Power grid operation control method based on knowledge graph rule extraction

Abstract

The invention provides a power grid operation management and control method based on knowledge graph rule extraction, and belongs to the technical field of power grid dispatching, and the method specifically comprises the steps: carrying out the reduction and optimization of information in a knowledge graph based on the similarity and fault probability for unrecorded power grid operation data, forming a corresponding association rule according to an association rule algorithm in combination with a pruning strategy, and carrying out the optimization of the corresponding association rule; association rules are extracted in combination with a decision tree algorithm, a corresponding rule decision tree is established to perform fault budgeting so as to perform exception handling, and fault budgeting and exception handling are performed on the recorded power grid operation data according to the fault case graph. According to the method, on the basis of constructing the decision tree based on the knowledge graph, information participating in decision tree construction can be efficiently determined by reducing and optimizing information in the knowledge graph and applying an association rule algorithm, interference of invalid information and information with low association degree is reduced, and the construction efficiency and accuracy of the decision tree are improved; and the efficiency and accuracy of subsequent fault budgeting and power grid dispatching are ensured.

Description

Technical Field

[0001] The present invention relates to the field of power grid dispatching technology, and in particular to a power grid operation control method based on knowledge graph rule extraction. Background Art

[0002] At present, the power grid operation data has the characteristics of diversity, massiveness and real-time. The effective management and analysis of these data are crucial to the safe, stable and efficient operation of the power grid. As a carrier that can efficiently store and query complex relational data, knowledge graphs are suitable for representing the complex relationships between power grid equipment, operating status, fault modes, etc., and are often used in power grid operation status monitoring. However, most existing knowledge graph applications focus on data display and simple queries, lacking in-depth data analysis and fault budgeting methods. Therefore, when using knowledge graphs for power grid operation status monitoring, decision tree algorithms are often used to use the classification ability of decision trees and the knowledge representation ability of knowledge graphs to jointly achieve accurate fault prediction to ensure the safety and reliability of power grid operation.

[0003] In the process of using the decision tree algorithm to build the corresponding decision tree for fault budgeting, it is necessary to select the features in the knowledge graph and select the features that have the greatest impact on the prediction results to build the decision tree. The knowledge graph has a huge number of nodes and edges, and there are complex relationships between the nodes. This also leads to a large amount of invalid data being mixed in when building the corresponding decision tree based on the information contained in the knowledge graph, which affects the construction efficiency and accuracy of the decision tree, and further affects the efficiency of subsequent fault budgeting and power grid scheduling. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art in which a large amount of invalid data is mixed in the construction process of the decision tree due to the complexity of the knowledge graph in the process of fault budgeting to achieve power grid dispatching, which greatly reduces the construction efficiency and accuracy of the decision tree, and further affects the subsequent fault budgeting and power grid dispatching efficiency. A power grid operation control method based on knowledge graph rule extraction is provided. On the basis of constructing a decision tree based on the knowledge graph for fault budgeting, the information in the knowledge graph involved in the construction of the decision tree is processed by an association rule algorithm combined with a pruning strategy to reduce the complexity of the knowledge graph and avoid invalid data irrelevant to the fault budget or data with a low degree of association from interfering with the construction of the decision tree. Before the association rule algorithm processes the information, the information in the knowledge graph is first reduced and optimized by similarity and fault probability to improve the processing efficiency of the association rule algorithm for large data sets, further improve the construction efficiency and accuracy of the decision tree, thereby improving the efficiency and accuracy of subsequent fault budgeting and power grid dispatching, and ensuring the safety and stability of power grid operation.

[0005] The purpose of the present invention is achieved through the following technical solutions: The power grid operation control method based on knowledge graph rule extraction includes: Obtain power grid operation data and identify the type of power grid operation data based on the knowledge graph; When the power grid operation data type is an unrecorded data type, the similarity between the power grid operation data and the historical data in the knowledge graph is calculated, and invalid data in the knowledge graph is filtered and removed according to the similarity calculation result; Combined with the failure probability, select the rule extraction range from the knowledge graph that removes invalid data, and obtain the rule extraction sub-graph for fault budgeting; Combined with the pruning strategy, the association rules of the rule extraction sub-graph are formed according to the association rule algorithm, and the association rules are extracted according to the decision tree algorithm to establish a rule decision tree; Based on the rule decision tree, fault budget is made according to the power grid operation data, and corresponding power grid dispatch strategy is formulated to handle exceptions; The fault case map is updated according to the power grid dispatching strategy, and when the power grid operation data type is the recorded data type, fault budget and exception processing are performed according to the fault case map.

[0006] While using the knowledge graph and decision tree algorithm to perform fault budgeting, the existing fault types are recorded in conjunction with the fault case graph, so that the fault type can be quickly identified through the fault case graph, and then the corresponding power grid dispatching strategy can be selected, avoiding repeated calculations of the same type of power grid operation data and optimizing the fault budgeting efficiency. When faced with an unrecorded fault type without a corresponding fault case graph, a decision tree is constructed based on the knowledge graph. On the basis of performing fault budgeting through the decision tree, association rules are first extracted from the knowledge graph based on the association rule algorithm combined with the pruning strategy, and then the corresponding decision tree is constructed based on the association rules to reduce the complexity of the knowledge graph, avoid invalid data that is not related to the fault budget or data with a low degree of association from interfering with the decision tree construction, and improve the construction efficiency and accuracy of the decision tree. Before extracting association rules, the information in the knowledge graph is first reduced and optimized through similarity and fault probability to improve the processing efficiency of the association rule algorithm for large data sets, so that the association rule algorithm can adapt to the information processing of knowledge graphs with large amounts of data, and ensure the accuracy and efficiency of the association rule algorithm in extracting association rules, so as to further improve the efficiency and accuracy of subsequent fault budgeting and power grid scheduling, and ensure the safety and stability of power grid operation.

[0007] Furthermore, the identifying of the power grid operation data type based on the knowledge graph includes: Determine the data source of the power grid operation data, and determine the entity information, attribute information and relationship information in the power grid operation data based on the data source; Match the entity information, attribute information and relationship information of the power grid operation data with the knowledge graph in sequence; When the entity information, attribute information and relationship information all match the corresponding information in the knowledge graph, the power grid operation data type is determined to be a recorded data type; In other cases, the power grid operation data type is determined to be an unrecorded data type.

[0008] Furthermore, the similarity between the power grid operation data and the historical data in the knowledge graph is calculated, and invalid data in the knowledge graph is filtered and removed according to the similarity calculation result, including: Obtain historical update information and historical fault information of the knowledge graph, match the historical fault information with the historical update information, determine the update information corresponding to each fault, and set the weights of entity information, attribute information, and relationship information based on the update information corresponding to each fault; The entity information, attribute information and relationship information corresponding to each entity are used as similarity calculation targets respectively, and the similarity between each entity in the knowledge graph and the power grid operation data is calculated in combination with the corresponding weights; Filter out entities whose similarity is lower than the preset similarity threshold to filter out the historical data corresponding to the entity in the knowledge graph as invalid data, and eliminate all invalid data in the knowledge graph.

[0009] Furthermore, the rule extraction range is selected from the knowledge graph that removes invalid data in combination with the fault probability to obtain a rule extraction sub-graph for fault budgeting, including: Determine all entities in the knowledge graph after removing invalid data, obtain the relationship information between each entity, and obtain the extraction scope of all candidate rules based on the relationship information; According to all entities within the extraction range of each candidate rule, the entity update frequency, attribute fluctuation range and relationship increase rate corresponding to each candidate rule extraction range are obtained; The failure probability score of each candidate rule extraction range is scored based on the entity update frequency, attribute fluctuation range, and relationship increase, and the comprehensive score of each candidate rule extraction range is calculated based on the similarity of the corresponding entities; The rule extraction scope is selected based on the comprehensive score, the extraction boundary is set based on the entities included in the rule extraction scope, and the rule extraction sub-graph for fault budgeting is obtained from the knowledge graph based on the extraction boundary.

[0010] Considering that the operating efficiency of the association rule algorithm is directly related to the size of the data set for rule extraction, the larger the data set, the more frequent item sets and candidate item sets need to be traversed, and the running time of the association rule algorithm will increase accordingly. Therefore, the invalid information in the knowledge graph is screened and eliminated according to the similarity of the power grid operation data, reducing the interference of repeated and useless item sets on the association rule extraction during the association extraction process, and ensuring the efficiency of the association rule algorithm for the extraction of association rules. Considering that multiple areas with similar lines or operation scenarios may be identified, the fault probability analysis of each similar area is further performed according to the fluctuation of entity, attribute and relationship information, and the area without recorded power grid operation data that is most likely to have a fault tendency is determined, and then the knowledge graph area closest to the corresponding scene of the power grid operation data is selected for subsequent rule extraction to ensure the efficiency and accuracy of association rule extraction.

[0011] Furthermore, the combination of the pruning strategy and the formation of association rules of the rule extraction sub-graph according to the association rule algorithm includes: Extract entities, relationships, and attributes in the sub-graph based on rules and build the corresponding transaction database; Set the support threshold and scan the transaction database to obtain the items and their occurrence counts in all transactions, and obtain the support of each item based on the corresponding occurrence count; Filter out the item sets whose support is less than the support threshold and delete them, and construct the initial frequent item sets based on the remaining items; Combined with the pruning strategy, candidate item sets are generated based on the initial frequent item sets and the corresponding support is calculated. The next level of frequent item sets are constructed based on the support of the candidate item sets. Repeat the generation of candidate item sets and the construction of frequent item sets until all frequent item sets in the transaction database are exhausted; Extract association rules of subgraphs based on frequent item set formation rules.

[0012] When extracting association rules based on the association rule algorithm, the pruning strategy is combined to simplify the algorithm execution process and further optimize the efficiency of association rule extraction.

[0013] Furthermore, the extraction of association rules according to the decision tree algorithm and the establishment of a rule decision tree of the rule extraction sub-graph include: Generate input features of decision tree algorithm based on formed association rules and their corresponding transactions, and select root node based on Gini coefficient; The corresponding child nodes are recursively constructed by calculating the corresponding Gini coefficient until the stopping condition is met, and the rule decision tree of the rule extraction sub-graph is obtained.

[0014] Furthermore, the extraction of association rules according to the decision tree algorithm and the establishment of a rule decision tree of the rule extraction sub-graph also include: The nodes of the rule decision tree are searched for similarity between the nodes, and the rule decision tree is supplemented with rules according to the search results.

[0015] Further, updating the fault case map according to the power grid dispatching strategy includes: Causal reasoning is performed based on the grid operation data and the corresponding grid dispatching strategy to generate the corresponding accident plan, and the fault case map is updated according to the accident plan.

[0016] Furthermore, the causal reasoning is performed by combining the power grid operation data with the corresponding power grid dispatching strategy to generate a corresponding accident plan, including: Extract text features of power grid dispatch strategy based on pre-trained language representation model, and build causal graph based on the extracted text features; Perform causal reasoning based on the causal graph, construct the accident scenario based on the causal reasoning results, and generate a text description corresponding to the accident scenario in combination with the corresponding text features; Generate accident plans based on accident scenarios and corresponding text descriptions.

[0017] Further, updating the fault case map according to the accident plan includes: Create or select the corresponding fault node according to the accident plan, match the relevant nodes in the fault case graph according to the text description of the accident plan, and connect the fault node and the relevant nodes; Create or select the cause node, impact node and measure node corresponding to the fault node according to the accident plan, and establish connections between the cause node, impact node and measure node and the fault node respectively.

[0018] The beneficial effects of the present invention are: (1) While using the knowledge graph and decision tree algorithm to perform fault budgeting, the existing fault types are recorded in conjunction with the fault case graph, so that the fault type can be quickly identified through the fault case graph, and then the corresponding power grid dispatch strategy can be selected, avoiding repeated calculations of the same type of power grid operation data and optimizing the fault budgeting efficiency. When faced with an unrecorded fault type without a corresponding fault case graph, a decision tree is constructed based on the knowledge graph. On the basis of performing fault budgeting through the decision tree, association rules are first extracted from the knowledge graph based on the association rule algorithm combined with the pruning strategy, and then the corresponding decision tree is constructed based on the association rules to reduce the complexity of the knowledge graph, avoid invalid data that is not related to the fault budget or data with a low degree of association from interfering with the decision tree construction, and improve the construction efficiency and accuracy of the decision tree. Before extracting association rules, the information in the knowledge graph is first reduced and optimized through similarity and fault probability to improve the processing efficiency of the association rule algorithm for large data sets, so that the association rule algorithm can adapt to the information processing of knowledge graphs with large amounts of data, and ensure the accuracy and efficiency of the association rule algorithm in extracting association rules, so as to further improve the efficiency and accuracy of subsequent fault budgeting and power grid scheduling, and ensure the safety and stability of power grid operation.

[0019] (2) Invalid information in the knowledge graph is screened and removed based on the similarity of the power grid operation data, reducing the interference of repeated and useless item sets on the extraction of association rules during the association extraction process, and ensuring the efficiency of the association rule algorithm in extracting association rules. Considering that multiple areas with similar lines or operation scenarios may be identified, the fault probability of each similar area is further analyzed based on the fluctuation of entity, attribute and relationship information, and the areas without recorded power grid operation data that are most likely to have fault tendencies are determined. Then, the knowledge graph area closest to the scene corresponding to the power grid operation data is selected for subsequent rule extraction to ensure the efficiency and accuracy of association rule extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of a process of the present invention; Figure 2 It is an overall flow chart of power grid operation control according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0022] Embodiment: A power grid operation control method based on knowledge graph rule extraction, such as Figure 1 As shown, including: Obtain power grid operation data and identify the type of power grid operation data based on the knowledge graph; When the power grid operation data type is an unrecorded data type, the similarity between the power grid operation data and the historical data in the knowledge graph is calculated, and invalid data in the knowledge graph is filtered and removed according to the similarity calculation result; Combined with the failure probability, select the rule extraction range from the knowledge graph that removes invalid data, and obtain the rule extraction sub-graph for fault budgeting; Combined with the pruning strategy, the association rules of the rule extraction sub-graph are formed according to the association rule algorithm, and the association rules are extracted according to the decision tree algorithm to establish a rule decision tree; Based on the rule decision tree, fault budget is made according to the power grid operation data, and corresponding power grid dispatch strategy is formulated to handle exceptions; The fault case map is updated according to the power grid dispatching strategy, and when the power grid operation data type is the recorded data type, fault budget and exception processing are performed according to the fault case map.

[0023] Knowledge graphs can organize data and the relationship between data through graph structures, making data processing and analysis more intuitive and efficient. Therefore, in power grid dispatching, knowledge graphs are constructed to express the relationship between entities in the form of "graphs" to better understand and process the complex knowledge in the power system, thereby improving the intelligence level of dispatching decisions.

[0024] Among them, the knowledge graph mainly includes three basic parts: entities, relationships and attributes. By extracting, integrating and representing concepts, entities, events and their relationships related to power grid dispatching, and further organizing them into a graphical structure, relevant information can be better extracted from it for subsequent power grid dispatching management.

[0025] Considering that the application of knowledge graphs is mostly concentrated on data display and simple query, further data mining and analysis methods are needed to identify and predict abnormal power grid operation in order to ensure the safety and reliability of power grid operation.

[0026] And the knowledge graph also includes multiple types according to its focus subjects, such as entity graphs that mainly focus on entities and the relationships between entities and fault graphs that represent fault scenarios. In this embodiment, these two types of knowledge graphs are specifically applied to construct corresponding knowledge graphs and fault case graphs to achieve efficient control of power grid operation.

[0027] Among them, for the knowledge graph of entity graph type that mainly focuses on entities and the relationships between entities, its entities are extracted from data such as equipment information, operating status, historical events, etc. collected from the power grid. The extracted entities include equipment such as generators, transformers, transmission lines, states such as voltage, current, power factor, and events such as failures and maintenance. The relationships are determined by the relationships between these entities, such as the connection relationship between devices, the correspondence between states and devices, and the association relationship between events and devices or states. Attributes provide detailed descriptions of these entities, such as equipment model, state parameters, and event descriptions.

[0028] For the fault case graph of the fault graph type representing the fault scenario, the entities are the components of the equipment, performance characteristics, fault status, etc. The relationships describe the associations between these entities, such as the composition relationship between components, the correspondence between performance characteristics and fault status, etc., while the attributes provide detailed descriptions of these entities, such as the model of the component, the specific value of the performance characteristics, etc.

[0029] For the establishment of the above two maps, the data required for establishment can be obtained through the power grid dispatching system. After the construction of the two maps is completed, the power grid operation status can be determined by matching the collected power grid operation data with the two maps.

[0030] The fault case map records relevant information about fault scenarios that have occurred. If the currently collected power grid operation data is data that has already been recorded, it can be matched with the fault case map to quickly obtain possible fault conditions without the need for analysis from scratch, thereby optimizing power grid management efficiency.

[0031] However, for unrecorded data, it is impossible to match it with the data recorded in the fault case graph or knowledge graph, so it is impossible to obtain its potential fault risk and determine whether it is in normal operation. This requires fault budgeting based on the current knowledge graph to respond to possible abnormal situations in a timely manner and reduce fault risks.

[0032] This embodiment specifically uses a decision tree algorithm to classify and identify the power grid operation information reflected in the knowledge graph to achieve fault budgeting for the power grid operation data.

[0033] Prior to this, it is necessary to identify the type of the collected power grid operation data. Specifically, the identification of the power grid operation data type based on the knowledge graph includes: Determine the data source of the power grid operation data, and determine the entity information, attribute information and relationship information in the power grid operation data based on the data source; Match the entity information, attribute information and relationship information of the power grid operation data with the knowledge graph in sequence; When the entity information, attribute information and relationship information all match the corresponding information in the knowledge graph, the power grid operation data type is determined to be a recorded data type; In other cases, the power grid operation data type is determined to be an unrecorded data type.

[0034] The grid operation data can come from various links of the power system, such as power generation, transmission, transformation, distribution, power consumption and dispatching. By tracing the source of the data, the grid operation data can be directly located to the corresponding grid equipment or events, thereby determining the corresponding entity information, attribute information and relationship information. Then match it with the corresponding information recorded in the knowledge graph. If the match can be completed, it proves that the relevant data has been recorded in the knowledge graph and is a recorded data type; otherwise, it is an unrecorded data type.

[0035] For the grid operation data of recorded data types, the fault type can be quickly identified through the fault case map, and then the corresponding grid dispatching strategy can be selected, avoiding repeated calculation of the same type of grid operation data and optimizing the fault budget efficiency.

[0036] For power grid operation data with unrecorded data types, it is necessary to combine the decision tree algorithm for fault budgeting. Before building the decision tree for fault budgeting, it is necessary to first extract association rules from the knowledge graph based on the association rule algorithm combined with the pruning strategy to reduce the interference of invalid information irrelevant to the fault budget or information with low correlation degree on the construction of the decision tree, and improve the construction efficiency and accuracy of the decision tree.

[0037] Among them, rule extraction mainly includes three aspects: entity extraction, attribute extraction and relationship extraction. Entity extraction is the first step of rule extraction, which can identify nouns or noun phrases with specific meanings from the text stored in the knowledge graph. These nouns or phrases usually represent specific objects or concepts in the real world, such as power grid equipment, power grid nodes, timestamps, etc. Attribute extraction is to further extract specific information or features associated with these entities after identifying the entities. These attributes describe certain aspects or characteristics of the entities, such as voltage level, current intensity, power value, etc. Relationship extraction is the last link of rule extraction. It aims to identify the association or interaction between entities or between entities and attributes, including the connection relationship between devices (such as the connection of a transformer to a certain line), the logical relationship between attributes (such as the balance relationship between voltage and current), and time series relationships (such as data changes at different time points).

[0038] Specifically, the association rules extracted in this implementation are of the form of X→Y, where X and Y are respectively called the predecessor and successor of the association rule, and the association rule algorithm Apriori is specifically used to find the relationship between item sets in the database through an iterative method of layer-by-layer search to form association rules.

[0039] Among them, the operation process of the association rule algorithm Apriori involves two parameters, namely support and confidence. Support represents the frequency of item sets appearing in all transactions, and confidence represents the probability of the subsequent item set appearing when the leading item set appears. The settings of support and confidence are directly related to the selection of frequent item sets and the generation of association rules by the association rule algorithm Apriori, and can be adjusted according to actual needs.

[0040] In addition to these two parameters, the rule extraction results of the association rule algorithm Apriori are also affected by the data set for rule extraction. The larger the data set, the more frequent item sets and candidate item sets need to be traversed. When the data set is large and sparse, it is also necessary to scan the entire data set multiple times to find frequent item sets, and the running time of the association rule algorithm will increase accordingly.

[0041] Therefore, invalid information in the knowledge graph is screened and eliminated according to the similarity of the power grid operation data, reducing the interference of repeated and useless item sets on the extraction of association rules during the association extraction process, and ensuring the efficiency of the association rule algorithm in extracting association rules. Even for large data sets, association rule extraction can be achieved efficiently and quickly, reducing the impact of the initial data volume on the operating efficiency of the association rule algorithm.

[0042] Specifically, the calculating the similarity between the power grid operation data and the historical data in the knowledge graph, and filtering and removing the invalid data in the knowledge graph according to the similarity calculation result, includes: Obtain historical update information and historical fault information of the knowledge graph, match the historical fault information with the historical update information, determine the update information corresponding to each fault, and set the weights of entity information, attribute information, and relationship information based on the update information corresponding to each fault; The entity information, attribute information and relationship information corresponding to each entity are used as similarity calculation targets respectively, and the similarity between each entity in the knowledge graph and the power grid operation data is calculated in combination with the corresponding weights; Filter out entities whose similarity is lower than the preset similarity threshold to filter out the historical data corresponding to the entity in the knowledge graph as invalid data, and eliminate all invalid data in the knowledge graph.

[0043] If the similarity is calculated directly using one of the entity information, attribute information, and relationship information, it is inevitable that similar entities will be misjudged or missed. Therefore, starting from the ultimate goal of fault identification, the update probability of each type of information is determined when a fault occurs, and the corresponding weight is assigned, and the similarity between each entity in the knowledge graph and the current power grid operation data is evaluated in multiple dimensions. For example, before and after the fault occurs, the probability of updating entity information is small, and the probability of updating attribute information and relationship information is large. Therefore, compared with attribute information and relationship information, a larger weight is set for entity information to avoid misjudgment and missed judgment of similar entities.

[0044] The historical update information of the knowledge graph includes the change records of entities, attributes, and relationships, as well as the timestamps of these changes. The historical fault information includes detailed information such as fault type, fault time, fault location, and affected equipment. The timestamp can be used to align the historical fault information with the historical update information to obtain the relevant update information before and after each fault occurs.

[0045] In this embodiment, the corresponding similarity calculation can be performed by using a similarity calculation method such as cosine similarity.

[0046] Taking into account that multiple areas with similar lines or operating scenarios may be identified at the same time, the corresponding entity update frequency, attribute fluctuation range and relationship increase rate are further extracted to determine the failure probability of the power grid operation data within each candidate rule extraction range, and a comprehensive score is given to each candidate rule extraction range based on the similarity to screen out the knowledge graph area with the highest similarity to the power grid operation data, further narrowing the scope of the knowledge graph involved in the construction of the decision tree, and ensuring a strong correlation between the data used in the decision tree construction process and the fault budget corresponding data of the power grid operation data.

[0047] The method of combining the fault probability, selecting a rule extraction range from the knowledge graph that removes invalid data, and obtaining a rule extraction sub-graph for performing fault budgeting includes: Determine all entities in the knowledge graph after removing invalid data, obtain the relationship information between each entity, and obtain the extraction scope of all candidate rules based on the relationship information; According to all entities within the extraction range of each candidate rule, the entity update frequency, attribute fluctuation range and relationship increase rate corresponding to each candidate rule extraction range are obtained; The failure probability score of each candidate rule extraction range is scored based on the entity update frequency, attribute fluctuation range, and relationship increase, and the comprehensive score of each candidate rule extraction range is calculated based on the similarity of the corresponding entities; The rule extraction scope is selected based on the comprehensive score, the extraction boundary is set based on the entities included in the rule extraction scope, and the rule extraction sub-graph for fault budgeting is obtained from the knowledge graph based on the extraction boundary.

[0048] The comprehensive score of each candidate rule extraction scope can be obtained by weighted summing the similarity and the probability of failure. The weights of the similarity and the probability of failure can also be adjusted according to actual needs. The candidate rule extraction scope with the highest comprehensive score is selected as the final rule extraction scope. The selected rule extraction scope contains the entity set that is most likely to extract useful rules.

[0049] Based on the entity set, the boundary entity of the rule extraction scope can be determined, so that according to the position of the boundary entity on the knowledge graph, the rule extraction subgraph containing the entity set can be obtained from the knowledge graph.

[0050] After obtaining the rule extraction sub-graph, combined with the pruning strategy, the association rules of the rule extraction sub-graph are formed according to the association rule algorithm, including: Extract entities, relationships, and attributes in the sub-graph based on rules and build the corresponding transaction database; Set the support threshold and scan the transaction database to obtain the items and their occurrence counts in all transactions, and obtain the support of each item based on the corresponding occurrence count; Filter out the item sets whose support is less than the support threshold and delete them, and construct the initial frequent item sets based on the remaining items; Combined with the pruning strategy, candidate item sets are generated based on the initial frequent item sets and the corresponding support is calculated. The next level of frequent item sets are constructed based on the support of the candidate item sets. Repeat the generation of candidate item sets and the construction of frequent item sets until all frequent item sets in the transaction database are exhausted; Extract association rules of subgraphs based on frequent item set formation rules.

[0051] In order to ensure the normal operation of the association rule algorithm Apriori, the information in the rule extraction sub-graph must be converted into a processable data type. Specifically, the entities, relationships and attributes in the rule extraction sub-graph are extracted and converted into the form of a transaction database, where each transaction represents an entity or a set of entities, and the items in the transaction represent the attributes or relationships related to the entity. Based on the generated transaction database, the association rule algorithm Apriori generates frequent item sets through the support threshold.

[0052] Among them, the support is set by the number of occurrences of the item, which can be expressed as: Support (X) = number of transactions containing item set X / total number of transactions.

[0053] Considering that even if the scale of the knowledge graph for association rule extraction is limited, the data cardinality for association rule extraction is still large due to the scale of power grid data, the frequent item set generation process of the association rule algorithm Apriori is optimized based on the pruning strategy principle that if an item set is a frequent item set, all its subsets are frequent item sets; if an item set is a non-frequent item set, all its parent sets are non-frequent item sets. This reduces useless candidate item subsets and enables the generation of the next level candidate frequent item sets without pruning, thereby saving the algorithm running time and improving mining efficiency.

[0054] Specifically, by constructing the initial frequent item sets and candidate item sets, the items or candidate item sets involved are sorted in ascending order, and the termination condition of the algorithm exhaustion is to determine whether all subsets of the generated candidate item sets are infrequent item sets, so as to optimize the frequent item set generation process of the association rule algorithm Apriori.

[0055] Through the frequent item set search of the association rule algorithm Apriori, the information related to the power grid operation data in the rule extraction sub-graph can be further extracted. In order to further realize the fault budget for the power grid operation data, the association strategies formed by further induction through the decision tree algorithm are used to generate readable rules and corresponding decision trees. Then, the new power grid operation data is analyzed through the decision tree to realize the fault budget.

[0056] In this embodiment, the CART decision tree algorithm is specifically selected to construct a rule-based decision tree.

[0057] Specifically, the extraction of association rules according to the decision tree algorithm and the establishment of a rule decision tree of the rule extraction sub-graph include: Generate input features of decision tree algorithm based on formed association rules and their corresponding transactions, and select root node based on Gini coefficient; The corresponding child nodes are recursively constructed by calculating the corresponding Gini coefficient until the stopping condition is met, and the rule decision tree of the rule extraction sub-graph is obtained.

[0058] Through the formed association rules and their transactions, an input feature data set of the decision tree algorithm can be constructed, which includes the corresponding features of the association rules and the corresponding features of the combination of items involved in their transactions.

[0059] By calculating the Gini coefficient of each feature in the input feature data set, the best feature is selected as the root node, where the root node is the feature with the smallest Gini coefficient in the input feature data set.

[0060] The input feature data set is divided into different subsets based on the feature value of the root node, and for each subset, the calculation of the Gini coefficient and the selection of the best feature are repeated to achieve recursive construction of the sub-nodes until the stopping condition is met, such as the size of the subset is less than the corresponding threshold, or the reduction of the Gini coefficient is less than the corresponding threshold, etc. After the recursion is completed, the value of the leaf node is set according to the features in the current subset to obtain the final rule decision tree.

[0061] The extracting of association rules according to the decision tree algorithm and establishing a rule decision tree of the rule extraction sub-graph also includes: The nodes of the rule decision tree are searched for similarity between the nodes, and the rule decision tree is supplemented with rules according to the search results.

[0062] In order to further improve the expressiveness and generalization capabilities of the rule decision tree, FAISS (Facebook AI Similarity Search) is used to perform similarity search on the nodes of the rule decision tree. Before performing the similarity search, it is necessary to first extract features from the nodes of the rule decision tree, which can be the corresponding subset statistics or the Gini coefficient. The extracted features are then converted into vector form through simple feature concatenation, using feature hash functions, or mapping the features to high-dimensional space for use by FAISS.

[0063] In the process of similarity search, the corresponding index needs to be built first. In the creation of the index, a brute force search method can be used. Then, for each node in the decision tree, the corresponding vector of the extracted features is used as the query vector, and the vector most similar to the query vector is searched in the index. Based on the search results, it is determined whether new rules need to be added to cover additional data subsets or improve existing classification results. Then, according to the judgment results, the decision path in the decision tree is updated to achieve rule supplementation.

[0064] The updating of the fault case map according to the power grid dispatching strategy includes: Causal reasoning is performed based on the grid operation data and the corresponding grid dispatching strategy to generate the corresponding accident plan, and the fault case map is updated according to the accident plan.

[0065] The causal reasoning is performed by combining the power grid operation data and the corresponding power grid dispatching strategy to generate the corresponding accident plan, including: Extract text features of power grid dispatch strategy based on pre-trained language representation model, and build causal graph based on the extracted text features; Perform causal reasoning based on the causal graph, construct the accident scenario based on the causal reasoning results, and generate a text description corresponding to the accident scenario in combination with the corresponding text features; Generate accident plans based on accident scenarios and corresponding text descriptions.

[0066] First, the information on causal relationships is stored and managed by constructing a causal graph. Then, the current set of events is determined through the current power grid operation data and the power grid dispatch strategy, so that the causal relationship can be explained and predicted based on the causal graph.

[0067] When constructing a causal graph, the causal relationship stored therein can be obtained by extracting text features, and the forms of causal relationships are diverse, including causal relationships between nouns or phrases, between sentences or clauses, and between events. In this embodiment, the extraction of text features is specifically achieved through the pre-trained BERT model, thereby achieving causal relationship analysis between the extracted text features.

[0068] After extracting text features, it is necessary to first define causal variables related to grid operation according to actual needs, such as voltage stability, frequency deviation or load demand, and then construct a causal graph based on the extracted specific text features. The constructed causal graph can represent the dependency and influence direction between each causal variable.

[0069] After the construction is completed, the appropriate causal reasoning algorithm can be selected according to the structure and characteristics of the causal diagram, such as Bayesian network reasoning, structural equation model, etc. The extracted text features are input into the causal reasoning algorithm, and reasoning is performed based on the causal diagram to obtain the status of each causal variable and its changing trend.

[0070] Based on the results of causal reasoning, the key variables of power grid accidents corresponding to the current power grid operation data and power grid dispatching strategy are identified, and then the corresponding accident scenarios are constructed according to the status of the key variables and their changing trends.

[0071] Finally, natural language generation technology is used to combine the extracted text features and the results of causal reasoning to generate a text description corresponding to the accident scenario, where the text description at least includes the scene characteristics and impact of the accident scenario. The accident plan includes response measures for the generated accident scenario, such as adjusting the load, removing the faulty equipment, etc., providing a data basis for the update of the fault case map.

[0072] The updating of the fault case map according to the accident plan includes: Create or select the corresponding fault node according to the accident plan, match the relevant nodes in the fault case graph according to the text description of the accident plan, and connect the fault node and the relevant nodes; Create or select the cause node, impact node and measure node corresponding to the fault node according to the accident plan, and establish connections between the cause node, impact node and measure node and the fault node respectively.

[0073] The fault case map contains multiple fault nodes, each of which corresponds to a fault type, and each fault node includes basic information about the fault, such as fault name, fault code, fault description, etc. Considering that the causes of the same fault type may also be different, the basic information of the corresponding fault of the corresponding accident scenario is first obtained through the accident plan, and it is matched with the existing fault nodes to determine whether there is a fault node of the same fault type. If so, the corresponding fault node can be directly selected. If not, the corresponding common fault node is selected.

[0074] In addition to the fault nodes that record fault information, the fault case map also includes other types of nodes that represent other power grid operation information such as equipment information. Through the text description of the accident plan, other nodes related to the fault node can be found in the fault case map, such as equipment nodes, geographic location nodes, time nodes, etc. to establish connections. The connection relationship can be defined by the text description. For example, you can choose to establish a connection based on keywords such as "affect", "located" or "occurs in".

[0075] For the fault causes, fault impacts and response measures mentioned in the accident plan, they can also be matched with the information stored in the fault case map through the corresponding text description. If it already exists, the existing information can be directly applied to avoid duplication of nodes in the fault case map. If it does not exist, the corresponding node is created.

[0076] By creating or updating fault nodes, related nodes, cause nodes, impact nodes and measure nodes, the faults that may correspond to the current power grid operation data are integrated into the fault case map from the aspects of fault cause, fault characteristics, countermeasures, etc. If the same type of data as the current power grid operation data appears again in the future, the fault budget can be executed quickly, and combined with the corresponding accident plan to achieve efficient processing, so as to ensure the safety and stability of power grid operation.

[0077] Considering that there are errors between the relevant data collected under abnormal conditions of the power grid and those under normal operating conditions, equipment replacement or adjustment may also occur after abnormal processing. If the knowledge graph is directly updated with the acquired power grid operation data, it may cause information deviation in the knowledge graph. Therefore, after completing the abnormal processing, the corresponding abnormal processing results, that is, the adjusted power grid situation, are combined to update the knowledge graph to ensure the accuracy of the content displayed in the knowledge graph and improve the management and control efficiency and accuracy of the power grid.

[0078] In summary, the overall process of power grid operation control described in this embodiment is as follows: Figure 2 As shown, for the constructed power grid dispatching strategy, the corresponding exception handling can be achieved through human-machine collaboration.

[0079] The above-described embodiment is only a preferred solution of the present invention and does not limit the present invention in any form. There are other variations and modifications without exceeding the technical solution described in the claims.

Claims

1. A power grid operation control method based on knowledge graph rule extraction, characterized in that: include: Obtain power grid operation data and identify the type of power grid operation data based on the knowledge graph; When the power grid operation data type is an unrecorded data type, the similarity between the power grid operation data and the historical data in the knowledge graph is calculated, and invalid data in the knowledge graph is filtered and removed according to the similarity calculation result; Combined with the failure probability, select the rule extraction range from the knowledge graph that removes invalid data, and obtain the rule extraction sub-graph for fault budgeting; Combined with the pruning strategy, the association rules of the rule extraction sub-graph are formed according to the association rule algorithm, and the association rules are extracted according to the decision tree algorithm to establish a rule decision tree; Based on the rule decision tree, fault budget is made according to the power grid operation data, and corresponding power grid dispatch strategy is formulated to handle exceptions; The fault case map is updated according to the power grid dispatching strategy, and when the power grid operation data type is the recorded data type, fault budget and exception processing are performed according to the fault case map.

2. The power grid operation control method based on knowledge graph rule extraction according to claim 1 is characterized in that: The identifying of power grid operation data types based on the knowledge graph includes: Determine the data source of the power grid operation data, and determine the entity information, attribute information and relationship information in the power grid operation data based on the data source; Match the entity information, attribute information and relationship information of the power grid operation data with the knowledge graph in sequence; When the entity information, attribute information and relationship information all match the corresponding information in the knowledge graph, the power grid operation data type is determined to be a recorded data type; In other cases, the power grid operation data type is determined to be an unrecorded data type.

3. The power grid operation control method based on knowledge graph rule extraction according to claim 2 is characterized in that: The calculating the similarity between the power grid operation data and the historical data in the knowledge graph, and filtering and removing the invalid data in the knowledge graph according to the similarity calculation result, includes: Obtain historical update information and historical fault information of the knowledge graph, match the historical fault information with the historical update information, determine the update information corresponding to each fault, and set the weights of entity information, attribute information, and relationship information based on the update information corresponding to each fault; The entity information, attribute information and relationship information corresponding to each entity are used as similarity calculation targets respectively, and the similarity between each entity in the knowledge graph and the power grid operation data is calculated in combination with the corresponding weights; Filter out entities whose similarity is lower than the preset similarity threshold to filter out the historical data corresponding to the entity in the knowledge graph as invalid data, and eliminate all invalid data in the knowledge graph.

4. The power grid operation control method based on knowledge graph rule extraction according to claim 3 is characterized in that: The method combines the fault probability, selects a rule extraction range from the knowledge graph excluding invalid data, and obtains a rule extraction sub-graph for performing fault budgeting, including: Determine all entities in the knowledge graph after removing invalid data, obtain the relationship information between each entity, and obtain the extraction scope of all candidate rules based on the relationship information; According to all entities within the extraction range of each candidate rule, the entity update frequency, attribute fluctuation range and relationship increase rate corresponding to each candidate rule extraction range are obtained; The failure probability score of each candidate rule extraction range is scored based on the entity update frequency, attribute fluctuation range, and relationship increase, and the comprehensive score of each candidate rule extraction range is calculated based on the similarity of the corresponding entities; The rule extraction scope is selected based on the comprehensive score, the extraction boundary is set based on the entities included in the rule extraction scope, and the rule extraction sub-graph for fault budgeting is obtained from the knowledge graph based on the extraction boundary.

5. The power grid operation control method based on knowledge graph rule extraction according to claim 4 is characterized in that: The combined pruning strategy forms association rules of the rule extraction sub-graph according to the association rule algorithm, including: Extract entities, relationships, and attributes in the sub-graph based on rules and build the corresponding transaction database; Set the support threshold and scan the transaction database to obtain the items and their occurrence counts in all transactions, and obtain the support of each item based on the corresponding occurrence count; Filter out the item sets whose support is less than the support threshold and delete them, and construct the initial frequent item sets based on the remaining items; Combined with the pruning strategy, candidate item sets are generated based on the initial frequent item sets and the corresponding support is calculated. The next level of frequent item sets are constructed based on the support of the candidate item sets. Repeat the generation of candidate item sets and the construction of frequent item sets until all frequent item sets in the transaction database are exhausted; Extract association rules of subgraphs based on frequent item set formation rules.

6. The power grid operation control method based on knowledge graph rule extraction according to claim 5 is characterized in that: The extraction of association rules according to the decision tree algorithm and the establishment of a rule decision tree of the rule extraction sub-graph include: Generate the input features of the decision tree algorithm based on the formed association rules and their corresponding transactions, and select the root node based on the Gini coefficient; The corresponding child nodes are recursively constructed by calculating the corresponding Gini coefficient until the stopping condition is met, and the rule decision tree of the rule extraction sub-graph is obtained.

7. The power grid operation control method based on knowledge graph rule extraction according to claim 6 is characterized in that: The extracting of association rules according to the decision tree algorithm and establishing a rule decision tree of the rule extraction sub-graph also includes: The nodes of the rule decision tree are searched for similarity between the nodes, and the rule decision tree is supplemented with rules according to the search results.

8. The power grid operation control method based on knowledge graph rule extraction according to claim 1 is characterized in that: The updating of the fault case map according to the power grid dispatching strategy includes: Causal reasoning is performed based on the grid operation data and the corresponding grid dispatching strategy to generate the corresponding accident plan, and the fault case map is updated according to the accident plan.

9. The power grid operation control method based on knowledge graph rule extraction according to claim 8 is characterized in that: The causal reasoning is performed by combining the power grid operation data and the corresponding power grid dispatching strategy to generate the corresponding accident plan, including: Extract text features of power grid dispatch strategy based on pre-trained language representation model, and build causal graph based on the extracted text features; Perform causal reasoning based on the causal graph, construct the accident scenario based on the causal reasoning results, and generate a text description corresponding to the accident scenario in combination with the corresponding text features; Generate accident plans based on accident scenarios and corresponding text descriptions.

10. The power grid operation control method based on knowledge graph rule extraction according to claim 9 is characterized in that: The updating of the fault case map according to the accident plan includes: Create or select the corresponding fault node according to the accident plan, match the relevant nodes in the fault case graph according to the text description of the accident plan, and connect the fault node and the relevant nodes; Create or select the cause node, impact node and measure node corresponding to the fault node according to the accident plan, and establish connections between the cause node, impact node and measure node and the fault node respectively.

Images