Whole-process AI intelligent agent management and control system based on integrated wheel set maintenance intelligent workshop

Abstract

The invention is suitable for the technical field of rail transit, and provides a whole-process AI agent management and control system based on an integrated wheel set maintenance intelligent workshop, and the system comprises a multi-source fusion agent module which is used for collecting the state and process data of equipment, achieving the cross-modal alignment through an AI model, generating a feature vector, and storing the feature vector into a database; the distributed knowledge graph is updated based on the time sequence diagram convolutional network; the role perception decision-making agent module is used for analyzing the user permission label and the instruction and generating a scheduling scheme in combination with an equipment load; the equipment conflict arbitration agent module is used for monitoring the spatial topology of the AGV and the manipulator; and the closed-loop optimization agent module is used for analyzing the wheel set press-fitting torque data and extracting historical cases of the knowledge graph to generate a process adjustment instruction. According to the method, the whole process from a data fusion agent and a dynamic decision agent to a process closed loop is intelligently increased.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, specifically to a full-process AI intelligent control system based on an integrated intelligent workshop for wheelset maintenance. Background Technology

[0002] Wheelset maintenance in rail transit is a crucial maintenance activity to ensure the safe and stable operation of trains. Wheelsets, consisting of wheels and axles, directly affect driving safety, comfort, and energy consumption.

[0003] Wheelset maintenance involves data such as equipment monitoring, process standards, and historical fault databases scattered across PMS (Production Management System), MES (Manufacturing Execution System), and individual sensors, easily creating information silos. Wheel inspection robots, AGVs, and other equipment operate independently, lacking unified scheduling and unable to dynamically adjust priorities based on the overall tasks of the workshop. Wheel axle flaw detection relies on technicians' visual inspection, and existing AI-assisted equipment is not linked to the maintenance knowledge base, making it impossible to optimize key parameters such as pressing torque in real time. In addition, new processes, equipment, and standards are constantly emerging, and current training relies on apprenticeship and static manuals, resulting in low knowledge transfer efficiency and an inability to keep up with the latest industry technical standards.

[0004] In summary, current wheelset maintenance workshops generally suffer from serious information silos, insufficient dynamic collaboration, delayed process control, and low knowledge transfer efficiency. Therefore, a full-process AI intelligent agent management and control system based on an integrated intelligent wheelset maintenance workshop is proposed to address the above-mentioned problems. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a full-process AI intelligent control system based on an integrated intelligent workshop for wheelset maintenance, so as to solve the problems existing in the background technology.

[0006] This invention is implemented as follows: a full-process AI intelligent agent control system based on an integrated intelligent workshop for wheelset maintenance, the system comprising:

[0007] The multi-source fusion intelligent agent module is used to collect equipment status and process data, generate feature vectors through cross-modal alignment using an AI model, and update a distributed knowledge graph based on a temporal graph convolutional network. The status and process data include flaw detection images and audio data.

[0008] The role-aware decision-making intelligent agent module is used to parse user permission tags and instructions, generate scheduling schemes in combination with device load, and send instructions to the device controller through edge nodes and store them on the blockchain.

[0009] A device conflict arbitration intelligent agent module is used for monitoring the spatial topology of AGVs and manipulators, and when trajectories overlap, a fusion instruction is generated through Raft consensus based on a timestamp verification version number, and an authority topology tree is updated, the spatial topology being the relative position relationship of AGVs and manipulators.

[0010] A closed-loop optimization intelligent agent module is used for analyzing wheel set press-fitting torque data, extracting historical cases of a knowledge graph to generate process adjustment instructions, and feeding the process adjustment instructions to PLCs through blockchain signature and feeding back optimization effects, the wheel set press-fitting torque data being torque applied in the press-fitting process and torque change trend.

[0011] As a further scheme of the application, the multi-source fusion intelligent agent module comprises:

[0012] A cross-modal alignment unit is used for inputting flaw detection images into a ResNet-50 model to extract crack morphology features, simultaneously converting audio data into 128-dimensional mel spectra through MFCC, and adopting an industrial CLIP model to map to a unified semantic space to generate a fusion feature vector carrying a reliability weight.

[0013] A graph dynamic evolution unit is used for constructing a three-layer relationship topology taking a wheel set serial number as a main entity, predicting an associated path of bearing wear and crack propagation based on a time series graph convolution network, and automatically expanding a subgraph structure when a new fault case is stored.

[0014] A storage optimization unit is used for storing knowledge graph subgraphs according to production line stations, dynamically migrating data to edge nodes according to a query frequency heat map, and realizing real-time relationship query through a graph database.

[0015] As a further scheme of the application, the role-aware decision intelligent agent module comprises:

[0016] An authority-driven analysis unit is used for identifying a role tag associated with a user badge two-dimensional code, parsing a natural language query into a structured operation chain and filtering super authority instructions.

[0017] A multi-objective optimization unit is used for establishing a multi-objective function of device utilization rate, order delay penalty coefficient and energy consumption threshold, and generating a scheduling scheme by using an AI optimization algorithm.

[0018] An edge collaboration unit is used for caching process standard subgraphs smaller than 50MB at a PAD end, decomposing complex requests into edge light computing and cloud heavy reasoning tasks, and enabling the system to directly connect to controllers of high-risk devices through a protocol.

[0019] As a further scheme of the application, the role-aware decision intelligent agent module further comprises:

[0020] The blockchain trusted record unit is used for signing the issued device emergency stop instruction and process parameter adjustment operation by an SM2 national secret algorithm, and synchronizing the hash value of the operation log to the Fabric alliance chain after IPFS fragmentation storage.

[0021] As a further scheme of the application, the device conflict arbitration intelligent agent module comprises:

[0022] The conflict prediction unit is used for calculating the path occupancy rate within a future setting time according to the spatial topology of the AGV and the manipulator, and marking a high-risk conflict when the path overlap rate is greater than 30%;

[0023] The consensus arbitration unit is used for adding a Lamport timestamp vector to the operation log of each device, and if the version offset exceeds a dynamic threshold, a master node is elected by an AI-driven dynamic consensus mechanism to generate a fusion instruction.

[0024] The permission topology update unit adjusts the device permission tree weight according to the arbitration result, and binds the AGV task delay record to the knowledge graph.

[0025] As a further scheme of the application, the closed-loop optimization intelligent agent module comprises:

[0026] The process deviation analysis unit is used for collecting wheel set press-in torque data in real time according to a set collection rate, and marking a key deviation event when the detection standard deviation exceeds the upper limit of the process by 5%;

[0027] The graph reverse optimization unit is used for extracting a historical similar case disposal subgraph in the knowledge graph, calculating the matching degree of a new scheme by a graph attention mechanism, and injecting a confidence rule;

[0028] The device parameter adaptive unit is used for generating a hydraulic compensation instruction with a specified precision, issuing the instruction to the PLC after blockchain signature, and feeding back the wheel set roundness error to the knowledge graph.

[0029] As a further scheme of the application, the multi-source fusion intelligent agent module further comprises an anti-magnetic RFID hardware group, which is composed of a high-temperature-resistant tag implanted in the inner ring of the wheel shaft bearing and a reader arranged at the maintenance station, the ID of the high-temperature-resistant tag is bound to the wheel set unique identification entity in the knowledge graph, and the hardware group is used for collecting the state and process data of the device.

[0030] Compared with the prior art, the application has the following advantages:

[0031] The application thoroughly eliminates information islands and realizes intelligent association and semantic penetration of heterogeneous data by deeply fusing equipment state monitoring data and process parameters through multi-source fusion intelligent agent module, and constructing a dynamically evolving knowledge graph using cross-modal alignment technology; The role perception decision-making intelligent agent module generates precise scheduling instructions based on real-time analysis of user permission levels and device operating load, and combines blockchain distributed notarization technology to ensure the auditability and security of the whole process operation, and to strengthen the abnormal response capability and multi-task coordination efficiency; The equipment conflict arbitration intelligent agent module innovatively resolves the trajectory overlap conflict of AGV and manipulator equipment through real-time perception and dynamic consensus mechanism based on spatial topology, realizes seamless cooperation and resource optimization configuration of multiple intelligent agents in complex scenarios; The closed-loop optimization intelligent agent module actively optimizes key process parameters such as wheel set press-in torque by deeply mining the knowledge of the historical case library, and continuously synchronizes the latest technical standards through the knowledge graph self-learning mechanism, which improves the assembly process precision and builds an ecological system of efficient knowledge circulation. Overall, the application overcomes the core bottlenecks of data fragmentation, low efficiency of cooperation, process lag and knowledge fault in the wheel set maintenance field for a long time, and realizes the intelligent leap of the whole process from data fusion, dynamic decision-making to process closed loop. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 FIG. 1 is a structural schematic diagram of a whole-process AI intelligent agent management and control system based on an integrated wheel set maintenance intelligent workshop.

[0033] Figure 2 FIG. 2 is a structural schematic diagram of a multi-source fusion intelligent agent module in the whole-process AI intelligent agent management and control system based on the integrated wheel set maintenance intelligent workshop.

[0034] Figure 3 FIG. 3 is a structural schematic diagram of a role perception decision-making intelligent agent module in the whole-process AI intelligent agent management and control system based on the integrated wheel set maintenance intelligent workshop.

[0035] Figure 4 FIG. 4 is a structural schematic diagram of an equipment conflict arbitration intelligent agent module in the whole-process AI intelligent agent management and control system based on the integrated wheel set maintenance intelligent workshop.

[0036] Figure 5 FIG. 5 is a structural schematic diagram of a closed-loop optimization intelligent agent module in the whole-process AI intelligent agent management and control system based on the integrated wheel set maintenance intelligent workshop.

[0037] Figure 6 FIG. 6 is a flowchart of a whole-process management and control method based on the integrated wheel set maintenance intelligent workshop. DETAILED DESCRIPTION

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0039] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0040] like Figure 1 As shown, this embodiment of the invention provides a full-process AI intelligent agent control system based on an integrated intelligent workshop for wheelset maintenance. The system includes:

[0041] The multi-source fusion intelligent agent module 100 is used to collect the status and process data of the equipment, generate feature vectors through cross-modal alignment using the Zhongke Zhiyuan AI model, and update the distributed knowledge graph based on the temporal graph convolutional network. The status and process data include flaw detection images and audio data.

[0042] The Role Perception and Decision-Making Intelligent Agent Module 200 is used to parse user permission tags and instructions, generate a scheduling scheme in combination with device load, and send instructions to the device controller through edge nodes and perform blockchain notarization.

[0043] The equipment conflict arbitration intelligent agent module 300 is used to monitor the spatial topology of AGV and robot. When the trajectories overlap, it generates a fusion instruction based on the timestamp verification version number through Raft consensus and updates the permission topology tree. The spatial topology is the relative positional relationship between AGV and robot.

[0044] The closed-loop optimization intelligent agent module 400 is used to analyze wheelset pressing torque data, extract historical cases from the knowledge graph to generate process adjustment instructions, send them to the PLC via blockchain signature, and provide feedback on the optimization effect. The wheelset pressing torque data refers to the torque applied during the pressing process and the torque change trend.

[0045] It should be noted that the multi-source fusion intelligent agent module includes an anti-magnetic RFID hardware group, which consists of a high-temperature resistant tag embedded in the inner ring of the wheel axle bearing and a reader / writer deployed at the maintenance station. The ID of the high-temperature resistant tag is bound to the unique wheel pair identifier entity in the knowledge graph, and the hardware group is used to collect equipment status and process data.

[0046] In the embodiment of the present application, the multi-source fusion intelligent agent module deeply fuses the equipment state monitoring data and process parameters, uses the cross-modal alignment technology to construct a dynamically evolving knowledge graph, completely eliminates information islands and realizes intelligent association and semantic penetration of heterogeneous data; the role perception decision intelligent agent module generates precise scheduling instructions based on the real-time analysis of user permission levels and device running load, combines the distributed storage technology of blockchain to ensure the auditability and security of the whole process operation, and strengthens the abnormal response capability and multi-task coordination efficiency; the equipment conflict arbitration intelligent agent module innovatively uses the spatial topology real-time perception and dynamic consensus mechanism to efficiently resolve the track overlap conflict of AGV and manipulator equipment, realizes seamless cooperation and resource optimization configuration of multiple intelligent agents in complex scenarios; the closed-loop optimization intelligent agent module deeply mines the knowledge of the historical case library, actively optimizes key process parameters such as wheelset press-in torque, and continuously synchronizes the latest technical standards through the knowledge graph self-learning mechanism, which improves the assembly process precision and builds an ecological system of efficient knowledge circulation. Overall, the present application overcomes the core bottlenecks of data fragmentation, low efficiency of cooperation, process lag and knowledge fault in the wheelset maintenance field for a long time, and realizes the intelligent leap of the whole process from data fusion, dynamic decision to process closed loop.

[0047] As shown in Figure 2 As a preferred embodiment of the present application, the multi-source fusion intelligent agent module 100 comprises:

[0048] The cross-modal alignment unit 101 is used for inputting the flaw detection image into the ResNet-50 model to extract the crack morphology features, converting the audio data into 128-dimensional mel spectrum through MFCC at the same time, and using the industrial CLIP model to map to the unified semantic space to generate a fusion feature vector carrying a reliability weight;

[0049] The graph dynamic evolution unit 102 is used for constructing a three-layer relationship topology taking the wheelset serial number as the main entity, predicting the associated path of bearing wear and crack propagation based on the time series graph convolution network, and automatically expanding the subgraph structure when a new fault case is stored;

[0050] The storage optimization unit 103 is used for storing the knowledge graph subgraph according to the production line station partition, dynamically migrating data to the edge node according to the query frequency heat map, and realizing real-time relationship query through the graph database.

[0051] In the embodiment of the application, the multi-source fusion intelligent agent module 100 realizes the deep integration of heterogeneous industrial data and dynamic knowledge management through three core units. The cross-modal alignment unit 101 uses the ResNet-50 model to analyze the crack morphology features (such as crack length, bifurcation angle, etc. geometric properties) in the flaw detection image, at the same time, converts the audio data such as bearing abnormal sound into 128-dimensional mel frequency spectrum representation through MFCC, and maps the image and audio features to a unified semantic space based on the CLIP model optimized for the industrial scene, generates a fusion feature vector carrying a confidence weight, for example, if the correlation degree between the crack picture feature of a certain bearing and the abnormal sound spectrum feature in the semantic space is greater than a certain value, a high weight alarm will be triggered; the graph dynamic evolution unit 102 constructs a three-layer relationship topology with the wheel pair serial number as the main entity, the physical layer is the attribute parameters of the bearing and the wheel, the process layer is the press fitting parameter and the flaw detection record, and the fault layer is some historical cases, using the time series graph convolution network (T-GCN) to predict the equipment state evolution path, which can identify the correlation between bearing wear degree and crack propagation rate, and automatically expand the subgraph structure when a new fault case is added; the storage optimization unit 103 stores the knowledge graph subgraph according to the production line station partition, dynamically migrates high-frequency data to the edge node according to the query heat map, and can realize millisecond-level relationship query through the Neo4j graph database, for example, it can retrieve all flaw detection records and maintenance work orders associated with a certain wheel pair serial number within two years.

[0052] As shown in Figure 3 , as a preferred embodiment of the application, the role-aware decision-making intelligent agent module 200 includes:

[0053] The permission-driven analysis unit 201 is used for identifying the role label associated with the user card two-dimensional code, parsing the natural language query into a structured operation chain and filtering super permission instructions;

[0054] The multi-objective optimization unit 202 is used for establishing a multi-objective function of device utilization rate, order delay penalty coefficient and energy consumption threshold, and generating a scheduling scheme by using an AI optimization algorithm;

[0055] The edge collaboration unit 203 is used for caching process standard subgraph with less than 50MB at the PAD end, decomposing complex requests into edge light computing and cloud heavy reasoning tasks, and connecting the controller of the high-risk device through the protocol to make the system control the high-risk device.

[0056] The blockchain trusted storage unit 204 is used for signing the device emergency stop instruction and process parameter adjustment operation by using the SM2 national secret algorithm, storing the operation log through IPFS fragmentation, and synchronizing the hash value to the Fabric consortium chain.

[0057] In the embodiment of the application, the role-aware decision-making agent module 200 realizes a safe, efficient and reliable intelligent decision-making closed loop through the cooperation of four core units. The permission-driven analysis unit 201 identifies the role label (such as senior technician, quality inspector or intern) of the user badge according to the two-dimensional code, converts the natural language instruction into a structured operation sequence in real time, and automatically intercepts the super permission request. For example, when a junior operator tries to modify the bearing press-fit parameters by skipping the approval process, the system will immediately terminate the instruction and trigger a permission warning, thereby eliminating the risk of misoperation from the source. The multi-objective optimization unit 202 will build a Pareto optimal model covering multiple constraints such as real-time utilization rate of equipment, order delivery delay penalty coefficient, and energy consumption threshold of production line, and dynamically generate a globally optimal scheduling scheme using AI optimization algorithm. For example, when a certain flaw detection station suddenly fails, the AGV path planning and shunting of the detected wheelset to the idle station are simultaneously optimized. The edge collaboration unit 203 pre-caches high-frequency process standard sub-atlas (such as wheel flaw detection threshold table) at the PAD end, decomposes complex requests into real-time calculation at the edge end and deep reasoning tasks in the cloud through lightweight container technology, and based on the OPC UA protocol, directly connects the unloading machine PLC controller to perform high-risk operations such as emergency stop, thereby ensuring low response delay of critical instructions. The blockchain trusted storage unit 204 digitally signs sensitive operations such as device emergency stop instructions and process parameter adjustments using the SM2 algorithm, and after encrypting the complete operation log through IPFS distributed storage fragmentation, synchronizes the hash fingerprint to each supervision node of the Fabric consortium chain, realizes the non-tamperable storage and full life cycle traceability of the operation process, and provides a trusted technical foundation for quality audit. Therefore, the management capability of the repair shop in terms of permission control, global resource scheduling, edge response and operation audit can be improved.

[0058] As shown in Figure 4 As a preferred embodiment of the application, the device conflict arbitration agent module 300 includes:

[0059] The conflict prediction unit 301 is configured to calculate the path occupancy rate within a future set time according to the spatial topology of the AGV and the robot, and mark a high-risk conflict when the path overlap rate is greater than 30%.

[0060] The consensus arbitration unit 302 is configured to add a Lamport timestamp vector to the operation log of each device, and if the version offset exceeds the dynamic threshold, a master node is elected to generate a fusion instruction through an AI-driven dynamic consensus mechanism.

[0061] The permission topology updating unit 303 adjusts the device permission tree weight according to the arbitration result, and binds the AGV task delay record to the knowledge graph.

[0062] In the embodiment of the application, the device conflict arbitration intelligent agent module 300 realizes device collaborative security in complex scenarios through a triple protection mechanism. The conflict prediction unit 301 constructs a real-time spatial topology graph based on AGV laser radar point cloud and a kinematics model of the robot, calculates the path occupancy rate in the next few seconds through a trajectory prediction algorithm, and immediately marks a high-risk conflict when detecting that the path overlap rate exceeds 30%, for example, a certain AGV and a certain robot are about to cross trajectories in a material transfer area. The consensus arbitration unit 302 adds a version vector containing a Lamport timestamp to the operation log of each device. When detecting that the version offset between devices exceeds a dynamic threshold (for example, there is a 300 ms deviation between the robot emergency stop instruction and the AGV avoidance signal), the Raft consensus algorithm is automatically triggered to elect a master node to generate a fusion instruction. At this time, the master node decides that the robot suspends work and authorizes the AGV to have priority in passing. The permission topology updating unit 303 dynamically adjusts the device permission tree weight according to the arbitration result (for example, a certain AGV that frequently triggers conflicts will have its path priority coefficient in the permission tree reduced, and the conflict event key parameters (such as delay duration and avoidance path) are bound to the fault mode library of the knowledge graph, providing historical basis for subsequent scheduling decisions. This module significantly shortens the device conflict response time, and through dynamic optimization of the permission tree, reduces the device collaborative failure rate.

[0063] As shown in Figure 5 , as a preferred embodiment of the application, the closed-loop optimization intelligent agent module 400 includes:

[0064] The process deviation analysis unit 401 is used to collect wheel set press-in torque data in real time according to a set collection rate, and when detecting that the standard deviation exceeds the upper limit of the process by 5%, it is marked as a key deviation event.

[0065] The graph reverse optimization unit 402 is used to extract a historical similar case disposal subgraph in the knowledge graph, calculate the matching degree of the new scheme through a graph attention mechanism, and inject rules with confidence.

[0066] The device parameter adaptive unit 403 is used to generate a hydraulic compensation instruction with a specified precision, which is signed by a blockchain and then issued to a PLC, and the wheel set roundness error is fed back to the knowledge graph.

[0067] In the embodiment of the present application, it is assumed that the process deviation analysis unit 401 samples the wheel set press-fitting torque data at a high frequency of 1000 times per second. When it is detected that the standard deviation of 10 consecutive sampling points exceeds the upper limit of the process by 5%, it is automatically marked as a key deviation event. For example, it is detected that the wheel set bearing press-fitting torque fluctuation of a certain bogie reaches ±8%, triggering a system alarm. At this time, the atlas reverse optimization unit 402 extracts the historical similar case disposal subgraph in the knowledge graph (such as searching the case library of torque exceeding the standard of the same type of bearing in the past three years), calculates the matching degree of the new disposal scheme and the historical optimal solution through the graph attention neural network, generates rules with injection confidence weight, for example, the scheme with a matching degree of 92% is recommended to increase the hydraulic compensation value by 0.3MPa, and the system will execute accordingly. The equipment parameter adaptive unit 403 generates a high-precision hydraulic compensation instruction based on the rules, signs it with the SM3 algorithm, and then issues it to the PLC executor through the OPC UA protocol, and feeds back the wheel set roundness error data after execution to the process optimization node of the knowledge graph. In summary, this module improves the wheel set assembly qualification rate and reduces the same type of process deviation every year through a continuous feedback learning mechanism.

[0068] As shown in Figure 6 The embodiment of the present application also provides a whole-process management and control method based on an integrated wheel set maintenance intelligent workshop, which comprises the following steps:

[0069] S100, collecting the state and process data of the equipment, realizing cross-modal alignment to generate a feature vector through an AI model of Zhongke Zhiyuan, and updating a distributed knowledge graph based on a time series graph convolution network, wherein the state and process data comprises flaw detection images and audio data;

[0070] S200, analyzing user permission tags and instructions, generating a scheduling scheme in combination with equipment load, issuing instructions to a device controller through an edge node, and performing blockchain notarization;

[0071] S300, monitoring the spatial topology of AGV and mechanical hands, generating a fusion instruction based on the timestamp verification version number through Raft consensus when the trajectories overlap, and updating the permission topology tree, wherein the spatial topology is the relative position relationship of AGV and mechanical hands;

[0072] S400, analyzing wheel set press-fitting torque data, extracting historical cases of the knowledge graph to generate process adjustment instructions, issuing the instructions to a PLC through blockchain signature, and feeding back the optimization effect, wherein the wheel set press-fitting torque data is the torque applied during the press-fitting process and the torque change trend.

[0073] In the embodiment of the present application, the device state monitoring data and process parameters are deeply integrated, the dynamic evolution knowledge graph is constructed by using cross-modal alignment technology, the information island is completely eliminated, and the intelligent association and semantic penetration of heterogeneous data are realized; based on the real-time analysis of user permission level and device running load, precise scheduling instructions are generated in collaboration, and the auditability and security of the whole process operation are ensured by combining with the distributed evidence technology of blockchain, the abnormal response capability and multi-task cooperation efficiency are strengthened; the trajectory overlap conflict of AGV and mechanical hand equipment is efficiently resolved through the real-time perception of space topology and dynamic consensus mechanism, seamless cooperation and resource optimization configuration of multiple intelligent agents in complex scenes are realized; the knowledge of historical case library is mined, key process parameters such as wheel set press-in torque are actively optimized, and the latest technical standards are continuously synchronized through the knowledge graph self-learning mechanism, which improves the assembly process precision and builds an ecological system of efficient knowledge circulation. Overall, the method solves the core bottlenecks of data fragmentation, low efficiency of cooperation, process lag and knowledge fault in the wheel set maintenance field for a long time by using systematic thinking, and realizes the overall improvement from data fusion, dynamic decision to process closed loop.

[0074] The above only describes the preferred embodiments of the present application in detail, and does not limit the present application, any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0075] It should be understood that although each step in the flowchart of each embodiment of the present application is displayed in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise stated herein, the execution of these steps has no strict sequence limitation, and these steps can be executed in other orders. Moreover, at least part of the steps in each embodiment can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.

[0076] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer readable storage medium, and when the program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, storage, databases, or other media in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0077] Other embodiments of the present disclosure will be apparent to those skilled in the art with the disclosure herein. The present application is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include known equivalents or other adaptations as are generally known in the art to which the present disclosure belongs. The present application is to be limited only by the claims and equivalents thereof.

Claims

1. A full-process AI intelligent agent control system based on an integrated intelligent workshop for wheelset maintenance, characterized in that, The system includes: The multi-source fusion intelligent agent module is used to collect equipment status and process data, generate feature vectors through cross-modal alignment using an AI model, and update a distributed knowledge graph based on a temporal graph convolutional network. The status and process data include flaw detection images and audio data. The role-aware decision-making intelligent agent module is used to parse user permission tags and instructions, generate scheduling schemes in combination with device load, and send instructions to the device controller through edge nodes and store them on the blockchain. The equipment conflict arbitration intelligent agent module is used to monitor the spatial topology of AGV and robot. When the trajectories overlap, it generates a fusion instruction based on the version number verified by the timestamp through Raft consensus and updates the permission topology tree. The spatial topology is the relative positional relationship between AGV and robot. The closed-loop optimization intelligent agent module is used to analyze wheelset pressing torque data, extract historical cases from the knowledge graph to generate process adjustment instructions, send them to the PLC via blockchain signature, and provide feedback on the optimization effect. The wheelset pressing torque data refers to the torque applied during the pressing process and the torque change trend.

2. The AI-powered intelligent control system for the entire process of an integrated wheelset maintenance workshop as described in claim 1, characterized in that, The multi-source fusion intelligent agent module includes: The cross-modal alignment unit is used to input the flaw detection image into the ResNet-50 model to extract crack morphology features. At the same time, the audio data is converted into a 128-dimensional mel spectrum through MFCC, and the industrial-grade CLIP model is used to map to a unified semantic space to generate a fusion feature vector carrying confidence weights. The graph dynamic evolution unit is used to construct a three-layer relational topology with wheelset serial numbers as the main entity, predict the association path between bearing wear and crack propagation based on the temporal graph convolutional network, and automatically expand the subgraph structure when new fault cases are added to the database. The storage optimization unit is used to store knowledge graph subgraphs by production line workstation partition, dynamically migrate data to edge nodes based on query frequency heatmaps, and realize real-time relationship queries through graph databases.

3. The AI-powered intelligent control system for the entire process of an integrated wheelset maintenance workshop as described in claim 1, characterized in that, The role-aware decision-making intelligent agent module includes: The permission-driven parsing unit is used to identify the role tags associated with the user's work badge QR code, parse natural language queries into structured operation chains, and filter out unauthorized instructions. The multi-objective optimization unit is used to establish a multi-objective function of equipment utilization, order delay penalty coefficient, and energy consumption threshold, and uses AI optimization algorithm to generate scheduling scheme; The edge collaboration unit is used to cache less than 50MB of process standard sub-maps on the PAD, decompose complex requests into edge light computing and cloud heavy inference tasks, and enable the system to directly connect to the controller of high-risk equipment via a protocol.

4. The full-process AI intelligent control system for integrated wheelset maintenance workshops as described in claim 3, characterized in that, The role-aware decision-making agent module also includes: The blockchain trusted evidence storage unit is used to sign the issued equipment emergency stop commands and process parameter adjustment operations using the SM2 national cryptographic algorithm, and synchronize the hash value of the operation log to the Fabric consortium chain after IPFS sharding and storage.

5. The full-process AI intelligent control system for integrated wheelset maintenance workshops according to claim 1, characterized in that, The device conflict arbitration intelligent agent module includes: The conflict prediction unit is used to calculate the path occupancy rate within a set time period based on the spatial topology of the AGV and the robot arm. When the path overlap rate is greater than 30%, a high-risk conflict is marked. The consensus arbitration unit is used to attach a Lamport timestamp vector to the operation log of each device. If the version offset exceeds the dynamic threshold, the master node is elected through the AI-driven dynamic consensus mechanism to generate fusion instructions. The permission topology update unit adjusts the weight of the device permission tree based on the arbitration result and binds the AGV task delay record to the knowledge graph.

6. The full-process AI intelligent control system for integrated wheelset maintenance workshops according to claim 1, characterized in that, The closed-loop optimization agent module includes: The process deviation analysis unit is used to collect wheelset pressing torque data in real time according to the set collection rate. When the detection standard deviation exceeds the process upper limit by 5%, it is marked as a critical deviation event. The graph reverse optimization unit is used to extract the historical similar case handling subgraph from the knowledge graph, calculate the matching degree of the new solution through the graph attention mechanism, and inject the confidence rule. The equipment parameter adaptive unit is used to generate hydraulic compensation commands with specified accuracy, which are then sent to the PLC after being signed by the blockchain, and the wheelset roundness error is fed back to the knowledge graph.

7. The full-process AI intelligent control system for integrated wheelset maintenance workshops according to claim 1, characterized in that, The multi-source fusion intelligent agent module also includes an anti-magnetic RFID hardware group, which consists of a high-temperature resistant tag embedded in the inner ring of the wheel axle bearing and a reader / writer deployed at the maintenance station. The ID of the high-temperature resistant tag is bound to the unique identification entity of the wheelset in the knowledge graph, and the hardware group is used to collect the status and process data of the equipment.

Images