学习序列长时间依赖的特性,建立了编码器-解码器预测模型;最后,以某型APU实时报文数据为例,通过与其他方法进行对比验证了模型的可行性,能够提高EGT预测的准确度。
|
| 关 键 词： | 辅助动力装置 排气温度 多尺度卷积神经网络 长短期记忆网络 编码器-解码器 |
APU Exhaust Gas Temperature Prediction Based on MSCNN-LSTM Encoder-Decoder Method |
| |
| BAI Chun-yuan,SUN You-chao.APU Exhaust Gas Temperature Prediction Based on MSCNN-LSTM Encoder-Decoder Method[J].Modular Machine Tool & Automatic Manufacturing Technique,2022(2). |
| |
| Authors: | BAI Chun-yuan SUN You-chao |
| |
| Affiliation: | (College of Civil Aviation,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China) |
| |
| Abstract: | To improve the accuracy of auxiliary power unit(APU)exhaust temperature(EGT)prediction,a model based on feature selection and MSCNN-LSTM encoder-decoder is proposed.Firstly,random forest method is employed to determine APU state monitoring parameters which is important to predict EGT.Secondly,since multiscale convo-lutional neural network has the property of capturing deep features of signals and LSTM network has the property of long time dependence,an encoder-decoder prediction model is established.Finally,it is proved by the real-time message data of a certain type of APU and compared with other methods that the proposed method has high accuracy of prediction. |
| |
| Keywords: | auxiliary power unit exhaust gas temperature multiscale convolutional neural network long short term memory network encoder-decoder |
| 本文献已被 维普 等数据库收录！ |
