基于多维极值参数的飞行风险量化评估方法

发布时间：2018-06-17 09:40

本文选题：多维极值参数 + 广义极值分布　；参考：《系统工程与电子技术》2015年01期

【摘要】：结合极值理论与多因素耦合系统建模仿真思路,提出了一种基于Copula的多维极值风险评估方法。针对飞行过程中的复杂随机性,基于蒙特卡罗法提取所需要的三维极值参数,验证了所提取极值参数具有和试验数据相同的分布形式,并构建了飞行风险发生的判定条件。在对一维极值参数符合广义极值分布的假设进行证明的基础上,提出了三维极值参数的四参数变权重(four adaptive weight parameters,FAWP),Copula模型利用自适应粒子群算法对一维和三维目标函数中的未知参数进行了辨识,对多种Copula辨识出的三维极值分布进行了拟合优度检验,结果表明FAWP Copula对三维极值参数分布形式的描述最为精确。利用FAWP Copula模型对尾流遭遇情形下的飞行风险概率进行了量化计算,所得指标可用来研究尾流场内的风险规避策略及算法。
[Abstract]:Combining extremum theory with multi-factor coupling system modeling and simulation, a multi-dimensional extreme risk assessment method based on Copula is proposed. In view of the complex randomness in flight process, the 3D extreme parameters are extracted based on Monte Carlo method, which verify that the extracted extremum parameters have the same distribution form as the experimental data, and construct the decision conditions of flight risk occurrence. On the basis of proving the hypothesis that the one-dimensional extremum parameter accords with the generalized extreme value distribution, An adaptive particle swarm optimization algorithm is proposed to identify unknown parameters in one-dimensional and three-dimensional objective functions by using the four-parameter variable weight four adaptive weight parameters Fawp Copula model of three dimensional extremum parameters. The results show that FAWP Copula is the most accurate way to describe the distribution form of 3D extremum parameter. Using FAWP Copula model, the flight risk probability in wake encounter is calculated quantitatively, and the obtained index can be used to study the risk aversion strategy and algorithm in wake field.
【作者单位】：空军工程大学航空航天工程学院;空军北京航空装备训练基地;
【基金】：国家自然科学基金(U1333131,61374145)资助课题
【分类号】：V328

【参考文献】