基于BP神经网络和飞行参数的航段风险评价

摘要

飞行运行风险的定量评估对于民航安全保障具有重要意义.设计并实现了一种基于BP神经网络算法的航段风险评价系统.该系统将大量飞行情景参数直接作为飞行风险评估的基础,通过飞行能力评价体系获得飞行运行风险度经验数据,并使用神经网络实现了任意飞行情景的风险度拟合运算.经过真实航班飞行数据的测试和飞行品质监控系统的验证表明,该系统可为进近着陆阶段任意时刻的飞行运行风险评估提供可靠的结果.%Quantitative risk assessment of flight operation is significant to flight safety guarantee. A risk assessment system was developed based on BP neural network. This system contains a flight capability index system to evaluate the risk of flight scenario, and a BP neural network model to simulate the risk value. Massive scenarios were used to train the model. Test results of real flight data during approach and landing was confirmed by FOQA program,and indicated that the proposed system can provide reliable flight operation risk assessment value of any moment.