Research On Human Factors Reliability Of Civil Aviation System Based On Deep Learning

With the rapid development of the civil aviation transportation industry,the amount of air traffic and the number of shifts have increased significantly.Human factors affect the safety of civil aviation,the safety of aircraft and operating costs.This paper establishes a framework system for human risk analysis and assessment of civil aviation system.Civil aviation maintenance system is taken as the research object to analyze and evaluate human risk,which can identify the weak links in operation,and formulate corresponding management policies.Therefore,studying the human factors in civil aviation maintenance has important practical significance for the civil aviation industry.The full text of the research includes:Firstly,from the basic concept of human factors,the advantages and disadvantages of each generation of HRA models are analyzed.The necessity and importance of deep learning in civil aviation maintenance system are proposed,and the human factor reliability framework for civil aviation system is established.Secondly,the concept of functional resonance network is used to analyze the human factors of civil aviation maintenance system,determine the functional units that may resonate and formulate relevant protective measures.The CREAM method is used to predict and analyze the failure probability of cognitive activities of accident instances.Thirdly,in the civil aviation maintenance system human risk analysis by deep learning,the unsafe event data caused by China’s civil aviation maintenance errors is collected as the risk analysis sample data,and then the risk level map is drawn.Considering the number of samples needed for deep learning,the Bootstrap method is used to expand the sample size.The risk of human factors between unsafe events and event rates is analyzed by deep learning.Then,the types of events with higher degree of influence are obtained,the mapping of human factors with higher impact types is constructed,and countermeasures are proposed to help quantify maintenance errors and prompt the system to recognize and identify human factors in maintenance errors.Finally,in the civil aviation maintenance system human risk assessment by deep learning,the number of civil aviation traffic in China and the number of maintenance accidents are counted to obtain the incident rate.And the incident rate is used as the object of human risk assessment.The deep BP neural network,LSTM method and ARIMA are used to evaluate and predict the human risk probability,and provide decision support for the human risk assessment of the maintenance system.