An integrated data management platform for the construction, management, and maintenance of rail transit infrastructure

Abstract

The present invention discloses an integrated data management platform for the construction, management, and maintenance of rail transit infrastructure, belonging to the technical field of rail transit facility management. It includes a facility parameter input module, a data acquisition module, a damage reporting module, a grouping planning module, a damage analysis module, and a facility management and maintenance module. The facility parameter input module is used to input the basic parameters of rail transit infrastructure, including track facilities and station facilities. The data acquisition module is used to collect influence data of different stations, and the influence data includes rainfall, passenger flow, and train trips. An integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to the present invention realizes real-time interaction and collaboration of multi-module data by performing parameter input, data acquisition, damage analysis, grouping planning, and maintenance management on rail transit infrastructure at different stations, improves the management efficiency and decision-making accuracy of rail transit, and improves the operation reliability of rail transit infrastructure.

Description

Technical Field

[0001] The present invention belongs to the technical field of rail transit facility management, and specifically relates to an integrated data management platform for the construction, management, and maintenance of rail transit infrastructure. Background Art

[0002] With the rapid development of urban rail transit, the scale of its infrastructure has been continuously expanding, covering numerous rail transit stations and a large number of facilities and equipment with different specifications and models. However, many problems have emerged in the traditional management mode of rail transit infrastructure, making it difficult to meet the current requirements of efficient and safe operation;

[0003] During the traditional management and maintenance process of rail transit facilities, routine inspections of rail transit infrastructure are usually carried out based on fixed inspection cycles, or maintenance is carried out based on the reported information when a fault occurs. There is a lack of comprehensive consideration of the specifications of different rail transit infrastructure in combination with the geographical environment information of the stations, and the damage of the infrastructure at each rail transit station cannot be comprehensively analyzed by combining the specifications, models, parameters, and geographical environment of the infrastructure, making it difficult to detect potential damage hazards in advance, resulting in problems of low practicability and functionality;

[0004] This case proposes an integrated data management platform for the construction, management, and maintenance of rail transit infrastructure to solve the above technical problems. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art. For this purpose, the present invention proposes an integrated data management platform for the construction, management, and maintenance of rail transit infrastructure, and solves the above technical problems by improving the detection method and processing method.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] An integrated data management platform for the construction, management, and maintenance of rail transit infrastructure, comprising a facility parameter input module, a data collection module, a damage reporting module, a grouping and planning module, a damage analysis module, and a facility management and maintenance module;

[0008] The facility parameter input module is used to input the basic parameters of rail transit infrastructure, including track facilities and station facilities;

[0009] The data collection module is used to collect influence data of different stations. The influence data includes rainfall, passenger flow, and train frequency, and comprehensively statistics the total value of influence data on a daily, weekly, monthly, quarterly, and annual basis;

[0010] The damage reporting module, based on the reported damage of rail transit infrastructure, integrates the parameters of the damaged facilities and the influence data of the damage reporting location, and transmits them to the damage analysis module;

[0011] The damage analysis module qualitatively determines the damage types for the reported damages of rail transit infrastructure, in combination with the impact parameters at the damage reporting locations, including accidental damages and structural damages. It does not record the impact parameters at the damage reporting locations of accidental damages, but records the impact parameters at the damage reporting locations of structural damages, analyzes the causes of damages, screens the impact parameters related to structural damages, and transmits them to the subsequent modules;

[0012] The grouping planning module divides the rail transit stations that use the same specification of infrastructure and are in the same type of climate zone into the same group according to the basic parameters of the rail transit infrastructure and in combination with the geographical location climate zones;

[0013] The facility management and maintenance module calculates the abnormal impact parameter range of the associated impact parameters according to the impact parameters related to structural damages, analyzes the impact data of the remaining stations in the same group as the station where the structural damage occurred, determines the similarity of the station impact parameters, marks the similar potential dangerous points, and manages and maintains the rail transit infrastructure.

[0014] Further, the facility parameter entry module establishes a rail transit infrastructure management database through MySQL, and respectively establishes city archives according to the cities where the rail transit is located. Inside the city archives, sub-archives are respectively established according to the names of the rail transit stations under the jurisdiction of the city, and the basic parameters of the rail transit infrastructure of each station are entered, including track facility parameters and station facility parameters;

[0015] The data collection module includes a passenger flow sensor, collects meteorological data in different station areas, obtains the passenger flow, train numbers, rainfall, temperature, and humidity data of different stations according to the train operation schedules of the stations, conducts daily, weekly, monthly, quarterly, and annual statistics on the collected rainfall, temperature, humidity, passenger flow, and train number data, calculates the total value, average value, maximum value, minimum value, and peak value within a unit time of each data item, and respectively records them inside the corresponding station sub-archives;

[0016] The grouping planning module obtains the climate zone information of the locations of each rail transit station through geographic information system technology, and based on the basic parameter information of the facilities in different sub-archives in the rail transit infrastructure management database, conducts infrastructure matching grouping on the sub-archives. The specific steps are as follows:

[0017] For categorical parameters, the sub-archives of the stations that use the same model and material of infrastructure are classified into the same category archive, and further grouping is conducted on the category archive by performing error matching on the numerical parameters inside the sub-archives of the stations in the same category archive. The steps are as follows:

[0018] Extract the numerical parameters of each site sub-file in the same category file, define one of the site sub-files as the target sub-file, and calculate the parameter matching degree between the target sub-file and other sub-files respectively:

[0019]

[0020] Among them, x and y represent the values of the same numerical parameter of the target site sub-file and the remaining site sub-files in the same category file respectively, θ represents the error range, and m represents the matching degree;

[0021] Based on the parameter matching degree of each parameter, calculate the overall parameter matching degree:

[0022]

[0023] In the above formula, n represents the number of facility parameters, m i represents the matching degree of the i-th parameter, ω i represents the corresponding weight, and M represents the overall parameter matching degree. When M≥λ, it means that the matching degree of the numerical parameters in the two site sub-files meets the requirements, and the files that meet the matching degree requirements with the target sub-file are recorded inside the target sub-file. When M<λ, it means that the matching degree of the numerical parameters in the two site sub-files does not meet the requirements and is not recorded;

[0024] For the remaining sub-files recorded in each target sub-file, based on the climate zone information of the current target sub-file, retain the remaining sub-files that are consistent with the climate zone of the target sub-file to obtain the associated sub-files of the current target sub-file.

[0025] Furthermore, the damage reporting module receives the rail transit infrastructure damage information reported from manual and automatic channels. The damage information includes the damage occurrence site information, the type of damaged infrastructure, the damage occurrence time, whether it involves human damage, and the damage location image. The impact data of the damage reporting location is retrieved through the rail transit infrastructure management database and combined with the damage information to form a damage report, which is transmitted into the damage analysis module;

[0026] The damage analysis module, according to the damage report transmitted by the damage reporting module, based on the characteristics and relevant information of the damage, qualitatively determines the damage type in the current damage report to determine whether the damage is accidental damage or structural damage;

[0027] For the qualitatively determined structural damage, record the impact parameters of the damage reporting location of the structural damage and analyze the cause of the damage, screen the structural damage associated impact parameters in the current damage report, and transmit them to the subsequent module.

[0028] Further, based on the damage report transmitted by the damage reporting module, qualitatively determine the damage type in the current damage report as accidental damage or structural damage according to the characteristics and relevant information of the damage, including the following steps:

[0029] When the damage report involves human damage, qualitatively determine the damage type in the current damage report as accidental damage;

[0030] When the damage report does not involve human damage, qualitatively determine the damage type in the current damage report based on the damaged infrastructure type. The specific steps are as follows:

[0031] Collect historical damage image data of different rail transit infrastructures, label the image data as structural damage or accidental damage, adjust the images to the same size, normalize the pixel values to the range [0, 1], and divide the processed historical damage image data of different rail transit infrastructures into training sets, validation sets, and test sets according to the infrastructure type. The data ratios are 70%, 15%, and 15% respectively. Build a convolutional neural network model through the training set, adjust the model hyperparameters through the validation set, and evaluate the final performance of the model through the test set;

[0032] Input the damage location image data according to the damaged infrastructure type in the damage report into the CNN determination model under the corresponding infrastructure type, output the determination result, and determine the damage type in the current damage report as accidental damage or structural damage.

[0033] Further, for the qualitatively determined structural damage, record the damage location impact parameters of the structural damage and analyze the damage cause, screen the structural damage associated impact parameters in the current damage report, and transmit them to the subsequent module, including the following steps:

[0034] For structural damage, obtain the impact data during the occurrence of the damage at the damage location through the damage report, including the number of people, train trips, rainfall, temperature, and humidity data within the station where the damage occurred on the day, as well as the number of people, train trips, rainfall, temperature, and humidity data at the current station on historical normal dates, and obtain the number of historical damage status record samples of the same infrastructure at the current station. Among them, the number of people, train trips, rainfall, temperature, and humidity data are X1, X2, X3, X4, X5 respectively, and the number of historical damage status record samples is Y;

[0035] Standardize the number of people, train trips, rainfall, temperature, and humidity data through the Z-score, screen the damage-related parameters through the Pearson correlation coefficient and the Spearman rank correlation coefficient, and perform multivariate feature selection:

[0036]

[0037] In the above formula, Yi represents the damage state of the i-th sample, β0 is the intercept term, β j represents the regression coefficient of the j-th parameter, is the regularization parameter, and the parameters with β j ≠0 are used as the structure damage associated influence parameters, X ij represents the value of the j-th parameter of the i-th sample.

[0038] Furthermore, the facility management and maintenance module calculates the abnormal influence parameter range of the associated influence parameters according to the structure damage associated influence parameters, analyzes the influence data of the remaining stations in the same group as the structure damage occurrence site, determines the similarity of the station influence parameters, marks the similar potential dangerous points, and manages and maintains the rail transit infrastructure. The specific steps are as follows:

[0039] For the obtained structure damage associated influence parameters, through univariate anomaly detection and multivariate anomaly detection, obtain the associated influence parameter range of the current structure damage;

[0040] According to the current damage report, determine the associated sub-file information in the current damaged station sub-file in the rail transit infrastructure management database, extract the structure damage associated influence parameters in the associated sub-file, determine the similarity of the associated influence parameters in the associated sub-file, and mark the similar potential dangerous points in the rail transit infrastructure management database to manage and maintain the rail transit infrastructure.

[0041] Furthermore, for the obtained structure damage associated influence parameters, through univariate anomaly detection and multivariate anomaly detection, obtain the associated influence parameter range of the current structure damage. The specific steps are as follows:

[0042] When the structure damage associated influence parameter is univariate, determine the structure damage associated influence parameter range based on the 3σ principle;

[0043] When the structure damage associated influence parameter is multivariate, determine the structure damage associated influence parameter range through the Mahalanobis distance. The specific steps are as follows:

[0044] D 2 =(G - μ)∑ -1 (G - μ) T ;

[0045] where G represents the vector containing multiple structure damage associated influence parameters, μ represents the mean vector of multiple parameters, that is, the average value of each parameter under normal conditions, and ∑ -1 represents the inverse matrix of the covariance matrix. When , it is determined that the multiple parameters are abnormal, where represents the chi-square distribution critical value, p is the number of parameters, and 0.95 is the confidence level.

[0046] Furthermore, based on the current loss report, determine the associated sub-file information in the current loss site sub-file in the rail transit infrastructure management database, extract the structural damage associated impact parameters in the associated sub-file, determine the similarity of the associated impact parameters in the associated sub-file, mark the similar potential hazard points in the rail transit infrastructure management database, and manage and maintain the rail transit infrastructure. The specific steps are as follows:

[0047] For the type of structural damage associated impact parameters of the current site sub-file, extract the corresponding associated impact parameter data records in the associated sub-file, and determine the potential hazard points of the site in the associated sub-file:

[0048] When the structural damage associated impact parameter is a single variable, count the structural damage associated impact parameter data in the associated sub-file. When there is any day in the associated sub-file where the structural damage associated impact parameter data is not within the 3σ range, it means that there is a potential infrastructure anomaly in the current associated sub-file. Mark the relevant infrastructure in the associated sub-file in red, and mark the status as to be maintained in the rail transit infrastructure management database;

[0049] When the structural damage associated impact parameter is a multi-variable, extract the corresponding associated impact parameter data records in the associated sub-file, and calculate D 2 , when there is any day in the historical dates when it means that there is a potential infrastructure anomaly in the current associated sub-file. Mark the relevant infrastructure in the associated sub-file in red, and mark the status as to be maintained in the rail transit infrastructure management database.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] 1. In the present invention, based on the geographical environment information of different rail transit stations, combined with the specifications, models, and parameters of the rail transit infrastructure, damage maintenance and divergence analysis are carried out on the rail transit infrastructure, realizing potential damage management and maintenance of the rail transit infrastructure, discovering potential damage in advance, reducing the operation safety risk of the rail transit infrastructure, and enhancing the practicability;

[0052] 2. In the present invention, by determining the damage types of the rail transit infrastructure and taking different management and maintenance measures for different types of damage, the structural damage of the infrastructure is discovered in time and the remaining rail transit stations in similar environments are maintained, warning potential hazard points in advance, reducing the risk of sudden accidents, and enhancing the functionality of the system;

[0053] 3. In the present invention, by using the infrastructure information of rail transit stations in combination with climate zone data and through the facility parameter matching degree algorithm, homogeneous station groups are intelligently divided to achieve "centralized control of similar risk stations", improve the pertinence and efficiency of maintenance resource allocation, realize real-time interaction and collaboration of multi-module data, and improve the management efficiency of rail transit infrastructure;

[0054] 4. In the present invention, by predicting the potential damage locations of rail transit infrastructure and conducting targeted inspection and maintenance, the service life of the infrastructure is extended through damage management and maintenance, the economic damage caused by the damage of rail transit infrastructure is reduced, integrated management and maintenance of rail transit infrastructure are realized, and the difficulty of rail transit maintenance management is reduced;

[0055] The integrated data management platform for the construction, management, and maintenance of the entire rail transit infrastructure realizes real-time interaction and collaboration of multi-module data through parameter entry, data collection, damage analysis, grouping planning, and maintenance management of rail transit infrastructure at different stations, improves the management efficiency and decision-making accuracy of rail transit, and improves the operation reliability of rail transit infrastructure. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a block diagram of an integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] Next, the technical solutions of the present invention will be clearly and completely described in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0058] Embodiment 1:

[0059] As Figure 1 shown, an integrated data management platform for the construction, management, and maintenance of rail transit infrastructure includes.

[0060] An integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to the present invention, when in use, includes a facility parameter entry module, a data collection module, a damage reporting module, a grouping planning module, a damage analysis module, and a facility management and maintenance module;

[0061] The facility parameter entry module establishes a rail transit infrastructure management database through MySQL and establishes city files according to the cities where the rail transit is located. Sub-files are respectively established according to the names of rail transit stations under the jurisdiction of the city in the city files, and the basic parameters of the rail transit infrastructure at each station are entered, including track facility parameters and station facility parameters;

[0062] It should be noted that the track facility parameters include track model, track material, catenary model, gauge, floor area, etc., and the station facility parameters include the model, quantity, running speed of elevators and escalators, the type and quantity of lamps in the lighting system, the refrigerating capacity, power of the air-conditioning system, the model of turnstiles, etc. The track facility parameters and station facility parameters are collected according to the construction design drawings of rail transit stations and recorded in relevant sub-files.

[0063] The data acquisition module, including a pedestrian flow sensor, collects meteorological data in different station areas. According to the train operation timetable of the station, it obtains the pedestrian flow, train number, rainfall, temperature, and humidity data of different stations, and conducts daily, weekly, monthly, quarterly, and annual statistics on the collected rainfall, temperature, humidity, pedestrian flow, and train number data, calculates the total value, average value, maximum value, minimum value, and peak value within a unit time of each item of data, and records them respectively inside the sub-files of the corresponding stations;

[0064] It should be noted that the unit time is 1 hour or half an hour, and the peak value within a unit time is the maximum value of the data statistical values in all 1-hour or half-hour time periods of this day, which is used to reflect the extreme situations of different collected data on the scale of this unit time, so as to analyze the maximum load or intensity of the pedestrian flow, train number, or rainfall in the station area within a specific short time.

[0065] The grouping and planning module, through geographic information system technology, obtains the climate zone information of the locations where each rail transit station is located, and based on the basic facility parameter information of different sub-files in the rail transit infrastructure management database, conducts infrastructure matching and grouping on the sub-files. The specific steps are as follows:

[0066] For categorical parameters, the sub-files of stations with the same model and material infrastructure are classified into the same category file, and the category file is further grouped by performing error matching on the numerical parameters in the sub-files of stations in the same category file. The steps are as follows:

[0067] Extract the numerical parameters of each station sub-file in the same category file, define one of the station sub-files as the target sub-file, and calculate the parameter matching degrees between the target sub-file and other sub-files respectively:

[0068]

[0069] Among them, x and y respectively represent the values of the same numerical parameter of the target station sub-file and the remaining station sub-files in the same category file, θ represents the error range, and m represents the matching degree;

[0070] It should be noted that by defining each site sub-file in the same category file as the target sub-file respectively, the relationship between each site sub-file and other site sub-files in the same category file can be obtained, which is convenient for subsequent facility management and maintenance. Among them, the numerical parameters include gauge, building area, platform length, platform height, axle load, voltage and current of power supply equipment, etc. The error ranges of different numerical parameters are different. For each numerical parameter, calculate its fluctuation range in the historical data, that is, the maximum and minimum values of the parameter under normal use, obtain the fluctuation interval of the parameter, and take 20% of the fluctuation range as the error range.

[0071] Based on the parameter matching degree of each parameter, calculate the overall parameter matching degree:

[0072]

[0073] In the above formula, n represents the number of facility parameters, m i represents the matching degree of the i-th parameter, ω i represents the corresponding weight, and M represents the overall parameter matching degree. When M≥λ, it means that the matching degree of the numerical parameters in the two site sub-files meets the requirements, and the file that meets the matching degree requirements with the target sub-file is recorded inside the target sub-file. When M<λ, it means that the matching degree of the numerical parameters in the two site sub-files does not meet the requirements and is not recorded;

[0074] It should be noted that the value of λ is set to 0.9, which means that there is 90% similarity between the two in terms of numerical parameters. ω i is evenly set to ensure that all parameter weights are the same. By adjusting the value of λ, the screening conditions can be further refined or blurred.

[0075] For the remaining sub-files recorded in each target sub-file, based on the climate zone information of the current target sub-file, retain the remaining sub-files with the same climate zone as the target sub-file to obtain the associated sub-files of the current target sub-file.

[0076] It should be noted that for stations in the same climate zone, their climates will have similarities, that is, the environmental impact factors are the same. Since China's territory is vast and covers multiple climate zones, classifying stations in the same climate zone provides data support for subsequent facility management and maintenance. The climate zones include cold temperate zone, middle temperate zone, warm temperate zone, north subtropical zone, central subtropical zone, south subtropical zone, marginal tropical zone, middle tropical zone, equatorial tropical zone, etc.

[0077] Example 2:

[0078] The damage reporting module receives the damage information of rail transit infrastructure reported from manual and automatic channels. The damage information includes the information of the station where the damage occurs, the type of damaged infrastructure, the time of damage occurrence, whether it involves human damage, and the image of the damage location. It retrieves the impact data of the damaged location from the rail transit infrastructure management database, combines it with the damage information to summarize a damage report, and transmits it into the damage analysis module;

[0079] It should be noted that the manual channels include the daily inspection reports of inspectors, the feedback from maintenance personnel, the feedback from passengers, etc., and the automatic channels include abnormal sensor monitoring data, etc. By receiving the damage information from manual and automatic channels, the information is accurately entered into the system by the staff, and it is recorded whether the damage involves human damage. The damage report includes the location of the current damage, the image of the damage location, the involved stations, the time of occurrence, the type of damaged infrastructure, whether it involves human damage, and the impact data during the damage occurrence at the damaged location.

[0080] The damage analysis module, based on the damage report transmitted by the damage reporting module, qualitatively determines the type of damage in the current damage report as accidental damage or structural damage according to the characteristics and relevant information of the damage, including the following steps:

[0081] When the damage report involves human damage, the type of damage in the current damage report is qualitatively determined as accidental damage;

[0082] When the damage report does not involve human damage, based on the type of damaged infrastructure, the type of damage in the current damage report is qualitatively determined. The specific steps are as follows:

[0083] Collect the historical damage image data of different rail transit infrastructures, label the image data as structural damage or accidental damage, adjust the images to the same size, normalize the pixel values to the range of [0, 1], and divide the processed historical damage image data of different rail transit infrastructures into a training set, a validation set, and a test set according to the infrastructure type respectively. The data ratios are 70%, 15%, and 15% respectively. Build a convolutional neural network model through the training set, adjust the model hyperparameters through the validation set, and evaluate the final performance of the model through the test set;

[0084] According to the type of damaged infrastructure in the damage report, input the damage location image data into the CNN determination model under the corresponding infrastructure type, output the determination result, and determine the type of damage in the current damage report as accidental damage or structural damage.

[0085] It should be noted that during the construction of the convolutional neural network model, for the image data input layer, the dimensions are the height, width, and number of channels of the image. The number of neurons in the output layer is set to 2, corresponding to two types of accidental damage and structural damage. The probability of each category is output through the softmax activation function. The cross-entropy loss function is used to measure the difference between the model prediction result and the true label. The Adam optimizer is used to update the parameters of the model. When a new damage report is received, the data in the damage report is input into the model corresponding to the infrastructure type, and the qualitative result of the damage type is output;

[0086] For the qualitative structural damage, record the damage location impact parameters of the structural damage and analyze the cause of the damage. Screen the structural damage associated impact parameters in the current damage report and transmit them to the subsequent module, including the following steps:

[0087] For the structural damage, obtain the impact data during the occurrence of the damage at the damage location through the damage report, including the number of people, train trips, rainfall, temperature, and humidity data within the site where the damage occurred on the day, as well as the number of people, train trips, rainfall, temperature, and humidity data at the current site during historical normal dates. Also obtain the number of historical damage status record samples of the same infrastructure at the current site, where the number of people, train trips, rainfall, temperature, and humidity data are X1, X2, X3, X4, X5 respectively, and the number of historical damage status record samples is Y;

[0088] It should be noted that Y is a binary variable, indicating whether structural damage occurred at the current site at a certain moment. If there is structural damage, it is 1; if there is no structural damage, it is 0.

[0089] Standardize the data of the number of people, train trips, rainfall, temperature, and humidity through the Z-score. Screen the damage-related parameters through the Pearson correlation coefficient and the Spearman rank correlation coefficient, and perform multivariate feature selection:

[0090]

[0091] In the above formula, Y i represents the damage status of the i-th sample, β0 is the intercept term, β j represents the regression coefficient of the j-th parameter, is the regularization parameter, and retain the parameter with β j ≠0 as the structural damage associated impact parameter, X ij represents the j-th parameter value of the i-th sample.

[0092] It should be noted that β j reflects the influence direction and intensity of the parameter X j on the damage status Y. When β j >0, X jThe greater the value, the higher the probability of damage, β j When < 0, X j The greater the value, the lower the probability of damage. If β j = 0, it means that X j has no significant effect on damage and needs to be excluded. For linearly related data, correlation analysis is performed based on the Pearson correlation coefficient. Its algorithm formula is:

[0093]

[0094] where r represents the correlation coefficient, represents the mean value of parameter X j The value of j ranges from 1 to 5, representing the data of passenger flow, train trips, rainfall, temperature, and humidity respectively.

[0095] For non-linearly related data, correlation analysis is performed through the Spearman rank correlation coefficient:

[0096]

[0097] where ρ represents the rank correlation coefficient, d represents the rank difference between parameter X j and Y, and n represents the sample size. Parameters with |ρ|, |r| > 0.3 are retained for multivariate feature selection.

[0098] Example 3:

[0099] The facility management and maintenance module calculates the abnormal influence parameter range of the associated influence parameter according to the structure damage associated influence parameter, analyzes the influence data of the remaining stations in the same group as the structure damage occurrence station, determines the similarity of the station influence parameters, marks the similar potential danger points, and manages and maintains the rail transit infrastructure. The specific steps are as follows:

[0100] For the obtained structure damage associated influence parameter, the associated influence parameter range of the current structure damage is obtained through univariate anomaly detection and multivariate anomaly detection. The specific steps are as follows:

[0101] When the structure damage associated influence parameter is univariate, the structure damage associated influence parameter range is determined based on the 3σ principle;

[0102] When the structure damage associated influence parameter is multivariate, the structure damage associated influence parameter range is determined through Mahalanobis distance. The specific steps are as follows:

[0103] D 2 =(G - μ)∑ -1 (G - μ) T ;

[0104] Among them, G represents a vector containing multiple structure damage associated influence parameters, μ represents the mean vector of multiple parameters, that is, the average value of each parameter under normal conditions, and ∑ -1 represents the inverse matrix of the covariance matrix. When , it is determined that the multiple parameters are abnormal, where represents the critical value of the chi-square distribution, p is the number of parameters, and 0.95 is the confidence level.

[0105] It should be noted that G represents a vector containing multiple structure damage associated influence parameters. The maximum case of G is G = [X1, X2, X3, X4, X5]. The normal range of the structure damage associated influence parameters obtained by the 3σ principle is [μ - 3σ, μ + 3σ].

[0106] According to the current damage report, determine the associated sub-file information in the current damage site sub-file in the rail transit infrastructure management database, extract the structure damage associated influence parameters in the associated sub-file, determine the similarity of the associated influence parameters in the associated sub-file, and mark the similar potential dangerous points in the rail transit infrastructure management database, and conduct management and maintenance on the rail transit infrastructure. The specific steps are as follows:

[0107] For the type of structure damage associated influence parameters of the current site sub-file, extract the corresponding associated influence parameter data records in the associated sub-file, and determine the potential dangerous points of the site in the associated sub-file:

[0108] When the structure damage associated influence parameter is a single variable, count the structure damage associated influence parameter data in the associated sub-file. When there is any day in the associated sub-file where the structure damage associated influence parameter data is not within the 3σ range, it means that there is a potential infrastructure abnormality in the current associated sub-file. Red mark the relevant infrastructure in the associated sub-file and mark the status as to be maintained in the rail transit infrastructure management database;

[0109] When the structure damage associated influence parameter is a multi-variable, extract the corresponding associated influence parameter data records in the associated sub-file and calculate D 2 , when there is any day in the historical dates , it means that there is a potential infrastructure abnormality in the current associated sub-file. Red mark the relevant infrastructure in the associated sub-file and mark the status as to be maintained in the rail transit infrastructure management database.

[0110] It should be noted that by classifying the sites in the similar environment, based on the structure damage associated influence parameters of a certain site in the category, determine the potential abnormal conditions of the remaining sites, and realize the preventive maintenance of the rail transit infrastructure.

[0111] In the embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of modules is only a logical function division, and there can be other division methods in actual implementation; the modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the method of this embodiment.

[0112] The above embodiments are only used to illustrate the technical methods of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical methods of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. An integrated data management platform for the construction, management, and maintenance of rail transit infrastructure, characterized in that: It includes a facility parameter entry module, a data collection module, a damage reporting module, a grouping planning module, a damage analysis module, and a facility management and maintenance module; The facility parameter entry module establishes a rail transit infrastructure management database through MySQL, and creates city archives according to the cities where the rail transit is located. Inside the city archives, sub-archives are created according to the names of rail transit stations under the jurisdiction of the city, and the basic parameters of the rail transit infrastructure at each station are entered, including track facility parameters and station facility parameters; The data collection module includes a passenger flow sensor, which collects meteorological data in different station areas. According to the train operation timetables of the stations, it obtains the passenger flow, train numbers, rainfall, temperature, and humidity data of different stations, and conducts daily, weekly, monthly, quarterly, and annual statistics on the collected rainfall, temperature, humidity, passenger flow, and train number data, calculates the total value, average value, maximum value, minimum value, and peak value within a unit time of each item of data, and records them respectively inside the corresponding station sub-archives; The damage reporting module integrates the parameters of the damaged facilities and the impact data of the damaged location based on the reported damage to the rail transit infrastructure, and transmits them to the damage analysis module; For the reported damage to the rail transit infrastructure, the damage analysis module combines the impact parameters of the damaged location to qualitatively determine the damage type, including accidental damage and structural damage. It does not record the impact parameters of the damaged location for accidental damage, records the impact parameters of the damaged location for structural damage and analyzes the cause of the damage, screens the impact parameters related to structural damage, and transmits them to the subsequent module; According to the damage report transmitted by the damage reporting module, based on the characteristics and relevant information of the damage, qualitatively determine the damage type in the current damage report, and determine whether the damage is accidental damage or structural damage; For the qualitatively determined structural damage, record the impact parameters of the damaged location of the structural damage and analyze the cause of the damage, screen the impact parameters related to structural damage in the current damage report, and transmit them to the subsequent module; The grouping planning module divides the rail transit stations that use the same specification infrastructure and are in the same type of climate zone into the same group according to the basic parameters of the rail transit infrastructure and in combination with the geographical location climate zone. Through geographic information system technology, it obtains the climate zone information of the locations where each rail transit station is located, and based on the facility basic parameter information of different sub-archives in the rail transit infrastructure management database, conducts infrastructure matching grouping for the sub-archives. The specific steps are as follows: For categorical parameters, classify the station sub-archives that adopt the same model and material infrastructure into the same category archive, and further group the category archive by performing error matching on the numerical parameters inside the station sub-archives in the same category archive. The steps are as follows: Extract the numerical parameters of each station sub-archive in the same category archive, define one of the station sub-archives as the target sub-archive, and calculate the parameter matching degrees between the target sub-archive and other sub-archives respectively: Among them, x and y respectively represent the same numerical parameter values of the target site sub-file and the remaining site sub-files in the same category file, θ represents the error range, and m represents the matching degree; Based on the parameter matching degrees of each parameter, the overall parameter matching degree is calculated: In the above formula, n represents the number of facility parameters, and m i represents the matching degree of the i-th parameter, and ω i represents the corresponding weight, and M represents the overall parameter matching degree. When M ≥ λ, it means that the matching degree of the numerical parameters in the two site sub-files meets the requirements, and the files that meet the matching degree requirements with the target sub-file are recorded inside the target sub-file. When M < λ, it means that the matching degree of the numerical parameters in the two site sub-files does not meet the requirements and no recording is made; For the remaining sub-files recorded in each target sub-file, based on the climate zone information of the current target sub-file, the remaining sub-files with the same climate zone as the target sub-file are retained to obtain the associated sub-files of the current target sub-file; The facility management and maintenance module calculates the abnormal influence parameter range of the associated influence parameters according to the structure damage associated influence parameters, analyzes the influence data of the remaining sites in the same group as the site where the structure damage occurs, determines the similarity of the site influence parameters, marks the similar potential dangerous points, and manages and maintains the rail transit infrastructure.

2. The integrated data management platform for construction, management, and maintenance of rail transit infrastructure according to claim 1, characterized in that: The damage reporting module receives the rail transit infrastructure damage information reported from manual and automatic channels. The damage information includes the damage occurrence site information, the type of damaged infrastructure, the damage occurrence time, whether it involves human damage, and the damage location image. The influence data of the reported damage location is retrieved through the rail transit infrastructure management database and combined with the damage information to form a damage report, which is transmitted into the damage analysis module.

3. The integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to claim 1, wherein: According to the damage report transmitted by the damage reporting module, based on the characteristics and relevant information of the damage, the damage type in the current damage report is qualitatively determined to determine whether the damage is accidental damage or structural damage, including the following steps: When the damage report involves human damage, the damage type in the current damage report is qualitatively determined as accidental damage; When the damage report does not involve human damage, based on the type of damaged infrastructure, the damage type in the current damage report is qualitatively determined. The specific steps are as follows: Collect historical damage image data of different rail transit infrastructures, and label the image data as structural damage or accidental damage. Adjust the images to the same size, normalize the pixel values to the range of [0, 1]. The processed historical damage image data of different rail transit infrastructures are divided into training sets, validation sets, and test sets according to the infrastructure type, and the data ratios are 70%, 15%, and 15% respectively. A convolutional neural network model is constructed through the training set, the model hyperparameters are adjusted through the validation set, and the final performance of the model is evaluated through the test set; According to the type of damaged infrastructure in the damage report, the damage location image data is input into the CNN determination model under the corresponding infrastructure type, and the determination result is output to determine whether the damage type in the current damage report is accidental damage or structural damage.

4. The integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to claim 1, characterized in that: For the qualitatively determined structural damage, record the influence parameters of the damage reporting location of the structural damage and analyze the cause of the damage, screen the structure damage associated influence parameters in the current damage report, and transmit them to the subsequent module, including the following steps: For structural damage, obtain the impact data during the damage occurrence at the damaged location through a damage report, including the number of people, train trips, rainfall, temperature, and humidity data within the damaged site on the day of damage occurrence, as well as the number of people, train trips, rainfall, temperature, and humidity data of the current site on historical normal dates, and obtain the number of historical damage status record samples of the same infrastructure at the current site, where the number of people, train trips, rainfall, temperature, and humidity data are X1, X2, X3, X4, X5 respectively, and the number of historical damage status record samples is Y; Standardize the data of the number of people, train trips, rainfall, temperature, and humidity through the Z-score, screen the damage-related parameters through the Pearson correlation coefficient and the Spearman rank correlation coefficient, and perform multivariate feature selection: In the above formula, Y i represents the damage state of the i-th sample, β0 is the intercept term, and β j represents the regression coefficient of the j-th parameter, is the regularization parameter, and the parameters with β j ≠ 0 are retained as the structure damage correlation influence parameters, and X ij represents the j-th parameter value of the i-th sample.

5. The integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to claim 4, characterized in that: The facility management and maintenance module calculates the abnormal impact parameter range of the associated impact parameters according to the structural damage-associated impact parameters, analyzes the impact data of the remaining sites in the same group as the site where the structural damage occurs, determines the similarity of the site impact parameters, marks the similar potential hazard points, and manages and maintains the rail transit infrastructure. The specific steps are as follows: For the obtained structural damage-associated impact parameters, obtain the associated impact parameter range of the current structural damage through univariate anomaly detection and multivariate anomaly detection; According to the current damage report, determine the associated sub-file information in the sub-file of the current damaged site in the rail transit infrastructure management database, extract the structural damage-associated impact parameters in the associated sub-file, determine the similarity of the associated impact parameters in the associated sub-file, mark the similar potential hazard points in the rail transit infrastructure management database, and manage and maintain the rail transit infrastructure.

6. The integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to claim 5, characterized in that: The step of obtaining the associated impact parameter range of the current structural damage through univariate anomaly detection and multivariate anomaly detection for the obtained structural damage-associated impact parameters is as follows: When the structural damage-associated impact parameter is univariate, determine the structural damage-associated impact parameter range based on the 3σ principle; When the structural damage-associated impact parameter is multivariate, determine the structural damage-associated impact parameter range through the Mahalanobis distance. The specific steps are as follows: D 2 = (G - μ) ∑ -1 (G - μ) T ; Among them, G represents a vector containing multiple structure damage associated influence parameters, μ represents the mean vector of multiple parameters, that is, the average value of each parameter under normal conditions, and ∑ -1 represents the inverse matrix of the covariance matrix. When holds, it is determined that the multiple parameters are abnormal, where represents the critical value of the chi-square distribution, p is the number of parameters, and 0.95 is the confidence level.

7. The integrated data management platform for the construction, management, and maintenance of rail transit infrastructure according to claim 6, characterized in that: The step of determining the associated sub-file information in the sub-file of the current damaged site in the rail transit infrastructure management database according to the current damage report, extracting the structural damage-associated impact parameters in the associated sub-file, determining the similarity of the associated impact parameters in the associated sub-file, marking the similar potential hazard points in the rail transit infrastructure management database, and managing and maintaining the rail transit infrastructure is as follows: For the type of structural damage-associated impact parameters in the current site sub-file, extract the corresponding associated impact parameter data records in the associated sub-file, and determine the potential hazard points of the sites in the associated sub-file: When the structure damage associated influence parameter is a single variable, the structure damage associated influence parameter data in the statistical association sub-file is counted. When there is structure damage associated influence parameter data in any daily association sub-file that is not within the 3σ range, it indicates that there is a potential infrastructure anomaly in the current association sub-file. The relevant infrastructure in the association sub-file is marked in red, and the status is marked as to be maintained in the rail transit infrastructure management database; When the structure damage correlation influence parameter is a multi-variable, extract the corresponding correlation influence parameter data record in the correlation sub-file, and calculate D 2 , when there is any day in the historical date , it represents that there is a potential infrastructure anomaly in the current correlation sub-file. Red-mark the relevant infrastructure in the correlation sub-file, and mark the status as to be maintained in the rail transit infrastructure management database.

Images