Urban rail transit vehicle system key component identification method and system

Abstract

The application provides a kind of urban rail transit vehicle system key component identification method and system, belongs to urban rail transit operation maintenance technical field, constructs city rail vehicle topology network model;Determine the function importance index of component;Determine the component structure importance index based on the pseudo Laplace centrality;According to the city rail vehicle accident data, establish the city rail vehicle system accident professional library;On the basis of Markov state transition thought, the component of city rail vehicle system accident professional library is regarded as key word, and the key component index of city rail vehicle component based on accident data is determined;Based on function importance index, structure importance index and key index, the key component of city rail vehicle system is identified.The application fully considers the influence of different risk factors on components during the operation of city rail vehicle, and identifies the key components of city rail vehicle system from the aspects of vehicle topology analysis and accident text, providing reliable theoretical method support for improving the risk identification of city rail vehicle system.

Description

Technical Field

[0001] This invention relates to the field of urban rail transit operation and maintenance technology, specifically to a method and system for identifying key components of urban rail transit vehicle systems. Background Technology

[0002] The vehicle system of an urban rail transit system is the main carrier for fulfilling transportation functions and carrying passengers, and is a key component of the urban rail operation system. Therefore, the vehicle system is a key focus in ensuring the safety of urban rail operation. Identifying the key components in the urban rail vehicle system is fundamental to implementing the process of ensuring the safety of vehicle system operation. The factors affecting the status of components in the urban rail vehicle system are generally divided into two categories: one is internal factors, namely the reliability of the component itself, the degree of dependence between components, and the structural position of the component, which are usually determined by the topology data of the vehicle system; the other is external factors, which are a comprehensive reflection of factors such as the operating environment of the vehicle and the operating line.

[0003] The importance of components in a vehicle system is related not only to internal factors but also to external factors. Existing critical component assessment methods typically consider only one or the other, resulting in inaccurate and unreliable assessment structures. This paper identifies critical components in urban rail vehicle systems from both internal and external perspectives, providing theoretical guidance for ensuring the safe operation of urban rail vehicles. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for identifying key components of urban rail transit vehicle systems, so as to solve at least one of the technical problems existing in the background art.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] On one hand, the present invention provides a method for identifying key components of an urban rail transit vehicle system, comprising:

[0007] A topology network model of urban rail vehicles is constructed using the components in the vehicle as nodes and the interactions between the components as connecting edges.

[0008] Based on the failure data, fit the reliability of the components, quantify the degree of dependence between components, and determine the functional importance index of the components;

[0009] Based on the urban rail vehicle topology network model, the component structural importance index based on quasi-Laplace centrality is determined;

[0010] Based on urban rail vehicle accident data, establish a specialized terminology database for urban rail vehicle system accidents;

[0011] Based on the Markov state transition concept, the components of the urban rail vehicle system accident terminology database are regarded as keywords to determine the key indicators of urban rail vehicle components based on accident data.

[0012] Based on functional importance indicators, structural importance indicators, and criticality indicators, key components of urban rail vehicle systems are identified.

[0013] Preferably, determining the functional importance index of a component includes: fitting a failure probability function based on the failure data of each component, finding the distribution with the smallest fitting result as the failure probability density function of this component; obtaining the parameter values ​​in the Copula function based on maximum likelihood estimation, and calculating the functional dependence between components; then calculating the LR value of each node based on the network adjacency matrix and the LeaderRank algorithm formula; finally, based on the LeaderRank algorithm, introducing a component functional importance evaluation index that comprehensively considers component functional dependence and component reliability.

[0014] Preferably, the determination of structural importance index includes: introducing quasi-Laplacian centrality to measure the structural importance of a node, considering the importance of the node itself and the importance of its neighboring components; the quasi-Laplacian centrality of a node is measured by the change in quasi-Laplacian energy caused by removing it, and the quasi-Laplacian centrality of a node is related not only to its own degree, but also to the degree of other nodes connected to it, and the structural importance of the component is obtained based on the change in quasi-Laplacian energy caused by removing the node.

[0015] Preferably, a professional thesaurus for urban rail vehicle system accidents is established, including: integrating and processing accident documents, selecting the precise mode of the Jieba word segmentation tool when segmenting text, customizing a commonly used dictionary for urban rail vehicle system accidents, and constructing a dictionary of accident causes for urban rail vehicle systems based on components and the devices to which they belong by extracting components appearing in accident reports and combining this with a review of vehicle system components.

[0016] Preferably, based on the Markov state transition concept, the components of the urban rail vehicle system accident professional terminology database are regarded as keywords, including: the components of the urban rail vehicle system accident professional terminology database are regarded as keywords, that is, the nodes in the constructed urban rail vehicle system accident text word graph model; the Markov state transition model is combined with the word graph, and each sentence is used as a window. When two words exist in the same sentence, the edge connecting the corresponding nodes in the word graph is meaningful, and the meaning of the edge is the number of times the words co-occur.

[0017] Preferably, key indicators for urban rail vehicle components based on accident data are determined, including:

[0018] The accident text of the urban rail vehicle system is segmented into sentences to obtain a set of sentences, and then jieba word segmentation and stop word removal are performed.

[0019] Number the sentences as units, and store the sentences after removing duplicate words;

[0020] The components of the urban rail vehicle system are used as keywords to determine the keyword set. The initial value of the keyword is the value of word frequency minus reverse document frequency. Here, word frequency represents the frequency of a word in the text, and reverse document frequency represents the logarithm of the ratio of the number of texts in the text set to the number of texts in which the word appears.

[0021] Iterate through each sentence. If a keyword exists in a sentence, store the sentence number in the keyword position table.

[0022] Calculate the probability that two keywords exist in the same sentence, and use the conditional probability as the node transition matrix, that is, the weight of the node edge;

[0023] The importance of keywords is determined iteratively based on the weights of the nodes, which is the importance of urban rail vehicle system components based on accident data.

[0024] Secondly, the present invention provides a key component identification system for urban rail transit vehicle systems, comprising:

[0025] The first construction module is used to build a topology network model of urban rail vehicles, with the components in the vehicle as nodes and the interaction relationships between the components as connecting edges.

[0026] The first calculation module is used to fit the reliability of components based on fault data, quantify the degree of dependence between components, and determine the functional importance index of components.

[0027] The second calculation module is used to determine the component structural importance index based on the quasi-Laplace centrality of the urban rail vehicle topology network model.

[0028] The second construction module is used to establish a professional terminology database for urban rail vehicle system accidents based on urban rail vehicle accident data.

[0029] The third calculation module is used to determine the key indicators of urban rail vehicle components based on accident data by taking the components in the professional terminology library of urban rail vehicle system accidents as keywords, based on the Markov state transition idea.

[0030] The identification module is used to identify key components of the urban rail vehicle system based on functional importance indicators, structural importance indicators, and criticality indicators.

[0031] Thirdly, the present invention provides a non-transitory computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the method for identifying key components of an urban rail transit vehicle system as described above.

[0032] Fourthly, the present invention provides a computer program product, including a computer program that, when run on one or more processors, is used to implement the method for identifying key components of an urban rail transit vehicle system as described above.

[0033] Fifthly, the present invention provides an electronic device, comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the method for identifying key components of urban rail transit vehicle systems as described above.

[0034] The beneficial effects of this invention are: it fully considers the impact of different risk factors on components during the operation of urban rail vehicles, and identifies key components of urban rail vehicle systems from both the perspectives of vehicle topology analysis and accident text analysis, providing theoretical and methodological support for improving risk identification of urban rail vehicle systems.

[0035] The advantages of additional aspects of the invention will be set forth more clearly in the following description or will be learned by practice of the invention. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is the topology network model of the traction system and bogie system described in the embodiments of the present invention.

[0038] Figure 2 This is a schematic diagram of the fault probability density curve fitting of the brake control unit according to an embodiment of the present invention.

[0039] Figure 3 This is a schematic diagram of the fault probability density curve fitting of the traction inverter according to an embodiment of the present invention.

[0040] Figure 4 This is a schematic diagram of the Copula function fitting results for the braking control unit and traction inverter described in an embodiment of the present invention.

[0041] Figure 5 This is a schematic diagram illustrating the functional dependency strength relationship between nodes of the traction subsystem and bogie subsystem as described in an embodiment of the present invention.

[0042] Figure 6 This is a numerical diagram of the topological parameters of the traction subsystem and bogie subsystem described in the embodiments of the present invention.

[0043] Figure 7 This is a diagram showing the TF-IDF values ​​of the traction subsystem and bogie subsystem nodes according to an embodiment of the present invention.

[0044] Figure 8 This is a schematic diagram of the word graph node transition matrix according to an embodiment of the present invention.

[0045] Figure 9 This is a word cloud diagram of keywords for the traction subsystem and bogie system described in the embodiments of the present invention.

[0046] Figure 10 This is a distribution diagram of various indicators of each component of the traction subsystem and bogie system described in the embodiments of the present invention. Detailed Implementation

[0047] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0048] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0049] To facilitate understanding of the present invention, the present invention will be further explained and described below with reference to the accompanying drawings and specific embodiments. However, the specific embodiments do not constitute a limitation on the embodiments of the present invention.

[0050] Those skilled in the art should understand that the accompanying drawings are merely schematic diagrams of embodiments, and the components in the drawings are not necessarily essential for implementing the present invention.

[0051] Example 1

[0052] This embodiment 1 provides a key component identification system for urban rail transit vehicle systems, including:

[0053] The first construction module is used to build a topology network model of urban rail vehicles, with the components in the vehicle as nodes and the interaction relationships between the components as connecting edges.

[0054] The first calculation module is used to fit the reliability of components based on fault data, quantify the degree of dependence between components, and determine the functional importance index of components.

[0055] The second calculation module is used to determine the component structural importance index based on the quasi-Laplace centrality of the urban rail vehicle topology network model.

[0056] The second construction module is used to establish a professional terminology database for urban rail vehicle system accidents based on urban rail vehicle accident data.

[0057] The third calculation module is used to determine the key indicators of urban rail vehicle components based on accident data by taking the components in the professional terminology library of urban rail vehicle system accidents as keywords, based on the Markov state transition idea.

[0058] The identification module is used to identify key components of the urban rail vehicle system based on functional importance indicators, structural importance indicators, and criticality indicators.

[0059] In this embodiment 1, the above system is used to implement a method for identifying key components of urban rail transit vehicle systems, including:

[0060] A topology network model of urban rail vehicles is constructed using the components in the vehicle as nodes and the interactions between the components as connecting edges.

[0061] Based on the failure data, fit the reliability of the components, quantify the degree of dependence between components, and determine the functional importance index of the components;

[0062] Based on the urban rail vehicle topology network model, the component structural importance index based on quasi-Laplace centrality is determined;

[0063] Based on urban rail vehicle accident data, establish a specialized terminology database for urban rail vehicle system accidents;

[0064] Based on the Markov state transition concept, the components of the urban rail vehicle system accident terminology database are regarded as keywords to determine the key indicators of urban rail vehicle components based on accident data.

[0065] Based on functional importance indicators, structural importance indicators, and criticality indicators, key components of urban rail vehicle systems are identified.

[0066] The determination of the functional importance index of a component includes: fitting a failure probability function based on the failure data of each component, finding the distribution with the smallest fitting result as the failure probability density function of this component; obtaining the parameter values ​​in the Copula function based on maximum likelihood estimation, and calculating the functional dependence between components; then calculating the LR value of each node based on the network adjacency matrix and the LeaderRank algorithm formula; finally, based on the LeaderRank algorithm, introducing a comprehensive evaluation index that considers both component functional dependence and component reliability to describe the functional importance assessment index of the component.

[0067] The determination of structural importance indicators includes: introducing quasi-Laplacian centrality to measure the structural importance of nodes, considering the importance of the node itself and the importance of its neighboring components; the quasi-Laplacian centrality of a node is measured by the change in quasi-Laplacian energy caused by removing it, and the quasi-Laplacian centrality of a node is related not only to its own degree, but also to the degree of other nodes connected to it. Based on the change in quasi-Laplacian energy caused by removing a node, the structural importance of the component is obtained.

[0068] The establishment of a specialized thesaurus for urban rail vehicle system accidents includes: integrating and processing accident documents; selecting the precise mode of the Jieba word segmentation tool when segmenting text; customizing a commonly used dictionary for urban rail vehicle system accidents; and constructing a dictionary of accident causes for urban rail vehicle systems by extracting components appearing in accident reports and combining this with a review of vehicle system components, based on the components and the devices to which they belong.

[0069] Based on the Markov state transition concept, the components of the urban rail vehicle system accident professional terminology are regarded as keywords, including: the components of the urban rail vehicle system accident professional terminology are regarded as keywords, which are the nodes in the constructed urban rail vehicle system accident text word graph model; by combining the Markov state transition model with the word graph, each sentence is used as a window. When two words exist in the same sentence, the edge connecting the corresponding nodes in the word graph is meaningful, and the meaning of the edge is the number of times the words co-occur.

[0070] Key performance indicators for urban rail vehicle components based on accident data were identified, including:

[0071] The accident text of the urban rail vehicle system is segmented into sentences to obtain a set of sentences, and then jieba word segmentation and stop word removal are performed.

[0072] Number the sentences as units, and store the sentences after removing duplicate words;

[0073] The components of the urban rail vehicle system are used as keywords to determine the keyword set. The initial value of the keyword is the value of word frequency minus reverse document frequency. Here, word frequency represents the frequency of a word in the text, and reverse document frequency represents the logarithm of the ratio of the number of texts in the text set to the number of texts in which the word appears.

[0074] Iterate through each sentence. If a keyword exists in a sentence, store the sentence number in the keyword position table.

[0075] Calculate the probability that two keywords exist in the same sentence, and use the conditional probability as the node transition matrix, that is, the weight of the node edge;

[0076] The importance of keywords is determined iteratively based on the weights of the nodes, which is the importance of urban rail vehicle system components based on accident data.

[0077] Example 2

[0078] This embodiment 2 provides a method for identifying key components of an urban rail transit vehicle system, including the following steps:

[0079] Step 1: Using the components in the vehicle as nodes and mechanical connections, electrical connections, and information connections as connecting edges, construct a topology network model for urban rail vehicles. This is an undirected and unweighted complex network. The urban rail vehicle system is divided into the urban rail vehicle body and interior subsystem, auxiliary power supply system, train control subsystem, traction subsystem, braking subsystem, and bogie system.

[0080] Step 2: Based on the LeaderRank algorithm, and taking into account the reliability of components and the functional dependence between components, a functional importance index for urban rail vehicle system components is proposed.

[0081] The iterative process of the LeaderRank algorithm consists of three parts:

[0082]

[0083]

[0084]

[0085]

[0086] Where, n lr LR is the total number of nodes in the topology network (excluding background nodes). i (k) represents node v i The score at time k, k c For LR i (k) When steady state is reached, LR i (k c ), LRg (k c ) represent the node v after the system reaches a stable state. i Background node v g The score; For node v i The degree of departure.

[0087] First, for the component's operational failure data, Minitab software was used to fit the component's failure probability density function using four distributions: normal, log-normal, Weibull, and exponential. The goodness-of-fit test (AD) was then used to compare the fitting effects of the four distributions. The AD statistic (AD) was obtained by calculating the difference between the actual and expected frequencies, squaring the differences, and summing them. * AD * The smaller the value, the more accurate the curve fitting. For mechanical components, the mileage in the fault data is treated as time t; for electrical components, time t is still used for fitting.

[0088] Secondly, in the urban rail transit topology network, the inter-component dependency strength is represented by the functional dependency between nodes. The functional dependency between nodes refers to the degree to which a node's failure affects the failures of other nodes; a higher functional dependency means a greater degree of influence between components. Based on the fitting of component failure data, the most suitable unreliability function is solved, and the copula function is solved through maximum likelihood estimation. Based on the AIC value, select an appropriate Copula function, and finally calculate the failure rate of a component when another component fails, which is the interdependence C between the components. xy That is, the inter-component dependency strength. The Copula function C satisfies formula (5).

[0089] H(x, y) = C(F(x), G(y)) (5)

[0090] Common Copula functions include three types, as shown in the table below:

[0091] Table 1. List of nodes in the topology network of the traction subsystem and bogie system.

[0092]

[0093] For random variables The joint distribution function is The marginal distribution function is Copula is... and The functions connected together are expressed as formula (6).

[0094]

[0095] In the formula n c θ represents the number of random variables; θ represents the correlation parameter of the Copula function.

[0096] Following the algorithm of edge density function, the density function of common Copula function is obtained according to formula (7).

[0097]

[0098] Before applying the AIC criterion, it is necessary to determine the parameter values ​​in each Copula function and use maximum likelihood estimation to obtain the estimated values ​​of the parameters, including the estimated value of ρ. To maximize L(ρ), The maximum likelihood function is:

[0099]

[0100] Take the logarithm of the likelihood function:

[0101]

[0102] Taking the derivative of lnL and setting it to zero, we get:

[0103]

[0104] Where F(x) and F(y) are the unreliability of the two components, and x and y are the lifetimes of the two different components in the system when they fail for the first time.

[0105] The likelihood functions of commonly used Copula functions are as follows:

[0106] The likelihood function of the Gumbel function:

[0107]

[0108] The likelihood function of the Clayton function:

[0109]

[0110] The likelihood function of the Frank function:

[0111]

[0112] The AIC criterion is used to select the Copula function between fitted components. The AIC criterion expression is as follows:

[0113]

[0114] Where L is the likelihood function of the Copula function, and n aWhere is the number of samples, and K is the number of parameters to be estimated.

[0115] The AIC value represents the fitting result of different Copula functions. The smaller the AIC value, the better the Copula function fits the data. The unreliability function of the component is F. x (x), F y (y). According to Sklar's theorem, assuming the edge distribution function f(x) of the component is continuous, there exists a Copula function such that...

[0116] F(x, y) = C(F x (x), F y (y))=C xy (16)

[0117] Specifically, step 2 includes:

[0118] Step 1: Fit a failure probability function based on the failure data of each component, and find the fitting result AD. * The minimum distribution is used as the failure probability density function for this component;

[0119] Step 2: Next, calculate the parameter values ​​in the Copula function based on maximum likelihood estimation, and then calculate the functional dependencies between components:

[0120] Step 3: Then, calculate the LR value of each node based on the network adjacency matrix and the LeaderRank algorithm formula;

[0121] Step 4: Finally, based on the LeaderRank algorithm, a metric for evaluating the importance of component functions is introduced that comprehensively considers both component functional dependency and component reliability.

[0122]

[0123] Among them, node v of Γ1(j) i The first-order neighbor set, C ij Represents a pair of nodes (v) i v j Functional dependency, LR i (k) represents node v i The fraction at time k.

[0124] Step 3: Introduce quasi-Laplacian centrality to measure the structural importance of nodes, comprehensively considering both the importance of the node itself and the importance of its neighbors. The quasi-Laplacian centrality of a node is measured by the change in quasi-Laplacian energy caused by removing it. The quasi-Laplacian centrality of a node is related not only to its own degree but also to the degrees of other connected nodes. Based on the change in quasi-Laplacian energy caused by removing a node, the structural importance of the component is obtained.

[0125] In the topology network diagram of an urban rail vehicle system, it is represented as G = (V, E). If node v i and node v j If connected, then a ij =1, when i=j, a ii =0, its adjacency matrix is:

[0126]

[0127] Where, n v This represents the total number of nodes in the network.

[0128] The degree matrix is ​​a matrix composed of the sum of the neighbors of each node, therefore its degree matrix is:

[0129]

[0130] Where, d i For node v i The degree, N(v i ) is node v i The neighbors.

[0131] For a topological network G, the quasi-Laplace matrix is ​​the sum of the adjacency matrix A(G) and the degree matrix D(G).

[0132] Q(G)=D(G)+A(G) (20)

[0133] The vehicle system's topology is an undirected, unweighted network, and each node has a degree of at least 1. Therefore, Q(G) is a real symmetric positive semi-definite matrix. Based on its properties, we can find that the eigenvalues ​​μ of Q(G) are... i ≥0. The quasi-Laplace energy is defined as, denoted as E. Q (G):

[0134]

[0135] To facilitate the calculation of quasi-Laplace energy, the concept of a line graph is introduced. The nodes of the line graph L(G) are the edges in the topological network G. Two nodes in L(G) are adjacent if and only if the corresponding edges in G are adjacent.

[0136] For a line graph L(G), the total number of edges is ε(L(G)). In the line graph, each node v i There are a total Edge.

[0137]

[0138] According to the properties of matrices, for a matrix B, the sum of its diagonals is the sum of its eigenvalues.

[0139]

[0140] μ is the eigenvalue of the quasi-Laplace matrix Q, and for any non-zero vector x, Qx = μx.

[0141] Q 2 x=Q(Qx)=Q(μx)=μ 2 x (24)

[0142] The above formula explains matrix Q 2 The eigenvalue is μ 2 .

[0143] Q 2 = (D+A)×(D+A)=D 2 +DA+AD+A 2 (25)

[0144] Based on the degree matrix D and the adjacency matrix A, tr(AD) = tr(DA) = 0.

[0145] Therefore, the quasi-Laplace energy can also be expressed as:

[0146]

[0147] As we know from graph theory, in a graph, the number of edges in the network is half the sum of the degrees of all nodes in the network.

[0148]

[0149] according to The pseudo-Laplace energy can be expressed as:

[0150]

[0151] For graph H to be a subgraph of graph G, then E Q (H)≤E Q (G) is defined if and only if, after removing the nodes of graph H, graph G is left with only isolated nodes.

[0152] The description of quasi-Laplacian energy shows that the quasi-Laplacian value of a node is related to the number of edges in both graphs and line graphs. The quasi-Laplacian centrality of a node is measured by the change in quasi-Laplacian energy caused by removing it; node v... i Quasi-Laplace centrality C Q (v i G) is:

[0153]

[0154] Among them, G i Delete node v from graph G i The subgraph obtained by connecting it to the edges.

[0155] Obviously, C Q (v i Since G is non-negative, in order to reduce the amount of computation, i.e. to avoid calculating the eigenvalues ​​of the quasi-Laplace matrix, the following derivation formula is introduced to simplify the calculation.

[0156]

[0157]

[0158]

[0159]

[0160]

[0161]

[0162] In summary, quasi-Laplace centrality, i.e., the structural importance of a component, can be expressed as:

[0163]

[0164] Wherein, N(v) i ) is node v i The neighboring node, d i For node v i The degree.

[0165] Step 4: First, the accident documents need to be integrated and processed. Word segmentation, stop word removal, and text representation are the three main processes of text preprocessing. The Jieba word segmentation tool's precise mode was used for word segmentation. Currently, there is no specific thesaurus for urban rail vehicle accidents, so a custom dictionary of commonly used accidents in urban rail vehicle systems needs to be created. Each urban rail company has a different level of granularity in its accident report descriptions. Therefore, when constructing a causal dictionary for urban rail vehicle system accidents, it should include the components and the devices to which they belong. By extracting the components appearing in the accident reports and combining this with a review of the vehicle system components, a causal dictionary for urban rail vehicle system accidents was constructed.

[0166] Step 5: Treat the components of the urban rail vehicle system accident professional terminology database as keywords, which are the nodes in the constructed urban rail vehicle system accident text word graph model; combine the Markov state transition model with the word graph, and treat each sentence as a window. When two words exist in the same sentence, the edge connecting the corresponding nodes in the word graph is meaningful, and the meaning of the edge is the number of times the words co-occur.

[0167] Preferably, step 5 includes:

[0168] Step 1: Segment the accident text of the urban rail vehicle system into sentences, obtaining a sentence set L = {L1, L2, ..., L...} m Then perform jieba word segmentation and remove stop words;

[0169] Step 2: Number the sentences and store them after removing duplicate words from the sentences;

[0170] Step 3: Components v in the urban rail vehicle system i As keywords, the keyword set is V = {v1, v2, ..., v...} n}, using the TF-IDF value as the initial value for the keyword;

[0171] Term frequency (TF) is the frequency with which words appear in a text.

[0172]

[0173] Where, n ij For word i in text d j The frequency of occurrence in n kj For text d j The total frequency of all words appearing in the text.

[0174] Inverse Document Frequency (IDF) is the ratio of the number of texts in a text set to the number of texts containing a word, then taking the logarithm of that ratio.

[0175]

[0176] Where |D| is the total number of texts, |{j:l i ∈d j}| is for words containing l i The number of texts is calculated by incrementing the denominator by one to prevent the denominator from being zero if some words are not present in the document.

[0177] The TF-IDF value is the product of TF and IDF:

[0178] TF-IDF = idf ij *tf ij (39)

[0179] Step 4: Iterate through each sentence. If a keyword exists in a sentence, store the sentence number in the keyword position table.

[0180] Step 5: Calculate the probability that there are two keywords in the same sentence, and use the conditional probability as the node transition matrix, that is, the weight of the node edge;

[0181] Step 6: Based on the calculation, iterate to determine the importance of the keywords, that is, the importance S(v) of the urban rail vehicle system components based on accident data. i ).

[0182]

[0183]

[0184] Based on the obtained keyword position table, compare keyword v i and v j Given the position vectors, find the number of identical terms, which is n. i，j (n i，j =n j，i ), TF i For the word w i The word frequency, W(i,j), refers to the frequency from node w. i To node w j The weight of the edges.

[0185] Step 6: Identify key components of the urban rail vehicle system by combining three categories of indicators: functional importance, structural importance, and criticality based on accident data.

[0186] In summary, this embodiment 2 proposes a method for identifying key components in urban rail vehicle systems from the perspective of networked operation, which has important theoretical and practical significance for improving the management of rail transit operation safety risks and ensuring safe and efficient operation.

[0187] Example 3

[0188] This embodiment 3 takes a specific urban rail transit system as an example and provides a method for identifying key components of a train system. The method includes the following specific steps:

[0189] Step 1: Summarize and extract the components from the traction subsystem and bogie subsystem. Based on the connection relationships between components, and assuming the connection edges are undirected and unweighted, construct a topological network model of the urban rail vehicle traction subsystem and bogie system. Blue nodes represent bogie subsystem components, and green nodes represent traction subsystem components. The node size is the degree. Figure 1 As shown. For ease of description, the components are numbered from v1 to V. 25 V is a node of the traction subsystem. 26 to v 61 The nodes of the bogie system are shown in Table 1.

[0190] Table 1. List of nodes in the topology network of the traction subsystem and bogie system.

[0191]

[0192] Step 2: Based on the LeaderRank algorithm, and taking into account the reliability of components and the functional dependence between components, a functional importance index for urban rail vehicle system components is proposed.

[0193] First, for the component's operational failure data, Minitab software was used to fit the component's failure probability density function using four distributions: normal, log-normal, Weibull, and exponential. The goodness-of-fit test (AD) was then used to compare the fitting effects of the four distributions. The AD statistic (AD) was obtained by calculating the difference between the actual and expected frequencies, squaring the differences, and summing them. * AD * The smaller the value, the more accurate the curve fitting. For mechanical components, the mileage in the fault data is treated as time t; for electrical components, time t is still used for fitting.

[0194] For mechanical components, taking the brake control unit as an example, firstly, fault information of the brake control unit is extracted from the fault data of a certain subway train in 2020. Then, the reliability probability distribution function described above is used for fitting to obtain the AD of each distribution. * .from Figure 2 As can be seen from this, the log-normal distribution's AD * The value is minimized, therefore a log-normal distribution reliability function is used to measure the state function of the braking control unit. For electrical components, taking the traction inverter as an example, fault data from a certain subway system in 2020 is fitted, from... Figure 3 As can be seen from this, the AD distribution of the normal distribution* Since the value is the smallest, a normally distributed reliability function is used to measure the state function of the traction inverter. Table 2 lists the reliability of each component of the traction subsystem and bogie subsystem after 50 days of operation.

[0195] Table 2. Normalized numerical values ​​of the functional importance of nodes in the traction subsystem and bogie system.

[0196]

[0197] Secondly, in the urban rail transit topology network, the inter-component dependency strength is represented by the functional dependency between nodes. The functional dependency between nodes refers to the degree to which a node's failure affects the failures of other nodes; a higher functional dependency means a greater degree of influence between components. Based on the fitting of component failure data, the most suitable unreliability function is solved, and the copula function is solved through maximum likelihood estimation. Based on the AIC value, select an appropriate Copula function, and finally calculate the failure rate of a component when another component fails, which is the interdependence C between the components. xy This refers to the interdependence strength between components.

[0198] For example, by fitting the failure probability density functions of the braking control unit and the traction inverter, the Gumbel Copula, Clayton Copula, and Frank Copula algorithms are used to fit the dependence strength between the two components, respectively. Figure 4 As shown, the estimated values ​​of the parameters are obtained based on the maximum likelihood estimation, and their AIC values ​​are compared. It is found that the FrankCopula function has a better fitting effect.

[0199] Based on the reliability functions of each component and the connection edges between components, Copula functions are fitted to the component pairs to obtain the parameter values ​​of each Copula function. The Copula function with the smallest AIC value is selected. Different component pairs have different Copula functions; for example, the Clayton Copula function is used between the traction motor and the traction converter, while the Frank Copula function is used between the frame and the traction control unit. The inter-component dependency strength is calculated for different component pairs, and the results are as follows: Figure 5 As shown.

[0200] Finally, based on the LeaderRank algorithm, the functional importance of each node was obtained, and the normalized values ​​are shown in Table 3. According to the functional importance, node v... 46 Functionality is the most important, followed by v 58 v 21 v3 and v 38These components rank highly in terms of functional importance and play a crucial role in the traction subsystem and bogie system.

[0201] Table 3. Normalized numerical values ​​of the functional importance of nodes in the traction subsystem and bogie system.

[0202]

[0203]

[0204] Step 3: Introduce quasi-Laplacian centrality to measure the structural importance of nodes, comprehensively considering both the importance of the node itself and the importance of its neighbors. The quasi-Laplacian centrality of a node is measured by the change in quasi-Laplacian energy caused by removing it. The quasi-Laplacian centrality of a node is related not only to its own degree but also to the degrees of other connected nodes. Based on the change in quasi-Laplacian energy caused by removing a node, the structural importance of the component is obtained.

[0205] Based on the topological network model of the traction subsystem and bogie subsystem of urban rail vehicles, the degrees, betweenness centrality, PageRank, and structural importance of nodes were calculated and normalized. The final results are shown in Table 4. Each node is plotted on the x-axis, and the normalized values ​​of each node's indicators are plotted on the y-axis, as shown in Table 4. Figure 6 As shown.

[0206] Table 4. List of Normalized Values ​​for Topology Indicators of Traction Subsystem and Bogie System

[0207]

[0208]

[0209] Quasi-Laplace centrality takes into account both the importance of a node itself and the importance of its neighbors. It is related not only to the degree of the node itself, but also to the degree of its neighboring nodes. Figure 6 Comparing this with classic node importance metrics, it can be seen that the results obtained by the four types of metrics are consistent, indicating that node v 46 Nodes v3 and v4 occupy the most important structural positions, playing a crucial role in the topological network. Nodes with the same degree value cannot distinguish their importance. Compared to betweenness centrality and the PageRank algorithm, quasi-Laplacian centrality has a lower time complexity, as its time complexity depends only on the number of nodes and edges. This advantage is even more pronounced in more complex networks.

[0210] Step 4: First, the accident documents need to be integrated and processed. Word segmentation, stop word removal, and text representation are the three main processes of text preprocessing. The Jieba word segmentation tool's precise mode was used for word segmentation. Currently, there is no specific thesaurus for urban rail vehicle accidents, so a custom dictionary of commonly used accidents in urban rail vehicle systems needs to be created. Each urban rail company has a different level of granularity in its accident report descriptions. Therefore, when constructing a causal dictionary for urban rail vehicle system accidents, it should include the components and the devices to which they belong. By extracting the components appearing in the accident reports and combining this with a review of the vehicle system components, a causal dictionary for urban rail vehicle system accidents was constructed.

[0211] Table 5. Dictionary of Accident Causes in Urban Rail Vehicle Systems (Partial)

[0212]

[0213] Step 5: Treat the components of the urban rail vehicle system accident professional terminology database as keywords, which are the nodes in the constructed urban rail vehicle system accident text word graph model; combine the Markov state transition model with the word graph, and treat each sentence as a window. When two words exist in the same sentence, the edge connecting the corresponding nodes in the word graph is meaningful, and the meaning of the edge is the number of times the words co-occur.

[0214] Based on the collected accident data related to the traction subsystem and bogie subsystem of urban rail vehicles from 2017 to 2020, the components of the traction subsystem and bogie system were used as keywords to calculate the TF-IDF value of each keyword in the text, such as... Figure 7 As shown. Using the keyword position table, keywords are the nodes in the word graph model, yielding the node transition matrix W(i, j), which represents the weights of the node edges, as shown... Figure 8 As shown.

[0215] Based on the TF-IDF values ​​and transition matrices of the keywords, the node relationships in the word graph are transformed into state transitions. Using an improved TI-TextRank algorithm, a node importance index based on accident data is obtained, and the normalized values ​​are shown in Table 6. Figure 9 As can be seen from the word cloud, regarding the key indicators of nodes based on accident data, node v 20 and v 21 Larger values ​​indicate that components such as the driver controller and temperature sensor are more likely to cause accidents.

[0216] Table 6. Improved TI-TextRank Normalized List of Traction Subsystem and Bogie System Nodes

[0217]

[0218] Step 6: Combine three categories of indicators: functional importance, structural importance, and criticality based on accident data, such as... Figure 10 As shown in Table 7, the key components of the traction subsystem and bogie system have been identified.

[0219] Table 7 Key Components of Traction Subsystem and Bogie System

[0220]

[0221] Example 4

[0222] Embodiment 4 of the present invention provides a non-transitory computer-readable storage medium for storing computer instructions. When the computer instructions are executed by a processor, a method for identifying key components of an urban rail transit vehicle system is implemented.

[0223] Example 5

[0224] Embodiment 5 of the present invention provides a computer program (product), including a computer program that, when run on one or more processors, is used to implement a method for identifying key components of an urban rail transit vehicle system.

[0225] Example 6

[0226] Embodiment 6 of the present invention provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, and the computer program is stored in the memory. When the electronic device is running, the processor executes the computer program stored in the memory to enable the electronic device to execute instructions for implementing a method for identifying key components of an urban rail transit vehicle system.

[0227] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that, based on the technical solutions disclosed in the present invention, various modifications or variations that can be made by those skilled in the art without creative effort should be included within the scope of protection of the present invention.

Claims

1. A method for identifying key components of an urban rail transit vehicle system, characterized in that, include: A topology network model of urban rail vehicles is constructed using the components in the vehicle as nodes and the interactions between the components as connecting edges. Based on the failure data, fit the reliability of the components, quantify the degree of dependence between components, and determine the functional importance index of the components; Determining the functional importance index of components includes: fitting a failure probability function based on the failure data of each component, finding the distribution with the smallest fitting result as the failure probability density function of this component; obtaining the parameter values ​​in the Copula function based on maximum likelihood estimation, and calculating the functional dependence between components; then calculating the LR value of each node based on the network adjacency matrix and the LeaderRank algorithm formula; finally, based on the LeaderRank algorithm, introducing a comprehensive evaluation index that considers both component functional dependence and component reliability to describe the functional importance assessment index of components. Based on the urban rail vehicle topology network model, a component structural importance index based on quasi-Laplacian centrality is determined. The determination of the structural importance index includes: introducing quasi-Laplacian centrality to measure the structural importance of nodes, considering both the importance of the node itself and the importance of its neighboring components; the quasi-Laplacian centrality of a node is measured by the change in quasi-Laplacian energy caused by removing it. The quasi-Laplacian centrality of a node is related not only to its own degree but also to the degrees of other connected nodes. Based on the change in quasi-Laplacian energy caused by removing a node, the structural importance of the component is obtained. Based on urban rail vehicle accident data, establish a specialized terminology database for urban rail vehicle system accidents; Based on the Markov state transition concept, this paper treats components of the urban rail vehicle system accident lexicon as keywords to determine key indicators of urban rail vehicle components based on accident data. The determination of these key indicators includes: segmenting accident texts of the urban rail vehicle system into sentence sets, followed by jieba word segmentation and stop word removal; numbering sentences and storing duplicate words within each sentence; using components of the urban rail vehicle system as keywords to determine the keyword set, with the term frequency minus the inverse file frequency as the initial keyword value; where term frequency represents the frequency of a word in the text, and inverse file frequency represents the logarithm of the ratio of the number of texts in the text set to the number of texts containing the word; iterating through each sentence, if a keyword exists in a sentence, storing that sentence in the keyword position table; calculating the probability of two keywords existing in the same sentence, using the conditional probability as the node transition matrix, i.e., the weight of the node edges; and iterating based on the weights of the node edges to determine the importance of the keywords, i.e., the importance of the urban rail vehicle system components based on accident data. Based on functional importance indicators, structural importance indicators, and criticality indicators, key components of urban rail vehicle systems are identified.

2. The method for identifying key components of an urban rail transit vehicle system according to claim 1, characterized in that, Establish a professional thesaurus for urban rail vehicle system accidents, including: integrating and processing accident documents, selecting the precise mode of the Jieba word segmentation tool when segmenting text, customizing a commonly used dictionary for urban rail vehicle system accidents, and constructing a dictionary of accident causes for urban rail vehicle systems based on components and the devices to which they belong by extracting components appearing in accident reports and combining this with a review of vehicle system components.

3. The method for identifying key components of an urban rail transit vehicle system according to claim 1, characterized in that, Based on the Markov state transition concept, the components of the urban rail vehicle system accident professional terminology are regarded as keywords, including: the components of the urban rail vehicle system accident professional terminology are regarded as keywords, which are the nodes in the constructed urban rail vehicle system accident text word graph model; by combining the Markov state transition model with the word graph, each sentence is used as a window. When two words exist in the same sentence, the edge connecting the corresponding nodes in the word graph is meaningful, and the meaning of the edge is the number of times the words co-occur.

4. A key component identification system for urban rail transit vehicle systems, characterized in that, include: The first construction module is used to build a topology network model of urban rail vehicles, with the components in the vehicle as nodes and the interaction relationships between the components as connecting edges. The first calculation module is used to fit the reliability of components based on fault data, quantify the degree of dependence between components, and determine the functional importance index of components. Determining the functional importance index of components includes: fitting a failure probability function based on the failure data of each component, finding the distribution with the smallest fitting result as the failure probability density function of this component; obtaining the parameter values ​​in the Copula function based on maximum likelihood estimation, and calculating the functional dependence between components; then calculating the LR value of each node based on the network adjacency matrix and the LeaderRank algorithm formula; finally, based on the LeaderRank algorithm, introducing a comprehensive evaluation index that considers both component functional dependence and component reliability to describe the functional importance assessment index of components. The second calculation module is used to determine the structural importance index of components based on the quasi-Laplacian centrality of the urban rail vehicle topology network model. The determination of the structural importance index includes: introducing quasi-Laplacian centrality to measure the structural importance of nodes, considering the importance of the node itself and the importance of its neighboring components; the quasi-Laplacian centrality of a node is measured by the change in quasi-Laplacian energy caused by removing it. The quasi-Laplacian centrality of a node is related not only to its own degree, but also to the degree of other nodes connected to it. The structural importance of the component is obtained based on the change in quasi-Laplacian energy caused by removing the node. The second construction module is used to establish a professional terminology database for urban rail vehicle system accidents based on urban rail vehicle accident data. The third calculation module, based on the Markov state transition concept, treats components from the urban rail vehicle system accident lexicon as keywords to determine key indicators of urban rail vehicle components based on accident data. This determination includes: segmenting accident texts of the urban rail vehicle system into sentence sets, performing jieba word segmentation and removing stop words; numbering sentences and storing duplicate words within each sentence; using components of the urban rail vehicle system as keywords to determine the keyword set, with the term frequency minus the reverse file frequency as the initial keyword value; where term frequency represents the frequency of a word in the text, and reverse file frequency represents the logarithm of the ratio of the number of texts in the text set to the number of texts containing the word; iterating through each sentence, storing the sentence number in the keyword position table if a keyword exists in a given sentence; calculating the probability of two keywords existing in the same sentence, using the conditional probability as the node transition matrix, i.e., the weight of the node edges; and iterating based on the weights of the node edges to determine the importance of the keywords, i.e., the importance of the urban rail vehicle system components based on accident data. The identification module is used to identify key components of the urban rail vehicle system based on functional importance indicators, structural importance indicators, and criticality indicators.

5. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions, which, when executed by a processor, implement the method for identifying key components of an urban rail transit vehicle system as described in any one of claims 1-3.

6. A computer program product, characterized in that, It includes a computer program, which, when run on one or more processors, is used to implement the method for identifying key components of an urban rail transit vehicle system as described in any one of claims 1-3.

7. An electronic device, characterized in that, include: The device includes a processor, a memory, and a computer program; wherein the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory to cause the electronic device to execute instructions for implementing the method for identifying key components of an urban rail transit vehicle system as described in any one of claims 1-3.

Images