Subway system toughness evaluation method

Abstract

The invention discloses a subway system toughness evaluation method, which comprises the following steps of: at an initial stage, comprehensively collecting historical card swiping data of a subway station, cleaning and preprocessing by using a professional technology, and building a foundation for subsequent analysis; in a deep analysis link, on one hand, key indexes such as daily average passenger flow volume and peak hour passenger flow volume are precisely selected, a time sequence and sliding window analysis method is combined, a time evolution track of site importance is precisely captured, and meanwhile, abnormity is sensitively detected by applying a threshold value and a clustering means; and on the other hand, a site association network is built, the association degree between sites is quantified by means of a cooperation coefficient and an influence index, and dynamic transition along with time is analyzed and insighted by means of a dynamic network. The method is characterized in that abnormal situations such as equipment faults are simulated, and influence rules and time trends of abnormal stations on indexes such as passenger flow volume and transfer volume of surrounding stations are deeply interpreted by means of a propagation model.

Description

Technical Field

[0001] The present invention belongs to the technical field related to urban transportation, and particularly relates to a method for evaluating the resilience of a subway system. Background Art

[0002] With the rapid development of the urbanization process, the subway has become one of the main ways for urban residents to travel. However, when subway stations face unexpected situations such as a sudden surge in passenger flow, equipment failures, and bad weather, safety hazards are likely to occur, seriously affecting the normal travel order of citizens. For example, during the morning rush hour in some big cities, the passenger flow at certain popular stations is too large, which may lead to overcrowding of passengers, difficulty in getting on and off the train, and even safety accidents; when a device failure or an unexpected event occurs at a certain station, it will not only affect the normal operation of that station, but may also have a chain reaction on the operation of other stations and the entire subway system through line conduction.

[0003] Currently, the evaluation of the subway system mainly focuses on traditional passenger flow statistics, equipment operation status monitoring, etc., lacking a comprehensive evaluation of the resilience of the subway system. Resilience evaluation aims to measure the resistance ability, recovery ability, and dynamic changes in system performance of the subway system when facing various disturbances. The existing evaluation methods cannot make full use of the historical card-swipe data of subway stations to deeply analyze the importance of stations and the mutual influence relationship between stations, making it difficult to achieve real-time monitoring and accurate evaluation of the operation status of the subway system, nor can it provide strong data support for the subway operation department and the public security organs to deploy security measures in advance and optimize subway lines. Therefore, developing an effective method for evaluating the resilience of the subway system has important practical significance. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for evaluating the resilience of a subway system to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] A method for evaluating the resilience of a subway system, including:

[0007] Data collection and preprocessing, importance and anomaly analysis of different stations evolving over time, correlation analysis of stations evolving over time, and impact analysis of an anomaly at a certain station on the evolution of other stations over time;

[0008] Among them, the importance and anomaly analysis of different stations evolving over time mainly includes importance index selection, time evolution analysis method, and anomaly analysis method;

[0009] The correlation analysis of stations evolving over time mainly includes constructing a station correlation network, correlation index selection, and time evolution analysis method;

[0010] The analysis of the impact of anomalies at a certain site on the time evolution of other sites mainly includes simulating anomaly scenarios and impact analysis methods.

[0011] Preferably, the data collection and preprocessing include collecting the historical card-swipe data of subway stations, including information such as passenger ID, entry time, exit time, entry station, exit station, etc., cleaning the data to remove duplicate records, error data, and outliers, ensuring the accuracy and integrity of the data, and providing a reliable data basis for subsequent analysis.

[0012] Preferably, the selection of importance indicators includes selecting multiple indicators that can reflect the importance of the site, such as the daily average passenger flow, which is used to measure the overall passenger flow scale of the site over a period of time; the peak-hour passenger flow, which highlights the carrying pressure of the site during the peak passenger flow period; the daily average transfer volume, which reflects the importance of the site as a transfer hub for the connectivity of the subway network; and the concentration and dispersion index, which evaluates the ability of the site to attract and disperse passengers by combining the population density, number of job positions in the surrounding area of the site, and information on the departure and destination of passengers.

[0013] Preferably, the time evolution analysis method includes using time series analysis methods to arrange the importance indicators of each site in chronological order to form a time series, and analyzing the trend, seasonality, and periodic changes of the time series through algorithms such as moving average and exponential smoothing to understand the long-term change law of site importance over time. At the same time, sliding window analysis is adopted, setting a sliding window with a certain time length, calculating the importance indicators of each site within the window, and capturing the dynamic changes of site importance in the short term in real time to timely discover potential anomalies.

[0014] Preferably, the anomaly analysis method includes: adopting a threshold-based anomaly detection method, setting thresholds for the normal range of each importance indicator according to historical data and experience, and determining that a site has an anomaly when the indicator value of a certain site exceeds the threshold. In addition, clustering analysis is used, taking the importance indicators of each site at different time points as feature vectors, and using clustering algorithms to divide the sites into different categories. If the category to which a certain site belongs at a certain moment is significantly different from the normal category, it may indicate that the site has an anomaly.

[0015] Preferably, the construction of the site association network includes: taking subway stations as nodes and the passenger flow connection between sites as edges to construct a subway station association network, and determining the weight of the edge according to indicators such as the passenger flow and transfer volume between sites, reflecting the tightness of the connection between sites, and intuitively showing the connection relationship of each site in the subway network.

[0016] Preferably, the selection of the correlation index includes: calculating a synergy coefficient to measure the similarity of the change trends of two stations in terms of passenger flow and transfer volume. The higher the synergy coefficient, the stronger the correlation between the two stations; by analyzing the impact of the passenger flow change of one station on other stations, the influence of this station is determined. For example, if an increase in the passenger flow of Station A leads to a significant increase in the passenger flow of Station B and this impact has strong persistence, it indicates that Station A has a greater influence on Station B.

[0017] Preferably, the time evolution analysis method includes: using the dynamic network analysis method to analyze the structural changes of the station association network over time, studying the changes in the importance of nodes, the increase and decrease of edges, and the evolution of community structure, etc. At the same time, performing time series correlation analysis on the correlation indexes between stations, calculating the changes in the synergy coefficients between two stations in different time periods, and judging whether the association relationship between them strengthens or weakens over time.

[0018] Preferably, the simulated abnormal scenarios include: setting different types of station abnormal scenarios according to historical data and actual situations, such as a sudden equipment failure at a certain station resulting in its closure or a sudden sharp increase in passenger flow. In the simulated scenarios, by adjusting relevant data parameters, the operating state of the subway system under abnormal conditions is simulated.

[0019] Preferably, the impact analysis method includes: after simulating the abnormal scenarios, analyzing the impact of the abnormal stations on the indexes such as the passenger flow, transfer volume, and train operation interval of other stations, using propagation models such as the cascading failure model to simulate the propagation process of the abnormality in the subway network, evaluating the impact range and degree of the abnormality on the entire subway system, and determining the impact rules of the abnormal stations on other stations at different distances and on different lines, as well as the evolution trend of the impact over time by comparing the changes in various indexes of other stations before and after the occurrence of the abnormality.

[0020] Compared with the prior art, the present invention provides a method for evaluating the resilience of a subway system, having the following

[0021] Beneficial effects:

[0022] Improve operation safety: By real-time monitoring and evaluating the operating state of the subway system, potential safety hazards can be detected in a timely manner, security measures can be deployed in advance, and safety accidents can be effectively prevented, ensuring the travel safety of passengers;

[0023] Optimize line planning: Deeply understand the importance of stations and the mutual influence relationship between stations, provide a scientific basis for the subway operation department to optimize the subway lines, improve the operation efficiency and service quality of the lines, and reduce the travel time and transfer times of passengers;

[0024] Enhance emergency response capabilities: In the face of sudden abnormal situations, it can quickly evaluate the impact of the anomaly on the entire subway system, provide data support for the subway operation department and the public security organs to formulate reasonable emergency response strategies, improve the emergency handling ability, and reduce the impact of abnormal events on citizens' travel. Brief Description of the Drawings

[0025] Figure 1 It is a schematic structural diagram of the present utility model invention.

[0026] Figure 2 It is a schematic structural diagram of the present utility model invention.

[0027] Figure 3 It is a schematic structural diagram of the present utility model invention. Detailed Embodiments

[0028] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0029] The present invention provides data collection and preprocessing as Figures 1-3 shown:

[0030] Obtain historical card-swipe data for a certain period (such as the past year) from the subway operation management system and store it in the database.

[0031] Write a data cleaning script and use the pandas library in Python to clean the data. For example, use the drop_duplicates() function to remove duplicate records, remove obviously incorrect data through conditional screening (such as data with an entry time later than the exit time), and handle outliers (such as replacing passenger flow data outside the reasonable range with missing values, and then filling them with the mean or other appropriate methods).

[0032] Importance and anomaly analysis of different stations evolving over time:

[0033] Calculation of importance indicators:

[0034] Average daily passenger flow: Use SQL query statements to group and count the card - swiping data by station and date, calculate the total passenger flow for each station every day, and then calculate the average daily passenger flow. For example: SELECT station_id,AVG(daily_flow)FROM(SELECT station_id,SUM(flow)AS daily_flow FROM swipe_dataGROUP BY station_id,date)GROUP BY station_id;

[0035] Peak - hour passenger flow: First, determine the morning peak (e.g., 7:00 - 9:00) and evening peak (e.g., 17:00 - 19:00) time periods, and then count the card - swiping data by station within these time periods to calculate the passenger flow of each station during peak hours.

[0036] Average daily transfer volume: By identifying the records of entering and leaving different stations within a short period (e.g., within 30 minutes) in the card - swiping data, count the average daily transfer volume of each station.

[0037] Concentration index: Combine Geographic Information System (GIS) data to obtain data on population density and the number of job positions in the areas surrounding the stations. Use the geopandas library in Python for spatial analysis, associate the departure and destination locations of passengers in the card - swiping data with the areas surrounding the stations, and calculate the concentration index.

[0038] Time - series analysis:

[0039] Use the statsmodels library in Python to perform time - series analysis on the importance indicators of each station. For example, for the time series of average daily passenger flow, use the sm.tsa.seasonal_decompose() function for seasonal decomposition to analyze trends, seasonality, and periodic changes.

[0040] Moving - window analysis: Set the moving - window size to 1 hour, and use the rolling() function in pandas to calculate the importance indicators of each station within the window to monitor the dynamic changes of the indicators in real - time.

[0041] Anomaly analysis:

[0042] Threshold - based anomaly detection: Calculate the mean and standard deviation of each importance indicator based on historical data, and set the threshold as the mean plus twice the standard deviation. When the indicator value of a certain station exceeds the threshold, it is marked as an anomaly.

[0043] Cluster analysis: The importance indicators of each station at different time points are used to form a feature matrix, and the K-Means clustering algorithm in the scikit-learn library is used for cluster analysis. For example: from sklearn.cluster import KMeans; kmeans = KMeans(n_clusters = 3); kmeans.fit(features_matrix); The anomalies are detected by comparing the changes in the clusters to which the stations belong.

[0044] Analysis of the correlation evolution between stations over time:

[0045] Constructing the station correlation network: Use the networkx library in Python to construct the subway station correlation network. The stations are used as nodes, and the weights of the edges are determined according to the passenger flow and transfer volume between the stations. For example: import networkx as nx; G = nx.Graph(); for index, row in connection_data.iterrows(): G.add_edge(row['station1'], row['station2'], weight = row['flow'] + row['transfer_flow']);

[0046] Calculation of correlation indicators:

[0047] Synergy coefficient: Use the Pearson correlation coefficient to calculate the synergy coefficient of two stations in terms of indicators such as passenger flow and transfer volume. For example: from scipy.stats import pearsonr; corr, _ = pearsonr(station1_data, station2_data);

[0048] Influence indicator: By establishing a regression model, analyze the degree of influence of the passenger flow change of one station on other stations to determine the influence indicator.

[0049] Analysis of time evolution:

[0050] Dynamic network analysis: Use the dynamic graph analysis function of the networkx library to update the structure of the station correlation network and the weights of the edges over time, and analyze the changes in the network structure.

[0051] Time series correlation analysis: Conduct time series correlation analysis on the correlation indicators between stations, and use the corrwith() function of pandas to calculate the correlation in different time periods.

[0052] Analysis of the impact of the anomaly of a certain station on the time evolution of other stations:

[0053] Simulate an abnormal scenario: Simulate a scenario where a certain site suddenly shuts down due to equipment failure in the database, and set the inbound and outbound data of this site to zero for a period of time.

[0054] Impact analysis method:

[0055] Apply a propagation model, such as the cascade failure model based on network propagation, and use the numpy and scipy libraries in Python to implement the model algorithm. By simulating the propagation process of the anomaly in the subway network, evaluate the impact of the anomaly on indicators such as the passenger flow and transfer volume of other stations.

[0056] Compare the changes in various indicators of other stations before and after the occurrence of the anomaly, and use SQL query statements and Python data analysis libraries to compare and analyze the data to determine the impact law and time evolution trend of the abnormal station on other stations.

[0057] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A subway system resilience assessment method, characterized in that: include: Data collection and preprocessing, importance and anomaly analysis of different sites over time, correlation analysis of sites over time, and impact of anomalies at one site on the evolution of other sites over time; The importance and anomaly analysis of different sites evolving over time mainly include the selection of importance indicators, time evolution analysis methods and anomaly analysis methods; The analysis of the correlation between sites over time mainly includes the construction of site correlation network, the selection of correlation index and the time evolution analysis method; The analysis of the impact of anomalies at a certain site on the evolution of other sites over time mainly includes simulation of abnormal scenarios and impact analysis methods.

2. A subway system resilience assessment method according to claim 1, characterized in that: The data collection and preprocessing includes collecting historical card swiping data of subway stations, including passenger ID, entry time, exit time, entry station, exit station and other information, cleaning the data, removing duplicate records, erroneous data and outliers, ensuring the accuracy and completeness of the data, and providing a reliable data basis for subsequent analysis.

3. A subway system resilience assessment method according to claim 1, characterized in that: The selection of importance indexes includes selecting multiple indexes that can reflect the importance of a site, such as average daily passenger flow, which is used to measure the overall passenger flow scale of a site over a period of time; Peak hour passenger flow highlights the load-bearing capacity of the station during peak passenger flow periods; average daily transfer volume reflects the importance of the station as a transfer hub to the connectivity of the metro network; The concentration index evaluates the station's ability to attract and evacuate passengers by combining the population density, number of jobs, and passenger origin and destination information in the area surrounding the station.

4. A subway system resilience assessment method according to claim 1, characterized in that: The time evolution analysis method includes using a time series analysis method to arrange the importance indicators of each site in chronological order to form a time series, and using algorithms such as moving average and exponential smoothing to analyze the trend, seasonality and cyclical changes of the time series to understand the long-term changes in the importance of the site over time. At the same time, a sliding window analysis is used to set a sliding window of a certain time length, calculate the importance indicators of each site within the window, capture the dynamic changes of the importance of the site in the short term in real time, and discover potential abnormal situations in time.

5. A subway system resilience assessment method according to claim 1, characterized in that: The anomaly analysis method adopts a threshold-based anomaly detection method. According to historical data and experience, a normal range threshold is set for each importance index. When the index value of a certain site exceeds the threshold, the site is judged to be abnormal. In addition, cluster analysis is used to take the importance index of each site at different time points as a feature vector, and the site is divided into different categories using a clustering algorithm. If the category to which a site belongs at a certain moment is significantly different from the category under normal circumstances, it may indicate that the site is abnormal.

6. A subway system resilience assessment method according to claim 1, characterized in that: The construction of the site association network includes: taking subway stations as nodes and passenger flow connections between stations as edges, constructing a subway station site association network, wherein the weight of the edge is determined according to indicators such as passenger flow and transfer volume between stations, reflecting the closeness of the connection between stations and intuitively displaying the connection relationship between stations in the subway network.

7. A subway system resilience assessment method according to claim 1, characterized in that: The selection of the correlation index includes: calculating the synergy coefficient, which is used to measure the similarity of the changing trends of the passenger flow and transfer volume indicators of the two stations. The higher the synergy coefficient, the stronger the correlation between the two stations; by analyzing the impact of the passenger flow changes of one station on other stations, the influence of the station is determined. For example, if an increase in the passenger flow of station A will lead to a significant increase in the passenger flow of station B, and this impact is highly sustained, it means that station A has a greater influence on station B.

8. A subway system resilience assessment method according to claim 1, characterized in that: The time evolution analysis method includes: using a dynamic network analysis method to analyze the structural changes of the site association network over time, studying the changes in the importance of nodes, the increase and decrease of edges, and the evolution of community structure, etc. At the same time, performing time series correlation analysis on the correlation indicators between each site, calculating the changes in the synergy coefficient between two sites in different time periods, and judging whether the correlation relationship between them increases or decreases over time.

9. A subway system resilience assessment method according to claim 1, characterized in that: The simulated abnormal scenarios include: setting different types of site abnormal scenarios based on historical data and actual conditions, such as a site being closed due to a sudden equipment failure or a sudden surge in passenger flow; in the simulated scenarios, adjusting relevant data parameters to simulate the operating status of the subway system under abnormal conditions.

10. A subway system resilience assessment method according to claim 1, characterized in that: The impact analysis method includes: after simulating the abnormal scenario, analyzing the impact of the abnormal site on indicators such as passenger flow, transfer volume, and train running interval of other sites, using a propagation model, such as a cascading failure model, to simulate the propagation process of the anomaly in the subway network, evaluating the scope and degree of the impact of the anomaly on the entire subway system, and by comparing the changes in various indicators of other sites before and after the anomaly occurs, determining the impact of the abnormal site on other sites at different distances and on different lines, as well as the evolution trend of the impact over time.

Images