基于地面保障流程的过站航班延误预测方法  

Delay Prediction Method of Transit Flight Based onGround Guarantee Process

作　　者：羊钊[1] 陈怡欣 张智杰 YANG Zhao;CHEN Yixin;ZHANG Zhijie(College of General Aviation and Flight,Nanjing University of Aeronautics&Astronautics,Liyang 213300,Jiangsu,China;College of Civil Aviation,Nanjing University of Aeronautics&Astronautics,Nanjing 211106,Jiangsu,China)

机构地区：[1]南京航空航天大学通用航空与飞行学院,江苏溧阳213300 [2]南京航空航天大学民航学院,江苏南京211116

出　　处：《重庆交通大学学报（自然科学版）》2023年第9期122-129,154,共9页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：国家自然科学基金项目(52172328)。

摘　　要：目前民航航班持续高位运行,航班过站时机场、空管、航司多方同时交织参与,高峰时段机场地面保障系统满负荷运转,航班正点率难以提升。为了提前感知过站航班在地面保障时多流程进行协作时产生的延误,针对航班作业流程节点的相关特征,提出基于地面保障流程的过站航班延误预测方法。将航班保障流程构建为图网络结构,采用各流程节点上处理后的时间特征作为图卷积神经网络的各节点特征,针对节点特征采用多种聚合传递方式并进行集成,实现航班延误预测精度的提升。结果表明,提出的航班延误预测方法的平均预测误差降低至7.11 min,具有更好的泛化能力。At present,civil aviation flights continue to operate at a high level,and the airport,air traffic control and airline companies participate in the flight transit at the same time.The airport ground support system operates at full load during peak hours,and the flight punctuality rate is difficult to improve.In order to perceive the delays caused by multi-process cooperation during ground support for transit flights in advance,aiming at the relevant characteristics of flight operation process nodes,the flight delay prediction method based on the ground guarantee process for transit flights was proposed.The flight guarantee process was constructed as a graph convolutional neural network structure,the processed time features on each process node were used as the node features of the graph neural network.For the node features on the graph,a variety of aggregation delivery methods were used and integrated to improve the accuracy of flight delay prediction.The results show that,compared with the comparison methods,the average prediction error of the proposed flight delay prediction method is reduced to 7.11 minutes,which has better generalization ability.

关 键 词：交通运输工程 航班保障流程 时间特征处理 图神经网络集成 过站航班 延误预测 

分 类 号：V351[航空宇航科学与技术—人机与环境工程]