航空学报 > 2025, Vol. 46 Issue (11): 531252-531252   doi: 10.7527/S1000-6893.2025.31252

面向适航要求的eVTOL航电系统安全调度模型

赵长啸1,2,3(), 孙亦轩1   

  1. 1.中国民航大学 安全科学与工程学院,天津 300300
    2.民航航空器适航审定技术重点实验室,天津 300300
    3.天津市航空装备安全性与适航技术创新中心,天津 300300
  • 收稿日期:2024-09-23 修回日期:2024-10-24 接受日期:2025-01-22 出版日期:2025-02-19 发布日期:2025-02-18
  • 通讯作者: 赵长啸 E-mail:cxzhao@cauc.edu.cn
  • 基金资助:
    国家自然科学基金(52402523);国家自然科学基金(U2133203);天津市高等学校研究生教育改革研究计划(TJYG135);天津市航空装备安全性与适航技术创新中心开放基金(JCZX-2024-KF-01)

A safe scheduling model for eVTOL avionics systems for airworthiness requirements

Changxiao ZHAO1,2,3(), Yixuan SUN1   

  1. 1.College of Safety Science and Engineering,Civil Aviation University of China,Tianjin 300300,China
    2.Key Laboratory of Civil Aircraft Airworthiness Technology,CAAC,Tianjin 300300,China
    3.Tianjin Aviation Equipment Safety and Airworthiness Technology Innovation Center,Tianjin 300300,China
  • Received:2024-09-23 Revised:2024-10-24 Accepted:2025-01-22 Online:2025-02-19 Published:2025-02-18
  • Contact: Changxiao ZHAO E-mail:cxzhao@cauc.edu.cn

摘要:

针对现有以性能设计为核心导向的电动垂直起降飞行器(eVTOL)航电系统易产生风险汇集,影响飞行安全的问题,提出了基于风险量化方案的eVTOL航电系统安全调度模型。建立了时间约束和资源约束下的eVTOL航电系统双层调度方案;通过分析不同局方eVTOL的适航规章,提出了基于专家打分和熵权法的航电功能安全关键度描述方案;使用深度确定性策略梯度(DDPG)算法对安全关键度均衡问题进行求解。仿真实验设置负载均衡组与安全均衡组,进行优化前后纵向对比、组间横向对比,实验结果表明,提出的eVTOL航电系统安全调度模型可提升分区间安全关键度均衡程度30.7%,同时风险汇集偏离系数降低10.8%;对比负载均衡组,在损失3.6%负载均衡的情况下获得了9.4%的分区间安全关键度均衡。该模型为eVTOL航电系统的进一步综合与大规模应用提供了理论支撑。

关键词: eVTOL航电, 安全调度模型, 适航规章, 强化学习, DDPG

Abstract:

The safety scheduling model of electric Vertical Take-Off and Landing (eVTOL) avionics system based on risk quantification scheme is proposed to solve the problem of risk aggregation in the existing performance-driven design of eVTOL aircraft avionics systems, which can impact flight safety. A two-layer scheduling scheme of eVTOL avionics system under time constraint and resource constraint is established. By analyzing the airworthiness regulationsof different eVTOL authorities, a description scheme of avionics functional safety critical degree based on expert score and entropy weight method is proposed. The Deep Deterministic Policy Gradient (DDPG) algorithm is used to solve the security criticality equilibrium problem. In the simulation experiment, the load balancing group and safety balancing group were set up, and the longitudinal comparison before and after optimization and the horizontal comparison between groups were carried out. The experimental results show that the proposed eVTOL safety scheduling model can improve the safety critical balance degree between regions by 30. 7%, and reduce the risk pooling deviation coefficient by 10. 8%. Compared with the load balancing group,9. 4% inter-zone safety criticality balance was obtained with 3. 6% load balancing loss. This model provides theoretical support for further integration and large-scale application of eVTOL avionics system.

Key words: eVTOL avionics, safety scheduling model, airworthiness regulations, reinforcement learning, DDPG

中图分类号: