Research On Fault Propagation Mechanism Of A320 Air Turbine Starter

Air Turbine Starter(ATS)is an important component of civil aircraft starting system,which provides auxiliary power for ground and air starting of the engine and ensures the safety of engine starting.It is of great significance to study the fault propagation mechanism of ATS for making scientific maintenance plan,improving troubleshooting efficiency and reducing maintenance cost.The main research work includes following aspects:Firstly,the basic structure of ATS and analyzation of propagation modes of typical fault in ATS.In this paper,the coupling relationship between the internal structure and components of GTCP131-9A/B air turbine starter for A320 are deeply studied.The physical structure of the starter is analyzed,and the typical fault propagation mode of the starter is studied.The analysis results of fault single step propagation mode and propagation path mode are summarized,which provides a theoretical basis for the subsequent research of fault propagation characteristics in ATS.Secondly,the fault propagation model of ATS based on Small World Network(SWN)is established.Based on the physical structure of starter,the topology network model of starter is established by SWN,and the small world characteristics of model are verified.The Ant Colony Optimization(ACO)algorithm is introduced to solve and analyze the high-risk propagation path in ATS by taking the fault propagation strength between nodes as the heuristic information parameter of the algorithm and setting the global propagation threshold of fault.Furthermore,the fault propagation model of starter based on Bayesian Network(BN)is established.According to the functional structure of the starter,the Most Weight Supported Tree(MWST)is used to describe the coupling relationship of components in the starter,and then it is transformed into BN structure.The Dual Strategy Ant Colony Optimization(DS-ACO)algorithm is proposed to optimize the BN structure,and then multiple candidate BN structures are evaluated.Finally,the Bayesian network model based on the starter is obtained.Through probability reasoning,the occurrence probability of each fault propagation path is calculated to solve the high-risk path in ATS.Finally,the performance of the two proposed fault propagation mechanism modeling methods is compared and verified.The experimental results show that the path learning speed of small world network framework is excellent,and the structural accuracy of Bayesian network framework is more prominent.The experimental results have reference value for the formulation of safety emergency maintenance schemes in different maintenance environments.