Network intelligent center based on rail transit large model data analysis

Abstract

The invention belongs to the technical field of data analysis, and discloses a network intelligent center based on rail transit large model data analysis. And the intelligent center comprises an equipment state prediction module, a toughness decision construction module, a resource dynamic balance module, a causal reasoning mechanism module and a man-machine mixed evaluation module, analyzes the feature vectors to obtain a track safety prediction value for quantifying and feeding back track safety, processes the track safety prediction value to obtain a track safety report, and provides the track safety report to the human-machine mixed evaluation module. Rail operation related energy consumption output is adjusted according to passenger flow data, an energy-saving optimization report is obtained, traceability decision-making reasons are reversely traced based on a rail safety report, a traceability explanation report is obtained in combination with the rail safety report, and a work order needs to be manually confirmed for output. The method has the remarkable advantages that the safety guarantee degree is high, the safety guarantee strengthening capacity is high, and the rail operation economic effect is good.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and more particularly to a network intelligent center based on rail transit large-scale model data analysis. Background Art

[0002] Rail transit refers to a type of transportation vehicle or transportation system in which operating vehicles need to travel on specific tracks. Thanks to the rapid development of science and technology, more and more rail transit departments choose to use large model data analysis instead of traditional manual data analysis.

[0003] The patent application with publication number CN119911306A discloses a large-scale model-based outdoor switch status monitoring and analysis system. By deeply analyzing multi-source heterogeneous data, it effectively improves the monitoring accuracy and fault prediction capabilities. Compared with traditional methods, the large-scale model can more accurately identify the switch status, predict faults in advance and adapt to different working conditions. In addition, combined with the Internet of Things, edge computing and digital twin technology, the large-scale model realizes efficient and intelligent switch monitoring and management, and optimizes the accuracy and timeliness of fault diagnosis and prediction.

[0004] However, the above-mentioned large-scale model-based outdoor switch state monitoring and analysis system, although it has achieved the purpose of efficient and intelligent switch monitoring and management, and optimized the accuracy and timeliness of fault diagnosis and prediction to a certain extent by conducting in-depth analysis of multi-source heterogeneous data based on large models, in the process of using rail transit large models, there is still a reliance on rule engines or shallow machine learning, which makes the large model lack the ability to adapt to complex scenarios, and therefore it is difficult to effectively ensure the safety of rail transit. At the same time, traditional rail transit energy-saving solutions rely on fixed strategies, such as adjusting rail transit operations according to peak and valley electricity prices, which makes it difficult to effectively balance the relative contradiction between demand and energy saving, resulting in a lack of flexibility in adjustment measures. Moreover, the black box model is prone to false alarms and missed alarms in the system, which leads to low staff trust in the decisions output by AI, and further leads to poor economic efficiency of the rail transit large model.

[0005] In view of this, the present invention proposes a network intelligent center based on rail transit large model data analysis to solve the above problems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solutions, including: Safety data acquisition module, used to collect safety data sets, including track vibration data, power supply stability data, bearing temperature data and traction energy consumption data; Multi-source data acquisition module, used to collect multi-source data sets, including comprehensive meteorological data, geological disaster data and passenger flow data; The data interaction and fusion module is used to preprocess the security data set and multi-source data set, extract feature data, and obtain feature vectors; Furthermore, the steps of preprocessing the security dataset and the multi-source dataset and extracting feature data include: Q1: Complete data cleaning of all sub-data items in the basic data set by removing outliers, and normalize all sub-data items in the basic data set to the range of [0, 1] according to the normalization formula; Q2: Substitute the track vibration data into the calculation formula: , get vibration characteristic data ,in, is the number of sampling points, For the Track vibration data of sampling points ; Q3: Stability feature extraction is performed based on power supply stability data. The specific calculation formula for stability feature extraction is: ; Get stable characteristic data ,in, Power supply stability data The standard deviation of is a constant term; Q4: By calculating the change rate of bearing temperature data, the specific formula is: ; Get change feature data ,in, Current time Bearing temperature data, For the last time Bearing temperature data, is the time interval; Q5: By calculating the traction energy consumption data The linear regression slope of ; Q6: Pack vibration feature data, stability feature data, change feature data, trend feature data, comprehensive meteorological data and geological disaster data to obtain feature vectors; The equipment status prediction module is used to analyze the characteristic vector to obtain the track safety prediction value for quantification and feedback of track safety; Furthermore, the step of analyzing the feature vector includes: Step 1: Obtain a set of historical feature vectors stored in the database, compare them with the current time based on the timestamp, and group the comparison results into corresponding groups from small to large. The labeled results are L1, L2, L3, ..., Ln, and the labeled results are used as the sample set; Step 2: Divide the sample set into a 70% training set, a 15% test set, and a 15% validation set, and establish a track safety prediction model based on the sample set; Step 3: Based on the eigenvector, perform polynomial expansion and interactive processing on the eigenvector. The specific calculation formula group for the processing is: ; Get the first interaction data respectively , second interaction data and third-party interaction data ,in, is the hyperbolic tangent function, 、 、 、 、 、 and is the feature enhancement weight parameter, Comprehensive meteorological data, For geological disaster data, is an exponential function; Step 4: Based on the first interaction data, the second interaction data, and the third interaction data in step 3, a preliminary prediction is made according to the multiple heterogeneous sub-model. The specific calculation formula for the prediction is: ; Get the first prediction value respectively , the second predicted value and the third predicted value ,in, is the sliding window size, is the historical forecast value attenuation factor, is the historical track safety prediction value, is the Sigmoid activation function, is the interaction vector, For the The time series model parameters of the interaction vector, For past time points No. interaction vectors, is the number of decision trees, No. The weight factor of each tree, is the number of leaf nodes in each tree, For the The first The output value of the leaf node, is the indicator function, For the The first The feature space division of leaf nodes, is the transpose of the weight matrix of the attention output layer, is the normalized exponential function, 、 and are the projection matrices for query, key, and value, respectively, is the scaling factor of the key vector; Step 5: Perform integrated re-verification on the first predicted value, the second predicted value, and the third predicted value in step 4. The specific calculation formula for the integrated re-verification is: ; Get the fourth prediction value ,in, 、 and is the sub-model weight factor, 、 and To verify the indicator function; Step 6: Based on the fourth prediction value in step 5, calculate the track safety prediction value. The specific calculation formula is: ; Get track safety prediction value ,in, is the prediction result weight factor, is the size of the historical lookback window, is the time decay weight factor, For historical prediction time points Track safety prediction value; Step 7: Output the track safety prediction value to the resilience decision building module; The resilience decision building module is used to process the track safety prediction value and obtain the track safety report; Furthermore, the track safety prediction value is processed in the following ways: Based on the safety threshold interval (R1, R2); When the track safety prediction value is greater than or equal to R2, a safety report is generated; when the track safety prediction value is greater than or equal to R1 and less than R2, a warning report and a speed reduction instruction are generated; when the track safety prediction value is less than R1, a danger report and a stop operation instruction are generated; The safety report includes a statement that the system predicts a low level of track safety risk and asks staff to conduct routine monitoring according to established operating procedures; The warning report includes a speed reduction instruction and explains the system's predicted track safety risk level. Please check whether the safety-related data is normal as soon as possible. The hazard report includes a call to stop operations and a statement stating that the predicted track safety risk is high and that personnel are requested to immediately conduct an on-site inspection. The speed reduction command includes a set of characters representing reducing the speed of the designated train to 90% of the current speed; The stop running instruction contains a group of characters representing that the running speed of the designated train is reset to 0; Pack safety reports, warning reports and hazard reports to obtain track safety reports; The resource dynamic balance module is used to adjust the energy consumption output related to rail operation based on passenger flow data and obtain energy-saving optimization reports; Furthermore, methods for adjusting rail operation-related energy consumption output based on passenger flow data include: Based on passenger flow threshold; When the passenger flow data is less than the passenger flow threshold, an energy-saving optimization report is generated; The energy saving optimization report includes an energy saving directive; The energy-saving instructions include a set of instructions that reduce the train's traction power by 20% and reduce the lighting brightness in the train area by 30%; The causal reasoning mechanism module is used to trace back the decision-making causes based on the track safety report and obtain a traceability explanation report in combination with the track safety report; Furthermore, the methods of tracing back the decision-making reasons based on the track safety report and combining the track safety report include: Based on track safety reports and feature vectors; When the track safety report is a warning report or a danger report, the feature vector is identified. The specific formula for identification is: ; Obtain causal features ,in, Maximize the function, is an index symbol and contains all sub-data items in the eigenvector, For the Item feature weight factor, For the Feature data in feature vectors; Integrate track safety reports, causal characteristics, and causal rules to obtain a traceability explanation report; The human-machine hybrid assessment module is used to identify the content of the traceability explanation report, analyze it based on the identification results, and output a work order that requires manual confirmation; Furthermore, methods for identifying the content of the traceability explanation report and analyzing it based on the identification results include: When there is a hazard report in the traceability explanation report, the manual confirmation mechanism is triggered and a work order requiring manual confirmation is generated. The manual confirmation mechanism includes displaying the work order requiring manual confirmation through a visual panel; Work orders that require manual confirmation include indicating whether to execute the stop operation instruction; When manually confirming the execution of the stop operation instruction, a confirmation execution instruction is generated; When the manual confirmation is not to execute the stop operation instruction, a cancel execution instruction is generated; Confirm that the execution instruction contains a set of characters representing a stop execution instruction for retaining the hazard report; The cancel execution instruction contains a group of characters representing the instruction to stop running in the elimination danger report; Data processing and output module, used to display the traceability explanation report through the visual panel, process the traceability explanation report and energy-saving optimization report, integrate system-related data and store them in the database; Furthermore, the methods for processing the traceability explanation report and energy-saving optimization report include: Identify traceability explanation reports and energy-saving optimization reports; When there is a speed reduction instruction, a stop operation instruction or an energy-saving instruction, the speed reduction instruction, the stop operation instruction or the energy-saving instruction is sent to the intelligent controller; Collect the security data set, multi-source data set, feature vector, traceability explanation report, energy-saving optimization report and manual confirmation return sheet to obtain the system data set, and store the system data set in the database; Furthermore, S1: collecting a safety data set, the safety data set including track vibration data, power supply stability data, bearing temperature data and traction energy consumption data; S2: Collect multi-source data sets, including comprehensive meteorological data, geological disaster data, and passenger flow data; S3: Preprocess the security dataset and multi-source dataset, extract feature data, and obtain feature vectors; S4: Analyze the characteristic vector to obtain the track safety prediction value that quantifies and feeds back the track safety; S5: Processing the track safety prediction value to obtain a track safety report; S6: Adjust the energy consumption output related to rail operation based on passenger flow data to obtain an energy-saving optimization report; S7: Based on the track safety report, reverse trace the decision-making reasons, and combine the track safety report to obtain a traceability explanation report; S8: Identify the content of the traceability explanation report, analyze it based on the identification results, and output a work order that requires manual confirmation; S9: Display the traceability explanation report through the visual panel, process the traceability explanation report and energy-saving optimization report, integrate system-related data and store them in the database.

[0007] The technical effects and advantages of the network intelligent center based on rail transit large model data analysis of the present invention are as follows: The present invention collects safety data sets, including track vibration data, power supply stability data, bearing temperature data and traction energy consumption data, collects multi-source data sets, including comprehensive meteorological data, geological disaster data and passenger flow data, pre-processes the safety data sets and multi-source data sets, extracts feature data to obtain feature vectors, analyzes the feature vectors, obtains track safety prediction values ​​that quantify and feedback track safety, processes the track safety prediction values ​​to obtain a track safety report, adjusts track operation-related energy consumption output according to passenger flow data, obtains an energy-saving optimization report, traces the decision-making reasons based on the track safety report, and obtains a tracing explanation report in combination with the track safety report, identifies the content of the tracing explanation report, analyzes it based on the identification results, outputs a work order that requires manual confirmation, displays the tracing explanation report through a visual panel, processes the tracing explanation report and the energy-saving optimization report, integrates system-related data and stores them in a database, so that the system can be based on multi-source data. Accurately quantifying and predicting the degree of rail transit safety risks has greatly improved the safety of rail transit operations and people's lives and property. In addition, the present invention also traces the main factors of the decision report output by the system and integrates them into the decision report, which not only effectively improves the staff's ability to respond quickly to safety risk factors, but also greatly improves the staff's trust in AI decisions, thereby further strengthening the safety of rail transit. At the same time, through the manual review mechanism implemented for key instructions, the impact of the system's false alarm output on large-scale decision-making can be effectively reduced, thereby enabling the system to have multiple safety assurance capabilities. Moreover, by dynamically adjusting the train traction strategy and station energy consumption output based on real-time passenger flow, the system can effectively achieve the goal of coordinated optimization of energy and capacity, thereby greatly reducing the relative contradiction between demand and energy saving. Overall, the present invention has the significant advantages of high safety assurance, strong safety assurance enhancement capability and good economic effect of rail operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A schematic diagram of a network intelligent center based on rail transit large model data analysis according to the present invention; Figure 2 Schematic diagram of a network method based on rail transit large model data analysis of the present invention. DETAILED DESCRIPTION

[0009] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0010] The terms used in the embodiments of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The singular forms "a," "an," "the," and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, and unless the context clearly indicates otherwise, "a plurality" generally includes at least two.

[0011] As used herein, the words "if" and "if" may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to the determination" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)," depending on the context.

[0012] In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.

[0013] In practice, the server-side devices deployed by the network intelligence center based on large-scale rail transit data analysis may be composed of one or more devices. The aforementioned network intelligence center based on large-scale rail transit data analysis can be implemented as: a service instance, a virtual machine, and hardware devices. For example, the network intelligence center based on large-scale rail transit data analysis can be implemented as a service instance deployed on one or more devices in a cloud node. Simply put, the network intelligence center based on large-scale rail transit data analysis can be understood as software deployed on a cloud node that provides a network intelligence center based on large-scale rail transit data analysis for each user terminal. Alternatively, the network intelligence center based on large-scale rail transit data analysis can be implemented as a virtual machine deployed on one or more devices in a cloud node. Application software for managing each user terminal is installed in the virtual machine. Alternatively, the network intelligence center based on large-scale rail transit data analysis can be implemented as a server-side device composed of numerous hardware devices of the same or different types, with one or more hardware devices configured to provide a network intelligence center based on large-scale rail transit data analysis for each user terminal.

[0014] In terms of implementation, the network intelligence center based on large-scale rail transit data analysis and the user end are mutually compatible. Specifically, the network intelligence center based on large-scale rail transit data analysis is an application installed on a cloud service platform, and the user end is a client that establishes a communication connection with the application. Alternatively, the network intelligence center based on large-scale rail transit data analysis is implemented as a website, and the user end is implemented as a webpage. Alternatively, the network intelligence center based on large-scale rail transit data analysis is implemented as a cloud service platform, and the user end is implemented as a mini-program within an instant messaging application.

[0015] like Figure 1 1 is a system architecture diagram of a network intelligence center based on rail transit large model data analysis provided by one embodiment of the present invention.

[0016] The network intelligence center based on the analysis of rail transit large model data described in the present invention can be set in a cloud server. In terms of implementation, it can be used as one or more service devices, or it can be installed as an application on the cloud (such as the server of the mobile service operator, server cluster, etc.), or it can be developed as a website. According to the functions implemented, the network intelligence center based on the analysis of rail transit large model data can include a security data acquisition module, a multi-source data acquisition module, a data interaction fusion module, an equipment status prediction module, a resilience decision-making construction module, a resource dynamic balance module, a causal reasoning mechanism module, a human-machine hybrid evaluation module and a data processing output module. The module described in the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0017] In an embodiment of the present invention, in a network intelligent center based on rail transit large-scale data analysis, each of the above modules can be independently implemented and called with other modules. The call here can be understood as a module that can connect to multiple modules of another type and provide corresponding services to the multiple modules connected to it. For example, the sharing evaluation module can call the same information acquisition module to obtain the information collected by the information acquisition module. Based on the above characteristics, in a network intelligent center based on rail transit large-scale data analysis provided by an embodiment of the present invention, the scope of application of the network intelligent center architecture based on rail transit large-scale data analysis can be adjusted by adding modules and directly calling them without modifying the program code, thereby realizing cluster-type horizontal expansion, so as to achieve the purpose of quickly and flexibly expanding the network intelligent center based on rail transit large-scale data analysis. In actual applications, the above modules can be set in the same device or different devices, or they can be set in virtual devices, such as service instances in cloud servers.

[0018] Example 1, please refer to Figure 1As shown, the network intelligent center based on rail transit large model data analysis described in this embodiment includes: The safety data acquisition module is used to collect safety data sets, including track vibration data, power supply stability data, bearing temperature data and traction energy consumption data; It should be explained that the vertical vibration amplitude of the specified track is collected through the piezoelectric acceleration sensor to obtain the track vibration data; the voltage fluctuation rate of the specified power supply grid is collected through the power quality analyzer to obtain the power supply stability data; the surface temperature of the specified train bearing is collected through the infrared thermal imager array to obtain the bearing temperature data; and the traction energy consumption of the specified train is collected through the smart meter to obtain the traction energy consumption data. The multi-source data acquisition module is used to collect multi-source data sets, which include comprehensive meteorological data, geological disaster data and passenger flow data; It should be explained that the micro-meteorological station collects the comprehensive meteorological risk index in the designated area to obtain comprehensive meteorological data. The comprehensive meteorological risk index refers to, for example, the weighted sum of rainfall, wind speed, and visibility divided by the corresponding standard values. The geological displacement sensor collects the displacement rate of the slope at the designated tunnel entrance to obtain geological disaster data. The binocular passenger flow counting camera collects the real-time passenger flow values ​​of the stations in the designated area to obtain passenger flow data. The data interaction and fusion module is used to preprocess the security data set and the multi-source data set, and extract feature data to obtain feature vectors; Furthermore, the steps of preprocessing the security dataset and the multi-source dataset and extracting feature data include: Q1: Complete data cleaning of all sub-data items in the basic data set by removing outliers, and normalize all sub-data items in the basic data set to the range of [0, 1] according to the normalization formula; It should be explained that removing outliers means, for example, that the bearing temperature data is negative; the basic data set includes the security data set and the multi-source data set; the specific expression formula of the normalization formula is: ,in is the normalized value, Any sub-data item of the basic data, is the historical maximum value of any sub-data item, is the historical minimum value of any sub-data item; normalization is used to eliminate the dimensions of all sub-data items in the basic data set; Q2: Substitute the track vibration data into the calculation formula: , get vibration characteristic data ,in, is the number of sampling points, For the Track vibration data of sampling points ; Q3: Stability feature extraction is performed based on power supply stability data. The specific calculation formula for stability feature extraction is: ; Get stable characteristic data ,in, Power supply stability data The standard deviation of is a constant term; Q4: By calculating the change rate of bearing temperature data, the specific formula is: ; Get change feature data ,in, Current time Bearing temperature data, For the last time Bearing temperature data, is the time interval; Q5: By calculating the traction energy consumption data The linear regression slope of ; Q6: Pack vibration feature data, stability feature data, change feature data, trend feature data, comprehensive meteorological data and geological disaster data to obtain feature vectors; It should be explained that the sub-data items used in the basic data involved in steps Q2 to Q6 and all subsequent modules (except the resource dynamic balance module) are all data values ​​that have been cleaned and normalized; The equipment status prediction module is used to analyze the characteristic vector to obtain a track safety prediction value that quantifies and feeds back track safety; Furthermore, the steps of analyzing the feature vector include: Step 1: Obtain a set of historical feature vectors stored in the database, compare them with the current time based on the timestamp, and group the comparison results into corresponding groups from small to large. The labeled results are L1, L2, L3, ..., Ln, and the labeled results are used as the sample set; Step 2: Divide the sample set into a 70% training set, a 15% test set, and a 15% validation set, and establish a track safety prediction model based on the sample set; Step 3: Based on the eigenvector, perform polynomial expansion and interactive processing on the eigenvector. The specific calculation formula group for the processing is: ; Get the first interaction data respectively , second interaction data and third-party interaction data ,in, is the hyperbolic tangent function, 、 and is the feature enhancement weight parameter, Comprehensive meteorological data, For geological disaster data, is an exponential function; It should be explained that the hyperbolic tangent function is used to constrain the output value in the brackets to be within the range of [-1, 1]; the exponential function is used to convert the output result of the calculation formula in the brackets into a probability value [0, 1]; Step 4: Based on the first interaction data, the second interaction data, and the third interaction data in step 3, a preliminary prediction is made according to the multiple heterogeneous sub-model. The specific calculation formula for the prediction is: ; Get the first prediction value respectively , the second predicted value and the third predicted value ,in, is the sliding window size, is the historical forecast value attenuation factor, is the historical track safety prediction value, is the Sigmoid activation function, is the interaction vector, For the The time series model parameters of the interaction vector, For past time points No. interaction vectors, is the number of decision trees, No. The weight factor of each tree, is the number of leaf nodes in each tree, For the The first The output value of the leaf node, is the indicator function, For the The first The feature space division of leaf nodes, is the transpose of the weight matrix of the attention output layer, is the normalized exponential function, 、 and are the projection matrices for query, key, and value, respectively, is the scaling factor of the key vector; It should be explained that the Sigmoid activation function is used to constrain the output value in the adjacent brackets to be within the range of (0, 1); the interaction vector is composed of the first interaction data, the second interaction data, and the third interaction data set; the indicator function is used to determine whether the conditions in the adjacent brackets are met, and when the conditions are met, the output value is set to 1; the normalized exponential function is used to convert each element in the adjacent brackets into a non-negative number and normalize these non-negative numbers so that the sum of all non-negative numbers is 1; Step 5: Perform integrated re-verification on the first predicted value, the second predicted value, and the third predicted value in step 4. The specific calculation formula for the integrated re-verification is: ; Get the fourth prediction value ,in, 、 and is the sub-model weight factor, 、 and To verify the indicator function; It should be explained that the verification indicator function is used to determine whether the predicted value is valid within the historical confidence interval. The historical confidence interval is , when the judgment is valid, the output value of the verification indicator function is 1, otherwise it is 0, where, is an index symbol, including 1, 2 and 3; Step 6: Based on the fourth prediction value in step 5, calculate the track safety prediction value. The specific calculation formula is: ; Get track safety prediction value ,in, is the prediction result weight factor, is the size of the historical lookback window, is the time decay weight factor, For historical prediction time points Track safety prediction value; Step 7: Output the track safety prediction value to the resilience decision building module; The resilience decision building module is used to process the track safety prediction value to obtain a track safety report; Furthermore, the track safety prediction value is processed in the following ways: Based on the safety threshold interval (R1, R2); It should be explained that the safety threshold interval is determined manually and output to the system; When the track safety prediction value is greater than or equal to R2, a safety report is generated; when the track safety prediction value is greater than or equal to R1 and less than R2, a warning report and a speed reduction instruction are generated; when the track safety prediction value is less than R1, a danger report and a stop operation instruction are generated; The safety report includes a statement that the system predicts a low level of track safety risk and asks staff to conduct routine monitoring according to established operating procedures; The warning report includes a speed reduction instruction and explains the system's predicted track safety risk level. Please check whether the safety-related data is normal as soon as possible. The hazard report includes a call to stop operations and a statement stating that the predicted track safety risk is high and that personnel are requested to immediately conduct an on-site inspection. The speed reduction command includes a set of characters representing reducing the speed of the designated train to 90% of the current speed; The stop running instruction contains a group of characters representing that the running speed of the designated train is reset to 0; Pack safety reports, warning reports and hazard reports to obtain track safety reports; The resource dynamic balancing module is used to adjust the energy consumption output related to rail operation according to passenger flow data and obtain an energy-saving optimization report; Furthermore, methods for adjusting rail operation-related energy consumption output based on passenger flow data include: Based on passenger flow threshold; It should be explained that the passenger flow threshold is the historical average passenger flow multiplied by 0.5; When the passenger flow data is less than the passenger flow threshold, an energy-saving optimization report is generated; The energy saving optimization report includes an energy saving directive; The energy-saving instructions include a set of instructions that reduce the train's traction power by 20% and reduce the lighting brightness in the train area by 30%; The causal reasoning mechanism module is used to reversely trace the decision-making reasons based on the track safety report and obtain a traceability explanation report in combination with the track safety report; Furthermore, the methods of tracing back the decision-making reasons based on the track safety report and combining the track safety report include: Based on track safety reports and feature vectors; When the track safety report is a warning report or a danger report, the feature vector is identified. The specific formula for identification is: ; Obtain causal features ,in, Maximize the function, is an index symbol and contains all sub-data items in the eigenvector, For the Item feature weight factor, For the Feature data in feature vectors; It should be explained that the maximization function is used to find the maximum value among adjacent calculated values ​​as the output value; Integrate track safety reports, causal characteristics, and causal rules to obtain a traceability explanation report; It needs to be explained that the causal rule means that, for example, if the causal feature is comprehensive meteorological data, the output is: comprehensive meteorological data is the main cause; The human-machine hybrid assessment module is used to identify the content of the traceability explanation report, analyze it based on the identification results, and output a work order that requires manual confirmation; Furthermore, methods for identifying the content of the traceability explanation report and conducting analysis based on the identification results include: When there is a hazard report in the traceability explanation report, the manual confirmation mechanism is triggered and a work order requiring manual confirmation is generated. The manual confirmation mechanism includes displaying the work order requiring manual confirmation through a visual panel; Work orders that require manual confirmation include indicating whether to execute the stop operation instruction; When manually confirming the execution of the stop operation instruction, a confirmation execution instruction is generated; When the manual confirmation is not to execute the stop operation instruction, a cancel execution instruction is generated; Confirm that the execution instruction contains a set of characters representing a stop execution instruction for retaining the hazard report; The cancel execution instruction contains a group of characters representing the instruction to stop running in the elimination danger report; The data processing and output module is used to display the traceability explanation report through a visual panel, process the traceability explanation report and energy-saving optimization report, integrate system-related data and store them in a database; Furthermore, the methods for processing the traceability explanation report and energy-saving optimization report include: Identify traceability explanation reports and energy-saving optimization reports; When there is a speed reduction instruction, a stop operation instruction or an energy-saving instruction, the speed reduction instruction, the stop operation instruction or the energy-saving instruction is sent to the intelligent controller; Collect the security data set, multi-source data set, feature vector, traceability explanation report, energy-saving optimization report and manual confirmation return sheet to obtain the system data set, and store the system data set in the database; This embodiment has the beneficial effects of collecting safety data sets, which include track vibration data, power supply stability data, bearing temperature data and traction energy consumption data, collecting multi-source data sets, which include comprehensive meteorological data, geological disaster data and passenger flow data, pre-processing the safety data sets and multi-source data sets, and extracting feature data to obtain feature vectors, analyzing the feature vectors to obtain track safety prediction values ​​that quantify and feedback track safety, processing the track safety prediction values ​​to obtain a track safety report, adjusting the track operation-related energy consumption output according to passenger flow data to obtain an energy-saving optimization report, tracing the decision reasons based on the track safety report, and obtaining a tracing explanation report in combination with the track safety report, identifying the content of the tracing explanation report, and analyzing it based on the identification results, outputting a work order that requires manual confirmation, displaying the tracing explanation report through a visual panel, processing the tracing explanation report and the energy-saving optimization report, integrating system-related data and storing them in a database, so that the system can be based on multiple The source data accurately quantifies and predicts the degree of safety risk in rail transit, greatly improving the safety of rail transit operation and people's lives and property. In addition, the present invention also traces the main factors of the decision report output by the system and integrates them into the decision report, which not only effectively improves the staff's ability to respond quickly to safety risk factors, but also greatly improves the staff's trust in AI decision-making, thereby further strengthening the safety of rail transit. At the same time, through the manual review mechanism implemented for key instructions, the impact of the system's false alarm output on large-scale decision-making can be effectively reduced, thereby enabling the system to have multiple safety assurance capabilities. Moreover, by dynamically adjusting the train traction strategy and station energy consumption output based on real-time passenger flow, the system can effectively achieve the goal of coordinated optimization of energy and capacity, thereby greatly reducing the relative contradiction between demand and energy saving. Overall, the present invention has the significant advantages of high safety assurance, strong safety assurance enhancement capability and good economic effect of rail operation.

[0019] Example 2, please refer to Figure 2 As shown, for parts not described in detail in this embodiment, please refer to the description of embodiment 1. A network method based on rail transit large model data analysis is provided, the method comprising: S1: collecting a safety data set, the safety data set comprising rail vibration data, power supply stability data, bearing temperature data, and traction energy consumption data; S2: Collect multi-source data sets, including comprehensive meteorological data, geological disaster data, and passenger flow data; S3: Preprocess the security dataset and multi-source dataset, extract feature data, and obtain feature vectors; S4: Analyze the characteristic vector to obtain the track safety prediction value that quantifies and feeds back the track safety; S5: Processing the track safety prediction value to obtain a track safety report; S6: Adjust the energy consumption output related to rail operation based on passenger flow data to obtain an energy-saving optimization report; S7: Based on the track safety report, reverse trace the decision-making reasons, and combine the track safety report to obtain a traceability explanation report; S8: Identify the content of the traceability explanation report, analyze it based on the identification results, and output a work order that requires manual confirmation; S9: Display the traceability explanation report through the visual panel, process the traceability explanation report and energy-saving optimization report, integrate system-related data and store them in the database.

[0020] Example 3 It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0021] Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0022] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0023] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. Terms such as "first" and "second" are used to indicate names and do not imply any particular order.

[0024] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A network intelligence center based on rail transit large model data analysis, characterized in that: The intelligent center includes: an equipment status prediction module, a resilience decision-making module, a resource dynamic balance module, a causal reasoning mechanism module, and a human-machine hybrid evaluation module, wherein: The equipment status prediction module is used to analyze the characteristic vector to obtain a track safety prediction value that quantifies and feeds back track safety; The resilience decision building module is used to process the track safety prediction value to obtain a track safety report; The resource dynamic balancing module is used to adjust the energy consumption output related to rail operation according to passenger flow data and obtain an energy-saving optimization report; The causal reasoning mechanism module is used to reversely trace the decision-making reasons based on the track safety report and obtain a traceability explanation report in combination with the track safety report; The methods of tracing back the decision-making reasons based on the track safety report and combining the track safety report include: Based on track safety reports and feature vectors; When the track safety report is a warning report or a danger report, the feature vector is identified. The specific formula for identification is: ； Obtain causal features ,in, Maximize the function, is an index symbol and contains all sub-data items in the eigenvector, For the Item feature weight factor, For the Feature data in feature vectors; Integrate track safety reports, causal characteristics and causal rules to obtain traceability explanation reports The human-machine hybrid assessment module is used to identify the content of the traceability explanation report, analyze it based on the identification results, and output a work order that requires manual confirmation.

2. The network intelligent center based on rail transit large model data analysis according to claim 1 is characterized in that: The intelligent center also includes: a security data acquisition module, a multi-source data acquisition module, a data interaction and fusion module, and a data processing and output module, wherein: The safety data acquisition module is used to collect safety data sets, including track vibration data, power supply stability data, bearing temperature data and traction energy consumption data; The multi-source data acquisition module is used to collect multi-source data sets, which include comprehensive meteorological data, geological disaster data and passenger flow data; The data interaction and fusion module is used to preprocess the security data set and the multi-source data set, and extract feature data to obtain feature vectors; The data processing and output module is used to display the traceability explanation report through a visual panel, process the traceability explanation report and the energy-saving optimization report, integrate system-related data and store them in a database.

3. The network intelligent center based on rail transit large model data analysis according to claim 2 is characterized in that: The steps for preprocessing the security dataset and multi-source dataset and extracting feature data include: Q1: Complete data cleaning of all sub-data items in the basic data set by removing outliers, and normalize all sub-data items in the basic data set to the range of [0, 1] according to the normalization formula; Q2: Substitute the track vibration data into the calculation formula: , get vibration characteristic data ,in, is the number of sampling points, For the Track vibration data of sampling points ; Q3: Stability feature extraction is performed based on power supply stability data. The specific calculation formula for stability feature extraction is: ； Get stable characteristic data ,in, Power supply stability data The standard deviation of is a constant term; Q4: By calculating the change rate of bearing temperature data, the specific formula is: ； Get change feature data ,in, Current time Bearing temperature data, For the last time Bearing temperature data, is the time interval; Q5: By calculating the traction energy consumption data The linear regression slope of ; Q6: Package vibration feature data, stability feature data, change feature data, trend feature data, comprehensive meteorological data and geological disaster data to obtain feature vectors.

4. The network intelligent center based on rail transit large model data analysis according to claim 1 is characterized in that: The steps to analyze the feature vector include: Step 1: Obtain a set of historical feature vectors stored in the database, compare them with the current time based on the timestamp, and group the comparison results into corresponding groups from small to large. The labeled results are L1, L2, L3, ..., Ln, and the labeled results are used as the sample set; Step 2: Divide the sample set into a 70% training set, a 15% test set, and a 15% validation set, and establish a track safety prediction model based on the sample set; Step 3: Based on the eigenvector, perform polynomial expansion and interactive processing on the eigenvector. The specific calculation formula group for the processing is: ； Get the first interaction data respectively , second interaction data and third-party interaction data ,in, is the hyperbolic tangent function, 、 、 、 、 、 and is the feature enhancement weight parameter, Comprehensive meteorological data, For geological disaster data, is an exponential function; Step 4: Based on the first interaction data, the second interaction data, and the third interaction data in step 3, a preliminary prediction is made according to the multiple heterogeneous sub-model. The specific calculation formula for the prediction is: ； Get the first prediction value respectively , the second predicted value and the third predicted value ,in, is the sliding window size, is the historical forecast value attenuation factor, is the historical track safety prediction value, is the Sigmoid activation function, is the interaction vector, For the The time series model parameters of the interaction vector, For past time points No. interaction vectors, is the number of decision trees, No. The weight factor of each tree, is the number of leaf nodes in each tree, For the The first The output value of the leaf node, is the indicator function, For the The first The feature space division of leaf nodes, is the transpose of the weight matrix of the attention output layer, is the normalized exponential function, 、 and are the projection matrices for query, key, and value, respectively, is the scaling factor of the key vector; Step 5: Perform integrated re-verification on the first predicted value, the second predicted value, and the third predicted value in step 4. The specific calculation formula for the integrated re-verification is: ； Get the fourth predicted value ,in, 、 and is the sub-model weight factor, 、 and To verify the indicator function; Step 6: Based on the fourth prediction value in step 5, calculate the track safety prediction value. The specific calculation formula is: ； Get track safety prediction value ,in, is the prediction result weight factor, is the size of the historical lookback window, is the time decay weight factor, For historical prediction time points Track safety prediction value; Step 7: Output the track safety prediction value to the resilience decision building module.

5. The network intelligent center based on rail transit large model data analysis according to claim 1 is characterized in that: The methods for processing track safety prediction values ​​include: Based on the safety threshold interval (R1, R2); When the track safety prediction value is greater than or equal to R2, a safety report is generated; when the track safety prediction value is greater than or equal to R1 and less than R2, a warning report and a speed reduction instruction are generated; when the track safety prediction value is less than R1, a danger report and a stop operation instruction are generated; The safety report includes a statement that the system predicts a low level of track safety risk and asks staff to conduct routine monitoring according to established operating procedures; The warning report includes a speed reduction instruction and explains the system's predicted track safety risk level. Please check whether the safety-related data is normal as soon as possible. The hazard report includes a call to stop operations and a statement stating that the predicted track safety risk is high and that personnel are requested to immediately conduct an on-site inspection. The speed reduction command includes a set of characters representing reducing the speed of the designated train to 90% of the current speed; The stop running instruction contains a group of characters representing that the running speed of the designated train is reset to 0; Package safety reports, warning reports and hazard reports to get track safety reports.

6. The network intelligent center based on rail transit large model data analysis according to claim 1 is characterized in that: Methods for adjusting rail operation-related energy consumption output based on passenger flow data include: Based on passenger flow threshold; When the passenger flow data is less than the passenger flow threshold, an energy-saving optimization report is generated; The energy saving optimization report includes an energy saving directive; The energy-saving instructions include a set of instructions to reduce the train's traction power by 20% and reduce the lighting brightness in the train area by 30%.

7. The network intelligent center based on rail transit large model data analysis according to claim 1 is characterized in that: Methods for identifying the contents of the traceability interpretation report and conducting analysis based on the identification results include: When there is a hazard report in the traceability explanation report, the manual confirmation mechanism is triggered and a work order requiring manual confirmation is generated. The manual confirmation mechanism includes displaying the work order requiring manual confirmation through a visual panel; Work orders that require manual confirmation include indicating whether to execute the stop operation instruction; When manually confirming the execution of the stop operation instruction, a confirmation execution instruction is generated; When the manual confirmation is not to execute the stop operation instruction, a cancel execution instruction is generated; Confirm that the execution instruction contains a set of characters representing a stop execution instruction for retaining the hazard report; The cancel command contains a set of characters that represents a command to stop the execution of the rejection hazard report.

8. The network intelligent center based on rail transit large model data analysis according to claim 2 is characterized in that: The methods for processing the traceability explanation report and energy-saving optimization report include: Identify traceability explanation reports and energy-saving optimization reports; When there is a speed reduction instruction, a stop operation instruction or an energy-saving instruction, the speed reduction instruction, the stop operation instruction or the energy-saving instruction is sent to the intelligent controller; The security data set, multi-source data set, feature vector, traceability explanation report, energy-saving optimization report and manual confirmation return sheet are collected to obtain the system data set, and the system data set is stored in the database.

9. A network method based on rail transit large model data analysis, implemented by a network intelligent center based on rail transit large model data analysis according to any one of claims 1 to 8, characterized in that: The following steps are included: S1: Collect safety data sets, including track vibration data, power supply stability data, bearing temperature data, and traction energy consumption data; S2: Collect multi-source data sets, including comprehensive meteorological data, geological disaster data, and passenger flow data; S3: Preprocess the security dataset and multi-source dataset, extract feature data, and obtain feature vectors; S4: Analyze the characteristic vector to obtain the track safety prediction value that quantifies and feeds back the track safety; S5: Processing the track safety prediction value to obtain a track safety report; S6: Adjust the energy consumption output related to rail operation based on passenger flow data to obtain an energy-saving optimization report; S7: Based on the track safety report, reverse trace the decision-making reasons, and combine the track safety report to obtain a traceability explanation report; S8: Identify the content of the traceability explanation report, analyze it based on the identification results, and output a work order that requires manual confirmation; S9: Display the traceability explanation report through the visual panel, process the traceability explanation report and energy-saving optimization report, integrate system-related data and store them in the database.

Images