Research On Construction And Classification Of Urban Rail Vehicle Risk Node Set Based On Multiple Data

As one of the core equipment systems to fulfill the mission of safe and efficient displacement of urban rail system,the vehicle system has the characteristics of complex operation environment,frequent start-up and highly concentrated parts.Once the parts fail,the safe operation of the vehicle may be affected.When studying the operation safety of urban rail vehicles,domestic and foreign scholars focus more on improving the reliability of components.However,urban rail vehicles do not exist independently,human,machines and environment will also affect its state in its operation process.Therefore,starting from the multiple data of urban rail vehicles,this thesis analyzes the interaction and location of components,and explore the impact of other risk factors on components,taking the components that affect the safety of vehicle operation as risk nodes to study the construction of urban rail vehicle risk node set and classification method.The main research contents are as follows:(1)The risk node set construction method of urban rail vehicle system is proposedBased on the extensive investigation of urban rail transit operating enterprises and vehicle design and manufacturing enterprises,this thesis analyzes the composition and working principle of urban rail vehicles,and forms the construction method of urban rail vehicle risk node set based on set theory.The research data related to the vehicle were summarized into topological data,fault data and accident data.Comprehensively analyzed the influence of the vehicle itself and other risk factors,three risk node classification indexes of component structural importance,component functional importance and component criticality based on accident data were designed.(2)The importance evaluation index of urban rail vehicle components based on topology is proposedTaking the components that affect vehicle safety as nodes and the interaction between components as connecting edges,the urban rail vehicle topological network model is constructed.Taking topological network model as the carrier,the reliability of components was obtained from the failure data,the dependency degree among components was quantified,and the function importance index based on Leader Rank component was proposed.Based on topological data,starting from the topological characteristics of the component,considering the importance of the component itself and its neighbors,the idea of "graph energy" is introduced,and on this basis,a component structural importance index based on quasi-Laplacian centrality is proposed.(3)The key index of urban rail vehicle components based on accident data is proposedTo study the influence of other risk factors on urban rail vehicle system,based on the accident data of urban rail vehicles,firstly filter the text data of the original accident by word segmentation and start-stop words,which provides a basis for mining potential accident information more effectively.TF-IDF algorithm is used to excavate the candidate keywords of vehicle system.On the basis of Markov state transfer idea,taking sentence as unit,TF-IDF value of nodes as initial value and edge weight between nodes as transfer matrix,an improved TI-Text Rank algorithm is proposed to obtain key indicators of urban rail vehicle components based on accident data.(4)The risk classification method of urban rail vehicles is proposedAccording to the three indexes of component functional importance,component structural importance and component criticality based on accident data,combined with the clustering idea,the value of each index was taken as the eigenvalue,and the optimal clustering category number and clustering result were obtained by the K-Means++algorithm based on genetic algorithm.The Gini importance of random forest was used to quantify the weight of the three indicators,and the corresponding relationship between clustering results and grading levels was calculated,and finally the risk grading results were obtained.Taking B-type car traction system and bogie system of an urban rail enterprise as an example,the method and model proposed above are applied and verified.The research shows that the method of constructing and classifying risk nodes set of urban rail vehicle system proposed in this thesis is practical and scientific,which opens up a new idea for the risk assessment of urban rail vehicle system.