Research On Test And Diagnosis Technology Of Aircraft System Based On System Principle And Multi-signal Flow

The aircraft system is a large-scale complex system,and its faults are characterized by multiple and dynamic characteristics.On the one hand,the increasingly complex structures and functions cause aircraft testability to fall,which results in the high costs for aircraft diagnosis and maintenance.On the other hand,higher requirements are put forward on the capability of aircraft system for fault diagnosis and prediction due to the comprehensive consideration of aircraft reliability and economy.In this paper,the test and diagnosis method based on the system principle and multi-signal flow is studied in the aircraft system.The main contents of this paper are as follows:(1)According to the requirements of aircraft system fault prediction and the existing limitations of test model,this paper proposes a new method of test-diagnosis modeling based on the system principle and multi-signal flow model.The information of the fault,test and diagnostic prediction in the real system are added to the test diagnostic model,which not only enables the test-diagnostic model to evaluate and optimize the diagnostic prediction ability of the system,but also its model information can better support failure prediction.(2)According to the characteristics of aircraft system test and diagnosis,an optimal method for diagnostic strategy and a dependency matrix based on multi-valued test is presented.A rollout algorithm is used to improve the fault diagnosis strategy generation algorithm based on information entropy in order to eliminate the existing problems of greedy algorithm and global optimization algorithm.The balance between computational complexity and result accuracy is achieved.(3)The method of fault state recognition and prediction based on test-diagnosis model is given.The improved Particle Swarm Optimization(PSO)algorithm is used to optimize the training process of Hidden Markov Markov Model,which improves the convergence ability in training process and the problem that hidden semi-Markov training can easily fall into the local extreme.(4)A software system for test-diagnosis model was developed based on the research of the above theoretical methods.The engine bleed air system is used as an example to verify the feasibility and effectiveness of this method.