A rail transit vehicle overhaul whole-process comprehensive management system and management method

Abstract

This invention relates to a comprehensive management system and method for the entire maintenance process of rail transit vehicles, belonging to the field of rail transit operation and maintenance and intelligent maintenance technology. The method includes: acquiring rail transit vehicle operation data; performing feature analysis on behavioral interactions during the maintenance process to generate risk pre-assessment parameters; iteratively optimizing the parallel relationships of maintenance tasks and equipment configuration paths to output dynamic maintenance operation plans, and responding to new maintenance tasks in real time based on an event-driven scheduling engine; structuring and unifying the operation data generated throughout the maintenance process to establish a comprehensive maintenance quality archive, tracing back historical testing data sources to obtain preventative maintenance strategies; and constructing a cloud-based maintenance synchronization platform with multi-terminal collaborative access to automatically execute maintenance field verification, and encrypting and archiving the maintenance field verification results to form a full lifecycle maintenance report.

Description

Technical Field

[0001] This invention belongs to the field of rail transit operation and maintenance and intelligent repair technology, specifically involving a comprehensive management system and management method for the entire process of rail transit vehicle maintenance. Background Technology

[0002] With the continuous expansion of the rail transit network and the increasing density of vehicle operation, its maintenance work is characterized by high task complexity, numerous collaborative links, and strong requirements for response time. Traditional rail transit vehicle maintenance management mainly relies on manual planning and scheduling and segmented operation processes. The degree of information exchange between various maintenance links is low, making it difficult to uniformly optimize and manage the parallel execution of multiple tasks. This results in low utilization of maintenance resources and frequent task conflicts and waiting phenomena.

[0003] Existing maintenance management systems typically schedule individual maintenance tasks independently, lacking the ability to integrate and analyze multi-source data throughout the entire maintenance process. They are unable to perform correlation modeling of vehicle operating status, fault evolution trends, and maintenance behavior, making it difficult to identify potential risks and make dynamic adjustments in a timely manner. Furthermore, during maintenance task execution, new maintenance requests or sudden equipment anomalies often require manual readjustment of the plan, resulting in delayed scheduling response and impacting overall maintenance efficiency and safety assurance capabilities.

[0004] Traditional maintenance data management methods primarily rely on static storage, lacking a unified, structured quality archive system. The correlation between maintenance process data, equipment status data, and historical maintenance records is low, hindering the development of full lifecycle traceability analysis. For root cause analysis, reliance is often placed on post-event experience-based judgment, lacking a causal deduction mechanism based on digital models, making it difficult to formulate reusable preventative maintenance strategies.

[0005] At the same time, with the development of cloud computing and Internet of Things technologies, rail transit maintenance is gradually evolving towards cloud collaboration and multi-terminal linkage. However, existing systems still have shortcomings in ensuring consistency of multi-terminal access, visual verification of maintenance processes, and secure data archiving. In particular, in complex maintenance scenarios, it is difficult to achieve real-time synchronization, unified verification, and reliable evidence storage.

[0006] Therefore, it is necessary to propose a comprehensive management method that can realize data fusion analysis, dynamic collaborative scheduling optimization, fault root cause tracing, and cloud-based collaborative verification throughout the entire maintenance process, so as to improve the intelligence level and overall operational efficiency of rail transit vehicle maintenance. Summary of the Invention

[0007] To address the aforementioned problems in the existing technology, this invention provides a comprehensive management method for the entire maintenance process of rail transit vehicles. The objective of this invention can be achieved through the following technical solutions: S1: Obtain rail transit vehicle operation data, perform feature analysis on behavioral interactions during the maintenance process, and generate risk pre-assessment parameters; S2: Based on the risk pre-assessment parameters, the parallel relationship of maintenance tasks and equipment configuration path are iteratively optimized through a multi-objective collaborative decision maintenance model to output a dynamic maintenance operation plan. Based on the event-driven scheduling engine, the execution status of the maintenance equipment is reconstructed in a rolling manner to respond to new maintenance tasks in real time. S3: Execute the dynamic maintenance operation plan, structure and unify the operation data generated throughout the maintenance process, establish a comprehensive maintenance quality archive, and trace back the historical detection data source based on the digital twin backtracking mechanism to perform correlation deduction on the root cause links that cause the fault maintenance event and obtain preventive maintenance strategies. S4: Driven by the goal of visual verification of maintenance, build a cloud-based maintenance synchronization platform with multi-terminal collaborative access, automatically perform maintenance field verification, and encrypt and archive the maintenance field verification results to form a full life cycle maintenance report.

[0008] Specifically, the feature parsing process is as follows: Maintenance tasks are formulated based on rail transit vehicle operation data, and interactive parameters in the maintenance tasks are collected synchronously to construct an initial behavioral interaction dataset. The rail transit vehicle operation data includes: vehicle speed data, traction and braking status data, and line section operation record data; By identifying graph associations, the event nodes in the initial behavioral interaction dataset are coupled and modeled. Based on normal and abnormal operation samples in the historical maintenance quality archive, abnormal boundary analysis is performed on cross-node collaborative behavior characteristics and time-series disturbance characteristics to generate risk pre-assessment parameters. The risk pre-assessment parameters include: equipment occupancy risk parameters and personnel skill matching deviation parameters.

[0009] Specifically, the method for constructing the multi-objective collaborative decision-making maintenance model is as follows: Regularly extract maintenance task time characteristics from the comprehensive maintenance quality archive, and after vectorizing the maintenance tasks, identify the task node set and the relationship edge set to construct an initial task collaboration network; Real-time detection of newly added maintenance tasks; dynamic adjustment of task priority coefficient set time limit constraint parameters based on risk pre-assessment parameters; partial reconstruction of the affected task set; and adaptive rearrangement optimization of process sorting results. The system calculates the execution stability of different maintenance tasks, cleans up long-term inefficient task collaboration relationships, merges and optimizes scheduling units with highly overlapping resource allocation paths, and synchronously updates the constraint parameters of the initial task collaboration network to construct a multi-objective collaborative decision-making maintenance model.

[0010] Specifically, the iterative optimization process for the parallel relationship of maintenance tasks and equipment configuration paths is as follows: Based on the initial task collaboration network in the multi-objective collaborative decision-making maintenance model, the set of maintenance task nodes is jointly encoded, and the risk pre-assessment parameters are used as dynamic penalty terms and input into the time-varying constraint optimization function for iterative optimization. In the initial stage of iterative optimization, the time-varying constraint optimization function generates a subset of feasible parallel tasks based on the task process dependency relationship, and performs initial resource mapping based on the equipment availability status and personnel skill matching degree to generate a basic feasible solution set. The basic feasible solution set includes: task execution order solution set, equipment allocation solution set, and resource allocation solution set; After each iteration, the conflict subgraph is remapped and paths are reallocated until the convergence condition is met or the number of iterations reaches a preset threshold. The globally optimal or near-optimal dynamic maintenance operation plan is then output, while historical optimal solutions are retained as a memory pool for subsequent iteration guidance.

[0011] Specifically, the event-driven scheduling engine includes: an event-aware processing unit and a dynamic scheduling constraint unit; The event perception and processing unit: performs unified access and semantic parsing of real-time event streams throughout the entire maintenance process, and, in conjunction with newly added maintenance task events and resource conflict events, classifies the urgency of maintenance events into multiple levels to form a structured scheduling trigger instruction set; The structured scheduling trigger instruction set includes: new maintenance task insertion instruction and emergency fault task priority scheduling instruction; The dynamic scheduling constraint unit: based on the scheduling trigger instruction set, it applies local or global constraints to the current maintenance task execution plan, performs incremental constraint updates on the affected task set, re-plans the maintenance tasks, and ensures that the scheduling results continuously converge within a preset time window.

[0012] Specifically, the process of responding to a newly added maintenance task includes the following steps: Local constraint scheduling is performed through an event-driven scheduling engine to perform topology remapping and constraint reallocation on the affected task subgraph, and new maintenance tasks are embedded into the optimal execution position according to the execution priority of the new tasks. After responding to a new task, the updated partial maintenance plan will be synchronized to the cloud-based maintenance synchronization platform to drive the rolling refresh of task progress on each terminal.

[0013] Specifically, the process of establishing the comprehensive maintenance quality archive includes the following steps: Execute dynamic maintenance operation plans and acquire structured maintenance operation data generated throughout the entire maintenance process; The structured maintenance operation data is classified and stored. High-frequency access real-time maintenance records are stored in a high-speed cache layer, and historical periodic maintenance data and image data are stored in a low-cost archiving layer to build a cross-dimensional maintenance quality correlation index. A unique version identifier is generated for all incoming data, and the entity that performs data addition, correction and archiving operations is recorded. Combined with the cross-dimensional maintenance quality correlation index, root cause retrieval support is provided to the outside world, and a comprehensive maintenance quality archive is established.

[0014] Specifically, the process of tracing back to historical detection data sources includes the following steps: The current fault repair event is reconstructed by mirror mapping, and the fault code and timestamp association information are extracted to build a multi-dimensional traceability trigger index. Based on the multidimensional tracing trigger index query, heterogeneous data sources are queried, and cross-domain time alignment is performed on the heterogeneous data sources to generate a standardized tracing data matrix; Based on the digital twin backtracking mechanism, the potential triggering event chain is searched backward along time from the current fault time, while the associated influencing nodes are searched along the structural coupling direction to identify historical detection data sources.

[0015] Specifically, the method for generating the preventive maintenance strategy is as follows: Based on the root cause analysis results of fault repair events, an initial fault trigger node, intermediate propagation nodes, and terminal manifestation nodes are set to construct a multi-level temporal causal relationship graph. Based on the multi-level temporal causal relationship graph, reverse time tracing and forward structural expansion are performed to automatically generate preventive maintenance tasks. The maintenance disturbance simulation is performed on the causal nodes and propagation boundary nodes. The root cause contribution weight of each node is calculated. The optimal maintenance window period is determined by combining the suppression effect of different maintenance intervention points on the fault evolution path, and a preventive maintenance strategy is generated.

[0016] Specifically, the cloud-based maintenance synchronization platform includes: a multi-terminal collaborative access structure and a maintenance visualization verification structure; The multi-terminal collaborative access structure: During the process of accessing the cloud-based maintenance synchronization platform, the permission roles are verified and a terminal connection table is established; The terminal connection table is used to record the terminal identification information and maintenance task number corresponding to each access terminal. The maintenance visualization verification structure: During the maintenance site verification stage, the consistency of the terminal connection table is verified, process skipping events are identified, and the maintenance completion status and safety confirmation records are visualized and mapped to generate a maintenance completion heatmap.

[0017] Specifically, the process of encrypting and archiving the inspection and verification results includes the following steps: Acquire multi-source result data generated during the maintenance and field verification phase, and establish a dataset to be archived. The dataset to be archived is restructured in a structured manner, and the operation permissions of employees are bound to form a controllable and shared encrypted archive payload. High-frequency access data is written to the fast storage layer, and long-term retained data is migrated to the low-cost storage layer. At the same time, when any storage node fails, replica takeover and index reconstruction are automatically triggered.

[0018] Specifically, a comprehensive management system for the entire maintenance process of rail transit vehicles includes: Risk pre-assessment parameter generation module: acquires rail transit vehicle operation data, performs feature analysis on behavioral interactions during the maintenance process, and generates risk pre-assessment parameters; Dynamic maintenance collaborative decision-making module: Based on the risk pre-assessment parameters, the module iteratively optimizes the parallel relationship of maintenance tasks and equipment configuration paths through a multi-objective collaborative decision-making maintenance model, outputs a dynamic maintenance operation plan, and performs rolling reconstruction of the execution status of maintenance equipment based on an event-driven scheduling engine to respond to new maintenance tasks in real time. Quality Archive and Maintenance Strategy Generation Module: Executes the dynamic maintenance operation plan, structures and unifies the operation data generated throughout the maintenance process, establishes a comprehensive maintenance quality archive database, and traces back historical detection data sources based on the digital twin backtracking mechanism, performs correlation deduction on the root cause links that cause fault maintenance events, and obtains preventive maintenance strategies. Cloud-based verification and encrypted archiving module: Driven by the goal of visual verification of maintenance, it builds a cloud-based maintenance synchronization platform with multi-terminal collaborative access, automatically performs maintenance field verification, and encrypts and archives the maintenance field verification results to form a full life cycle maintenance report.

[0019] The beneficial effects of this invention are as follows: This invention provides a comprehensive management method and system for the entire maintenance process of rail transit vehicles. By fusing and analyzing vehicle operation data and maintenance behavior interaction data, it can effectively identify and quantitatively assess maintenance risks in advance, improving the scientific rigor and foresight of maintenance planning. By constructing a multi-objective collaborative decision-making maintenance model, it achieves unified optimization scheduling of parallel maintenance tasks and resource allocation paths, thereby significantly improving the efficiency of maintenance resource utilization, reducing task conflicts and waiting time, and enhancing overall maintenance execution efficiency.

[0020] Meanwhile, this invention introduces an event-driven scheduling mechanism, which can dynamically respond to new maintenance tasks and emergencies, enabling local or global adaptive adjustments to the maintenance plan and enhancing the system's real-time scheduling capabilities and robustness in complex scenarios. Throughout the maintenance process, the structured and unified management of work data and the establishment of a comprehensive maintenance quality archive database ensure the traceability and reusability of maintenance data, providing a data foundation for subsequent analysis.

[0021] By correlating and extrapolating historical testing data using a digital twin backtracking mechanism, multi-dimensional tracing analysis of fault root causes can be achieved, generating targeted preventative maintenance strategies that effectively reduce the probability of similar faults recurring. Supported by a cloud-based maintenance synchronization platform, multi-terminal collaborative access and visualized verification of the maintenance process are enabled, improving the transparency and standardization of the maintenance process.

[0022] Finally, by encrypting and archiving the inspection and verification results and managing their hierarchical storage, the security and integrity of the inspection data are enhanced, and closed-loop management of the entire life cycle of rail transit vehicle maintenance is achieved. Attached Figure Description

[0023] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0024] Figure 1 This is a schematic diagram of the framework of a comprehensive management system for the entire maintenance process of rail transit vehicles according to the present invention.

[0025] Figure 2 This is a schematic diagram of the software architecture of a comprehensive management system for the entire maintenance process of rail transit vehicles according to the present invention.

[0026] Figure 3 This is a flowchart of the maintenance workstation flow information in a comprehensive management system for the entire maintenance process of rail transit vehicles according to the present invention.

[0027] Figure 4 This is a schematic diagram of the iterative optimization process described in the integrated management system for the entire maintenance process of rail transit vehicles according to the present invention. Detailed Implementation

[0028] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0029] Please see Figure 1 This invention also provides a comprehensive management system for the entire maintenance process of rail transit vehicles, specifically including: Risk pre-assessment parameter generation module: acquires rail transit vehicle operation data, performs feature analysis on behavioral interactions during the maintenance process, and generates risk pre-assessment parameters; Dynamic maintenance collaborative decision-making module: Based on the risk pre-assessment parameters, the module iteratively optimizes the parallel relationship of maintenance tasks and equipment configuration paths through a multi-objective collaborative decision-making maintenance model, outputs a dynamic maintenance operation plan, and performs rolling reconstruction of the execution status of maintenance equipment based on an event-driven scheduling engine to respond to new maintenance tasks in real time. Quality Archive and Maintenance Strategy Generation Module: Executes the dynamic maintenance operation plan, structures and unifies the operation data generated throughout the maintenance process, establishes a comprehensive maintenance quality archive database, and traces back historical detection data sources based on the digital twin backtracking mechanism, performs correlation deduction on the root cause links that cause fault maintenance events, and obtains preventive maintenance strategies. Cloud-based verification and encrypted archiving module: Driven by the goal of visual verification of maintenance, it builds a cloud-based maintenance synchronization platform with multi-terminal collaborative access, automatically performs maintenance field verification, and encrypts and archives the maintenance field verification results to form a full life cycle maintenance report.

[0030] In this embodiment, taking the routine maintenance scenario of an urban subway depot as an example, the specific implementation process is as follows: Risk pre-assessment parameter generation module: The rail transit vehicle operation data acquisition system collects real-time vibration data of the running gear, current and voltage curves of the traction system, pressure and timing data of the braking system, and logs from onboard sensors.

[0031] The maintenance dispatch center formulates a preliminary maintenance task list based on the day's operating mileage and historical fault records.

[0032] Simultaneously collect behavioral interaction parameters during the execution of maintenance tasks, including: the operation trajectory of the operator's handheld terminal, equipment usage logs, and spatiotemporal interaction data between personnel, equipment, and vehicles, to construct an initial behavioral interaction dataset.

[0033] By using graph association recognition technology, event nodes (such as "wheelset disassembly", "bearing inspection" and "electrical insulation test") in the initial behavioral interaction dataset are coupled and modeled.

[0034] Meanwhile, by combining a large number of normal and abnormal operation samples in the comprehensive maintenance quality archive, abnormal boundary analysis is conducted on cross-node collaborative behavior characteristics (such as the timing coordination of multi-person collaborative operations) and timing disturbance characteristics (such as the chain reaction caused by operation delays), and finally, risk pre-assessment parameters (including high-risk node early warning) are generated for each maintenance task.

[0035] Dynamic maintenance collaborative decision-making module: Periodically extract maintenance task time characteristics from the comprehensive maintenance quality archive, vectorize the maintenance tasks, identify the task node set and relationship edge set, and construct an initial task collaboration network.

[0036] When a new emergency maintenance task is detected (such as a sudden braking system alarm on a train), the task priority coefficient and time constraint parameters are dynamically adjusted according to the risk pre-assessment parameters, the affected task set is partially reconstructed, and the process sorting results are adaptively rearranged and optimized.

[0037] The model further calculates the execution stability of different maintenance tasks, cleans up the failures of long-term inefficient task coordination relationships, merges and optimizes scheduling units with highly overlapping resource allocation paths, and updates constraint parameters in sync, thus completing the construction of a multi-objective collaborative decision-making maintenance model.

[0038] During the iterative optimization phase, the set of maintenance task nodes is jointly encoded based on the initial task collaboration network, and the risk pre-assessment parameters are used as dynamic penalty terms and input into the time-varying constraint optimization function for iterative solution.

[0039] In the initial stage, a subset of feasible parallel tasks is generated, and initial resource mapping is performed based on equipment availability and personnel skill matching.

[0040] After each iteration, the conflict subgraph is remapped and paths are redistributed until convergence or the preset number of iterations is reached. The globally optimal or near-optimal dynamic maintenance operation plan is output, while historical optimal solutions are retained as a memory pool for subsequent iteration guidance.

[0041] The event-driven scheduling engine operates in real time: the event awareness and processing unit uniformly accesses and semantically parses the real-time event stream throughout the entire maintenance process, classifies new maintenance task events and resource conflict events into multi-level urgency categories, and forms a structured scheduling trigger instruction set.

[0042] The dynamic scheduling constraint unit updates the current plan with local or global constraints based on the trigger command, re-plans the maintenance tasks, and continues to converge within the preset time window.

[0043] When a new maintenance task arrives, the system performs local constraint scheduling through the event-driven scheduling engine, performs topology remapping and constraint reallocation on the affected task subgraph, and embeds the new task into the optimal execution position according to its execution priority.

[0044] The updated partial maintenance plan is synchronized to the cloud-based maintenance synchronization platform, driving the rolling refresh of task progress on each terminal.

[0045] Quality record and maintenance strategy generation module: After executing the dynamic maintenance operation plan, the system classifies and stores the structured operation data (including operation video clips, test reports, sensor readings, personnel operation records, etc.) generated throughout the maintenance process: high-frequency access real-time maintenance records are stored in the high-speed cache layer, historical periodic maintenance data and image data are stored in the low-cost archiving layer, and a cross-dimensional maintenance quality correlation index is constructed.

[0046] Generate a unique version identifier for all incoming data, record the operating entity, and establish a comprehensive maintenance quality archive.

[0047] When a fault repair event occurs, the system performs mirror mapping reconstruction of the current fault event, extracts fault codes and timestamp association information, and constructs a multi-dimensional traceability trigger index.

[0048] Then, heterogeneous data sources are queried, cross-domain time alignment is performed, and a standardized traceability data matrix is ​​generated.

[0049] Based on the digital twin backtracking mechanism, the potential triggering event chain is searched backward along time from the current fault time, while the associated influencing nodes are searched along the structural coupling direction to identify historical detection data sources.

[0050] Furthermore, based on the root cause analysis results of fault maintenance events, initial fault triggering nodes, intermediate propagation nodes, and terminal manifestation nodes are set to construct a multi-level temporal causal relationship graph. Through reverse time tracing and forward structural expansion, preventive maintenance tasks are automatically generated.

[0051] The root cause contribution weight of each node is calculated by simulating maintenance disturbances at the causal nodes and propagation boundary nodes. The optimal maintenance window period is determined by combining the suppression effect of different maintenance intervention points on the fault evolution path, and finally a preventive maintenance strategy is generated (such as "adding special flaw detection inspections on a specific bearing every 8,000 kilometers of operation").

[0052] Cloud-based verification and encrypted archiving module: Driven by the goal of visualized maintenance verification, a cloud-based maintenance synchronization platform with multi-terminal collaborative access is constructed.

[0053] The multi-terminal collaborative access structure verifies permissions and roles during the access process, establishes a terminal connection table, and records the terminal identification information and maintenance task number corresponding to each access terminal.

[0054] During the factory verification phase, the maintenance visualization verification structure performs consistency verification on the terminal connection table, identifies process skipping events, and visualizes the maintenance completion status and safety confirmation records to generate a maintenance completion heatmap, allowing managers to intuitively grasp the overall progress.

[0055] After the maintenance and on-site verification is completed, the system acquires multi-source verification result data and establishes a dataset to be archived. This dataset is then structurally reorganized, and employee access permissions are linked to create a controllable and shared encrypted archive payload.

[0056] High-frequency access data is written to the fast storage layer, long-term retained data is migrated to the low-cost storage layer, and replica takeover and index reconstruction are automatically triggered when any storage node fails, ultimately forming a full lifecycle maintenance report.

[0057] Example 1: Comparison of multi-task parallelism and emergency response; This embodiment demonstrates a comparison of how the same batch of maintenance tasks (including newly added tasks) were handled before and after the application of this method in the same depot.

[0058] Without this method, maintenance plans use static scheduling, and the insertion of unexpected tasks leads to delays in multiple tasks, frequent resource conflicts, and root cause tracing relies on manual experience, which is time-consuming and has low accuracy.

[0059] After applying this method, the system, through a multi-objective collaborative decision-making model and an event-driven scheduling engine, quickly performs local reconstruction and embedding optimization of sudden tasks. Parallel relationship optimization significantly improves the efficiency of overlapping execution of multiple maintenance tasks. The digital twin backtracking mechanism can complete root cause chain deduction within minutes, generating accurate preventive strategies. The cloud synchronization platform enables real-time visual collaboration across multiple terminals (handheld PDAs, depot large screens, and remote monitoring centers), improving the automation rate of maintenance field verification.

[0060] like Figure 2 As shown, the software architecture adopts a layered design, building a robust system from bottom to top. The infrastructure layer provides support through databases and file systems, while the hardware device layer enables physical interaction through sensors, PLCs, and various other devices. The service layer revolves around application services, domains, infrastructure, and general components. Application services cover functions such as JWT authentication and authorization, the domain layer focuses on core logic such as domain objects and data warehousing, the infrastructure layer ensures system operation through ORM and task scheduling, and general components provide basic capabilities such as localization and logging. The interface layer enables interaction between upper and lower layers through WebAPI. Each layer has a clear division of labor and collaborates efficiently, providing a reliable, flexible, and easily scalable software support architecture for the project.

[0061] In this embodiment, the calculation process for the execution stability of the different maintenance tasks is as follows: Collect actual execution data for each maintenance task within a preset statistical period. The actual execution data includes task start time, completion time, working time deviation, resource usage, number of task interruptions, number of reworks, and quality acceptance results. By comparing the actual execution data with the planned execution parameters of the corresponding tasks, evaluation indicators for each maintenance task are obtained in dimensions such as progress achievement rate, working hour fluctuation rate, resource matching degree, execution continuity and quality pass rate. The evaluation indicators of each dimension are standardized and weighted according to preset weights to form the stability evaluation value of the corresponding maintenance task. Based on the trend of stability evaluation values ​​over multiple consecutive statistical periods, maintenance tasks with large long-term fluctuations, persistently low execution efficiency, or high rework frequency are identified and classified as low-stability tasks. The execution stability results of different maintenance tasks are then output for subsequent optimization of task collaboration, adjustment of resource scheduling, and cleanup of failure associations.

[0062] Example 2: Traceability Management of the Maintenance Process; like Figure 3 As shown, a unique RFID tag traceability code enables precise tracking of component maintenance lifecycle data, ensuring quality control and problem traceability. A unique traceability code is assigned to each wheel, running through each workstation in the maintenance process; it can track maintenance times, operators, replacement parts, and maintenance records, establishing an electronic maintenance history; once a quality problem is discovered, the source of the problem can be located with a single click.

[0063] The system mainly realizes the process control and traceability management functions of each workstation in the maintenance process. Starting from the arrival of the wheel for maintenance, it monitors the maintenance process and status, records the relevant personnel information, operation status, quality inspection status, parts (must-replace parts and occasional replacement parts) replacement status, abnormal situations and other information in each area during the maintenance process, and realizes the whole process information traceability system of product maintenance operation information, material information and quality information.

[0064] When a wheel arrives at its respective workstation, the RFID reader automatically reads the wheel's tag traceability code and transmits it to the control system. The system then automatically retrieves the wheel's basic information from the database and subsequently provides corresponding work instructions and records the maintenance data for each workstation.

[0065] The system collects equipment status information, personnel information, material information, parameter information, assembly process operation information, test data, and test results information in real time, stably, and securely, and stores them on the server. This enables information interconnection and paperless archiving between workstations. Managers can log in to the system at any time via the web to query production process traceability records, realize forward and reverse traceability of the production process, and perform statistical analysis on production data to generate electronic production process history reports.

[0066] In this embodiment, as Figure 4 As shown, the iterative optimization process for the parallel relationship of maintenance tasks and equipment configuration paths is as follows: Based on the initial task collaboration network in the multi-objective collaborative decision-making maintenance model, the set of maintenance task nodes is jointly encoded, and the risk pre-assessment parameters are used as dynamic penalty terms and input into the time-varying constraint optimization function for iterative optimization. In the initial stage of iterative optimization, the time-varying constraint optimization function generates a subset of feasible parallel tasks based on the task process dependency relationship, and performs initial resource mapping based on the equipment availability status and personnel skill matching degree to generate a basic feasible solution set. After each iteration, the conflict subgraph is remapped and paths are reallocated until the convergence condition is met or the number of iterations reaches a preset threshold. The globally optimal or near-optimal dynamic maintenance operation plan is then output, while historical optimal solutions are retained as a memory pool for subsequent iteration guidance.

[0067] Specific examples are as follows: When a city rail transit depot was carrying out a level-three maintenance task on a train, the tasks to be performed included 12 items such as braking system testing, bogie inspection, door mechanism debugging, air conditioning system maintenance, electrical circuit testing, and on-board communication module verification.

[0068] There are dependencies between the various tasks. For example, electrical circuit testing must be performed after the door mechanism debugging is completed, and vehicle communication module verification must be performed after the electrical circuit testing is completed. On-site available resources include 2 lifting devices, 3 testing stations, and 8 professional maintenance personnel, among whom there are differences in skill levels.

[0069] Based on the initial task collaboration network in the multi-objective collaborative decision-making maintenance model, the set of 12 maintenance task nodes is jointly encoded, and the task execution order, resource occupancy status and equipment path information are converted into optimization variables. At the same time, risk pre-assessment parameters are used as dynamic penalty terms to input the time-varying constraint optimization function. The risk pre-assessment parameters include the high historical failure rate coefficient of the braking system, the operation delay coefficient of the air conditioning system, and the anomaly detection coefficient of the communication module, which are used to improve the priority and resource guarantee weight of the corresponding tasks.

[0070] In the initial stage of iterative optimization, feasible parallel task subsets are automatically generated based on task process dependencies. For example, brake system inspection and air conditioning system maintenance can be performed in parallel, and bogie inspection and door mechanism debugging can be performed in parallel. Subsequently, initial resource mapping is performed based on equipment availability and personnel skill matching. Lifting equipment is preferentially assigned to bogie inspection tasks, and maintenance personnel with electrical qualifications are assigned to line testing tasks, forming a basic feasible solution set.

[0071] After each iteration, the system checks for resource conflicts. When it is found that air conditioning system maintenance and door mechanism debugging are occupying the same inspection station at the same time, the conflict sub-graph is remapped, the air conditioning system maintenance is moved to the backup station, and the work paths of relevant personnel are replanned.

[0072] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A comprehensive management method for the entire maintenance process of rail transit vehicles, characterized in that, include S1: Obtain rail transit vehicle operation data, perform feature analysis on behavioral interactions during the maintenance process, and generate risk pre-assessment parameters; S2: Based on the risk pre-assessment parameters, the parallel relationship of maintenance tasks and equipment configuration path are iteratively optimized through a multi-objective collaborative decision maintenance model to output a dynamic maintenance operation plan. Based on the event-driven scheduling engine, the execution status of the maintenance equipment is reconstructed in a rolling manner to respond to new maintenance tasks in real time. S3: Execute the dynamic maintenance operation plan, structure and unify the operation data generated throughout the maintenance process, establish a comprehensive maintenance quality archive, and trace back the historical detection data source based on the digital twin backtracking mechanism to perform correlation deduction on the root cause links that cause the fault maintenance event and obtain preventive maintenance strategies. S4: Driven by the goal of visual verification of maintenance, build a cloud-based maintenance synchronization platform with multi-terminal collaborative access, automatically perform maintenance field verification, and encrypt and archive the maintenance field verification results to form a full life cycle maintenance report.

2. The method according to claim 1, characterized in that, The feature parsing process is as follows: Maintenance tasks are formulated based on rail transit vehicle operation data, and interactive parameters in the maintenance tasks are collected synchronously to construct an initial behavioral interaction dataset. By identifying graph associations, event nodes in the initial behavioral interaction dataset are coupled and modeled. Based on normal and abnormal operation samples in the historical maintenance quality archive, abnormal boundary analysis is performed on cross-node collaborative behavior characteristics and time-series disturbance characteristics to generate risk pre-assessment parameters.

3. The method according to claim 1, characterized in that, The method for constructing the multi-objective collaborative decision-making maintenance model is as follows: Regularly extract maintenance task time characteristics from the comprehensive maintenance quality archive, and after vectorizing the maintenance tasks, identify the task node set and the relationship edge set to construct an initial task collaboration network; Real-time detection of newly added maintenance tasks; dynamic adjustment of task priority coefficient set time limit constraint parameters based on risk pre-assessment parameters; partial reconstruction of the affected task set; and adaptive rearrangement optimization of process sorting results. The system calculates the execution stability of different maintenance tasks, cleans up long-term inefficient task collaboration relationships, merges and optimizes scheduling units with highly overlapping resource allocation paths, and synchronously updates the constraint parameters of the initial task collaboration network to construct a multi-objective collaborative decision-making maintenance model.

4. The method according to claim 1, characterized in that, The process of iteratively optimizing the parallel relationships of maintenance tasks and equipment configuration paths is as follows: Based on the initial task collaboration network in the multi-objective collaborative decision-making maintenance model, the set of maintenance task nodes is jointly encoded, and the risk pre-assessment parameters are used as dynamic penalty terms and input into the time-varying constraint optimization function for iterative optimization. In the initial stage of iterative optimization, the time-varying constraint optimization function generates a subset of feasible parallel tasks based on the task process dependency relationship, and performs initial resource mapping based on the equipment availability status and personnel skill matching degree to generate a basic feasible solution set. After each iteration, the conflict subgraph is remapped and paths are reallocated until the convergence condition is met or the number of iterations reaches a preset threshold. The globally optimal or near-optimal dynamic maintenance operation plan is then output, while historical optimal solutions are retained as a memory pool for subsequent iteration guidance.

5. The method according to claim 1, characterized in that, The event-driven scheduling engine includes: an event-aware processing unit and a dynamic scheduling constraint unit; The event perception and processing unit: performs unified access and semantic parsing of real-time event streams throughout the entire maintenance process, and, in conjunction with newly added maintenance task events and resource conflict events, classifies the urgency of maintenance events into multiple levels to form a structured scheduling trigger instruction set; The dynamic scheduling constraint unit: based on the scheduling trigger instruction set, it applies local or global constraints to the current maintenance task execution plan, performs incremental constraint updates on the affected task set, re-plans the maintenance tasks, and ensures that the scheduling results continuously converge within a preset time window.

6. The method according to claim 1, characterized in that, The process of responding to newly added maintenance tasks includes the following steps: Local constraint scheduling is performed through an event-driven scheduling engine to perform topology remapping and constraint reallocation on the affected task subgraph, and new maintenance tasks are embedded into the optimal execution position according to the execution priority of the new tasks. After responding to a new task, the updated partial maintenance plan will be synchronized to the cloud-based maintenance synchronization platform to drive the rolling refresh of task progress on each terminal.

7. The method according to claim 1, characterized in that, The process of establishing the comprehensive maintenance quality archive includes the following steps: Execute dynamic maintenance operation plans and acquire structured maintenance operation data generated throughout the entire maintenance process; The structured maintenance operation data is classified and stored. High-frequency access real-time maintenance records are stored in a high-speed cache layer, and historical periodic maintenance data and image data are stored in a low-cost archiving layer to build a cross-dimensional maintenance quality correlation index. A unique version identifier is generated for all incoming data, and the entity that performs data addition, correction and archiving operations is recorded. Combined with the cross-dimensional maintenance quality correlation index, root cause retrieval support is provided to the outside world, and a comprehensive maintenance quality archive is established.

8. The method according to claim 1, characterized in that, The process of tracing back to historical detection data sources includes the following steps: The current fault repair event is reconstructed by mirror mapping, and the fault code and timestamp association information are extracted to build a multi-dimensional traceability trigger index. Based on the multidimensional tracing trigger index query, heterogeneous data sources are queried, and cross-domain time alignment is performed on the heterogeneous data sources to generate a standardized tracing data matrix; Based on the digital twin backtracking mechanism, the potential triggering event chain is searched backward along time from the current fault time, while the associated influencing nodes are searched along the structural coupling direction to identify historical detection data sources.

9. The method according to claim 1, characterized in that, The method for generating the preventive maintenance strategy is as follows: Based on the root cause analysis results of fault repair events, an initial fault trigger node, intermediate propagation nodes, and terminal manifestation nodes are set to construct a multi-level temporal causal relationship graph. Based on the multi-level temporal causal relationship graph, reverse time tracing and forward structural expansion are performed to automatically generate preventive maintenance tasks. The maintenance disturbance simulation is performed on the causal nodes and propagation boundary nodes. The root cause contribution weight of each node is calculated. The optimal maintenance window period is determined by combining the suppression effect of different maintenance intervention points on the fault evolution path, and a preventive maintenance strategy is generated.

10. The method according to claim 1, characterized in that, The cloud-based maintenance synchronization platform includes: a multi-terminal collaborative access structure and a maintenance visualization verification structure; The multi-terminal collaborative access structure: During the process of accessing the cloud-based maintenance synchronization platform, the permission roles are verified and a terminal connection table is established; The terminal connection table is used to record the terminal identification information and maintenance task number corresponding to each access terminal. The maintenance visualization verification structure: During the maintenance site verification stage, the consistency of the terminal connection table is verified, process skipping events are identified, and the maintenance completion status and safety confirmation records are visualized and mapped to generate a maintenance completion heatmap.

11. The method according to claim 1, characterized in that, The process of encrypting and archiving the inspection and verification results includes the following steps: Acquire multi-source result data generated during the maintenance and field verification phase, and establish a dataset to be archived. The dataset to be archived is restructured in a structured manner, and the operation permissions of employees are bound to form a controllable and shared encrypted archive payload. High-frequency access data is written to the fast storage layer, and long-term retained data is migrated to the low-cost storage layer. At the same time, when any storage node fails, replica takeover and index reconstruction are automatically triggered.

12. A comprehensive management system for the entire maintenance process of rail transit vehicles, used to execute the method as described in any one of claims 1-11, characterized in that, Risk pre-assessment parameter generation module: acquires rail transit vehicle operation data, performs feature analysis on behavioral interactions during the maintenance process, and generates risk pre-assessment parameters; Dynamic maintenance collaborative decision-making module: Based on the risk pre-assessment parameters, the module iteratively optimizes the parallel relationship of maintenance tasks and equipment configuration paths through a multi-objective collaborative decision-making maintenance model, outputs a dynamic maintenance operation plan, and performs rolling reconstruction of the execution status of maintenance equipment based on an event-driven scheduling engine to respond to new maintenance tasks in real time. Quality Archive and Maintenance Strategy Generation Module: Executes the dynamic maintenance operation plan, structures and unifies the operation data generated throughout the maintenance process, establishes a comprehensive maintenance quality archive database, and traces back historical detection data sources based on the digital twin backtracking mechanism, performs correlation deduction on the root cause links that cause fault maintenance events, and obtains preventive maintenance strategies. Cloud-based verification and encrypted archiving module: Driven by the goal of visual verification of maintenance, it builds a cloud-based maintenance synchronization platform with multi-terminal collaborative access, automatically performs maintenance field verification, and encrypts and archives the maintenance field verification results to form a full life cycle maintenance report.

Images