论文部分内容阅读
QAR海量数据分析是航空安全管理的一个重要课题,准确而高效地挖掘其中蕴含的飞行操作模式、评估其风险水平,是切实提高飞行操作品质的关键。然而,飞行过程中飞行员需要根据动态变化的环境条件、飞机状态,对油门、杆、盘、舵等设备进行操作,是一个需要不断调整、相互配合的过程,这增加了对飞行操作模式及其风险进行分析的难度。本文从飞行参数多元时间序列数据的特征提取及相似性度量入手,通过聚类挖掘蕴含的飞行操作模式,进而,通过分析飞行操作模式与QAR监控指标值之间的关联关系,量化不同飞行操作模式的风险水平。最后,本文以首都机场特定机型着陆阶段QAR实际数据为例,验证了模型的有效性。
QAR mass data analysis is an important issue in aviation safety management. To accurately and efficiently mine the flight operation modes contained therein and evaluate its risk level is the key to effectively improve the quality of flight operations. However, during the flight, pilots need to constantly adjust and cooperate with each other according to the dynamic conditions of the environment, the status of the aircraft, and the operation of the devices such as the throttle, the lever, the disk and the rudder. This increases the requirement for the flight operation mode and its Risk analysis of the difficulty. This paper starts from the feature extraction and similarity measurement of multivariate time series data of flight parameters, and uses the clustering to mine the flight operation modes involved. Furthermore, by analyzing the correlation between flight operation modes and QAR monitoring index values, the paper quantifies different flight operation modes The level of risk. Finally, this paper validates the validity of the model by taking the QAR actual data of a specific aircraft at the Capital Airport as an example.