A method and system for maintenance scheme and spare parts prediction based on a knowledge graph
Patent Information
- Application Number
- CN202610928825.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-25
- Publication Date
- 2026-09-29
AI Technical Summary
[0005]本申请针对现有技术中设备故障诊断结果无法自动转化为定制化检修方案及备件需求预测的技术问题，提供一种基于知识图谱的检修方案与备件预测方法及系统
本申请提供一种基于知识图谱的检修方案与备件预测方法，本发明通过构建包含故障实体、检修工序实体、备件实体及工具实体并定义其映射关系与依赖关系的运维知识图谱，将诊断数据作为查询条件在知识图谱数据中进行路径推理与检索，自动生成包含检修工序序列、备件清单、工具清单及预估工时信息的定制化检修方案，再通过解析备件清单与外部备件库存管理系统的数据交互获取库存状态数据，并基于计划领用量与安全库存量的比较判断生成提前采购预警信息。这一完整的五步闭环架构，将原本依赖人工查阅资料、人工编制方案、人工查询库存的串行流程，转化为由系统自动完成的并行处理流程，使检修方案制定时间从数小时缩短至分钟级，有效避免了人工易遗漏关键工序或备件的问题，实现了从“诊断”到“行动”的全流程智能化贯通。
Smart Images
Figure FT_1 
Figure SMS_1 
Figure SMS_8
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent operation and maintenance technology for industrial equipment, specifically to a maintenance scheme and spare parts prediction method and system based on knowledge graphs. Background Technology
[0002] In capital-intensive industrial sectors such as thermal power, nuclear power, and petrochemicals, the safe and stable operation of equipment is the core guarantee for production activities. Taking thermal power units as an example, their operating environment is characterized by high temperature, high pressure, and high speed. The equipment is diverse, the mechanisms are complex, the failure frequency is high, and the economic losses caused by a single unplanned shutdown are enormous.
[0003] Currently, the industry's work on developing maintenance plans after equipment failures relies entirely on manual maintenance experience, resulting in a very low level of intelligence and automation in the overall process. Once the equipment monitoring and diagnostic system outputs a clear fault diagnosis, maintenance personnel must rely on their personal experience to manually review various equipment technical manuals, maintenance procedures, and work standards—both paper and electronic documents—to determine the appropriate maintenance procedures. Based on these procedures, they manually compile and verify the specific models, specifications, quantities, and accompanying tools for the required spare parts. Then, using past experience, they roughly estimate the overall maintenance time. Finally, they manually check the inventory management system to verify spare parts inventory levels. If a shortage is found, they manually initiate the procurement approval process. The entire maintenance decision-making and resource preparation process is entirely dependent on manual execution.
[0004] Existing traditional maintenance decision-making models and support systems suffer from numerous substantial technical deficiencies, making them ill-suited to the demands of efficient operation and maintenance in modern industry. Firstly, the manual process of reviewing data, outlining procedures, and compiling resource statistics is extremely inefficient, with a single maintenance plan decision taking hours or even days. Furthermore, manual operation is prone to oversights, potentially leading to omissions of critical maintenance procedures and incomplete spare parts statistics, directly resulting in maintenance delays and work stoppages. Secondly, data fragmentation across various business systems is severe. Fault diagnosis systems, maintenance plan generation systems, and spare parts inventory management systems operate independently, forming typical data silos. Fault diagnosis conclusions cannot be automatically converted into standardized, implementable maintenance procedures, nor can spare parts requirements and maintenance time be automatically quantified, completely disrupting the closed-loop link from fault diagnosis to on-site operation and creating a significant gap in business implementation. Thirdly, existing support systems can only output generalized, standardized maintenance suggestions, unable to dynamically adapt to specific fault types, severity levels, and real-time equipment operating status. They lack the customized generation capabilities for maintenance procedure reorganization, intelligent spare parts matching, and accurate time inference. Fourth, the disconnect between spare parts demand forecasting and inventory management systems makes it impossible to achieve dynamic early warning and advance procurement of spare parts based on fault diagnosis results. This frequently leads to spare parts shortages after maintenance starts, significantly extending the duration of unplanned equipment downtime and hindering the improvement of industrial production and maintenance efficiency. Establishing an intelligent decision-making and execution closed loop from fault diagnosis to on-site maintenance has become a key technical challenge that urgently needs to be solved in this field. Summary of the Invention
[0005] This application addresses the technical problem in the prior art that equipment fault diagnosis results cannot be automatically converted into customized maintenance plans and spare parts demand predictions, by providing a knowledge graph-based method and system for maintenance plans and spare parts prediction.
[0006] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application provides a maintenance scheme and spare parts prediction method based on knowledge graphs, including the following steps: S1, Obtain diagnostic results information from the equipment monitoring and diagnostic system to obtain diagnostic data including faulty equipment identification, fault type and fault severity; S2, construct and store the operation and maintenance knowledge graph to obtain knowledge graph data containing fault entities, maintenance process entities, spare parts entities and tool entities, wherein the operation and maintenance knowledge graph defines the mapping relationship between faults and maintenance processes, as well as the dependency relationship between maintenance processes and required spare parts and tools. S3, using the diagnostic data as query conditions, perform path reasoning and retrieval in the knowledge graph data to generate a customized maintenance plan that includes a maintenance process sequence, a spare parts list and tool list associated with each maintenance process, and estimated working hours. S4. Parse the generated spare parts list, obtain the model and quantity of each spare part, and interact with the external spare parts inventory management system to obtain the inventory status data of each spare part. S5. Based on the obtained model and quantity of the spare parts and the inventory status data, determine whether the current inventory of any spare part is lower than the sum of the planned requisition quantity and the safety stock quantity triggered by the customized maintenance plan. If so, generate and output advance procurement warning information for the spare part.
[0007] Furthermore, the specific process of S1 is as follows: The system establishes a connection with the device monitoring and diagnostic system through the application programming interface, receives the diagnostic results information output by the device monitoring and diagnostic system in real time, and parses the faulty device identifier, fault type and fault severity from the diagnostic results information as the diagnostic data.
[0008] Furthermore, the specific process of S2 is as follows: extracting triplet information from equipment historical maintenance records, equipment technical manuals and expert experience rules using natural language processing technology to define fault entities, maintenance process entities, spare parts entities and tool entities, as well as the relationships between each entity; storing the extracted triplet information in a graph database and establishing a semantic index for each entity and relationship to obtain the knowledge graph data.
[0009] Furthermore, the specific process of S3 is as follows: The fault types in the diagnostic data are matched with the fault entity nodes in the operation and maintenance knowledge graph to determine the starting node; Starting from the starting node, one or more paths composed of maintenance procedure entity nodes are deduced based on the mapping relationship between the fault and the maintenance procedure, and the optimal path is selected as the maintenance procedure sequence according to the preset priority rules. The spare parts entity nodes and tool entity nodes associated with each maintenance process entity node are queried along the optimal path, and the query results are aggregated to generate the spare parts list and the tool list; The preset standard working time attribute values of each maintenance process entity node in the optimal path are obtained and accumulated. Then, the preset weighting coefficient is called to correct the estimated working time information based on the severity of the fault in the diagnostic data. The maintenance procedure sequence, the spare parts list, the tool list, and the estimated working hours information are combined to form the customized maintenance plan.
[0010] Furthermore, the specific process of S4 is as follows: Extract the spare parts list from the customized maintenance plan; The spare parts list is parsed to obtain the name, model and planned requisition quantity of each spare part, and a connection is established with the external spare parts inventory management system through the application programming interface. Send a query request containing the models of each spare part to the external spare parts inventory management system; The system receives the current inventory, in-transit quantity, and safety stock threshold of each spare part from the external spare parts inventory management system as the inventory status data.
[0011] Furthermore, the specific process of S5 is as follows: The total demand for a future period is calculated based on the planned requisition quantity of each spare part in the spare parts list in the customized maintenance plan and the preset equipment deterioration trend correction factor; the current inventory, in-transit quantity and safety stock threshold of each spare part are queried and obtained in real time through the application programming interface to establish a connection with the external spare parts inventory management system. The advance procurement warning information is generated when it is determined that the sum of the current inventory and the quantity in transit is less than the sum of the total demand and the safety stock threshold. The warning information includes the spare parts name, model, recommended purchase quantity, and recommended delivery date.
[0012] Secondly, this application provides a maintenance scheme and spare parts prediction system based on knowledge graphs, including: The data acquisition module is used to acquire diagnostic results information from the equipment monitoring and diagnostic system, and obtain diagnostic data including faulty equipment identification, fault type and fault severity. The knowledge graph construction and storage module is used to construct and store the operation and maintenance knowledge graph, and obtain knowledge graph data containing fault entities, maintenance process entities, spare parts entities and tool entities. The operation and maintenance knowledge graph defines the mapping relationship between faults and maintenance processes, as well as the dependency relationship between maintenance processes and required spare parts and tools. The maintenance plan generation module is connected to the data acquisition module and the knowledge graph construction and storage module, respectively. It is used to use the diagnostic data as query conditions to perform path reasoning and retrieval in the knowledge graph data, and generate a customized maintenance plan that includes a maintenance process sequence, a spare parts list and a tool list associated with each maintenance process, and estimated working hours. The spare parts demand prediction module is connected to the maintenance plan generation module. It is used to parse the spare parts list in the customized maintenance plan, obtain the model and quantity of each spare part, and interact with the external spare parts inventory management system to obtain the inventory status data of each spare part. The early warning generation module is used to determine, based on the model and quantity of the spare parts and the inventory status data, whether the current inventory of any spare part is lower than the sum of the planned requisition quantity and the safety stock quantity triggered by the customized maintenance plan. If so, it generates and outputs an early warning information for the early procurement of the spare part.
[0013] Furthermore, the data acquisition module establishes a connection with the device monitoring and diagnostic system through an application programming interface, receives the diagnostic result information output by the device monitoring and diagnostic system in real time, and parses the faulty device identifier, fault type and fault severity from the diagnostic result information as the diagnostic data.
[0014] Furthermore, the knowledge graph construction and storage module includes: an information extraction unit, used to extract triple information from equipment historical maintenance records, equipment technical manuals, and expert experience rules using natural language processing technology to define fault entities, maintenance process entities, spare parts entities, tool entities, and the relationships between these entities; and a graph database management unit, connected to the information extraction unit, used to store the extracted triple information into a graph database and establish a semantic index for each entity and relationship to obtain the knowledge graph data.
[0015] Furthermore, the maintenance plan generation module includes: The path matching unit is used to match the fault types in the diagnostic data with the fault entity nodes in the operation and maintenance knowledge graph to determine the starting node; The sequence reasoning unit is used to deduce one or more paths consisting of maintenance process entity nodes from the starting node based on the mapping relationship between the fault and the maintenance process, and select the optimal path as the maintenance process sequence according to the preset priority rules. The resource aggregation unit is used to query the spare parts entity nodes and tool entity nodes associated with each maintenance process entity node along the optimal path and aggregate the query results to generate the spare parts list and the tool list. The time calculation unit is used to obtain the preset standard time attribute values of each maintenance process entity node in the optimal path, accumulate them, and then call the preset weighting coefficient to correct them according to the severity of the fault in the diagnostic data to obtain the estimated time information. It also combines the maintenance process sequence, the spare parts list, the tool list and the estimated time information into the customized maintenance plan.
[0016] Compared with the prior art, this application has the following beneficial effects: This application provides a knowledge graph-based maintenance plan and spare parts prediction method. The invention constructs an operation and maintenance knowledge graph containing fault entities, maintenance process entities, spare parts entities, and tool entities, defining their mapping relationships and dependencies. Diagnostic data is used as query conditions to perform path reasoning and retrieval within the knowledge graph data, automatically generating a customized maintenance plan containing maintenance process sequences, spare parts lists, tool lists, and estimated working hours. Then, by parsing the spare parts list and interacting with data from an external spare parts inventory management system, inventory status data is obtained, and advance procurement warning information is generated based on a comparison of planned usage and safety stock. This complete five-step closed-loop architecture transforms the original sequential process of manually reviewing materials, manually compiling plans, and manually querying inventory into a parallel processing flow automatically completed by the system. This reduces maintenance plan development time from hours to minutes, effectively avoiding the problem of human error in overlooking key processes or spare parts, and achieving intelligent integration of the entire process from "diagnosis" to "action."
[0017] Specifically, this invention utilizes natural language processing technology in the knowledge graph construction process to extract triplet information from historical maintenance records, technical manuals, and expert experience rules, and stores it in a graph database to establish a semantic index. This allows fragmented expert experience and historical knowledge to be structurally accumulated and reused, effectively solving the problems of knowledge gaps and experience dependence in traditional operation and maintenance. Furthermore, in the maintenance plan generation process, the starting node is determined by matching the fault type with the fault entity node, the path is inferred based on the mapping relationship, and the optimal path is selected according to priority rules. At the same time, spare parts and tool information is aggregated along the optimal path, realizing the customization of the plan under different maintenance objectives. In particular, by obtaining and accumulating the preset standard working hour attribute values of each process, and then correcting them according to the severity of the fault by calling a weighting coefficient, the estimated working hours are obtained, solving the problem of inaccurate estimation of traditional fixed working hours. Attached Figure Description
[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0019] Figure 1 This is a schematic diagram of a maintenance scheme and spare parts prediction method based on knowledge graphs according to this application. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0021] See Figure 1 This application provides a maintenance scheme and spare parts prediction method based on knowledge graphs, including the following steps: S1, Obtain diagnostic results information from the equipment monitoring and diagnostic system to obtain diagnostic data including faulty equipment identification, fault type and fault severity; Specifically, step S1 involves establishing a connection with the equipment monitoring and diagnostic system via an application programming interface (API), receiving diagnostic results from the system in real time, and parsing the faulty equipment identifier, fault type, and fault severity from these results as diagnostic data. Standardized API access and a unified diagnostic data format (including equipment identifier, fault type, and severity) provide standardized and structured input for subsequent maintenance plan generation.
[0022] In some embodiments of this application, the equipment monitoring and diagnosis system is an integrated intelligent monitoring and diagnosis management system for thermal power unit equipment based on a time-series large model. The output of this system includes fault equipment identification (e.g., "#3 turbine"), fault type (e.g., "bearing wear"), and fault severity (e.g., mild, moderate, severe, or expressed as a quantitative value, such as 0-100%).
[0023] S2, construct and store the operation and maintenance knowledge graph to obtain knowledge graph data containing fault entities, maintenance process entities, spare parts entities and tool entities, wherein the operation and maintenance knowledge graph defines the mapping relationship between faults and maintenance processes, as well as the dependency relationship between maintenance processes and required spare parts and tools. Specifically, step S2 involves extracting triplet information from historical equipment maintenance records, equipment technical manuals, and expert experience rules using natural language processing technology. This triplet information is used to define fault entities, maintenance process entities, spare parts entities, tool entities, and the relationships between these entities. The extracted triplet information is then stored in a graph database, and a semantic index is established for each entity and relationship to obtain the knowledge graph data.
[0024] In some embodiments of this application, the fault entity includes at least fault type entities of the boiler body, turbine, feedwater pump, fan, or coal mill in a thermal power unit, such as "bearing wear," "rotor imbalance," or "valve jamming." The maintenance process entity includes at least disassembly, inspection, replacement, or reassembly processes for the aforementioned faults. The spare parts entity includes attributes such as spare parts name, model, and specifications. The tool entity includes attributes such as tool name and specifications.
[0025] The specific relationships include: the mapping relationship between the fault entity and the maintenance process entity, for example, bearing wear is mapped to the process sequence of disassembling the bearing end cover → checking the bearing clearance → replacing the bearing → reassembling the bearing end cover; the dependency relationship between the maintenance process entity and the spare parts entity, for example, the bearing replacement process depends on the bearing (model SKF6308) spare parts; and the dependency relationship between the maintenance process entity and the tool entity, for example, the bearing end cover disassembly process depends on the Allen wrench (specification 8mm) tool.
[0026] The constructed operations and maintenance knowledge graph transforms scattered, unstructured maintenance knowledge into structured, computable knowledge representations, solving the problems of difficulty in accumulating expert experience and knowledge gaps in existing technologies. Through graph database storage and semantic indexing, subsequent path reasoning and retrieval can be completed in milliseconds, significantly improving knowledge utilization efficiency.
[0027] S3, using the diagnostic data as query conditions, perform path reasoning and retrieval in the knowledge graph data to generate a customized maintenance plan that includes a maintenance process sequence, a spare parts list and tool list associated with each maintenance process, and estimated working hours. Specifically, the process of step S3 is as follows: The fault types in the diagnostic data are matched with the fault entity nodes in the operation and maintenance knowledge graph to determine the starting node; through fault entity node matching and path reasoning, the abstract fault types are transformed into specific and orderly maintenance procedure sequences, which solves the problem of low efficiency of manual procedure review.
[0028] Starting from the initial node, one or more paths composed of maintenance procedure entity nodes are deduced based on the mapping relationship between the fault and the maintenance procedure. The optimal path is selected as the maintenance procedure sequence according to preset priority rules. The preset priority rules include the principle of highest historical maintenance success rate, the principle of lowest total maintenance cost, or the principle of shortest planned downtime. By aggregating spare parts and tool information along the path, a complete resource list is automatically generated, solving the problem of easy omissions when manually searching for spare parts.
[0029] The spare parts entity nodes and tool entity nodes associated with each maintenance process entity node are queried along the optimal path, and the query results are aggregated to generate the spare parts list and the tool list; The system obtains and accumulates the preset standard working time attribute values of each maintenance process entity node in the optimal path, and then corrects them according to the severity of the fault in the diagnostic data by calling a preset weighting coefficient to obtain the estimated working time information. By introducing fault severity weighted correction of the standard working time, the working time estimation is more realistic, solving the problem of inaccurate traditional fixed working time estimation.
[0030] Specifically, the mathematical expression for the estimated working hours is:
[0031] Indicates the estimated total working hours; This indicates the total number of procedures in the maintenance procedure sequence; Indicates the first The standard working time attribute value is preset for each maintenance process entity node, in hours; This indicates a preset weighting coefficient applied based on the severity of the fault. When the fault severity is mild, The value ranges from 0.8 to 1.0; for moderate cases, the value ranges from 1.0 to 1.2; and for severe cases, the value ranges from 1.2 to 1.5.
[0032] The maintenance procedure sequence, the spare parts list, the tool list, and the estimated working hours information are combined to form the customized maintenance plan.
[0033] S4. Parse the generated spare parts list, obtain the model and quantity of each spare part, and interact with the external spare parts inventory management system to obtain the inventory status data of each spare part. In a more specific embodiment provided in this application, step S4 is performed as follows: Extract the spare parts list from the customized maintenance plan; The spare parts list is parsed to obtain the name, model and planned requisition quantity of each spare part, and a connection is established with the external spare parts inventory management system through the application programming interface. Send a query request containing the models of each spare part to the external spare parts inventory management system; The system receives the current inventory, in-transit quantity, and safety stock threshold of each spare part from the external spare parts inventory management system as the inventory status data.
[0034] S5. Based on the obtained model and quantity of the spare parts and the inventory status data, determine whether the current inventory of any spare part is lower than the sum of the planned requisition quantity and the safety stock quantity triggered by the customized maintenance plan. If so, generate and output advance procurement warning information for the spare part.
[0035] In a more specific embodiment provided in this application, step S5 is performed as follows: The total demand over a future period is calculated based on the planned usage of each spare part in the spare parts list of the customized maintenance plan, combined with a preset equipment deterioration trend correction factor.
[0036] Specifically, the mathematical expression for total demand is:
[0037] in, This indicates the total demand over a future period of time. This indicates the planned usage quantity of the spare part in the spare parts list; This represents the preset equipment degradation trend correction factor, with a value greater than or equal to 1. When the equipment is in a rapid degradation phase... Choose a larger value (e.g., 1.2-1.5) when the equipment is in a stable operating period. The value is 1.0.
[0038] The system establishes a connection with the external spare parts inventory management system through the application programming interface (API) to query and obtain the current inventory, in-transit quantity, and safety stock threshold of each spare part in real time.
[0039] The advance procurement warning information is generated when it is determined that the sum of the current inventory and the quantity in transit is less than the sum of the total demand and the safety stock threshold.
[0040] Specifically, the mathematical expression for early warning judgment is: like This will trigger an early procurement warning.
[0041] in, Indicates the current inventory of spare parts; This indicates the quantity of spare parts in transit (the quantity that has been purchased but not yet put into storage). This indicates the total demand for spare parts; This indicates the safety stock threshold for spare parts.
[0042] The warning information includes the spare part name, model, suggested purchase quantity, and suggested delivery date. The mathematical expression for the suggested purchase quantity is:
[0043] The recommended delivery date is calculated based on the current date and the planned maintenance start date, using the following formula:
[0044] in To improve the lead time for spare parts procurement, This is the buffer time.
[0045] This application also provides a knowledge graph-based maintenance scheme and spare parts prediction system, which is used to perform the steps in the above method embodiments.
[0046] Specifically, the system includes: The data acquisition module is used to execute step S1, acquiring diagnostic result information from the equipment monitoring and diagnostic system to obtain diagnostic data containing the faulty equipment identifier, fault type, and fault severity. The data acquisition module establishes a connection with the equipment monitoring and diagnostic system through an application programming interface (API), receives the diagnostic result information output by the equipment monitoring and diagnostic system in real time, and parses the faulty equipment identifier, fault type, and fault severity from the diagnostic result information as the diagnostic data.
[0047] The knowledge graph construction and storage module is used to execute step S2, construct and store the operation and maintenance knowledge graph, and obtain knowledge graph data containing fault entities, maintenance process entities, spare parts entities, and tool entities. The knowledge graph construction and storage module includes: an information extraction unit, used to extract triplet information from historical equipment maintenance records, equipment technical manuals, and expert experience rules using natural language processing technology to define fault entities, maintenance process entities, spare parts entities, tool entities, and the relationships between these entities; and a graph database management unit, connected to the information extraction unit, used to store the extracted triplet information in a graph database and establish a semantic index for each entity and relationship to obtain the knowledge graph data.
[0048] The knowledge graph construction and storage module is used to execute step S2, construct and store the operation and maintenance knowledge graph, and obtain knowledge graph data containing fault entities, maintenance process entities, spare parts entities, and tool entities. The knowledge graph construction and storage module includes: an information extraction unit, used to extract triplet information from historical equipment maintenance records, equipment technical manuals, and expert experience rules using natural language processing technology to define fault entities, maintenance process entities, spare parts entities, tool entities, and the relationships between these entities; and a graph database management unit, connected to the information extraction unit, used to store the extracted triplet information in a graph database and establish a semantic index for each entity and relationship to obtain the knowledge graph data.
[0049] The spare parts demand prediction module is connected to the maintenance plan generation module and is used to execute step S4, parse the spare parts list in the customized maintenance plan, obtain the model and quantity of each spare part, and interact with the external spare parts inventory management system to obtain the inventory status data of each spare part.
[0050] The early warning generation module, connected to the spare parts demand prediction module, is used to execute step S5. Based on the model and quantity of the spare parts and the inventory status data, it determines whether the current inventory of any spare part is lower than the sum of the planned requisition quantity and the safety stock quantity triggered by the customized maintenance plan. If so, it generates and outputs an early warning information for the early procurement of the spare part.
[0051] The specific workflows of each of the above modules have been described in detail in the method embodiments, and will not be repeated here.
[0052] The above-described embodiments are merely preferred embodiments of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.
Claims
1. A maintenance scheme and spare parts prediction method based on knowledge graph, characterized in that, Includes the following steps: S1, Obtain diagnostic results information from the equipment monitoring and diagnostic system to obtain diagnostic data including faulty equipment identification, fault type and fault severity; S2, construct and store the operation and maintenance knowledge graph to obtain knowledge graph data containing fault entities, maintenance process entities, spare parts entities and tool entities, wherein the operation and maintenance knowledge graph defines the mapping relationship between faults and maintenance processes, as well as the dependency relationship between maintenance processes and required spare parts and tools. S3, using the diagnostic data as query conditions, perform path reasoning and retrieval in the knowledge graph data to generate a customized maintenance plan that includes a maintenance process sequence, a spare parts list and tool list associated with each maintenance process, and estimated working hours. S4. Parse the generated spare parts list, obtain the model and quantity of each spare part, and interact with the external spare parts inventory management system to obtain the inventory status data of each spare part. S5. Based on the obtained model and quantity of the spare parts and the inventory status data, determine whether the current inventory of any spare part is lower than the sum of the planned requisition quantity and the safety stock quantity triggered by the customized maintenance plan. If so, generate and output advance procurement warning information for the spare part.
2. The maintenance scheme and spare parts prediction method based on knowledge graphs according to claim 1, characterized in that, The specific process of S1 is as follows: The system establishes a connection with the device monitoring and diagnostic system through the application programming interface, receives the diagnostic results information output by the device monitoring and diagnostic system in real time, and parses the faulty device identifier, fault type and fault severity from the diagnostic results information as the diagnostic data.
3. The maintenance scheme and spare parts prediction method based on knowledge graphs according to claim 1, characterized in that, The specific process of S2 is as follows: extract triplet information from the equipment's historical maintenance records, equipment technical manuals, and expert experience rules using natural language processing technology to define fault entities, maintenance process entities, spare parts entities, tool entities, and the relationships between entities. Store the extracted triplet information in a graph database and establish a semantic index for each entity and relationship to obtain the knowledge graph data.
4. The maintenance scheme and spare parts prediction method based on knowledge graphs according to claim 1, characterized in that, The specific process of S3 is as follows: The fault types in the diagnostic data are matched with the fault entity nodes in the operation and maintenance knowledge graph to determine the starting node; Starting from the starting node, one or more paths composed of maintenance procedure entity nodes are deduced based on the mapping relationship between the fault and the maintenance procedure, and the optimal path is selected as the maintenance procedure sequence according to the preset priority rules. The spare parts entity nodes and tool entity nodes associated with each maintenance process entity node are queried along the optimal path, and the query results are aggregated to generate the spare parts list and the tool list; The preset standard working time attribute values of each maintenance process entity node in the optimal path are obtained and accumulated. Then, the preset weighting coefficient is called to correct the estimated working time information based on the severity of the fault in the diagnostic data. The maintenance procedure sequence, the spare parts list, the tool list, and the estimated working hours information are combined to form the customized maintenance plan.
5. The maintenance scheme and spare parts prediction method based on knowledge graphs according to claim 1, characterized in that, The specific process of S4 is as follows: Extract the spare parts list from the customized maintenance plan; The spare parts list is parsed to obtain the name, model and planned requisition quantity of each spare part, and a connection is established with the external spare parts inventory management system through the application programming interface. Send a query request containing the models of each spare part to the external spare parts inventory management system; The system receives the current inventory, in-transit quantity, and safety stock threshold of each spare part from the external spare parts inventory management system as the inventory status data.
6. The maintenance scheme and spare parts prediction method based on knowledge graphs according to claim 1, characterized in that, The specific process of S5 is as follows: The total demand for a future period is calculated based on the planned usage of each spare part in the spare parts list in the customized maintenance plan, combined with the preset equipment deterioration trend correction factor. The system establishes a connection with the external spare parts inventory management system through the application programming interface to query and obtain the current inventory, in-transit quantity, and safety stock threshold of each spare part in real time. The advance procurement warning information is generated when it is determined that the sum of the current inventory and the quantity in transit is less than the sum of the total demand and the safety stock threshold. The warning information includes the spare parts name, model, recommended purchase quantity, and recommended delivery date.
7. A maintenance scheme and spare parts prediction system based on knowledge graphs, characterized in that, include: The data acquisition module is used to acquire diagnostic results information from the equipment monitoring and diagnostic system, and obtain diagnostic data including faulty equipment identification, fault type and fault severity. The knowledge graph construction and storage module is used to construct and store the operation and maintenance knowledge graph, and obtain knowledge graph data containing fault entities, maintenance process entities, spare parts entities and tool entities. The operation and maintenance knowledge graph defines the mapping relationship between faults and maintenance processes, as well as the dependency relationship between maintenance processes and required spare parts and tools. The maintenance plan generation module is connected to the data acquisition module and the knowledge graph construction and storage module, respectively. It is used to use the diagnostic data as query conditions to perform path reasoning and retrieval in the knowledge graph data, and generate a customized maintenance plan that includes a maintenance process sequence, a spare parts list and a tool list associated with each maintenance process, and estimated working hours. The spare parts demand prediction module is connected to the maintenance plan generation module. It is used to parse the spare parts list in the customized maintenance plan, obtain the model and quantity of each spare part, and interact with the external spare parts inventory management system to obtain the inventory status data of each spare part. The early warning generation module is used to determine, based on the model and quantity of the spare parts and the inventory status data, whether the current inventory of any spare part is lower than the sum of the planned requisition quantity and the safety stock quantity triggered by the customized maintenance plan. If so, it generates and outputs an early warning information for the early procurement of the spare part.
8. The maintenance scheme and spare parts prediction system based on knowledge graphs according to claim 7, characterized in that, The data acquisition module establishes a connection with the device monitoring and diagnostic system through the application programming interface, receives the diagnostic result information output by the device monitoring and diagnostic system in real time, and parses the faulty device identifier, fault type and fault severity from the diagnostic result information as the diagnostic data.
9. The maintenance scheme and spare parts prediction system based on knowledge graphs according to claim 7, characterized in that, The knowledge graph construction and storage module includes: an information extraction unit, used to extract triple information from equipment historical maintenance records, equipment technical manuals, and expert experience rules using natural language processing technology to define fault entities, maintenance process entities, spare parts entities, tool entities, and the relationships between entities; and a graph database management unit, connected to the information extraction unit, used to store the extracted triple information into the graph database and establish a semantic index for each entity and relationship to obtain the knowledge graph data.
10. A maintenance scheme and spare parts prediction system based on knowledge graphs according to claim 7, characterized in that, The maintenance plan generation module includes: The path matching unit is used to match the fault types in the diagnostic data with the fault entity nodes in the operation and maintenance knowledge graph to determine the starting node; The sequence reasoning unit is used to deduce one or more paths consisting of maintenance process entity nodes from the starting node based on the mapping relationship between the fault and the maintenance process, and select the optimal path as the maintenance process sequence according to the preset priority rules. The resource aggregation unit is used to query the spare parts entity nodes and tool entity nodes associated with each maintenance process entity node along the optimal path and aggregate the query results to generate the spare parts list and the tool list. The time calculation unit is used to obtain the preset standard time attribute values of each maintenance process entity node in the optimal path, accumulate them, and then call the preset weighting coefficient to correct them according to the severity of the fault in the diagnostic data to obtain the estimated time information. It also combines the maintenance process sequence, the spare parts list, the tool list and the estimated time information into the customized maintenance plan.