A method, storage medium and device for checking power outage range of power grid operation ticket

Abstract

The present application relates to the field of power grid safety technology, and specifically to a method, storage medium, and device for verifying the power outage range of a power grid operation ticket. The present application performs attribute enhancement on subgraph nodes and edges, and then performs graph synchronization to establish a rich power grid topology map. A large language model is used to parse the operation ticket and maintenance application form. Finally, the actual power outage range is analyzed according to the model and compared with the operation ticket to complete the verification. The present application can verify whether the operation ticket matches the maintenance application, and at the same time identify the status of the equipment, filling the gap in the risk prevention of the anti-error system.

Description

Technical Field

[0001] The present application relates to the field of power grid safety technology, and in particular to a method, storage medium and device for verifying the power outage range of a power grid operation ticket. Background Art

[0002] The operation object and equipment status are the primary tasks of reviewing the dispatching operation ticket. During the review process, it is necessary to first confirm whether the operation task is consistent with the power outage scope specified in the maintenance application. During the peak maintenance period, especially during the overhaul of the substation, the power outage tasks are intensive and overlapping. The starting and target states of the auxiliary equipment of the same main component are inconsistent, and the power outage sequences of the line, busbar, and main transformer overlap before and after, which will lead to multiple maintenance applications with different power outage scopes and durations at the same time. Therefore, the above situation will make it extremely difficult to sort out the operation tickets, which will lead to the risk of equipment accidentally stopping power or missing power outages.

[0003] The traditional error prevention system only verifies the content of the operation ticket. It cannot determine whether the operation ticket matches the maintenance application, nor can it identify the status of the equipment. Therefore, the traditional error prevention system has gaps in risk prevention. Summary of the invention

[0004] The present application provides a method for verifying the power outage range of a power grid operation ticket, which can verify whether the power outage range involved in the operation ticket matches the power outage range involved in the maintenance application, and at the same time identify the status of the equipment, thereby filling the gap in risk prevention of the anti-error system.

[0005] The technical solution of this application is as follows:

[0006] A method for checking the power outage range of a power grid operation ticket, the method constructs a power grid topology map based on a VF2 algorithm, and specifically includes the following steps:

[0007] S1. Expand the attributes of nodes and edges of the power grid topology subgraph;

[0008] S2. Add node attribute matching items to the node matching metric function to perform node matching. The node attribute matching items are as follows:

[0009] ;

[0010] in, k Indicates the order of attributes; Represents a collection of node attributes; represents the indicator function; Representation Node and No. k Whether the attributes are consistent: if they are consistent, it takes 1, if they are inconsistent, it takes 0;

[0011] Add edge attribute matching items to the edge matching metric function to perform edge matching. The edge attribute matching items are as follows:

[0012] ;

[0013] in, Represents edge and j No. p Whether the attributes are consistent: if they are consistent, it takes 1, if they are inconsistent, it takes 0; Represents a set of edge attributes.

[0014] S3, performing power grid topology subgraph matching based on the node matching metric function and the edge matching metric function in S2 to construct a power grid topology graph;

[0015] S4. Use the trained large language model to read the operation ticket and extract the operation object, operation action and operation time in the operation ticket;

[0016] S5. Use the trained large language model to traverse the maintenance application form database, and extract the maintenance application forms of other operation objects that belong to the same station as the operation object and have overlapping maintenance time and operation time and have a connection relationship with the operation object in sequence according to the power grid topology map as target maintenance application forms;

[0017] S6. Use the trained large language model to read the target maintenance application form to obtain the power outage scope, and extract the operation objects within the power outage scope and the states of other operation objects connected to it at each operation time based on the power grid topology and semantic understanding;

[0018] S7. Compare the operation objects extracted by the large language model and their status at each operation time with the operation objects contained in the operation ticket and their status at each operation time. When the two corresponding items are consistent, it means that the project has passed the verification. Otherwise, a feedback alarm is sent to the dispatcher.

[0019] Furthermore, the node attributes expanded in step S1 are Including: equipment type, electrical parameters, voltage level, station, equipment name, equipment number, wiring method;

[0020] The edge connection types in step S1 include: transmission line connection, protective grounding connection, and interconnection switch connection. The edge attributes after the type of each edge is expanded are: Including: voltage level, station, and line model.

[0021] Furthermore, in step S1, the connection modes include: double bus connection, 2 / 3 connection, single bus connection, and double bus segmented connection.

[0022] Further, in step S3, the order of pruning of node matching is: node attribute, node degree;

[0023] In the node properties, the pruning order is: device type, wiring method;

[0024] In step S3, the edge matching pruning sequence is: connection type, edge attribute.

[0025] Furthermore, when it is determined in step S5 whether the maintenance time and the operation time overlap, the operation time is extended forwards and backwards by 24 hours each.

[0026] Furthermore, in step S6, the status of the operation time includes: equipment operation, hot standby, cold standby, and maintenance.

[0027] The present application also provides a power grid operation ticket power outage range verification device, including a processor and a memory storing program instructions, and the processor is configured to execute the above-mentioned power grid operation ticket power outage range verification method when running the program instructions.

[0028] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the method described above is implemented.

[0029] Due to the adoption of the above technical solution, the beneficial effects of this application are as follows:

[0030] 1. This application expands the functionality of the power grid topology knowledge graph. Traditional knowledge graphs simply display node identifiers and whether nodes are connected. Considering the various types of connection relationships between electrical equipment in the power grid, this application adds node attributes that can characterize technical parameters to the power grid topology graph, making the nodes in the graph functional. Furthermore, various types of edge attributes are added to the power grid topology graph to visualize the series connection of various functions, and finally a power grid topology graph that can be used to check operation tickets is constructed.

[0031] 2. This application has developed electrical connection rules that are consistent with the actual grid connection status through an improved VF2 algorithm, so that the relationship between electrical devices is transformed from a formal connection to a functional match. At the same time, this application constrains the connection of nodes and edges to further ensure the rationality of the match, and also adds wiring method constraints, further clarifying the connection relationship between electrical devices, laying the foundation for providing a more reasonable power outage range in the future.

[0032] 3. This application uses the trained large language model to parse the dispatch information in the operation ticket, filters the target maintenance application form that matches the dispatch information in the maintenance application form database layer by layer, and uses the large language model again to parse the power outage range according to the target maintenance application form. Within the power outage range, each operation object and its status are extracted according to the power grid topology map, and compared with the content of the operation ticket to realize the power outage range verification. In other words, this application realizes the prior verification of the operation ticket, jumping out of the traditional level of only verifying the content of the operation ticket, and thoroughly verifies it from the basis of the proposed ticket, which not only helps to improve the accuracy of the verification, but also helps the dispatcher understand the dispatch logic and improve his work ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The drawings described herein are used to provide further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0034] Figure 1 A flow chart of a method for verifying the power outage range of a power grid operation ticket is provided for this application;

[0035] Figure 2 This is the flow chart of the large language model training in this application;

[0036] Figure 3 A flowchart for obtaining the target application inspection form in this application;

[0037] Figure 4 This is a grid topology diagram of an embodiment of the present application. DETAILED DESCRIPTION

[0038] Based on the background technology, the traditional error prevention system only verifies the content of the operation ticket, and cannot verify the substantive content of the operation ticket according to the proposed ticket logic. Based on the above problems, this application provides a method for verifying the power outage range of the power grid operation ticket. Figure 1 , this method reconstructs the subgraph based on the VF2 algorithm, and specifically includes the following steps:

[0039] S1. Expand the attributes of the nodes and edges of the subgraph.

[0040] The power grid topology can be represented in the form of a knowledge graph. The nodes in the traditional knowledge graph are basic identifiers, and the edges represent the connection between nodes. That is, the traditional knowledge graph does not have the ability to identify deep content. The VF2 algorithm is used to reconstruct the graph to match nodes and edges between subgraphs. In the traditional VF2 algorithm, the node attributes and edge attributes are single, but in the power grid architecture, there are many parameters between each electrical device and they have different functions. The constraints when matching nodes are more complicated and cannot be replaced by a single attribute. The edge attributes also have the above characteristics, and the edge connection is an important part of the power grid topology.

[0041] The power grid topology is constructed based on CIME files, which include equipment details such as substations, lines, switches, buses, and equipment IDs, and the interconnection relationship between each device can be found through the equipment ID. After data cleaning and standardization, the technical parameters, location, and association relationship of the equipment are converted into graph structure data. Each group of equipment and its connection in the power grid topology forms a subgraph object, and the connection relationship of all subgraphs constitutes the power grid topology.

[0042] In this embodiment, the expanded node attributes Including: equipment type, electrical parameters, voltage level, station, equipment name, equipment number, wiring method. Wiring methods also include: double busbar wiring, 2 / 3 wiring, single busbar. Edges have connection types, and each connection type has attributes. Edge connection types include: transmission line connection, protective grounding connection, interconnection switch connection, and the edge attributes after each edge type is expanded Including: voltage level, station, and line model.

[0043] The VF2 algorithm determines whether there is a subgraph isomorphism relationship between two graphs by judging the node similarity of the two graphs, and transforms the graph isomorphism matching problem into a state search problem in the state space through state space search. The VF2 algorithm mainly prunes nodes based on the similarity between nodes to reduce the search time and space. In the node matching process, the VF2 algorithm uses the graph node labels and adjacency information to calculate the node similarity, and gives priority to matching nodes with high similarity. The VF2 algorithm can improve the matching efficiency through subgraph isomorphism judgment and pruning techniques.

[0044] S2. Set up a power grid topology diagram , N is a node set, , each , i =1,2,3… m Contains node attributes ; E is the edge set, , each element represents the connection between two devices. n nodes, then its adjacency matrix can be expressed as a The matrix , if there is an edge between two nodes, its element value is 1, if not, it is 0.

[0045] Use the node matching metric function to perform node matching:

[0046] , , ;

[0047] In the formula, , Matching metric function for nodes; , Representation Node and Are the degrees (number of edges connected to a node) equal? ​​If the node labels are different, the similarity is 0; if the node labels are the same but the adjacency (whether the nodes are directly connected) is different, the similarity is 1; if both the labels and adjacency are the same, the similarity is 2; , They are respectively the node degree equality judgment function and the weight coefficient of the node attribute matching item.

[0048] The above node matching metric function has node attribute matching items, and the node attribute matching items are as follows:

[0049] ;

[0050] in, k Indicates the order of attributes; Represents a collection of node attributes; represents the indicator function; Representation Node and No. k Whether the attributes are consistent: if they are consistent, it takes 1, if they are inconsistent, it takes 0.

[0051] The same node matching metric function is used to perform edge matching using the edge matching metric function:

[0052] , , ;

[0053] In the formula, is the edge matching metric function; , Represents edge , Type matching function; add edge attribute matching items to the edge matching metric function to perform edge matching. The edge attribute matching items are as follows:

[0054] ;

[0055] in, Represents edge and j No. pWhether the attributes are consistent: if they are consistent, it takes 1, if they are inconsistent, it takes 0. and Respectively represent the weight coefficients of edge type matching items and edge attribute matching items.

[0056] S3. Construct a power grid topology graph based on the node matching metric function and the edge matching metric function in S2.

[0057] When performing node matching, the degree of the node is considered step by step, and the attributes of the node must also be considered. If two nodes do not match in attributes, even if their degrees are the same, they are not judged to be graph isomorphic. Based on this, the search strategy for graph isomorphism can be optimized by pruning node attributes, thereby improving the efficiency of the algorithm. The optimized graph isomorphism search strategy in this application can exclude impossible matching situations in advance, reduce unnecessary calculations, compress the search space, and speed up the execution of the algorithm, which is especially suitable for the complex, staggered and overlapping characteristics of power grid layout.

[0058] In step S3, the order of pruning node matching is: node attribute, node degree;

[0059] In the node properties, the pruning order is: device type, wiring method.

[0060] The pruning mode is to determine whether the indicator function is 1 or 0. The node attribute sequence of this embodiment only shows the pruning sequence of the first two node attributes, and the pruning sequence of other attributes is not specifically limited.

[0061] In this embodiment, the nodes may be sorted by degree, and the matching strategy may be improved by trying to match nodes with high degrees first and then gradually diverging to nodes with low degrees.

[0062] In step S3, the edge matching pruning sequence is: connection type, edge attribute.

[0063] The pruning strategy of edge matching is similar to the node attribute pruning strategy. When matching edges, the connection type is determined first. If the connection types are different, they are pruned directly.

[0064] Considering the complexity and scale of the power grid topology, graph isomorphism matching takes a long time. When running the VF2 algorithm, a parallel computing optimization solution is adopted to accelerate the isomorphism matching process through multi-threaded computing. At the same time, the large power grid topology can be divided into several subgraphs for separate processing, and then the results are finally merged after isomorphism matching.

[0065] S4. Use the trained large language model to read the operation ticket and extract the operation object, operation action and operation time in the operation ticket.

[0066] The operation ticket is semi-structured data, including the operation task, operation equipment, operation type, operation time and specific operation instructions. In the scenario of power outage scope verification, only the operation equipment, operation time and the starting state and target state within the operation time in the operation ticket need to be extracted. The status of the equipment includes equipment operation, hot standby, cold standby and maintenance.

[0067] The training process of the large language model can be realized based on existing technologies. Figure 2 The process includes collecting training sets, data preprocessing, entity recognition, relationship extraction, action extraction, context understanding, and output. The training set uses historical operation ticket data. When selecting, try to cover various instruction formats. At the same time, construct some wrong operation ticket samples for model testing. Data processing includes text cleaning, word segmentation and data standardization. Word segmentation and tagging use word segmentation tools to break down instructions into words or subwords, and perform part-of-speech tagging, marking nouns, verbs, etc. Data standardization is corrected according to the operation ticket operation specifications and standard dispatching terms to eliminate ambiguities caused by different expressions. Relation extraction and extraction of associations between equipment, such as "Wuxu Station 220kVXX Line 201 Switch", first need to be read in and parsed by the large language model "Wuxu Station 220kVXX Line 201 Switch", to obtain key information, including the equipment site is "Wuxu Station", the equipment voltage level is "220kV", the equipment name is "XX Line", and the equipment number is "201 Switch". Then the large language model inputs "Wuxu Station" into the power grid topology map to search for the station, obtain the station topology information, and then inputs the equipment voltage level, equipment name, and equipment number into the power grid physical topology knowledge map to locate the specific equipment. If the relevant equipment can be accurately found in the map, the association between the entities "Wuxu Station", "220kVXX Line", and "201 Switch" can be accurately confirmed. Instruction parsing and action extraction extract the specific operations to be performed, such as "disable", "convert to cold standby", etc., according to the syntactic structure and context information, and associate them with the corresponding equipment objects. Context understanding relies on the context understanding and memory ability of the large language model to infer the logic before and after different operations. The output is to output the extracted key information, including equipment station, voltage level, equipment name, equipment number, and equipment status in a predetermined format.

[0068] S5. Use the trained large language model to traverse the maintenance application form database, and extract the maintenance application forms of other operation objects that belong to the same station as the operation object and whose maintenance time overlaps with the operation time and have a connection relationship with the operation object in sequence according to the power grid topology map as target maintenance application forms.

[0069] See attached Figure 3In the extraction process, the maintenance application forms under the three conditions are extracted in sequence, and the search range is gradually compressed. If they are not satisfied, they are discarded in time. In one embodiment, when determining whether the maintenance time and the operation time overlap, the operation time is expanded by 24 hours each.

[0070] S6. Use the trained large language model to read the target maintenance application form to obtain the power outage scope, and extract the operation objects within the power outage scope and the states of other operation objects connected to it at each operation time based on the power grid topology diagram and semantic understanding. The power outage scope describes the specific equipment names that need to be operated and the status information reached by the equipment.

[0071] After the large language model obtains the target maintenance application, it needs to obtain its power outage scope and perform semantic understanding, extract the equipment involved in the power outage scope and the change of the equipment status, and finally confirm the status that all the equipment involved in the maintenance application should maintain at the time of the operation ticket operation. If there is a device that is still before the end of the maintenance plan at the estimated operation time of the operation ticket, the device should be locked in the maintenance state and no state change should occur.

[0072] In this embodiment, see the attached Figure 4 In order to assist the large model to accurately extract the equipment involved in the power outage scope and the topological relationship between the equipment, firstly, according to the "equipment name + station" of the maintenance application form, the equipment is searched in the power grid topology map, the equipment is located and other equipment related to the equipment is obtained. Then, according to the equipment status parsed from the power outage scope by the large model, the status of the equipment involved in the power outage scope and its related equipment at the time of the operation ticket operation is finally determined. Figure 4 , the line (main equipment) in the figure is the operation object extracted from the operation ticket. Taking the power supply operation of the equipment as an example, before compiling the operation ticket, it is necessary to confirm the initial state and final state of the tie switch (accessory equipment) and the side switch (accessory equipment) associated with the line (main equipment). The main method of determination is to find the state requirements of other equipment that are electrically related to the tie switch (accessory equipment) and the side switch (accessory equipment) operation according to the above topology diagram, such as confirming whether the #1 and #2 busbars need to be kept in the maintenance state, and whether the transformer (other main equipment) needs to be kept in the maintenance state. If it is confirmed according to the maintenance application form that the transformer (other main equipment) and its switch need to be kept in the maintenance state, because the transformer and the line (main equipment) share the tie switch (accessory equipment), the tie switch (accessory equipment) cannot perform the power supply operation, and only the side switch (accessory equipment) can be powered.

[0073] S7. Compare the operation objects extracted by the large language model and their status at each operation time with the operation objects contained in the operation ticket and their status at each operation time. When the two corresponding items are consistent, it means that the project has passed the verification. Otherwise, a feedback alarm is sent to the dispatcher.

[0074] The embodiment of the present disclosure also provides a power grid operation ticket power outage range verification device, including a processor and a memory. Optionally, the device may also include a communication interface and a bus. The processor, the communication interface, and the memory may communicate with each other through the bus. The communication interface may be used for information transmission. The processor may call the logic instructions in the memory to execute the power grid operation ticket power outage range verification method of the above embodiment.

[0075] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.

[0076] The memory, as a computer-readable storage medium, can be used to store software programs and computer executable programs, such as program instructions / modules corresponding to the method in the embodiment of the present disclosure. The processor executes the function application and data processing by running the program instructions / modules stored in the memory, that is, implements a method for checking the power outage range of a power grid operation ticket in the above embodiment.

[0077] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory may include a high-speed random access memory and may also include a non-volatile memory.

[0078] An embodiment of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured to execute the above-mentioned method for verifying the power outage range of a power grid operation ticket.

[0079] The computer-readable storage medium mentioned above may be a transient computer-readable storage medium or a non-transitory computer-readable storage medium.

[0080] The technical solution of the embodiment of the present disclosure can be embodied in the form of a software product, which is stored in a storage medium and includes one or more instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiment of the present disclosure. The aforementioned storage medium may be a non-transient storage medium, including: a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a disk or an optical disk, and other media that can store program codes, or a transient storage medium.

[0081] Anything not described in this application can be achieved by adopting or drawing on existing technologies.

[0082] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for checking the power outage range of a power grid operation ticket, characterized in that: This method constructs a power grid topology map based on the VF2 algorithm, which specifically includes the following steps: S1. Expand the attributes of nodes and edges of the power grid topology subgraph; S2. Add node attribute matching items to the node matching metric function to perform node matching. The node attribute matching items are as follows: ； in, k Indicates the order of attributes; Represents a collection of node attributes; represents the indicator function; Representation Node and No. k Whether the attributes are consistent: if they are consistent, it takes 1, if they are inconsistent, it takes 0; Add edge attribute matching items to the edge matching metric function to perform edge matching. The edge attribute matching items are as follows: ； in, Represents edge and j No. p Whether the attributes are consistent: if they are consistent, it takes 1, if they are inconsistent, it takes 0; Represents a set of edge attributes; S3, performing power grid topology subgraph matching based on the node matching metric function and the edge matching metric function in S2 to construct a power grid topology graph; S4. Use the trained large language model to read the operation ticket and extract the operation object, operation action and operation time in the operation ticket; S5. Use the trained large language model to traverse the maintenance application form database, and extract the maintenance application forms of other operation objects that belong to the same station as the operation object and have overlapping maintenance time and operation time and have a connection relationship with the operation object in sequence according to the power grid topology map as target maintenance application forms; S6. Use the trained large language model to read the target maintenance application form to obtain the power outage scope, and extract the operation objects within the power outage scope and the states of other operation objects connected to it at each operation time based on the power grid topology map and semantic understanding; S7. Compare the operation objects extracted by the large language model and their status at each operation time with the operation objects contained in the operation ticket and their status at each operation time. When the two corresponding items are consistent, it means that the project has passed the verification. Otherwise, a feedback alarm is sent to the dispatcher.

2. A method for checking the power outage range of a power grid operation ticket according to claim 1, characterized in that: Node attributes after expansion in step S1 Including: equipment type, electrical parameters, voltage level, station, equipment name, equipment number, wiring method; The edge connection types in step S1 include: transmission line connection, protective grounding connection, and interconnection switch connection. The edge attributes after the type of each edge is expanded are: Including: voltage level, station, and line model.

3. A method for checking the power outage range of a power grid operation ticket according to claim 2, characterized in that: In step S1, the connection modes include: double bus connection, 2 / 3 connection, single bus connection, and double bus segmented connection.

4. A method for checking the power outage range of a power grid operation ticket according to any one of claims 2 or 3, characterized in that: In step S3, the order of pruning node matching is: node attribute, node degree; In the node properties, the pruning order is: device type, wiring method; In step S3, the edge matching pruning sequence is: connection type, edge attribute.

5. A method for checking the power outage range of a power grid operation ticket according to claim 4, characterized in that: When it is determined in step S5 whether the maintenance time and the operation time overlap, the operation time is extended forward and backward by 24 hours each.

6. A method for checking the power outage range of a power grid operation ticket according to claim 4, characterized in that: In step S6, the status of the operation time includes: equipment operation, hot standby, cold standby, and maintenance.

7. A device for checking the power outage range of a power grid operation ticket, comprising a processor and a memory storing program instructions, characterized in that: The processor is configured to execute a method for verifying the power outage range of a power grid operation ticket as described in claim 6 when running the program instructions.

8. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the method as claimed in claim 6 is implemented.

Images