Font Size: a A A

Research On The Fault Diagnosis Method Of Turnout

Posted on:2018-11-25Degree:MasterType:Thesis
Country:ChinaCandidate:J YangFull Text:PDF
GTID:2352330512476590Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
Switch,a key part of the railway signal equipment,not only is responsible for switching track line but also ensures operation safety.With the increase of railway lines and traffic density,the failure frequency of turnout equipment has become increasingly high.Therefore,it is of great practical significance to study the fault diagnosis methods to improve the automation level of diagnosis.In this paper,based on summarizing domestic and foreign research,switch fault diagnosis methods are studied from quantitative and qualitative aspects.Firstly,structure of switch equipment is introduced in detail,as well as control principle and operation process.Based on summarization and classification of common switch fault patterns,several faults are focused on in this paper and detailed causes of them are given.Secondly,operation signal collection system is designed and experiments of field data acquisition are conducted.According to the requirement analysis of the system,hardware design and selection are finished and functional software modules are completed on the LabVIEW platform.Detailed introduction of on-site installation and collection process of switch operation signal are given.Then,switch fault diagnosis method based on DWT and C-SVM is proposed from a quantitative angle.In this method,cluster analysis based on K-Means is conducted to determine the type of signal which contains the most fault feature and DWT is used to extract fault feature of the signal.Then the normalized failure feature data is utilized for fault identification in order to realize switch fault diagnosis.The cross validation experiments are performed with the measured fault signals and the results verify the accuracy and effectiveness of the proposed method.Finally,switch fault diagnosis method based on qualitative trend analysis is proposed from a qualitative angle.In this method,the trend of the reference operation signal in the different state of the switch is extracted by using the interval-halving method,and the knowledge base of fault diagnosis is established.Then the trend of signal to be diagnosed is extracted and the matching values of trend sequence and all the fault trend rules is calculated,which are ranked and compared comprehensively to realize switch fault diagnosis.The experimental results show that the presented method has good accuracy and is superior to the proposed quantitative fault diagnosis method in terms of distinguishing single fault.
Keywords/Search Tags:Switch, Fault diagnosis, Wavelet transform, Support vector machine, Qualitative trend analysis
PDF Full Text Request
Related items