Method, device, equipment and medium for recovering abnormal equipment failure based on large model

Abstract

The present application relates to a method, apparatus, device, and medium for recovering abnormal equipment faults based on a large model, wherein the method includes: constructing a RAG knowledge base, and constructing a dictionary corresponding to equipment fault codes and fault descriptions to generate a target fault dictionary; collecting fault signals of logistics automation equipment in real time, and preprocessing the fault signals to generate target fault signals; analyzing and classifying the target fault signals based on the RAG knowledge base using a pre-trained natural language processing large model to generate fault analysis results and fault classification results; determining the MCP service category based on the fault classification results, and recovering the faults of the logistics automation equipment based on the MCP service category; verifying and recording the fault recovery of the logistics automation equipment, obtaining verification results and recording results, and feeding the verification results and recording results back to the RAG knowledge base system. The present application improves the operating efficiency and reliability of logistics automation equipment.

Description

Technical Field

[0001] The present application relates to the technical field of automated logistics integration and stereoscopic warehouses, and in particular to a method, device, equipment and medium for recovering equipment abnormal faults based on a large model. Background Art

[0002] With the widespread adoption of automated logistics systems in e-commerce, manufacturing, retail, and other sectors, the stable operation of logistics automation equipment is crucial to the normal operation of businesses. In modern high-bay warehouses and automated logistics centers, sorting systems, stacking machines, conveyor lines, and other equipment form a complex logistics network. Any failure in these devices can disrupt the entire logistics system, resulting in significant economic losses. Currently, troubleshooting for logistics automation equipment primarily relies on manual intervention. When a device malfunctions, the system issues an alarm signal, and maintenance personnel determine the cause based on the alarm information before conducting on-site inspections and remediation. This traditional troubleshooting approach has several limitations: First, it relies heavily on specialized technicians, requiring maintenance personnel to possess extensive equipment maintenance experience and expertise. Second, fault recovery takes a long time, often extending equipment downtime from the time a fault occurs to the time maintenance personnel arrive and complete the repair. Third, manual troubleshooting makes it difficult to effectively accumulate and reuse troubleshooting experience, and the methods and results of different maintenance personnel can vary significantly. Finally, as the scale and complexity of logistics automation equipment expand, maintenance costs continue to rise, and the economic benefits of manual troubleshooting are gradually declining.

[0003] Large models combined with RAG technology have potential for application in equipment fault diagnosis and predictive maintenance, but there is still a lack of systematic solutions for automatic fault recovery in logistics automation equipment. In existing technologies, some automated equipment has simple self-diagnosis and self-recovery functions, but these functions are usually based on preset rules and processes, lack intelligence and adaptability, and are difficult to deal with complex and changing fault conditions. On the other hand, although there are studies that apply machine learning technology to equipment fault diagnosis, these methods often focus only on fault identification and classification, without involving the automatic recovery process of faults. In addition, most existing fault handling systems are designed for specific types of equipment or faults, lack versatility and scalability, and are difficult to adapt to the needs of different equipment and diverse fault scenarios. Summary of the Invention

[0004] The purpose of the embodiments of the present application is to propose a method, device, equipment and medium for equipment abnormal fault recovery based on a large model, so as to realize the intelligent identification, classification and automatic recovery of logistics automation equipment faults, so as to improve the operating efficiency and reliability of logistics automation equipment.

[0005] In order to solve the above technical problems, the present invention provides a method for recovering from device abnormality failures based on a large model, including:

[0006] Build a RAG knowledge base and a dictionary of device fault codes and fault descriptions to generate a target fault dictionary;

[0007] Collect fault signals of logistics automation equipment in real time, pre-process the fault signals and generate target fault signals;

[0008] Analyze and classify the target fault signal based on the RAG knowledge base through a pre-trained natural language processing large model to generate a fault analysis result and a fault classification result;

[0009] Determine an MCP service category according to the fault classification result, and perform restoration processing on the fault of the logistics automation equipment based on the MCP service category;

[0010] The fault recovery of the logistics automation equipment is verified and recorded, and the verification results and the recording results are obtained, and the verification results and the recording results are fed back to the RAG knowledge base system.

[0011] In order to solve the above technical problems, the present invention provides a device abnormal fault recovery device based on a large model, comprising:

[0012] The knowledge base construction module is used to build the RAG knowledge base and the corresponding dictionary of equipment fault codes and fault descriptions to generate the target fault dictionary;

[0013] A fault signal acquisition module is used to collect fault signals of logistics automation equipment in real time, pre-process the fault signals, and generate target fault signals;

[0014] A fault classification module is used to analyze and classify the target fault signal based on the RAG knowledge base using a pre-trained natural language processing model to generate a fault analysis result and a fault classification result;

[0015] a fault recovery module, configured to determine an MCP service category according to the fault classification result, and perform recovery processing on the fault of the logistics automation equipment based on the MCP service category;

[0016] The fault recovery verification module is used to verify and record the fault recovery of the logistics automation equipment, obtain verification results and record results, and feed back the verification results and record results to the RAG knowledge base system.

[0017] In order to solve the above technical problems, a technical solution adopted by the present invention is: to provide a computer device, including one or more processors; a memory for storing one or more programs, so that the one or more processors can implement any one of the above-mentioned large model-based device abnormal fault recovery methods.

[0018] In order to solve the above technical problems, a technical solution adopted by the present invention is: a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements any one of the above-mentioned large model-based device abnormal fault recovery methods.

[0019] The embodiments of the present invention provide a method, apparatus, device, and medium for recovering abnormal equipment faults based on a large model. The method includes: constructing a RAG knowledge base, and constructing a dictionary corresponding to equipment fault codes and fault descriptions to generate a target fault dictionary; collecting fault signals of logistics automation equipment in real time, and preprocessing the fault signals to generate target fault signals; analyzing and classifying the target fault signals based on the RAG knowledge base using a pre-trained natural language processing large model to generate fault analysis results and fault classification results; determining the MCP service category based on the fault classification results, and recovering the fault of the logistics automation equipment based on the MCP service category; verifying and recording the fault recovery of the logistics automation equipment to obtain verification results and recording results, and feeding the verification results and recording results back to the RAG knowledge base system. The embodiment of the present invention constructs a RAG knowledge base and a target fault dictionary. By collecting fault signals in real time, performing fault analysis and fault classification based on the fault signals, and then determining the MCP service category based on the fault analysis and fault classification results, the fault is restored and verified based on the MCP service category. This realizes the intelligent identification, classification and automatic recovery of logistics automation equipment faults, solves the problems of high manual dependence, long fault recovery time, and difficulty in accumulating and reusing experience in logistics automation equipment fault handling, and is conducive to improving the operating efficiency and reliability of logistics automation equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the solutions in this application, a brief introduction will be given below to the drawings required for use in the description of the embodiments of this application. Obviously, the drawings described below are some embodiments of this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0021] Figure 1 This is a flowchart of the implementation process of the device abnormal fault recovery method based on the large model provided in the embodiment of the present application;

[0022] Figure 2 This is a flowchart for implementing the first sub-process in the large model-based device abnormal fault recovery method provided in an embodiment of the present application;

[0023] Figure 3 This is a flowchart for implementing the second sub-process in the large model-based device abnormal fault recovery method provided in an embodiment of the present application;

[0024] Figure 4 This is a flowchart for implementing the third sub-process in the large model-based device abnormal fault recovery method provided in an embodiment of the present application;

[0025] Figure 5 This is a flowchart for implementing the fourth sub-process in the large model-based device abnormal fault recovery method provided in an embodiment of the present application;

[0026] Figure 6 This is a flowchart for implementing the fifth sub-process in the large model-based device abnormal fault recovery method provided in an embodiment of the present application;

[0027] Figure 7 This is a flowchart for implementing the sixth sub-process in the large model-based device abnormal fault recovery method provided in an embodiment of the present application;

[0028] Figure 8 This is a schematic diagram of a device abnormal fault recovery device based on a large model provided in an embodiment of the present application;

[0029] Figure 9 It is a schematic diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0030] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which this application belongs. The terms used in the specification of the application are for the purpose of describing specific embodiments only and are not intended to limit this application. The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0031] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0032] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0033] The present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0034] It should be noted that the device abnormality fault recovery method based on the large model provided in the embodiment of the present application is generally executed by a server. Accordingly, the device abnormality fault recovery device based on the large model is generally configured in the server.

[0035] The large-model-based device fault recovery method provided in this application is applied to an abnormal fault recovery system. The overall architecture of this abnormal fault recovery system primarily includes a host computer system, a large-model (LLM) fault analysis system, a RAG knowledge base system, an MCP service module, and logistics automation equipment. The host computer system is responsible for monitoring the operating status of logistics automation equipment. When a device experiences an abnormality, it collects fault signals, including information such as the device number, current status, fault code, and, for some devices, a fault description. It should be noted that normal operating data does not flow through the large model; only fault signals are transmitted to the large model for processing. The large-model fault analysis system uses natural language processing technology to intelligently analyze and classify collected fault signals. This system maintains a two-way interaction with the RAG knowledge base system, obtaining domain knowledge support from the knowledge base and feeding back processing results to the knowledge base, continuously enriching and optimizing the knowledge content. The RAG knowledge base system uses retrieval enhancement generation technology to embed with the large model, storing the structural information, functional parameters, fault types, fault characteristics, processing methods and other contents of logistics automation equipment, and specially establishes a corresponding dictionary of equipment fault codes and fault descriptions to provide domain knowledge support for the large model. The MCP service module contains multiple types of services, each of which corresponds to a different equipment fault handling method. The MCP service contains a series of equipment fault recovery business and logic, simulating the manual operation and processing method under normal procedures. The MCP service communicates with the equipment through the TCP / IP protocol and uses customized communication messages to send control instructions. Logistics automation equipment is the application object of this application, mainly including stackers, logistics conveyor lines and AGV robots. The logistics automation equipment receives the instructions sent by the MCP service, performs the corresponding recovery operations, and feeds back the execution results to the host computer system to form a closed-loop control.

[0036] See also Figure 1 , Figure 1 A specific implementation of a method for recovering from equipment abnormality failure based on a large model is shown.

[0037] It should be noted that the method of the present invention is not limited to the method of Figure 1 The process sequence shown is limited to the following steps:

[0038] S1: Build a RAG knowledge base and a dictionary of device fault codes and fault descriptions to generate a target fault dictionary.

[0039] Specifically, the RAG knowledge base refers to a retrieval-enhanced generation knowledge base that contains information about the structure, functional parameters, fault types, fault characteristics, and treatment methods of logistics automation equipment. The target fault dictionary includes information such as fault code, fault name, fault description, possible cause, severity, and recommended treatment methods, providing the foundation for the large model to accurately understand the meaning of the fault.

[0040] See also Figure 2 , Figure 2 A specific implementation of step S1 is shown, which is described in detail as follows:

[0041] S11: Obtain equipment technical documents, historical fault records, and expert experience rule data;

[0042] S12: Perform knowledge extraction on the historical fault records to obtain knowledge extraction data, and construct a corresponding dictionary of equipment fault codes and fault descriptions based on the knowledge extraction data to generate the target fault dictionary. The knowledge extraction data includes fault combination patterns, fault causes, fault handling priorities, and associated equipment modules;

[0043] S13: uniformly processing the equipment technical documents, the historical fault records, and the expert experience rule data into text formats to generate a target text;

[0044] S14: converting the target text into a preset dimension vector;

[0045] S15: construct a vector index, and perform retrieval based on a preset dimension vector using the vector index, and construct the RAG knowledge base according to the index content. The RAG knowledge base includes the structural information, functional parameters, fault types, fault characteristics, and processing methods of the logistics automation equipment.

[0046] Specifically, equipment technical documentation, historical fault records, and expert experience rule data are acquired. Equipment technical documentation describes the logistics automation equipment, including its description, functional parameters, and operating principles. Historical fault records collect historical fault cases, each containing information such as fault description, fault characteristics, handling methods, and handling results. The handling method library stores handling methods for various faults, including operating steps, parameter settings, and precautions. Expert experience rule data, contained in the expert rule library, includes fault handling rules and experience summarized by domain experts. The acquired data can be cleaned to remove invalid data. Knowledge extraction is then performed on the historical fault records. This knowledge extraction involves analyzing the historical fault records, extracting high-frequency fault combination patterns, and extracting three types of manually annotated key data. These three key attributes include the cause of the fault, fault handling priority (e.g., P0 Emergency, P1 Important, P2 Alert), and associated device modules (e.g., positioning system, drive system, and control system).

[0047] Convert equipment technical documentation, historical fault records, and expert experience rule data into a unified text format to generate the target text. The processed text is then converted to a vector using the BGE-large series of models. The vector is expanded from 1024 dimensions to 1536 dimensions using linear interpolation techniques, and the 1536-dimensional vector is normalized. Preset models can include the BGE-large model, text-embedding, or M3E. A vector index is then established, and the converted vector is searched based on the vector index to obtain the index content. Finally, the RAG knowledge base is constructed based on the index content. The vector index can be constructed using methods such as Milvus, Chroma, and Qdrant.

[0048] Regarding knowledge retrieval and update mechanisms, a hybrid retrieval strategy is employed, combining keyword matching, vector similarity calculation, and semantic understanding to achieve fast and accurate knowledge retrieval. Furthermore, continuous knowledge updates are supported, including both periodic and real-time updates. Periodic updates involve the system regularly conducting comprehensive checks and updates of the knowledge base; real-time updates involve the system immediately adding new knowledge points to the knowledge base after resolving each issue, ensuring the timeliness and integrity of the knowledge base.

[0049] S2: real-time collection of fault signals from logistics automation equipment, and pre-processing of the fault signals to generate target fault signals;

[0050] Specifically, the host computer system monitors the operating status of logistics automation equipment and collects fault signals when an abnormality occurs. These fault signals include the device number, current status, fault code, and, for some devices, a description of the fault. After preprocessing the fault signals, such as data cleaning, the target fault signal is generated.

[0051] It should be noted that normal operating data will not flow through the natural language processing model. Only fault signals will be transmitted to the natural language processing model for processing.

[0052] S3: Analyze and classify the target fault signal based on the RAG knowledge base through a pre-trained natural language processing large model to generate a fault analysis result and a fault classification result;

[0053] Specifically, the target fault signal is input into a pre-trained natural language processing large model (LLM). This natural language processing large model is combined with the target fault dictionary and domain knowledge in the RAG knowledge base to perform intelligent analysis of the fault content, infer the possible cause of the fault, and classify the fault into different processing types, determining the MCP service category that needs to be called.

[0054] See also Figure 3 , Figure 3 A specific implementation of step S3 is shown, which is described in detail as follows:

[0055] S31: Retrieving knowledge data corresponding to the target fault signal from the RAG knowledge base through the pre-trained natural language processing large model;

[0056] S32: performing fault analysis based on the knowledge data and the target fault signal to generate the fault analysis result;

[0057] S33: Perform fault classification based on the fault analysis result to generate the fault classification result.

[0058] Specifically, the natural language processing (NLP) large model receives a target fault signal from the host system. It then uses RAG technology to retrieve knowledge related to the current fault from the RAG knowledge base. Specifically, it obtains detailed descriptions of the fault code from the target fault dictionary to generate knowledge data. This knowledge data provides the NLP large model with crucial context for understanding the fault. Combining the target fault signal with the retrieved knowledge data, the NLP large model conducts an in-depth analysis of the fault content, infers the possible causes of the fault, and generates a fault analysis result. This analysis process leverages the large model's powerful natural language understanding and reasoning capabilities, enabling it to handle complex fault scenarios. Finally, based on the fault analysis results, the model performs fault classification and generates a fault classification result.

[0059] S4: Determine an MCP service category according to the fault classification result, and perform restoration processing on the fault of the logistics automation equipment based on the MCP service category;

[0060] Specifically, the MCP (Large Model Context Protocol) service category is determined based on the fault classification results. MCP services encompass a series of equipment fault recovery operations and logic, simulating manual handling methods under normal processes. Different MCP service types correspond to different fault handling strategies, such as task reassignment, material discharge and equipment status restoration, and exception reporting and notification. The selected MCP service communicates with the logistics automation equipment via the TCP / IP protocol, using customized communication messages to send corresponding control instructions. For faults that can be automatically recovered, the MCP service directly performs recovery operations. For faults that cannot be automatically recovered (such as power outages or network interruptions), the MCP service reports the fault via email or to a higher-level system to ensure timely resolution. Communication messages consist of a message header, a command code, a parameter area, and a checksum. The command code specifies the specific operation type, and the parameter area contains the detailed parameters required for the operation. The communication process supports a request-response model to ensure reliable command transmission and execution.

[0061] See also Figure 4 , Figure 4 A specific implementation of step S4 is shown, which is described in detail as follows:

[0062] S41: Determine the MCP service category according to the fault classification result;

[0063] Specifically, MCP service categories include task reassignment services, material removal and equipment status recovery services, and exception reporting and notification services. The MCP protocol is a general context protocol designed for interaction between large models and external environments. It defines the standard process for large models to interpret equipment failure information and invoke corresponding services for processing. The protocol specifies data formats, interaction methods, and service invocation rules to ensure effective communication between large models and devices.

[0064] S42: If the MCP service category is the task re-issuance service, regenerate and issue a task instruction based on the current state of the logistics automation equipment and the task requirements, so that the logistics automation equipment returns to a normal working state;

[0065] Specifically, the task re-issuance service is primarily used to handle failures caused by task interruptions or incorrect task parameters. When logistics automation equipment stops functioning due to task-related issues, this service regenerates and reissues task instructions based on the equipment's current status and task requirements, restoring normal operation. This service automatically adjusts and optimizes task parameters internally to ensure that newly issued tasks are correctly executed by the equipment.

[0066] S43: If the MCP service category is the material discharge and equipment status recovery service, controlling the logistics automation equipment to discharge the abnormal material to a preset abnormality processing port, and resetting the status parameters of the logistics automation equipment to restore the logistics automation equipment to a normal working state;

[0067] Specifically, the material discharge and equipment status recovery services are primarily used to address failures caused by material jams or abnormal equipment status. This service first controls the equipment to discharge the abnormal material to a designated abnormal handling port, then resets the equipment status parameters, restoring the logistics automation equipment to normal operation. This service internally implements intelligent material discharge routing and safe equipment status recovery, preventing secondary failures.

[0068] S44: If the MCP service category is the abnormality reporting and notification service, the fault analysis result and the fault classification result are sent to a preset user terminal and reported to a preset system;

[0069] Specifically, exception reporting and notification services are primarily used to handle faults that cannot be automatically resolved, such as device power outages and network interruptions. This service sends fault information to relevant personnel via email, SMS, or other notification methods, and simultaneously escalates the fault to higher-level systems to ensure timely resolution. The service implements a multi-channel notification and fault escalation mechanism to improve fault response speed.

[0070] See also Figure 5 , Figure 5 A specific implementation of step S41 is shown, which is described in detail as follows:

[0071] S4111: If the logistics automation equipment is a stacker, the fault classification result includes a positioning system fault, a drive system fault, and a control system fault;

[0072] Specifically, logistics automation equipment includes sorting systems, stackers, conveyor lines, shuttles, elevators, AGVs, and other equipment. If the logistics automation equipment is a stacker, the fault classification results include positioning system failures, drive system failures, and control system failures. These failures can cause the stacker to stop working, affecting the normal operation of the logistics system.

[0073] S4112: If the fault is the positioning system fault, the MCP service category is the material discharge and equipment status recovery service;

[0074] Specifically, for positioning system failures, select the material discharge and equipment status recovery service to control the stacker to safely return the goods to their original position, and then perform position recalibration operations.

[0075] S4113: If the fault is the drive system fault, use the MCP service type as the task re-issuance service;

[0076] Specifically, for a drive system failure, choose to re-issue the task service, first clear the alarm, and then re-issue the operation instruction.

[0077] S4114: If the fault is the control system fault, use the MCP service category as the abnormality reporting and notification service;

[0078] Specifically, for control system failures, if they are problems that cannot be automatically recovered, such as communication interruption, the exception reporting and notification service is selected to send the fault information to maintenance personnel.

[0079] The automatic recovery method for stacker crane faults in the embodiment of the present application is targeted, effective, and adaptive, can handle most common faults, significantly reduce the need for manual intervention, and improve equipment availability.

[0080] See also Figure 6 , Figure 6 Another specific implementation of step S41 is shown, which is described in detail as follows:

[0081] S4121: If the logistics automation equipment is a logistics conveyor line, the fault classification results include mechanical jamming fault, motor fault and sensor fault.

[0082] Specifically, common faults in logistics conveyor lines include mechanical jams (such as material blockages and belt slippage), motor failures (such as overload protection and startup failures), and sensor failures (such as false triggering of photoelectric switches and failure of proximity switches). These failures can cause the conveyor line to stop operating, disrupting the flow of logistics.

[0083] S4122: If the fault is the mechanical jam fault, the MCP service category is the material discharge and equipment status recovery service;

[0084] Specifically, for mechanical jam failures, select material discharge and equipment status recovery services to control the conveyor line to run in reverse for a short time, and then resume forward operation to release the jam.

[0085] S4123: If the fault is the motor fault, use the MCP service category as the task re-issuance service;

[0086] Specifically, for motor failure, select the re-issuance task service, first perform a power cycle or reset operation, and then restart the conveyor line.

[0087] S4124: If the fault is a sensor fault, use the MCP service category as the abnormality reporting and notification service;

[0088] Specifically, for sensor failures, if the sensor is dirty or falsely triggered, select the material discharge and equipment status recovery service to perform sensor recalibration operations; if the problem is that the sensor is damaged or cannot be automatically recovered, select the abnormality reporting and notification service to notify maintenance personnel to replace the sensor.

[0089] See also Figure 7 , Figure 7 Another specific implementation of step S41 is shown, which is described in detail as follows:

[0090] S4131: If the logistics automation equipment is an AGV robot, the fault classification results include navigation type fault, mechanical type fault, electrical and battery fault, communication and software fault, and safety and protection fault;

[0091] Specifically, MCP service categories also include navigation recalibration, path replanning, drive parameter adjustment, load balancing, charging task prioritization, charging docking adjustment, communication recovery, task reset, and parameter recovery. If the logistics automation equipment is an AGV robot, the fault classification results include navigation faults, mechanical faults, electrical and battery faults, communication and software faults, and safety and security faults.

[0092] S4132: If the fault is the navigation type fault, adopting the MCP service category of the navigation recalibration service and the path replanning service;

[0093] Specifically, if the fault is a navigation fault, the MCP service categories used are the navigation recalibration service and the path replanning service. Navigation faults include laser navigation sensor failure, positioning deviation or loss, path planning anomalies, navigation map anomalies, obstacle avoidance sensor failure, and path obstruction.

[0094] For positioning deviations or loss failures, select the Navigation Recalibration service to control the AGV robot to rescan environmental feature points for position calibration. This service first causes the AGV robot to stop its current task, then rotate in place one circle to collect environmental feature point data, match it with the navigation map, re-determine its precise position, and finally resume task execution. For path planning anomalies or path blockage failures, select the Path Replanning service to generate a new feasible path for the AGV robot. This service first analyzes the blockage of the current path, then calculates an alternative path based on the latest environmental information, and sends the new path to the AGV robot for execution. For laser navigation sensor failures caused by dirt or slight obstruction, select the Sensor Self-Check service to control the AGV robot to perform a sensor status check and a simple self-cleaning procedure. If self-checking and self-cleaning cannot resolve the problem, the system will escalate the fault level and notify maintenance personnel for manual intervention.

[0095] S4133: If the fault is a mechanical fault, use the MCP service category of the drive parameter adjustment service and the load balancing service;

[0096] Specifically, if the fault is mechanical, the MCP service category used is the drive parameter adjustment service and load balancing service. Mechanical faults primarily include drive wheel failure (slipping, stuck), steering system failure, lift mechanism failure, mechanical structure damage, and load anomalies (overload, unbalanced load).

[0097] For drive wheel slippage, the Drive Parameter Adjustment service adjusts the AGV's speed, acceleration, and torque parameters to adapt to the current ground conditions and restore normal operation. This service first reduces the AGV's speed and acceleration, then gradually adjusts the parameters until the optimal combination for the current conditions is found. For minor load anomalies (such as slight overload), the Load Balancing service controls the AGV to adjust its lifting mechanism or driving posture to balance the load distribution. For severe load anomalies (such as significant overload), the Task Termination service brings the AGV to a safe stop and notifies the upper-level system for manual intervention. For faults that cannot be resolved through software, such as mechanical damage, stuck drive wheels, or lifting mechanism failures, the Abnormal Reporting and Notification service sends the fault information to maintenance personnel and marks the AGV as unavailable to prevent the system from assigning further tasks to the device.

[0098] S4134: If the fault is the electrical and battery fault, adopt the MCP service category of the charging task priority service and the charging docking adjustment service;

[0099] Specifically, if the fault is an electrical or battery failure, the MCP service categories are charging task priority service and charging docking adjustment service. Electrical and battery failures mainly include motor controller failure, line short circuit or open circuit, sensor interface failure, power module failure, low battery charge, poor charging contact, battery aging or damage, and abnormal charging station docking.

[0100] For low battery failures, the Charging Task Priority service is selected to interrupt the AGV's current mission and immediately direct it to the nearest charging station for recharging. This service calculates whether the AGV's current battery level can support the mission. If not, charging is scheduled immediately; if it can, charging is scheduled afterward. For poor charging contact failures, the Charging Docking Adjustment service is selected to direct the AGV to re-docking. This service first removes the AGV from the charging station, then adjusts docking parameters and posture, and re-attempts docking. If docking fails after multiple attempts, the system notifies maintenance personnel for inspection and manual intervention. For partial sensor interface failures or temporary electrical anomalies, the System Reset service is selected to instruct the AGV to perform a soft reboot or reset specific modules. This service ensures safety and orderly shuts down and restarts the AGV's related systems to restore normal functionality. For faults that cannot be resolved through software, such as motor controller hardware failures, short circuits or open circuits, or battery damage, the Abnormal Reporting and Notification service sends fault information to maintenance personnel and simultaneously marks the AGV as unavailable.

[0101] S4135: If the fault is the communication and software fault, adopting the MCP service category of the communication recovery service, the task reset service, and the parameter recovery service;

[0102] Specifically, if the fault is a communication or software failure, the MCP service categories used are communication recovery service, task reset service, and parameter recovery service. Communication and software failures mainly include WiFi signal interruption, communication module failure, data transmission anomalies, server connection failure, task scheduling anomalies, system freeze or crash, software version incompatibility, and parameter configuration errors.

[0103] In the event of a temporary Wi-Fi signal outage or data transmission anomalies, the system will activate the Communication Recovery service, controlling the AGV to switch to a backup communication channel or reestablish a connection. This service will first attempt to reconnect through the primary communication channel. If this fails, it will switch to the backup channel. Simultaneously, the AGV will maintain basic functionality according to its pre-set offline operation strategy until communication is restored. In the event of a task scheduling anomaly, the system will activate the Task Reset service, clearing the AGV's current task queue and re-assigning tasks from the upper-level system. This service ensures task data consistency and prevents the AGV from executing unnecessary or incorrect operations due to incorrect task information. In the event of a parameter configuration error, the Parameter Recovery service will restore the AGV's key parameters to their default values ​​or the last known correct configuration. This service will adjust relevant parameters based on the specific symptoms of the fault, rather than simply resetting all parameters. For complex software failures such as system freezes, crashes, or software version incompatibility, the Software Restart service will activate the AGV to perform a complete software system reboot. If a software reboot fails to resolve the issue, the system will escalate the fault severity and notify the upper-level system for manual intervention.

[0104] S4136: If the fault is a safety and protection fault, use the MCP service category as an exception reporting and notification service;

[0105] Specifically, if the fault is a safety and protection fault, the MCP service category is abnormal reporting and notification services. Safety and protection faults mainly include emergency stop button triggering, collision protection triggering, safety light curtain triggering, and overspeed protection triggering.

[0106] In the event of an emergency stop button or collision protection trigger failure, the system will directly report it to the upper-level system, notifying maintenance personnel to conduct an on-site inspection to prevent special circumstances. In the event of activation of other safety mechanisms, such as safety light curtain triggering or overspeed protection triggering, the corresponding safety recovery service will be selected to gradually restore the AGV robot to normal operation while ensuring environmental safety. These services typically include steps such as environmental re-testing, parameter readjustment, and graded recovery. Because safety and protection failures are directly related to the safety of personnel and equipment, the system adopts a more conservative handling strategy for such failures, relying more on manual confirmation and intervention to ensure safety first.

[0107] S5: Verify and record the fault recovery of the logistics automation equipment, obtain verification results and record results, and feed back the verification results and record results to the RAG knowledge base system.

[0108] Specifically, the process and results of each troubleshooting operation are recorded and evaluated, including information such as fault characteristics, handling methods, execution results, and time consumption. Evaluation results are categorized as "successful," "partially successful," and "failed" for subsequent learning and optimization. The RAG knowledge base is dynamically updated based on feedback from troubleshooting. Successful handling cases are added to the fault case library to enrich the case resources. For partially successful or failed cases, the causes are analyzed, the handling methods are adjusted, and new knowledge points are generated. The system also regularly updates the fault code dictionary to ensure its accuracy and completeness. Knowledge updates utilize incremental learning to maintain knowledge continuity and consistency. The system regularly uses new fault cases and handling experience to fine-tune and optimize the large natural language processing model, enhancing its fault analysis and classification capabilities. The optimization process utilizes a combination of transfer learning and incremental learning to continuously improve model performance while maintaining model stability. Furthermore, the system adjusts RAG technology parameters and strategies based on actual application results to optimize knowledge retrieval and integration. The system self-learning and optimization mechanism of this application is adaptive, continuous and evolutionary, and can continuously improve the intelligence level and processing capabilities of the system as practical experience accumulates.

[0109] In one embodiment, during the stacker crane recovery verification process, the system continuously monitors the crane's operating status to verify whether the fault has been effectively resolved. If the crane resumes normal operation, the system marks the fault as successfully resolved. If the fault persists, the system attempts other recovery strategies or escalates the fault severity, notifying personnel for intervention.

[0110] During the restoration verification of the material conveyor line, the system continues to monitor the operating status of the material conveyor line to verify whether the fault has been effectively resolved. If the conveyor line resumes normal operation, the fault is marked as successfully resolved. If the fault persists, the system will try other recovery strategies or escalate the fault level to notify manual intervention.

[0111] During AGV recovery verification, after the AGV completes its fault recovery operations, the system verifies the recovery results to confirm whether the fault has been effectively resolved. Verification methods include status parameter checks, functional tests, and short-distance trial runs. If the verification results indicate the fault has been successfully resolved, the system determines whether to continue the original task, reassign a new task, or schedule a maintenance inspection based on the AGV's current status and task queue. For tasks interrupted by a fault, the system evaluates whether to restart from the beginning or resume from the point of interruption. If the verification results indicate the fault persists, the system attempts alternative recovery strategies or escalates the fault severity, notifying manual intervention. The system also marks the AGV as unavailable and reassigns its original tasks to other available AGVs, ensuring that the overall workflow is not significantly impacted by a single device failure. During task reallocation, the system considers factors such as each AGV's current location, battery level, load capacity, and task priority to optimize the allocation and improve overall efficiency. The AGV robot fault automatic recovery method of the present application is intelligent, targeted and efficient. It can handle most common faults, significantly reduce the need for manual intervention, improve the availability and work efficiency of AGV robots, and provide a more reliable mobile execution unit for automated logistics systems.

[0112] In an embodiment of the present application, a RAG knowledge base is constructed, and a corresponding dictionary of equipment fault codes and fault descriptions is constructed to generate a target fault dictionary; fault signals of logistics automation equipment are collected in real time, and the fault signals are preprocessed to generate target fault signals; the target fault signals are analyzed and classified based on the RAG knowledge base through a pre-trained natural language processing large model to generate fault analysis results and fault classification results; the MCP service category is determined according to the fault classification result, and the fault of the logistics automation equipment is restored based on the MCP service category; the fault restoration of the logistics automation equipment is verified and recorded to obtain verification results and recording results, and the verification results and recording results are fed back to the RAG knowledge base system. The embodiment of the present invention constructs a RAG knowledge base and a target fault dictionary. By collecting fault signals in real time, performing fault analysis and fault classification based on the fault signals, and then determining the MCP service category based on the fault analysis and fault classification results, the fault is restored and verified based on the MCP service category. This realizes the intelligent identification, classification and automatic recovery of logistics automation equipment faults, solves the problems of high manual dependence, long fault recovery time, and difficulty in accumulating and reusing experience in logistics automation equipment fault handling, and is conducive to improving the operating efficiency and reliability of logistics automation equipment.

[0113] The embodiments of the present application can significantly reduce the need for manual intervention. Through the combination of large models, RAG knowledge bases and MCP services, automatic identification, classification and processing of equipment faults are achieved, which greatly reduces the dependence on professional and technical personnel and enables the equipment to autonomously resume normal operation without human supervision. The embodiments of the present application shorten the fault recovery time. From the occurrence of the fault to the system response, analysis, selection of MCP services and execution of recovery operations, the entire process can be completed in a short time, significantly reducing equipment downtime and improving the overall operating efficiency of the logistics system. The embodiments of the present application improve equipment availability. Through fast and accurate fault handling, the unplanned downtime of the equipment is reduced, the availability and utilization of the equipment are improved, and the processing capacity and stability of the logistics system are enhanced. The embodiments of the present application reduce operation and maintenance costs. Reducing manual intervention and shortening fault recovery time directly reduce labor costs and economic losses caused by equipment downtime, and improve the operating efficiency of the enterprise.

[0114] Please refer to Figure 8 , as a response to the above Figure 1 The present application provides an embodiment of a device for recovering from abnormal faults of equipment based on a large model. Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0115] like Figure 8 As shown, the large model-based device abnormal fault recovery device of this embodiment includes: a knowledge base construction module 61, a fault signal acquisition module 62, a fault classification module 63, a fault recovery module 64 and a fault recovery verification module 65, wherein:

[0116] The knowledge base construction module 61 is used to construct a RAG knowledge base and a corresponding dictionary of device fault codes and fault descriptions to generate a target fault dictionary;

[0117] The fault signal acquisition module 62 is used to collect fault signals of logistics automation equipment in real time, and pre-process the fault signals to generate target fault signals;

[0118] A fault classification module 63 is configured to analyze and classify the target fault signal based on the RAG knowledge base using a pre-trained natural language processing model, and generate a fault analysis result and a fault classification result;

[0119] a fault recovery module 64, configured to determine an MCP service category according to the fault classification result, and perform recovery processing on the fault of the logistics automation equipment based on the MCP service category;

[0120] The fault recovery verification module 65 is used to verify and record the fault recovery of the logistics automation equipment, obtain verification results and record results, and feed back the verification results and record results to the RAG knowledge base system.

[0121] Furthermore, the knowledge base construction module 61 includes:

[0122] Data acquisition unit, used to obtain equipment technical documents, historical fault records and expert experience rule data;

[0123] a dictionary construction unit, configured to perform knowledge extraction on the historical fault records to obtain knowledge extraction data, and construct a corresponding dictionary of equipment fault codes and fault descriptions based on the knowledge extraction data to generate the target fault dictionary, wherein the knowledge extraction data includes fault combination patterns, fault causes, fault handling priorities, and associated equipment modules;

[0124] a target text generating unit, configured to uniformly process the equipment technical documents, the historical fault records, and the expert experience rule data into text formats to generate a target text;

[0125] A vector conversion unit, configured to convert the target text into a vector of a preset dimension;

[0126] An index construction unit is used to construct a vector index, and to search based on a preset dimension vector through the vector index, and to construct the RAG knowledge base according to the index content, wherein the RAG knowledge base includes the structural information, functional parameters, fault types, fault characteristics and processing methods of the logistics automation equipment.

[0127] Furthermore, the fault classification module 63 includes:

[0128] A data retrieval unit, configured to retrieve knowledge data corresponding to the target fault signal from the RAG knowledge base using the pre-trained natural language processing large model;

[0129] a fault analysis unit, configured to perform fault analysis based on the knowledge data and the target fault signal, and generate the fault analysis result;

[0130] The fault classification result generating unit is configured to perform fault classification based on the fault analysis result and generate the fault classification result.

[0131] Furthermore, the MCP service categories include task re-issuance services, material discharge and equipment status recovery services, and exception reporting and notification services; the fault recovery module 64 includes:

[0132] a service category generating unit, configured to determine the MCP service category according to the fault classification result;

[0133] A first restoration unit is configured to regenerate and issue a task instruction based on the current state of the logistics automation equipment and the task requirements if the MCP service category is the task reissuance service, so as to restore the logistics automation equipment to a normal working state;

[0134] a second restoration unit, configured to, if the MCP service category is the material discharge and equipment status recovery service, control the logistics automation equipment to discharge the abnormal material to a preset abnormality processing port, and reset the status parameters of the logistics automation equipment, so that the logistics automation equipment returns to a normal working state;

[0135] The third restoring unit is configured to send the fault analysis result and the fault classification result to a preset user terminal and report to a preset system if the MCP service category is the abnormality reporting and notification service.

[0136] Furthermore, the service category generating unit includes:

[0137] A stacker crane fault classification unit, for, if the logistics automation equipment is a stacker crane, the fault classification result includes a positioning system fault, a drive system fault, and a control system fault;

[0138] a first stacker service determination unit, configured to, if the fault is a positioning system fault, use the MCP service category of the material discharge and equipment status recovery service;

[0139] a second stacker service determination unit, configured to, if the fault is a drive system fault, adopt the MCP service category as the task re-issuance service;

[0140] The third stacker service determination unit is configured to, if the fault is the control system fault, use the MCP service category as the abnormality reporting and notification service.

[0141] Furthermore, the service category generating unit further includes:

[0142] A logistics conveyor line fault classification unit, for if the logistics automation equipment is a logistics conveyor line, the fault classification result includes mechanical jamming fault, motor fault and sensor fault;

[0143] A first logistics conveyor line service determination unit is configured to, if the fault is the mechanical jam fault, use the MCP service category as the material discharge and equipment status recovery service;

[0144] A second logistics conveyor line service determination unit is configured to, if the fault is the motor fault, use the MCP service category as the task re-issuance service;

[0145] The third logistics conveyor line service determination unit is configured to, if the fault is the sensor fault, use the MCP service category as the abnormality reporting and notification service.

[0146] Furthermore, the MCP service category also includes navigation recalibration service, path replanning service, drive parameter adjustment service, load balancing service, charging task priority service, charging docking adjustment service, communication recovery service, task reset service, parameter recovery service and safety detection service; the service category generation unit also includes:

[0147] A robot fault classification unit, for when the logistics automation equipment is an AGV robot, wherein the fault classification results include navigation type fault, mechanical type fault, electrical and battery fault, communication and software fault, and safety and protection fault;

[0148] a first robot service determination unit, configured to, if the fault is the navigation type fault, adopt the MCP service category of the navigation recalibration service and the path replanning service;

[0149] a second robot service determination unit, configured to, if the fault is the mechanical type fault, adopt the MCP service category to provide the drive parameter adjustment service and the load balancing service;

[0150] a third robot service determination unit, configured to, if the fault is the electrical and battery fault, adopt the MCP service category to be the charging task priority service and the charging docking adjustment service;

[0151] a fourth robot service determination unit, configured to, if the fault is the communication and software fault, adopt the MCP service categories of the communication recovery service, the task reset service, and the parameter recovery service;

[0152] The fifth robot service determination unit is configured to adopt the MCP service category as an abnormality reporting and notification service if the fault is the safety and protection fault.

[0153] To solve the above technical problems, the present application also provides a computer device. Figure 9 , Figure 9 This is a basic structural block diagram of the computer device in this embodiment.

[0154] The computer device 7 includes a memory 71, a processor 72, and a network interface 73 that are interconnected through a system bus. It should be noted that Figure 9Only a computer device 7 having three components, memory 71, processor 72, and network interface 73, is shown. However, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead. It should be understood by those skilled in the art that a computer device herein is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0155] Computer devices can be desktop computers, laptops, PDAs, cloud servers, etc. Computer devices can interact with users through keyboards, mice, remote controls, touchpads, or voice-activated devices.

[0156] Memory 71 includes at least one type of readable storage medium, including flash memory, hard disks, multimedia cards, card-type memories (e.g., SD or DX memories), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, magnetic disks, optical disks, and the like. In some embodiments, memory 71 may be an internal storage unit of computer device 7, such as the hard disk or internal memory of computer device 7. In other embodiments, memory 71 may also be an external storage device of computer device 7, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a flash memory card, and the like. Of course, memory 71 may also include both internal and external storage devices of computer device 7. In this embodiment, memory 71 is typically used to store the operating system and various application software installed on computer device 7, such as the program code for a large-scale model-based device failure recovery method. In addition, the memory 71 can also be used to temporarily store various types of data that have been output or are to be output.

[0157] In some embodiments, processor 72 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, a GPU processor, or other data processing chip. Processor 72 is generally used to control the overall operation of computer device 7. In this embodiment, processor 72 is used to execute program code stored in memory 71 or process data, such as executing the program code of the aforementioned large-scale model-based device abnormality fault recovery method to implement various embodiments of the large-scale model-based device abnormality fault recovery method.

[0158] The network interface 73 may include a wireless network interface or a wired network interface. The network interface 73 is generally used to establish a communication connection between the computer device 7 and other electronic devices.

[0159] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores a computer program, and the computer program can be executed by at least one processor to enable the at least one processor to perform the steps of the above-mentioned large model-based device abnormal fault recovery method.

[0160] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of this application, or the part that contributes to the existing technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods of each embodiment of this application.

[0161] Obviously, the embodiments described above are only some of the embodiments of the present application, rather than all of the embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions described in the aforementioned specific embodiments, or to make equivalent replacements for some of the technical features therein. Any equivalent structure made using the contents of the present application specification and the accompanying drawings, directly or indirectly used in other related technical fields, is also within the scope of protection of the present application.

Claims

1. A method for recovering from equipment abnormality based on a large model, characterized in that: include: Build a RAG knowledge base and a dictionary of device fault codes and fault descriptions to generate a target fault dictionary; Collect fault signals of logistics automation equipment in real time, pre-process the fault signals and generate target fault signals; Analyze and classify the target fault signal based on the RAG knowledge base through a pre-trained natural language processing large model to generate a fault analysis result and a fault classification result; Determine an MCP service category according to the fault classification result, and perform restoration processing on the fault of the logistics automation equipment based on the MCP service category; Verify and record the fault recovery of the logistics automation equipment, obtain verification results and record results, and feed the verification results and record results back to the RAG knowledge base system; The MCP service categories include task re-issuance services, material discharge and equipment status recovery services, and exception reporting and notification services; Determining the MCP service category according to the fault classification result, and performing restoration processing on the fault of the logistics automation equipment based on the MCP service category includes: Determine the MCP service category according to the fault classification result; If the MCP service category is the task re-issuance service, then regenerate and issue task instructions based on the current status of the logistics automation equipment and task requirements, so that the logistics automation equipment can be restored to a normal working state; If the MCP service category is the material discharge and equipment status recovery service, the logistics automation equipment is controlled to discharge the abnormal material to the preset abnormal processing port, and the status parameters of the logistics automation equipment are reset to restore the logistics automation equipment to a normal working state; If the MCP service category is the abnormality reporting and notification service, the fault analysis result and the fault classification result are sent to a preset user terminal and reported to a preset system.

2. The method for recovering from equipment abnormality and failure based on a large model according to claim 1, characterized in that: The construction of the RAG knowledge base, the construction of a dictionary corresponding to device fault codes and fault descriptions, and the generation of a target fault dictionary include: Obtain equipment technical documentation, historical fault records, and expert experience rule data; Performing knowledge extraction on the historical fault records to obtain knowledge extraction data, and constructing a corresponding dictionary of equipment fault codes and fault descriptions based on the knowledge extraction data to generate the target fault dictionary, wherein the knowledge extraction data includes fault combination patterns, fault causes, fault handling priorities, and associated equipment modules; The equipment technical documents, the historical fault records and the expert experience rule data are processed into a unified text format to generate a target text; Converting the target text into a vector of preset dimensions; Construct a vector index, and use the vector index to search based on a preset dimension vector, and construct the RAG knowledge base according to the index content, wherein the RAG knowledge base includes the structural information, functional parameters, fault types, fault characteristics and processing methods of the logistics automation equipment.

3. The method for recovering from equipment abnormality and failure based on a large model according to claim 1, characterized in that: The pre-trained natural language processing large model analyzes and classifies the target fault signal based on the RAG knowledge base to generate a fault analysis result and a fault classification result, including: Retrieving knowledge data corresponding to the target fault signal from the RAG knowledge base through the pre-trained natural language processing large model; Performing fault analysis based on the knowledge data and the target fault signal to generate the fault analysis result; Fault classification is performed based on the fault analysis result to generate the fault classification result.

4. The method for recovering from equipment abnormality and failure based on a large model according to claim 1, characterized in that: Determining the MCP service category according to the fault classification result includes: If the logistics automation equipment is a stacker, the fault classification results include positioning system failure, drive system failure and control system failure; If the fault is a fault of the positioning system, the MCP service category used is the material discharge and equipment status recovery service; If the fault is a drive system fault, the MCP service category is the task re-issuance service; If the fault is the control system fault, the MCP service category used is the abnormality reporting and notification service.

5. The method for recovering from equipment abnormality and failure based on a large model according to claim 1, characterized in that: Determining the MCP service category according to the fault classification result further includes: If the logistics automation equipment is a logistics conveyor line, the fault classification results include mechanical jamming fault, motor fault and sensor fault; If the fault is the mechanical jam fault, the MCP service category used is the material discharge and equipment status recovery service; If the fault is the motor fault, the MCP service category is used as the re-issuance task service; If the fault is a sensor fault, the MCP service category used is the abnormality reporting and notification service.

6. The method for recovering from equipment abnormality and failure based on a large model according to claim 1, characterized in that: The MCP service category also includes navigation recalibration service, path replanning service, drive parameter adjustment service, load balancing service, charging task priority service, charging docking adjustment service, communication recovery service, task reset service and parameter recovery service; the MCP service category determined according to the fault classification result also includes: If the logistics automation equipment is an AGV robot, the fault classification results include navigation type fault, mechanical type fault, electrical and battery fault, communication and software fault, and safety and protection fault; If the fault is the navigation type fault, the MCP service category used is the navigation recalibration service and the path replanning service; If the fault is the mechanical type fault, the MCP service category used is the drive parameter adjustment service and the load balancing service; If the fault is the electrical and battery fault, the MCP service category used is the charging task priority service and the charging docking adjustment service; If the fault is the communication and software fault, the MCP service categories used are the communication recovery service, the task reset service, and the parameter recovery service; If the fault is the safety and protection fault, the MCP service category is the abnormality reporting and notification service.

7. A device for recovering from equipment abnormality based on a large model, characterized in that: include: The knowledge base construction module is used to build the RAG knowledge base and the corresponding dictionary of equipment fault codes and fault descriptions to generate the target fault dictionary; A fault signal acquisition module is used to collect fault signals of logistics automation equipment in real time, pre-process the fault signals, and generate target fault signals; A fault classification module is used to analyze and classify the target fault signal based on the RAG knowledge base using a pre-trained natural language processing model to generate a fault analysis result and a fault classification result; a fault recovery module, configured to determine an MCP service category according to the fault classification result, and perform recovery processing on the fault of the logistics automation equipment based on the MCP service category; A fault recovery verification module is used to verify and record the fault recovery of the logistics automation equipment, obtain verification results and record results, and feed the verification results and record results back to the RAG knowledge base system; The MCP service categories include task re-issuance services, material discharge and equipment status recovery services, and exception reporting and notification services; The fault recovery module includes: a service category generating unit, configured to determine the MCP service category according to the fault classification result; A first restoration unit is configured to regenerate and issue a task instruction based on the current state of the logistics automation equipment and the task requirements if the MCP service category is the task reissuance service, so as to restore the logistics automation equipment to a normal working state; a second restoration unit, configured to, if the MCP service category is the material discharge and equipment status recovery service, control the logistics automation equipment to discharge the abnormal material to a preset abnormality processing port, and reset the status parameters of the logistics automation equipment, so that the logistics automation equipment returns to a normal working state; The third restoring unit is configured to send the fault analysis result and the fault classification result to a preset user terminal and report to a preset system if the MCP service category is the abnormality reporting and notification service.

8. A computer device, characterized in that: The device comprises a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the device abnormal fault recovery method based on a large model as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the large model-based device abnormal fault recovery method according to any one of claims 1 to 6.

Images