Risk Analysis Based On Operational Data Of Air Traffic Control

With rapid development of civil aviation industry of China at present,the number of annual flight has been increasing year by year,and the safety level has a leading position in the world.As the core of the civil aviation system,The analysis of ATC security has always been the focus of civil aviation industry.The identification of risk sources based on unsafe incidents,the construction of evaluation indexes and the risk assessment are the main work of existing civil aviation risk analysis,which are subjective and lack of forward-looking.Each flight will produce a large number of operational data in different types and formats,which not only record the entire process of aircraft missions,but also carry the basic data and decisions made by various departments.Based on the historical operation data,we can discover the problems existing in the actual running process and eliminate the hidden danger based on the problems,which can make the safety work more initiative.Therefore,it is necessary to analyze the elements of civil aviation unsafe incidents based on historical traffic data of ATC and find out the potential value and effective information from historical data.In this paper,a potential conflict detection algorithm based on historical trajectory data and four-dimensional trajectory prediction is proposed by analyzing the process of conflict predicting by ATC.The conflict scenarios are defined to describe the position and state of the aircraft when a potential conflict occurs,and the K-means clustering method is used to classify the conflict scenarios.Based on potential conflict detection,this paper defines the concept of ATC risk scenario,which is the case of frequent potential conflicts in a specific period and in a specific area.According to the definition,the identification process of risk scenario is clarified,and the identification of risk scenario and the spatio-temporal analysis of risk scenario are realized through examples.Based on the identified risk scenarios,a quantitative analysis method of conflict risk based on Monte Carlo simulation is proposed,and the risk of conflicts in the risk scenario is evaluated through examples.In view of the reports of ATC unsafe events and knowledge of ATC,this paper constructs the vocabulary of ATC security,and participle based on the vocabulary is realized by using the participle package called Rwordseg.Word frequency statistics and word cloud drawing are achieve on the basis of word segmentation,and the correlation analysis of the reports of ATC unsafe events based on keywords is implied by using text mining package.In line with the conceptual model and relational model of ATC security ontology established in this paper,we use the ontology construction tool Protege to create the ATC security ontology,and then realize the association analysis based on ontology of ATC security.Compared with the association analysis based on keywords,the results of association analysis based on ATC security ontology are more general.