Research For Optimization Of Troubleshooting In Civil Aircraft Air-condition System Based On Combination Analysis Method

Modern civil aircraft troubleshooting is a process which through the intermittent implementation of the test of diagnosis and maintenance operations.As one of the most important modern civil aircraft system,air-conditioning system’ internal and components exist various uncertainties and there are a variety of observations,maintenance operations,and questions related to troubleshooting during the interactive troubleshooting process.Therefore,the main problem for troubleshooting of civil air-conditioning system is how to get the final cause of failure from the uncertain and multi-source heterogeneous information.Aiming at the uncertainties of air-conditioning system failure and the requirement of interactive fault isolation and diagnosis of civil aircraft,this paper explores a diagnosis and troubleshooting method based on Bayesian network combined with multi-step observation model,and then to achieve rapid diagnosis and intelligent troubleshooting for the civil air-conditioning system.Though the in-depth analysis of actual troubleshooting process,the physical and mathematical models of air-conditioning system diagnosis and troubleshooting are established,which leads to the associated failture as the core factor that leads to the difficulty of diagnosis,and gives the knowledge expression which can be used in the diagnosis and troubleshooting of air-conditioning system with Bayesian network.On the basis of the analysis of diagnosis and troubleshooting model,aiming at the characteristics of air conditioning system failture,an expert medeling method based on the multi-level stratification is proposed.A conditional probability solving algorithm based on LeakyNoisyOr fuzzy evalution model,which combined with Bayesian network LeakyNoisyOr model,fuzzy mathematic and analytic hierarchy process,is proposed for the objective factors of lack of fault sample data in air-conditioning system.Based on the correlation analysis of the target symptom of the A320 fleet air conditioning system fault information table,the multi-level stratified expert modeling method is used to realize the expert modeling of the target fault.LeakyNoisyOr fuzzy evaluation model is used to solve the conditional probability.A method of extracting air conditioningsystem component data using ACMS is given and the model is verified using the acquired data.The diagnosis model is transformed into the troubleshooting model and combined with multi-step observation model,used to integrate observations,costs,maintenance operations and other molti-source information.The effectiveness of the method in the rapid diagnosis and intelligent troubleshooting of civil air-conditioning system is verified by simulation.C#and matlab mixed programming is used to design the diagnosis and troubleshooting system.