Research On Fault Diagnosis Of Railway Turnout Based On Fuzzy Cognitive Map

With the rapid development of China’s economy and transportation,people put forward higher level requirements for railway traffic safety.Turnout is the key basic equipment of railway signal system.The action state of turnout is closely related to the safe operation and transportation efficiency of train.The actual reasons for the failure of turnout and turnout conversion equipment are very complex.At present,the operation status information of turnout action is mainly monitored by the computer monitoring system,and the initial turnout failure judgment is made by the computer monitoring staff in China.The maintenance staff will further locate the turnout failure according to the site conditions during the skylight time.However,this method can’t diagnose the fault on-line in real time,and can’t locate the fault location accurately when the fault occurs,which requires the maintenance personnel to have rich working experience.In view of the above problems,this paper applies the theory of fuzzy cognitive map to the fault diagnosis of railway turnout,the main research contents are as follows:Firstly,the research status of railway turnout fault diagnosis at home and abroad is analyzed,this paper takes the most widely used ZD6 turnout system as the research object,combines with the structural composition and operation principle of turnout and its switch equipment,summarizes seven typical turnout fault modes,and analyzes its corresponding curve trend and fault causes.Secondly,since the action state of turnout will be directly reflected on the action current curve of turnout,a method of extracting energy features by wavelet packet decomposition is proposed,and three methods based on time fixed segmentation,turnout conversion state and wavelet packet energy decomposition are used to extract the action current feature vector of turnout,which are conducive to reducing the feature input dimension and improving the performance and operation efficiency of diagnosis model.Then,this paper summarizes the development of fuzzy cognitive map theory at home and abroad,describes the reasoning mechanism of fuzzy cognitive map applied to classification problems in detail,and builds a turnout fault diagnosis classification model based on fuzzy cognitive map.Genetic algorithm and particle swarm optimization algorithm are used to learn the weights respectively,so as to obtain the initial weight matrix of fuzzy cognitive map.Each node of fuzzy cognitive map is updated iteratively through the reasoning mechanism,and the system finally reaches a stable state.At this time,the sample category can be determined by the subscript of the maximum category node value.Through obtaining and sorting out the sample data of turnout fault action current curve,the simulation experiment is carried out under the MATLAB environment.The results show that it can achieve good results in multi-classifiers evaluation indexes of accuracy,error rate,macro weighted F1,micro F1 and algorithm running time.The results prove the feasibility and reliability of fuzzy cognitive map theory applied to the turnout fault diagnosis and classification problems,and the test results are basically consistent with the field maintenance results.Finally,based on the established turnout fault diagnosis classification model of fuzzy cognitive map,with the help of laboratory simulation test environment,visual studio programming is used to achieve the framework design of the whole system software,including network communication design,sample management,model management,fault diagnosis results and visual interface.It is verified by many groups of test samples,and the diagnosis results show that the software can effectively and quickly realize the fault diagnosis of ZD6 turnout.There are 58 figures,23 tables and 63 references.