基于贝叶斯网络的民航安全分析研究

安全在民用航空领域占有举足轻重的地位,直接影响民航企业的生存和发展。科学合理的安全分析便于及时发现民航领域存在的安全问题,预防航空事故诱发因素的萌生,提高民航安全管理的科学性和可操作性。根据民航安全的特点,建立了基于故障树的多态贝叶斯网络系统安全分析模型,综合运用诊断推理和因果推理形式,实现了定性与定量相结合的民航安全分析,通过航空灾难的实例验证了方法的有效性。在此基础上,提出采用区间型贝叶斯网络进行民航安全分析,克服了其他方法只能分析计算单点概率值的局限性,比传统的安全评估更客观,更有说服力。

Study on Bayesian Networks based Aviation Safety Analysis

Security plays significant role in aviation field which directly influences the survival and development of the aviation enterprises.Appropriate scientific security analysis may help find existed security problems in time and prevent initiation factors happening which may induce serious aviation accidents.By utilizing the analysis method it can enhance the scientific and maneuverability nature of the aviation security management.A FT-based multi-modal BN security analysis system is presented in this paper based on aviation security characteristics.A comprehensive application of diagnostic reasoning and causal reasoning is used to achieve a combination of qualitative and quantitative analysis of aviation security.A real case of aviation accident has been investigated to the method validity.On this basis,an interval Bayesian networks analysis model of aviation security is used to overcome the disadvantages of the limitation of the single-node probability calculation.Compared with the traditional security assessment system,a more objective and more persuasive one is presented.