Fault Analysis Of Locomotive Network Control System Based On Bayesian Network

Electric locomotive is a system with very complicated structure,and it is in working state most of the time,which requires high reliability of the components of electric locomotive.However,as its core component,the network control system has functions such as control and monitoring,which plays a decisive role in the efficient operation of the locomotive.Therefore,this thesis takes the network control system of the HXD2 electric locomotive as the research object.The main research contents are as follows :First,Research the shortcomings of fault diagnosis in current network control system,and propose two methods of network control system analysis based on fault tree and Bayesian network.Among them,it mainly introduces the basic theory of Bayesian network method,qualitative analysis and quantitative analysis of fault tree analysis method,fuzzy representation of Delphi method and fault tree is converted to a Bayesian network method,and studies the internal composition structure of HXD2 electric locomotive network control system.The expert experience and the fault record form are used as the support of the knowledge base to sort out the fault data,and specifically divide the data into a fault location,fault type,and fault cause three-level fault system.Then,use the fault tree analysis method and the Bayesian network method to obtain the fault data in the fault tree drawing and analysis software,Python software for modeling,simulation and example verification.Comparing these two experimental results,it is found that the fault tree analysis method is suitable for analyzing simple models,and it is not easy to update complex models.Bayesian network can not only update the complex model through code,but also realize layer-by-layer inference.However,considering the influence of the occurrence of the parent node on its child nodes,the parameters of the parent node in the Bayesian network need to be further optimized.It can be seen that the optimized Bayesian network not only improves the efficiency of fault diagnosis,but also makes the diagnosis of the cause of the fault More convincing.Finally,the Bayesian network method is used to realize the fault diagnosis of the network control system.Use the Py Charm tool in Python to conduct Bayesian network modeling and analysis,and use Visual Studio software to implement a visual interface.The interface mainly includes a login interface,an initialization device management module,a diagnosis casemanagement module,a model generation module,and a maintenance decision module.