Work ticket generation method and system based on power grid knowledge logic and topology constraint

Abstract

The invention provides a power grid knowledge logic and topology constraint-based work ticket generation method and system, and belongs to the field of power system maintenance, and the method specifically comprises the following steps: preloading a knowledge document which records a maintenance operation safety measure thinking chain in a natural language form; receiving a maintenance task, and performing knowledge retrieval and matching on equipment involved in the maintenance task to obtain knowledge fragments; fusing the knowledge fragments to generate a preliminary logic framework of the maintenance task work ticket; constructing a power grid topological graph model by taking the power equipment as a node and the electrical connection between the equipment as an edge, and calculating a topological analysis boundary of the target equipment of the maintenance task; and generating a structured prompt in combination with the preliminary logic framework and the topology analysis boundary, sending the structured prompt to the large language model, and generating a work ticket. According to the method, the limitation of a general artificial intelligence model in the professional field is effectively overcome, and the safety and credibility level and the generation efficiency of automatically generating the work ticket are improved.

Description

Technical Field

[0001] This invention belongs to the field of power system maintenance, specifically relating to a method and system for generating work tickets based on power grid knowledge logic and topology constraints. Background Technology

[0002] Traditional work permits rely on the professional knowledge and on-site judgment of staff. Staff need to manually look up the detailed information of the equipment and fill it out one by one according to the safety measures. This manual input system adapts to different work categories and power outage methods through preset structured forms. Although it realizes the electronic circulation of work permits, the degree of automation is limited. At present, the technology for automatically generating work permits is developing in a diversified direction, mainly focusing on three technical paths. To improve efficiency, the typical ticket database and historical ticket reuse technology have emerged. This technology builds a searchable database by accumulating standard ticket samples and historical approval tickets for common work scenarios. When faced with highly repetitive or similar maintenance tasks, users can retrieve similar tickets, make partial modifications, and apply them directly, thereby significantly reducing repetitive work. However, with the continuous upgrading of power grid equipment and the increasing complexity of maintenance scenarios, the ticket database requires a large amount of manpower for editing and version iteration. When encountering special scenarios such as equipment movement or atypical faults, the system's adaptive capability is clearly insufficient. The automatic ticket generation technology based on expert systems builds a rule base for the wiring method of specific substations and automatically generates the operation sequence of load safety procedures with the help of a logic reasoning engine. The expert system can reliably output work tickets based on equipment conversion rules and five-prevention logic constraints. However, the construction of the rule base often depends closely on the wiring method of specific substations and has poor universality. Developing a unified rule system applicable to multiple voltage levels and multiple wiring types will significantly increase costs. Moreover, existing expert systems have obvious deficiencies in acoustic testing of secondary equipment safety measures. The collaborative application of knowledge graphs and large language models (LLMs) has become a research hotspot in recent years. Encoders integrate graph entity relationships into text representations, and graph neural networks enable bidirectional interaction between text and graphs. A retrieval-enhanced generation framework combining parametric LLMs and non-parametric knowledge bases dynamically associates external knowledge to generate work orders. However, the one-way interaction mechanism of the knowledge-injected architecture limits the efficiency of knowledge flow. The fragmented processing of text and graphs makes it difficult to effectively support collaborative reasoning in complex power logic, resulting in poor knowledge fusion. While the retrieval-enhanced generation framework performs well in open-domain question answering, its vector retrieval tends to deviate from core professional concepts when facing power grid-specific terminology. Furthermore, existing technologies generally lack topology awareness, and LLMs struggle to understand the decisive impact of equipment connection relationships on safety measures, potentially generating content that violates the five-prevention rules. Summary of the Invention

[0003] To address the problems in the prior art, this invention proposes a work ticket generation method and system based on power grid knowledge logic and topological constraints. The technical solution of the present invention is as follows: In a first aspect, the present invention provides a method for generating work tickets based on power grid knowledge logic and topology constraints, comprising the following steps: Preload knowledge documents, which record the thought chain of maintenance operation safety measures in natural language; Upon receiving a maintenance task, the system performs knowledge retrieval and matching on the equipment involved in the maintenance task within the knowledge document to obtain knowledge fragments. By integrating knowledge fragments, a preliminary logical framework for generating maintenance task work orders is generated; By treating power equipment as nodes and electrical connections between equipment as edges, a power grid topology model is constructed, and the topology analysis boundary of the target equipment for the maintenance task is calculated. A structured prompt is generated by combining the initial logical framework and the topological analysis boundary, and then sent to the large language model to generate a work ticket.

[0004] Furthermore, the preliminary logical framework for generating the maintenance task work order includes the following steps: Based on the equipment attribute information of the target circuit breaker for the maintenance task, determine the bay type of the target circuit breaker; Based on the interval type, locate the disconnecting switch, grounding switch, and busbar associated with the target circuit breaker; Expand the circuit breakers associated with the busbar and recursively search for associated disconnect switches, grounding switches and busbars; All related disconnect switches and grounding switches are grouped into an operating equipment set to form a preliminary logical framework.

[0005] Furthermore, when there is an alternative power supply path for the circuit breaker associated with the bus, the equipment associated with the alternative power supply path and in a live or running state is marked as a running-keeping device and is not included in the set of operating devices when generating the preliminary logic framework.

[0006] Furthermore, the topology analysis boundary of the target equipment for the computational maintenance task adopts an adaptive multi-hop search strategy. Based on the function and electrical influence range of different equipment types within the power grid, the topology analysis boundary is dynamically determined. Specifically: For busbar power outages and maintenance, the topology analysis scope is set to include all associated devices within at least four hops. For transformers undergoing maintenance after power outages, the topology analysis scope is set to all associated devices within the three-hop range; When a circuit breaker is de-energized and switched to maintenance, the topology analysis range is set to all associated devices within a two-hop range.

[0007] Furthermore, the structured prompts include a preliminary logical framework, a list of devices in the topology analysis boundary, device relationships, and a standardized output format; The preliminary logical framework provides a content template for generating work orders; The device list and device relationships in the topology analysis boundary provide objective constraints for work order generation; The standardized output format explicitly requires that the content generated by the work order must follow a chapter division and itemized list format.

[0008] Furthermore, the work order generation method based on power grid knowledge logic and topology constraints also includes generating a security constraint mechanism based on the power grid topology model. When it is detected that the operation content of the work order generated by the large language model is inconsistent with the rules of the security constraint mechanism, the output of the work order result is blocked, the operation content of the work order is corrected to conform to the security constraint mechanism, and the work order result is output.

[0009] Furthermore, the safety measures thought chain for maintenance operations includes the equipment that needs to be disconnected during maintenance operations, the electrical connection relationships between equipment, the sequence that must be followed during maintenance operations, and auxiliary operations that need to be performed simultaneously.

[0010] Secondly, the present invention provides a work order generation system based on power grid knowledge logic and topology constraints, comprising: a knowledge storage module for storing knowledge documents; The knowledge retrieval module is used to receive maintenance tasks and perform knowledge retrieval and matching on the equipment involved in the maintenance tasks in the knowledge documents. The logic framework generation module is used to integrate the knowledge fragments retrieved from the knowledge retrieval and generate a preliminary logical framework for the maintenance task work order. The topology analysis module is used to construct a power grid topology model and calculate the topology analysis boundary of the target equipment for maintenance tasks. The structured hint generation module is used to generate structured hints by combining the initial logical framework and the topology analysis boundary; The work order generation module generates work orders based on structured prompts and a large language model.

[0011] Thirdly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the computer program is executed by the processor to perform the method described in any of the preceding claims.

[0012] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the preceding claims.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention integrates the objective world model of power grid physical topology with the subjective experience and knowledge of human experts. On top of a general artificial intelligence big language model, it injects structured domain professional knowledge and operating rules of the power system, and finally obtains topology content output. A three-in-one work order generation system was constructed. Through knowledge logic enhancement, it provides power system professional knowledge and reliable standards for the generation process; through topological analysis of power grid structure data, it lays an objective physical foundation and safety boundary for the generation process; through intelligent content generation and formatted output, it transforms the internal logic and external constraints into directly applicable maintenance operation guidance, effectively overcoming the limitations of general artificial intelligence large language models in professional fields, and providing technical support and feasible path for intelligent operation and maintenance of power systems. Attached Figure Description

[0014] Figure 1 This is a framework diagram of a work ticket generation method based on power grid knowledge logic and topological constraints. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1 This embodiment provides a method for generating work tickets based on power grid knowledge logic and topology constraints, specifically including the following steps: Preload knowledge documents, which record the thought chain of maintenance operation safety measures in natural language; Upon receiving a maintenance task, the system performs knowledge retrieval and matching on the equipment involved in the maintenance task within the knowledge document to obtain knowledge fragments. By integrating knowledge fragments, a preliminary logical framework for generating maintenance task work orders is generated; By treating power equipment as nodes and electrical connections between equipment as edges, a power grid topology model is constructed, and the topology analysis boundary of the target equipment for the maintenance task is calculated. A structured prompt is generated by combining the initial logical framework and the topological analysis boundary, and then sent to the large language model to generate a work ticket.

[0017] like Figure 1The diagram shows the framework of a work order generation method based on power grid knowledge logic and topological constraints. Users request maintenance task work orders. The constraints on the large language model for generating work orders are divided into two parts. First, the specific equipment information involved in the maintenance task is queried from the knowledge document. Through the text parsing module, the equipment to be operated is obtained. Using the equipment to be operated as the source node in the power equipment graph, a Subgraph is formed using the K-hop algorithm to obtain the equipment topology information, thus constraining the large language model for generating work orders. Second, expert logic is generated by combining expert logic guidance and external data from typical historical work orders to enhance the knowledge of the large language model.

[0018] Preferably, the safety measures thought chain for maintenance operations includes the equipment that needs to be disconnected during maintenance operations, the electrical connection relationships between the equipment, the sequence that must be followed during maintenance operations, and the auxiliary operations that need to be performed simultaneously; For example, when dealing with the operation of busbar maintenance, the knowledge document will not only list the circuit breakers and disconnectors that need to be disconnected, but also provide their electrical connection relationship with the busbar, further explaining why these circuit breakers and disconnectors need to be disconnected. The knowledge document will also provide the disconnection sequence of these circuit breakers and disconnectors, as well as the auxiliary operations that need to be performed simultaneously, such as disconnecting the control power supply and disconnecting the motor power supply.

[0019] Preferably, the preliminary logical framework for generating the maintenance task work order includes the following steps: Based on the equipment attribute information of the target circuit breaker for the maintenance task, determine the bay type of the target circuit breaker; Based on the interval type, locate the disconnecting switch, grounding switch, and busbar associated with the target circuit breaker and incorporate them into the operation logic; Extend the circuit breakers associated with the busbar and recursively search for associated disconnect switches, grounding switches and busbars, incorporating them into the operation logic; All associated disconnect switches and grounding switches are grouped into an operating equipment set to form a preliminary logical framework. Preferably, when there is an alternative power supply path for the circuit breaker associated with the busbar, the equipment associated with the alternative power supply path and in a live or operating state is marked as a running-holding device and is not included in the operating equipment set when generating the preliminary logical framework.

[0020] Taking the 220kV line 268 switch of a 220kV substation undergoing maintenance as an example, the preliminary logic framework is generated through the following steps: Determine whether the circuit breaker is a busbar compartment switch. If so, include the adjacent isolating switch and grounding switch in the operation logic. Determine whether the circuit breaker is a transformer interval switch. If so, include the adjacent isolating switch and grounding switch in the operation logic. Locate the busbar adjacent to the circuit breaker, locate all circuit breakers connected to that busbar, and recursively search for the devices associated with these circuit breakers. Add all isolating switches and grounding switches adjacent to the circuit breaker to the operation set; If the circuit breaker attribute information contains the phrase "use bypass instead", the disconnection operation of the compartment switch connected to the bypass bus will not be performed; When the circuit breaker is a trolley switch, pull the adjacent switch trolley to the test position and locate the grounding switch adjacent to the isolating switch to perform the disconnection operation.

[0021] Preferably, the knowledge retrieval and matching of equipment involved in the maintenance task specifically includes: The knowledge retrieval system automatically identifies the equipment involved in the maintenance task, searches the entire knowledge base, and locates all operational logic fragments related to the equipment. The search process is based on semantic understanding and association, and can locate aliases and model variants of the equipment.

[0022] The initial logical framework ensures that subsequent work order generation steps are strictly limited within the safety boundaries and operating procedures of the power industry, eliminating the risk of fabricating work order content out of thin air using ordinary language models.

[0023] Preferably, the topology analysis boundary of the target equipment for the maintenance task adopts an adaptive multi-hop search strategy, which dynamically determines the topology analysis boundary based on the function and electrical influence range of different equipment types in the power grid. For busbar power outage and maintenance, the topology analysis scope is set to include all associated devices within at least four hops. When a transformer is de-energized and switched to maintenance, its impact is usually limited to the voltage level side directly connected to it. The topology analysis scope is set to all associated equipment within the three-trip range. For circuit breakers that are de-energized and switched to maintenance, the topology analysis range is set to all associated devices within a two-hop range. This dynamic topology analysis strategy balances computational efficiency with security integrity.

[0024] Preferably, the structured prompts include a preliminary logical framework, a list of devices in the topology analysis boundary, device relationships, and a standardized output format; The preliminary logical framework provides a content template for generating work orders; The device list and device relationships in the topology analysis boundary provide objective constraints for work order generation; The standardized output format explicitly requires that the content generated by the work order must follow a chapter division and itemized list format.

[0025] Preferably, the work order generation method based on power grid knowledge logic and topology constraints further includes generating a security constraint mechanism based on the power grid topology model. When it is detected that the operation content of the work order generated by the large language model is inconsistent with the rules of the security constraint mechanism, the output of the work order result is blocked, the operation content of the work order is corrected to conform to the security constraint mechanism, and the work order result is output.

[0026] Example 2 This embodiment provides a work ticket generation system based on power grid knowledge logic and topological constraints, characterized in that it includes: The knowledge storage module is used to store knowledge documents; The knowledge retrieval module is used to receive maintenance tasks and perform knowledge retrieval and matching on the equipment involved in the maintenance tasks in the knowledge documents. The logic framework generation module is used to integrate the knowledge fragments retrieved from the knowledge retrieval and generate a preliminary logical framework for the maintenance task work order. The topology analysis module is used to construct a power grid topology model and calculate the topology analysis boundary of the target equipment for maintenance tasks. The structured hint generation module is used to generate structured hints by combining the initial logical framework and the topology analysis boundary; The work order generation module generates work orders based on structured prompts and a large language model.

[0027] Example 3 This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the method described in Embodiment 1.

[0028] Example 4 This embodiment provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in Embodiment 1.

[0029] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

Claims

1. A method for generating work tickets based on power grid knowledge logic and topological constraints, characterized in that, Includes the following steps: Preload knowledge documents, which record the thought chain of maintenance operation safety measures in natural language; Upon receiving a maintenance task, the system performs knowledge retrieval and matching on the equipment involved in the maintenance task within the knowledge document to obtain knowledge fragments. By integrating knowledge fragments, a preliminary logical framework for generating maintenance task work orders is generated; By treating power equipment as nodes and electrical connections between equipment as edges, a power grid topology model is constructed, and the topology analysis boundary of the target equipment for the maintenance task is calculated. A structured prompt is generated by combining the initial logical framework and the topological analysis boundary, and then sent to the large language model to generate a work ticket.

2. The work ticket generation method based on power grid knowledge logic and topology constraints according to claim 1, characterized in that, The preliminary logical framework for generating maintenance task work orders includes the following steps: Based on the equipment attribute information of the target circuit breaker for the maintenance task, determine the bay type of the target circuit breaker; Based on the interval type, locate the disconnecting switch, grounding switch, and busbar associated with the target circuit breaker; Expand the circuit breakers associated with the busbar and recursively search for associated disconnect switches, grounding switches and busbars; All related disconnect switches and grounding switches are grouped into an operating equipment set to form a preliminary logical framework.

3. The work ticket generation method based on power grid knowledge logic and topological constraints according to claim 2, characterized in that, When there is an alternative power supply path for the circuit breaker associated with the bus, the equipment associated with the alternative power supply path and in a live or running state is marked as a running-keeping device and is not included in the set of operating devices when generating the initial logic framework.

4. The work ticket generation method based on power grid knowledge logic and topology constraints according to claim 1, characterized in that, The topology analysis boundary for the target equipment in the maintenance task employs an adaptive multi-hop search strategy. This strategy dynamically determines the topology analysis boundary based on the function and electrical influence range of different equipment types within the power grid. Specifically: For busbar power outages and maintenance, the topology analysis scope is set to include all associated devices within at least four hops. For transformers undergoing maintenance after power outages, the topology analysis scope is set to all associated devices within the three-hop range; When a circuit breaker is de-energized and switched to maintenance, the topology analysis range is set to all associated devices within a two-hop range.

5. The work ticket generation method based on power grid knowledge logic and topology constraints according to claim 1, characterized in that, The structured prompts include a preliminary logical framework, a list of devices in the topology analysis boundary, device relationships, and a standardized output format; The preliminary logical framework provides a content template for generating work orders; The device list and device relationships in the topology analysis boundary provide objective constraints for work order generation; The standardized output format explicitly requires that the content generated by the work order must follow a chapter division and itemized list format.

6. The work ticket generation method based on power grid knowledge logic and topology constraints according to claim 1, characterized in that, It also includes a safety constraint mechanism based on the power grid topology model. When the operation content of the work order generated by the large language model is found to be inconsistent with the rules of the safety constraint mechanism, the output of the work order result is blocked, the operation content of the work order is corrected to conform to the safety constraint mechanism, and the work order result is output.

7. The work ticket generation method based on power grid knowledge logic and topology constraints according to claim 1, characterized in that, The safety measures thought chain for maintenance operations includes the equipment that needs to be disconnected during maintenance operations, the electrical connection relationships between equipment, the order that must be followed during maintenance operations, and auxiliary operations that need to be performed simultaneously.

8. A work order generation system based on power grid knowledge logic and topological constraints, characterized in that, include: The knowledge storage module is used to store knowledge documents; The knowledge retrieval module is used to receive maintenance tasks and perform knowledge retrieval and matching on the equipment involved in the maintenance tasks within the knowledge document library. The logic framework generation module is used to integrate the knowledge fragments retrieved from the knowledge retrieval and generate a preliminary logical framework for the maintenance task work order. The topology analysis module is used to construct a power grid topology model and calculate the topology analysis boundary of the target equipment for maintenance tasks. The structured hint generation module is used to generate structured hints by combining the initial logical framework and the topology analysis boundary; The work order generation module generates work orders based on structured prompts and a large language model.

9. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored on the memory and running on the processor, wherein the computer program is executed by the processor to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 7.

Images