基于BP神经网络模型的飞机起降冲偏出跑道预测研究

The Prediction of Runway Overrun and Excursion Event Based on BP Neural Network

飞机冲偏出跑道是飞机起降阶段高发事件之一,为了促进跑道安全,就飞机起降阶段冲偏出跑道事故进行预测研究.建立了飞机起降冲偏出跑道的BP神经网络模型,并根据2007-2016年与飞机冲偏出跑道有关的运行数据,运用BP神经网络多层感知器对2007-2020年飞机冲偏出跑道的累计次数和事故征候次数进行预测,结果显示:到2020年飞机冲偏出跑道累计次数将达到95次,事故征候次数将达到2861次,未来几年冲偏出跑道累计次数和事故征候累计次数均不断增加,但增幅有所下降.

One of the high-risk events during the take-off and landing phase is the aircraft’s overrun and excursion out of the runway.In order to promote the safety of the runway,the predictive study is conducted on the accident of overrun and excursion the runway during the takeoff and landing of the aircraft.A BP neural network model for aircraft overrun and excursion runway during taking off and landing off is established,the article also uses the BP neural network multi-layer perceptron to predict the number of cumulative drifting out of the runway accidents and the number of incidents during 2007-2020,based on the operational data related to the aircraft’s drift overrun from 2007 to 2016.The results show that:By2020,the cumulative number of aircraft overrun and excursion the runway will reach 95,and the number of incidents will reach 2,861.The number of accumulated deviations from the runway and the cumulative number of accidents will both increase in the next few years,but the growth rate has declined.