Research On Risk Assessment Technology Of Aeroengine Control System For Continued Airworthiness

Aeroengine control system is the core subsystem of engine,which has high failure rate and great safety impact.Although the aeroengine control system has met the airworthiness requirements and allowed to be put into use at the stage of type certification,the actual risk in operation is always higher than the expected risk due to uncertain factors such as undetected changes in airworthiness standard formulation and standard compliance,design and manufacture deficiencies,and unanticipated operating conditions or environmental conditions.Although the actual risk of aeroengine control system is often unpredictable,there are some symptoms before unacceptable accident consequences occur.By means of risk management,the actual risk level can be greatly reduced and the continuous operation safety of aircraft can be guaranteed.Based on this,this paper takes the aeroengine control system as the research object,and carries out the research on the risk management process,hazard analysis,risk identification and risk analysis and evaluation of the aeroengine control system.Starting from the study of risk management process of aeroengine control system,this paper combines the life-cycle safety management theory of aeronautical products with the characteristics of aeroengine control system and identifies three elements of risk management process of aeroengine control system,which are hazard analysis based on fault tree model,risk identification based on event information and machine learning,and risk analysis and assessment considering the situation of the fleet.Based on this,the risk management process of the aeroengine control system is determined.According to the process of risk management,the hazard sources of aeroengine control system are analyzed at first.A fault tree with LOTC as the top event is established to analyze the basic hazards that may cause the top failure of the aeroengine control system.The minimum cut sets of the fault tree and the probabilistic importance of each underlying event are further calculated.Minimum cut sets and probabilistic importance are the inputs of risk identification and analysis and evaluation process,which lay the foundation for the follow-up risk management process.Secondly,the risk identification of aeroengine control system based on event information is studied with machine learning methods.According to the characteristics of the unsafe events information of aeroengine control system,the evaluation index and risk categories are determined.The risk identification model based on Support Vector Machine(SVM)is established by selecting analysis samples from Service Difficulty Report(SDR)System,maintenance records and Aviation Safety Reporting System(ASRS).The average impact value(MIV)algorithm is used to choose risk features,and particle swarm optimization(PSO)algorithm is used to optimize the parameters of the SVM model.The recognition results show that the risk identification method based on PSO-MIV-SVM algorithm has higher recognition efficiency and accuracy.Finally,the identified potential unsafe issues are analyzed and assessed.Taking LOTC as an intermediate event,the Bow-tie model of aeroengine control system is established.On the basis of fault tree analysis and event chain analysis,the individual risk assessment model and fleet risk assessment model are established.In order to verify the validity of the risk analysis model,a case study and model validation of the adjustable stator blade(VSV)actuator of the CFM56-7B engine control system is carried out.