基于时间序列的飞机燃油测量方法  

A Fuel Measurement Method Based on Time Series Data

作　　者：郭毅博 李舜尧 陈艳华 尚志浩 潘俊[2] 刘祎 朱虹宇 徐明亮[1] Guo Yibo;Li Shunyao;Chen Yanhua;Shang Zhihao;Pan Jun;Liu Yi;Zhu Hongyu;Xu Mingliang(Zhengzhou University,Zhengzhou 450001,China;Aviation Key Laboratory of Science and Technology on Aero Electromechanical System Integration,AVIC Nanjing Engineering Institute of Aircraft Systems,Nanjing 211106,China)

机构地区：[1]郑州大学,河南郑州450001 [2]航空工业南京机电液压工程研究中心航空机电系统综合航空科技重点实验室,江苏南京211106

出　　处：《航空科学技术》2022年第5期63-68,共6页Aeronautical Science & Technology

基　　金：国家自然科学基金(61602421);航空科学基金(201828X4001)。

摘　　要：为满足现代飞机对燃油测量精度以及稳定性的需求,对飞机燃油测量领域进行探索,提出了一种基于时间序列的飞机燃油测量方法,将燃油数据视为多维时序数据,充分被利用数据信息,同时使用Transformer与LSTM将电容式传感器获取的数据作为时序数据处理,并通过贝叶斯优化合并两个模型减少了整体误差,增强了测量稳定性与可靠性。该方法在不同飞机姿态下都有稳定的表现,证明了可以通过基于时序数据的燃油测量方法获取燃油数据与飞机油量之间较为复杂的映射关系。In order to meet the requirements of modern aircraft for fuel measurement accuracy and stability,the field of aircraft fuel measurement is explored,and a time series-based aircraft fuel measurement method is proposed,which treats fuel data as multi-dimensional time series data and the data information is fully utilized.Using Transformer and LSTM at the same time to process the data obtained by capacitive sensor into time series data and the two models are merged through Bayesian optimization,reduce the overall error,and enhance the stability and reliability of the measurement.This method has stable performance in different aircraft attitudes,It is proved that the complex mapping relationship between fuel data and aircraft fuel quantity can be obtained by the fuel measurement method based on time series data.

关 键 词：油量测量 TRANSFORMER LSTM 时间序列数据 贝叶斯优化 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]