Research On Fault Diagnosis And Maintenance Decision-making Method For Avionics Integrated Test

As the avionics system tends to be more and more integrated and modularized,each subsystem is no longer working separately.There is a high degree of information interaction between the systems,and there are complex forms such as multiple faults and associated faults,which brings challenges to the fault diagnosis of avionics system.In order to shorten the development cycle of avionics system,avionics integrated testing technology is usually used to test the functions and performance indexes of each subsystem to find out the defects of system design in time.Based on the demand of airlines for faster and more accurate diagnosis of airline maintenance,this paper takes the avionics integrated test system as the platform and aims to realize rapid reasoning and dynamic diagnosis of avionics system faults.The method of fault diagnosis and maintenance decision of avionics system based on Bayesian network combined with association rules is proposed.Firstly,the method of constructing the diagnosis model of avionics system based on Bayesian network is proposed.The fault tree is established by analyzing the historical maintenance data and the fault isolation manual(TSM),etc.According to the conversion rules,the fault tree is transformed into the avionics system diagnosis model structure based on the Bayesian network,and BITE information is extended into the diagnosis model as the nodes.With the Bayesian network as the framework,the qualitative representation of complex forms such as multiple faults and associated faults of avionics system is realized,so as to ensure that the diagnostic model is easy to update,and the structure and parameters of the diagnostic model are optimized continuously with the accumulation of historical data.Secondly,an implementation method of avionics system diagnosis model parameter learning is proposed.For avionics systems,it is difficult to obtain complete failure sample data,and the authenticity of simulation data is difficult to guarantee.Based on the incomplete historical maintenance data,first use the association rule Apriori algorithm to mine the strong association rules contained in the historical maintenance data,and then integrate the expert experience to complete the learning of conditional probability parameters.With the increase of historical maintenance data,it is ensured that the parameters are more inclined to the actual failure frequency of the specific aircraft,and the dynamic learning of the parameters is realized.Finally,the simulation of fault diagnosis in the avionics integrated test platform is verified.A fault diagnosis and maintenance decision system is designed and applied in the avionics integrated test platform.The maintenance information transmitted by ARINC664 bus is obtained automatically.Meanwhile,reasoning analysis is conducted with other observation evidences to verify the effectiveness of the fault diagnosis and maintenance decision method.The simulation result shows that the fault diagnosis and maintenance decision method proposed in this paper can effectively diagnose the fault information in the avionics system,realize the dynamic diagnosis process,simplify and optimize the maintenance troubleshooting work,and improve the maintenance efficiency.It has a certain theoretical value for the real airline maintenance.