Research On Fault Diagnosis Of Vehicle On-board Signal Equipment Based On Knowledge Graph

Vehicle on-board equipment(VOBE)of high speed railway is an important part of Chinese Train Control System,its structure is complex,and the internal modules are closely connected.During operation of the train in practice,VOBE in case of failure,the fault phenomenon and the cause of the fault presents complex and diverse,and directly affects the safety and efficiency of train operation.Therefore,researching intelligent methods to discover the internal relationship between the fault phenomenon and the cause of the fault will help to improve the efficiency and accuracy of fault diagnosis,shorten the maintenance work time,and restore the train to normal in time.Fault maintenance log of VOBE is a detailed text about the failure of VOBE recorded by the electricians during daily maintenance work.These text messages contain a wealth of fault knowledge,but due to the limitations of recording methods and storage formats,it is difficult to mine effective information.In view of the characteristics of the fault maintenance log and the needs of intelligent fault diagnosis of VOBE,this thesis uses text mining technology to process the fault maintenance logs,build a fault knowledge map,excavate the internal connection of the faults of VOBE,and improve the efficiency of fault diagnosis.The specific research content is as follows:(1)In view of the characteristics of the fault maintenance log,this thesis combines the word length characteristics of words and the weighted information entropy characteristics of words to improve the Text Rank algorithm to extract keywords in the fault text data.Experiments show that the improved algorithm has higher accuracy,recall and F-mean than Text Rank algorithm and other keyword extraction algorithms.(2)Based on keyword extraction,this thesis proposes a method for extracting fault phrases in fault text.First extract the keywords phrase with the extracted keywords;then use text similarity to cluster the keywords phrase to determine the target fault phrase;and finally use semantic similarity to extract the fault phenomenon phrase and fault cause phrase in the fault text data.(3)Use a top-down approach to build a fault knowledge graph.First,determine the fault ontology library by the structure of the fault maintenance log and establish the data pattern diagram;then,based on the extracted fault phrases,build a fault knowledge map;and finally use the fault knowledge map to show the internal connection of the fault.(4)Using fault knowledge graph to design fault diagnosis method of VOBE.Fault diagnosis results include: top-level fault cause,faulty device location and fault root cause,and visualize the fault diagnosis results.Finally,the experiment proves that the fault diagnosis method based on the fault knowledge graph can provide help for the actual diagnosis.