航空企业基于SHEL模型的神经网络安全评价研究

从风险管理的角度,针对航空工业企业安全系统的构成因素,对企业安全系统进行分析和深入研究。基于SHEL模型,通过引入人工神经网络评估方法,选择前馈型BP网络模型,建立网路结构,将企业事故风险系统划分人-硬件、人-软件、人-环境、人-人4个子系统和事故综合评估总系统,辅以DPS软件,以某航空企业为例,进行训练和测试,并分别进行拟合量化分析,得出量化的评估结果及该模型的适用性、可行性和有效性。为航空企业管理者切实改进和完善安全管理提供理论和技术支持。

Neural Network Safety Assessment Method for Aviation Enterprises Based on SHEL Model

According to the constituent elements of enterprises safety system,a deep analysis is made on these elements from risk management view.Based on SHEL (software-hardware-environment-livewire) model and artificial neural network assessment method,a feed-forward BP network model is constructed.In this model,the enterprise accident systems are divided into a total accident integrated assessment system and four subsystems including human-hardware subsystem,human-software subsystem,human-environment subsystem,and human-human subsystem.With the data from an aviation enterprise as an example,the network model is trained and tested by using USP software.Fitting quantitative analysis and assessment result show that the model is applicable,feasible and effective,and can provide a theoretical basis and decision support for enterprise managers to improve and perfect their safety management.