Cause Analysis Of Flight Areas Unsafe Events Based On Spatio-temporal Mining

With the rapid development of the civil aviation,various unsafe events in the flight area occur frequently.At present,the control of unsafe events is mostly conducted after-the-fact supervision and qualitative risk source investigation,and cannot be controlled from the source of unsafe events.Therefore,quantitative analysis of unsafe events is urgently needed to improve the safety management level of the flight areas.Based on the idea of space-time mining,this paper studies the causes of unsafe events in the flight areas.Firstly,the flight areas of the airport and its common unsafe events are introduced.The macro and micro causes of unsafe events are analyzed.On this basis,the relationship between the macro-causes analysis method of unsafe events based on time mining and the micro-cause analysis method of unsafe events based on time-space mining is proposed.Then,a VAR model is established with unsafe events as the dependent variable and macro cause as the independent variables.The results show that the three indicators of consumption level of resident,civil aircraft number and civil aviation employment have significant impact on the number of unsafe incidents.Then the VECM model was established and the Granger causality test was used to determine the short-term factors: the number of unsafe events in the previous period and the passenger turnover.Furthermore,by means of impulse response analysis and variance decomposition,it is found that the indicators that contribute more to the dynamic changes of unsafe events is the number of civil aircrafts.Finally,this paper chooses bird strike as the typical case of micro analysis of unsafe events.The DENCLUE clustering algorithm is used to obtain the hotspot area of the bird strike risk,and the spatial distribution characteristics are analyzed.A GTWR model is proposed to eliminate the influence of variable space-time non-stationarity.At last,based on the spatio-temporal variation of the variable coefficients,the following conclusions are drawn: Population density,water source distribution and grass species distribution were positively correlated with bird strike risk;vegetation distribution,bird repellent equipment distribution and patrol route were negatively correlated with bird strike risk;population density was delayed;water source distribution was immediacy.This paper analyzes the causes of unsafe events in the flight areas,and can quantitatively analyze the deep relationship between unsafe events and causes,make up for the shortcomings of existing qualitative methods,improve the level of safety control in flight areas.