Construction and updating method of electric power knowledge graph, product and storage medium

Abstract

The invention discloses a construction and updating method of an electric power knowledge graph, a product and a storage medium, and relates to the technical field of electric power system management. The method comprises the following steps: acquiring an original power data set in a power system; preprocessing the original power data set to obtain a target power data set; according to the target power data set, an initial power knowledge graph is constructed, the initial power knowledge graph comprises a basic framework and a graph database, the basic framework is used for representing the architecture of power knowledge contained in the power system, and the graph database is used for storing and querying the power knowledge based on a graph structure; and updating the initial power knowledge graph based on an incremental learning algorithm to obtain a target power knowledge graph. The technical problem that safe and stable operation of a power system is affected due to the fact that an isolated database and a manual analysis method are adopted for power dispatching information management in the prior art is large in error and low in efficiency is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system management, and in particular to a method, product, and storage medium for constructing and updating a power knowledge graph. Background Art

[0002] As the scale of power systems continues to expand and become more complex, the amount and types of data faced by power dispatching systems have increased dramatically. Traditional power dispatching information management methods mostly use isolated databases and manual analysis methods, which leads to serious data islands, making it difficult to integrate and share data, and limiting the effective use of information. At the same time, the power dispatching process involves a large number of regulations, accident plans and other documents and materials, and the management and use of these unstructured data are very difficult. In addition, the operating status of the power system is dynamic and real-time, and traditional static data analysis methods cannot meet the needs of real-time updates and dynamic decision-making.

[0003] From the above analysis, it can be seen that there is currently no effective solution to the problem that the above-mentioned existing technologies use isolated databases and manual analysis methods to manage power dispatching information, resulting in large errors and low efficiency, which affects the safe and stable operation of the power system. Summary of the invention

[0004] The embodiments of the present invention provide a method, product, and storage medium for constructing and updating an electric power knowledge graph, so as to at least solve the technical problem that the prior art uses isolated databases and manual analysis methods to manage electric power dispatching information, resulting in large errors and low efficiency, thereby affecting the safe and stable operation of the power system.

[0005] According to one aspect of an embodiment of the present invention, a method for constructing and updating an electric power knowledge graph is provided, comprising:

[0006] Collect original power data sets in the power system, wherein the original power data sets include multiple categories of power data, and different categories of power data have different sources; preprocess the original power data sets to obtain a target power data set; construct an initial power knowledge graph based on the target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, the basic framework is used to characterize the architecture of power knowledge contained in the power system, and the graph database is used to store and query power knowledge based on a graph structure; update the initial power knowledge graph based on an incremental learning algorithm to obtain a target power knowledge graph.

[0007] Optionally, preprocessing the original power data set to obtain a target power data set includes: supplementing the missing data in the original power data set based on an interpolation method to obtain a first data set; identifying and removing abnormal data in the first data set based on a box plot to obtain a second data set; standardizing data of the same type in the second data set to obtain a third data set; and converting unstructured data in the third data set into structured data to obtain a target power data set.

[0008] Optionally, constructing an initial power knowledge graph based on the target power data set includes: using natural language processing technology to extract data from the target power data set to obtain a fourth data set, wherein the fourth data set includes text data in the target power data set; performing word segmentation and part-of-speech tagging on the fourth data set to obtain a fifth data set; using a deep learning model to perform entity recognition on the fifth data set to obtain a sixth data set, wherein the sixth data set includes multiple categories of entities and attribute information of any entity; using a relationship extraction model to identify the relationship between any two entities in the sixth data set to obtain a seventh data set, wherein the seventh data set includes association information between any two entities; and constructing an initial power knowledge graph based on the sixth and seventh data sets.

[0009] Optionally, performing word segmentation and part-of-speech tagging on the fourth data set to obtain the fifth data set includes: using a word segmentation processing tool to perform word segmentation on any text data in the fourth data set to obtain an eighth data set, wherein the eighth data set includes a word sequence corresponding to any text data; using a part-of-speech tagging tool to perform part-of-speech tagging on any word sequence in the eighth data set to obtain the fifth data set.

[0010] Optionally, constructing an initial power knowledge graph based on the sixth data set and the seventh data set includes: taking the entities in the sixth data set as the nodes of the basic framework, and taking the association information between any two entities in the seventh data set as the association information between the corresponding nodes of any two entities, to construct a basic framework; according to a preset storage method, storing the node information in the basic framework in the nodes of the graph structure, and storing the association information between any two nodes in the basic framework in the edges of the graph structure, to construct a graph database.

[0011] Optionally, the initial power knowledge graph is updated based on the incremental learning algorithm to obtain the target power knowledge graph, including: collecting a new power data set of the power system based on the data stream processing architecture; using a recursive neural network to perform incremental learning on the new power data set to obtain a learning result, wherein the learning result is used to determine the entities and / or associated information involved in any new data in the new power data set; querying the entity and / or associated information in the graph database to obtain a query result, wherein the query result is used to determine whether the entity is a newly added entity and / or whether the associated information is newly added associated information; updating the initial power knowledge graph according to the query result and the learning result to obtain the target power knowledge graph.

[0012] Optionally, the initial power knowledge graph is updated according to the query results and the learning results to obtain a target power knowledge graph including: in response to determining that the entity is a newly added entity and / or the associated information is newly added associated information according to the query results, storing the newly added entity and / or the newly added associated information in the graph database, and adding nodes corresponding to the newly added entity and / or edges corresponding to the newly added associated information in the graph structure according to the learning results; in response to determining that the entity is an existing entity and / or the associated information is existing associated information according to the query results, updating the nodes corresponding to the existing entities and / or the edges corresponding to the existing associated information in the graph structure according to the learning results.

[0013] Optionally, the above-mentioned method for constructing and updating the electric power knowledge graph also includes: using a web page ranking algorithm to evaluate the importance of any node in the graph structure to obtain a first evaluation result, wherein the first evaluation result is used to determine the importance of any node in the graph structure; using a community discovery algorithm to cluster the nodes and edges in the graph structure to obtain a clustering result, wherein the clustering result is used to determine the nodes and edges in the graph structure whose correlation degree is higher than a preset value; using a graph neural network to analyze and decide on the first evaluation result and the clustering result to obtain a processing result, wherein the processing result is used to determine the predicted fault events of the power system; using a Bayesian network to perform a probability assessment on the predicted fault events to obtain a second evaluation result, wherein the second evaluation result is used to determine the probability of the predicted fault events occurring.

[0014] According to another aspect of an embodiment of the present invention, there is also provided a device for constructing and updating an electric power knowledge graph, including:

[0015] An acquisition module is used to acquire the original power data set in the power system, wherein the original power data set includes multiple categories of power data, and the sources of different categories of power data are different; a preprocessing module is used to preprocess the original power data set to obtain a target power data set; a construction module is used to construct an initial power knowledge graph based on the target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, the basic framework is used to characterize the architecture of the power knowledge contained in the power system, and the graph database is used to store and query power knowledge based on a graph structure; an update module is used to update the initial power knowledge graph based on an incremental learning algorithm to obtain a target power knowledge graph.

[0016] Optionally, the above-mentioned preprocessing module is also used to: supplement the missing data in the original power data set based on the interpolation method to obtain a first data set; identify and remove abnormal data in the first data set based on the box plot to obtain a second data set; standardize the data of the same type in the second data set to obtain a third data set; convert the unstructured data in the third data set into structured data to obtain a target power data set.

[0017] Optionally, the above-mentioned construction module is also used to: use natural language processing technology to extract data from the target power data set to obtain a fourth data set, wherein the fourth data set includes text data in the target power data set; perform word segmentation and part-of-speech tagging on the fourth data set to obtain a fifth data set; use a deep learning model to perform entity recognition on the fifth data set to obtain a sixth data set, wherein the sixth data set includes multiple categories of entities and attribute information of any entity; use a relationship extraction model to identify the relationship between any two entities in the sixth data set to obtain a seventh data set, wherein the seventh data set includes association information between any two entities; and construct an initial power knowledge graph based on the sixth data set and the seventh data set.

[0018] Optionally, the above-mentioned construction module is also used to: use a word segmentation processing tool to perform word segmentation processing on any text data in the fourth data set to obtain an eighth data set, wherein the eighth data set includes a word sequence corresponding to any text data; use a part-of-speech tagging tool to perform part-of-speech tagging on any word sequence in the eighth data set to obtain a fifth data set.

[0019] Optionally, the above-mentioned construction module is also used to: use the entities in the sixth data set as the nodes of the basic framework, and use the association information between any two entities in the seventh data set as the association information between the corresponding nodes of any two entities to construct a basic framework; store the node information in the basic framework in the nodes of the graph structure, and store the association information between any two nodes in the basic framework in the edges of the graph structure according to a preset storage method, to construct a graph database.

[0020] Optionally, the above-mentioned update module is also used to: collect new power data sets of the power system based on the data stream processing architecture; use a recursive neural network to perform incremental learning on the new power data set to obtain learning results, wherein the learning results are used to determine the entities and / or associated information involved in any new data in the new power data set; query the entity and / or associated information in the graph database to obtain the query results, wherein the query results are used to determine whether the entity is a newly added entity and / or whether the associated information is newly added associated information; update the initial power knowledge graph according to the query results and the learning results to obtain the target power knowledge graph.

[0021] Optionally, the above-mentioned update module is also used to: in response to determining that the entity is a newly added entity and / or the associated information is newly added associated information based on the query results, store the newly added entity and / or the newly added associated information in the graph database, and add nodes corresponding to the newly added entity and / or edges corresponding to the newly added associated information in the graph structure based on the learning results; in response to determining that the entity is an existing entity and / or the associated information is existing associated information based on the query results, update the nodes corresponding to the existing entities and / or the edges corresponding to the existing associated information in the graph structure based on the learning results.

[0022] Optionally, the above-mentioned power knowledge graph construction and update device also includes: a prediction module, which is used to use a web page ranking algorithm to perform an importance evaluation on any node in the graph structure to obtain a first evaluation result, wherein the first evaluation result is used to determine the importance of any node in the graph structure; clustering the nodes and edges in the graph structure using a community discovery algorithm to obtain a clustering result, wherein the clustering result is used to determine the nodes and edges in the graph structure whose correlation degree is higher than a preset value; using a graph neural network to analyze and decide on the first evaluation result and the clustering result to obtain a processing result, wherein the processing result is used to determine the predicted fault event of the power system; using a Bayesian network to perform a probability evaluation on the predicted fault event to obtain a second evaluation result, wherein the second evaluation result is used to determine the probability of the predicted fault event occurring.

[0023] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, which, when executed by a processor, implements any one of the aforementioned methods for constructing and updating an electric power knowledge graph.

[0024] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the aforementioned methods for constructing and updating an electric power knowledge graph.

[0025] In an embodiment of the present invention, firstly, the original power data set in the power system is collected, wherein the original power data set includes multiple categories of power data, and the sources of different categories of power data are different. Then, the original power data set is preprocessed to obtain the target power data set, and then the initial power knowledge graph is constructed according to the target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, wherein the basic framework is used to characterize the architecture of the power knowledge contained in the power system, and the graph database is used to store and query power knowledge based on the graph structure, and finally the initial power knowledge graph is updated based on the incremental learning algorithm to obtain the target power knowledge graph. By using multiple categories of power data in the power system to construct a power knowledge graph based on a graph structure, and using the incremental learning algorithm to dynamically update the power knowledge graph, the purpose of improving the expression accuracy and richness of the power knowledge graph and improving the maintainability of the power knowledge graph is achieved, thereby achieving the technical effect of improving the dispatching management efficiency of the power system to improve the stability and safety of the power system, thereby solving the technical problem that the existing technology uses an isolated database and manual analysis method for power dispatching information management, resulting in large errors and low efficiency, thus affecting the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0027] Figure 1 It is a hardware structure block diagram of a computing terminal for an optional method for constructing and updating an electric power knowledge graph according to an embodiment of the present invention;

[0028] Figure 2 is a flow chart of a method for constructing and updating an electric power knowledge graph according to an embodiment of the present invention;

[0029] Figure 3 It is a structural block diagram of a device for constructing and updating an electric power knowledge graph according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] According to an embodiment of the present invention, a method embodiment of a method for constructing and updating an electric power knowledge graph is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0033] Figure 1 is a hardware structure block diagram of a computing terminal for an optional method for constructing and updating an electric power knowledge graph according to an embodiment of the present invention, such as Figure 1 As shown, the computing terminal (or mobile device) may include one or more processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor (Microcontroller Unit, MCU) or a programmable logic device (Field Programmable Gate Array, FPGA)), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display device 110, an input and output device 108 (i.e., an IO device), a Universal Serial Bus (USB) port (which may be included as one of the ports of a computer bus, not shown in the figure), a network interface (not shown in the figure), a power supply (not shown in the figure) and / or a camera (not shown in the figure). It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above computing terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0034] It should be noted that the one or more processors 102 and / or other data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be fully or partially integrated into any of the other components in the computing terminal (or mobile device).

[0035] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the method for constructing and updating the electric power knowledge graph in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the method for constructing and updating the electric power knowledge graph described above is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computing terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0036] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computing terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0037] Under the above operating environment, the embodiment of the present invention provides the following Figure 2 The construction and update method of the power knowledge graph shown in FIG. Figure 2 is a flow chart of a method for constructing and updating an electric power knowledge graph according to an embodiment of the present invention. Figure 2 As shown, the method includes the following implementation steps:

[0038] Step S201, collecting an original power data set in the power system, wherein the original power data set includes multiple categories of power data, and different categories of power data have different sources;

[0039] Step S202, preprocessing the original power data set to obtain a target power data set;

[0040] Step S203, constructing an initial power knowledge graph according to the target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, the basic framework is used to characterize the architecture of power knowledge contained in the power system, and the graph database is used to store and query power knowledge based on the graph structure;

[0041] Step S204: Update the initial power knowledge graph based on the incremental learning algorithm to obtain the target power knowledge graph.

[0042] The various categories of power data contained in the above-mentioned original power data set may include but are not limited to: equipment operation data, such as the voltage, current, and power of the substation; basic equipment data, such as the model, specification, and status of the equipment; regulations, which are used to describe the operation rules and operating procedures of the power system; accident plans, which are used to describe the emergency response and handling measures of the power system under different circumstances. It should be noted that the equipment operation data can be collected in real time by data acquisition terminals such as remote terminal units (RTU) and phasor measurement units (PMU); basic equipment data can be obtained from the equipment management system; regulations can be extracted from text files and databases of the power management system; accident plans can be extracted from the fault management system. It should be noted that the accident plan may include information such as the type of fault, fault handling steps, and emergency contacts of power safety accidents.

[0043] The present invention can ensure the consistency of power data in the target power data set by performing preprocessing operations such as data cleaning, data standardization, and data format conversion on the original power data set. The power knowledge graph can be used to record all power knowledge related to the power system, and the power knowledge graph can also record and characterize the correlation between different power knowledge.

[0044] The above-mentioned incremental learning algorithm can be used to update the initial power knowledge graph in real time using large-scale new power data collected, without the need to retrain the power knowledge graph using a large amount of power data, thereby reducing computing costs and improving the updating efficiency of the knowledge graph.

[0045] In an embodiment of the present invention, firstly, the original power data set in the power system is collected, wherein the original power data set includes multiple categories of power data, and the sources of different categories of power data are different. Then, the original power data set is preprocessed to obtain the target power data set, and then the initial power knowledge graph is constructed according to the target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, wherein the basic framework is used to characterize the architecture of the power knowledge contained in the power system, and the graph database is used to store and query power knowledge based on the graph structure, and finally the initial power knowledge graph is updated based on the incremental learning algorithm to obtain the target power knowledge graph. By using multiple categories of power data in the power system to construct a power knowledge graph based on a graph structure, and using the incremental learning algorithm to dynamically update the power knowledge graph, the purpose of improving the expression accuracy and richness of the power knowledge graph and improving the maintainability of the power knowledge graph is achieved, thereby achieving the technical effect of improving the dispatching management efficiency of the power system to improve the stability and safety of the power system, thereby solving the technical problem that the existing technology uses an isolated database and manual analysis method for power dispatching information management, resulting in large errors and low efficiency, thus affecting the safe and stable operation of the power system.

[0046] The above method of the embodiment of the present invention is further introduced below.

[0047] In an optional embodiment, in step S202, the original power data set is preprocessed to obtain the target power data set including:

[0048] Step S221, supplementing the missing data in the original power data set based on the interpolation method to obtain a first data set;

[0049] Step S222, identifying and removing abnormal data in the first data set based on the box plot to obtain a second data set;

[0050] Step S223, performing standardization processing on the data of the same type in the second data set to obtain a third data set;

[0051] Step S224: convert the unstructured data in the third data set into structured data to obtain a target power data set.

[0052] In the present invention, the interpolation method is used to supplement the missing data in the original power data set as shown in the following formula (1):

[0053]

[0054] In the above formula (1), x(t) is the interpolation value corresponding to time t, x i Indicates time t i The data point, x i+1 Indicates time ti+1 data point, i represents the data point x i The corresponding index.

[0055] Furthermore, the power data in the interpolated data set (first data set) is unified, for example, all voltage data in the first data set is converted into kilovolts (kV), and all current data in the first data set is converted into amperes (A). Then, a boxplot is used to identify and remove abnormal data in the first data set to obtain a second data set.

[0056] Next, the standard score (Z-Score) standardization method shown in the following formula (2) is used to standardize the unified power data so that the power data conforms to the standard normal distribution:

[0057]

[0058] In the above formula (2), x represents the power data before standardization, z represents the data after standardization, μ refers to the mean of the same type of power data, and σ refers to the standard deviation of the same type of power data.

[0059] The last step of the preprocessing process is to convert the unstructured data (such as the fault log of the power equipment) in the third data set obtained by the standardization process into structured data to obtain the target power data set.

[0060] In an optional embodiment, in step S203, constructing an initial power knowledge graph according to the target power data set includes:

[0061] Step S231, extracting data from the target power data set using natural language processing technology to obtain a fourth data set, wherein the fourth data set includes text data in the target power data set;

[0062] Step S232, performing word segmentation and part-of-speech tagging on the fourth data set to obtain a fifth data set;

[0063] Step S233, performing entity recognition on the fifth data set using a deep learning model to obtain a sixth data set, wherein the sixth data set includes entities of multiple categories and attribute information of any entity;

[0064] Step S234, using the relationship extraction model to identify the relationship between any two entities in the sixth data set to obtain a seventh data set, wherein the seventh data set includes association information between any two entities;

[0065] Step S235, constructing an initial power knowledge graph based on the sixth data set and the seventh data set.

[0066] Among them, in step S232, word segmentation processing and part-of-speech tagging are performed on the fourth data set, and the fifth data set obtained includes:

[0067] Step S2321: Use a word segmentation processing tool to perform word segmentation processing on any piece of text data in the fourth data set to obtain an eighth data set, where the eighth data set includes a word sequence corresponding to any piece of text data.

[0068] Step S2322: Use a part-of-speech tagging tool to perform part-of-speech tagging on any word sequence in the eighth data set to obtain the fifth data set.

[0069] In the present invention, natural language processing technologies (such as: Tokenization, POS Tagging) can be used to extract text data from the target power data set to obtain the fourth data set. Then, the Jieba word segmentation processing tool or the Natural Language Toolkit (NLTK) can be used to identify and mark text elements such as words and phrases in the text data, and the text data can be segmented into word sequences according to the marking results to obtain the eighth data set. Further, the Stanford Natural Language Processing tool or the Spacy part-of-speech tagging tool can be used to tag the part of speech (such as nouns, adjectives, etc.) of each word in the eighth data set to obtain the fifth data set. In addition, according to the identification and marking results of the Jieba word segmentation processing tool, meaningless words such as "de" and "shi" in the eighth data set can be removed to improve the accuracy of subsequent data processing.

[0070] After performing part-of-speech tagging on the data, a deep learning model is used to identify entities in the fifth data set and the attribute information of each entity. The types of entities can include, but are not limited to: substations, equipment, regulations, accidents, operations. Different entities have different attribute information. For example, the attribute information of a substation includes name, location, capacity, etc. It should be noted here that the deep learning model can be as shown in the following formula (3). This model is a combination of a Bidirectional Long Short-Term Memory (BiLSTM) and a Conditional Random Field (CRF) model, and this model can effectively capture and label the feature of the data sequence.

[0071]

[0072] In the above formula (3), x represents the input sequence of the deep learning model, i represents the position of the current index in the data sequence x, n represents the total number of indexes contained in the data sequence x, and y i represents the true label sequence corresponding to the i-th index in the data sequence x output by the deep learning model, ψ is the feature function, y ′ i Represents all possible label sequences corresponding to the i-th index in the data sequence x.

[0073] After performing entity recognition to obtain the sixth data set, the relational extraction model is used to extract the association information between any two entities from the entities contained in the sixth data set. It should be noted here that the relational extraction model can be a recurrent neural network (RNN) or a bidirectional encoder based on a transformer (BERT). The relational extraction model can represent entity pairs (any two entities) and contextual information of entity pairs by constructing feature vectors (such as word embedding and position embedding). When performing relational extraction, the contextual information between entity pairs needs to be considered. The identified association information can include but is not limited to "include", "measurement", "dependency", etc. The association relationship between entities is such as "substation-include-equipment", "equipment-measurement-parameter", etc. The relational extraction process can be shown as the following formula (4).

[0074]

[0075] In the above formula (4), σ is the activation function (such as sigmoid or softmax), W r represents the weight of relation r, b r represents the bias of relation r, is the feature vector of entity e1, is the feature vector of entity e2.

[0076] In an optional embodiment, in step S235, constructing an initial power knowledge graph according to the sixth data set and the seventh data set includes:

[0077] Step S2351, using entities in the sixth data set as nodes of the basic framework, and using association information between any two entities in the seventh data set as association information between nodes corresponding to any two entities, to construct a basic framework;

[0078] Step S2352, according to a preset storage method, the node information in the basic framework is stored in the nodes of the graph structure, and the association information between any two nodes in the basic framework is stored in the edges of the graph structure, so as to construct a graph database.

[0079] In the present invention, the basic framework of the electric power knowledge graph can be predefined to include nodes and relationships, the entities in the sixth data set can be used as nodes of the basic framework, and the association information between any two entities (or entity pairs) in the seventh data set can be used as the relationship between the corresponding two nodes. It should be noted that in addition to considering the actual operating logic of the power system, the definition of the relationship must also ensure the directionality and uniqueness of the relationship, and the hierarchical structure of nodes and relationships must be considered to facilitate hierarchical management of the electric power knowledge graph and rapid query of nodes and / or relationships from the electric power knowledge graph. For example, equipment nodes can be divided into primary equipment and secondary equipment, and procedure nodes can be divided into operating procedures and emergency procedures.

[0080] After constructing the basic framework of the power knowledge graph, a graph database can be used to store power knowledge in the power system. The graph database relies on a graph structure containing nodes and edges for data storage. Each node and each edge has attribute information and indexes. For example, the attribute information of a substation node includes name, location, capacity, etc., the attribute information of a device node includes name, model, status, etc., and the attribute information of an edge may include the relationship between nodes, weights, etc. In the present invention, each node is designed to have a unique identifier to ensure the uniqueness of each node, thereby improving data query and indexing.

[0081] The above-mentioned preset storage method can be, but is not limited to, a storage method based on an adjacency list or a storage method based on an adjacency matrix. In the present invention, in order to improve the efficiency of data query, an index can be established for high-frequency query attribute information (such as substation name, equipment model, etc.). In addition, an access interface for a graph database can be designed, and a query language (such as Cypher) that supports complex queries (such as path queries, subgraph matching, etc.) can be configured for the graph database to enhance the flexibility of the power knowledge graph.

[0082] In an optional embodiment, in step S204, the initial power knowledge graph is updated based on the incremental learning algorithm to obtain the target power knowledge graph including:

[0083] Step S241, collecting a new power data set of the power system based on the data stream processing architecture;

[0084] Step S242, using a recursive neural network to perform incremental learning on the new power data set to obtain a learning result, wherein the learning result is used to determine an entity and / or associated information involved in any new data in the new power data set;

[0085] Step S243, querying the entity and / or associated information in the graph database to obtain a query result, wherein the query result is used to determine whether the entity is a newly added entity and / or whether the associated information is newly added associated information;

[0086] Step S243, update the initial power knowledge graph according to the query results and learning results to obtain the target power knowledge graph.

[0087] Among them, in step S243, the initial power knowledge graph is updated according to the query results and the learning results, and the target power knowledge graph is obtained, including:

[0088] Step S2431, in response to determining that the entity is a newly added entity and / or the associated information is newly added associated information according to the query result, storing the newly added entity and / or the newly added associated information in the graph database, and adding a node corresponding to the newly added entity and / or an edge corresponding to the newly added associated information in the graph structure according to the learning result;

[0089] Step S2432, in response to determining that the entity is an existing entity and / or the associated information is existing associated information according to the query result, update the nodes corresponding to the existing entity and / or the edges corresponding to the existing associated information in the graph structure according to the learning result.

[0090] In the present invention, the data stream processing architecture based on Apache Kafka can be used to collect new power data in the power system in real time. The data sources of the new power data may include but are not limited to: equipment operation data, equipment status data, event alarm data, etc. Then, the new power data can be processed according to the above-mentioned preprocessing method, and the processed data can be stored in a temporary database (such as Redis). The temporary database can support fast access to data, and the data in the temporary database has been verified and cleaned to ensure high data quality.

[0091] Since the data stream processing architecture based on Apache Kafka is a distributed stream processing platform, it has high throughput, low latency and high scalability, thus meeting the real-time requirements of power data collection.

[0092] Furthermore, a recursive neural network RNN ​​shown in the following formula (5) can be used to perform incremental learning on the processed new power data extracted from the temporary database.

[0093] h t =σ(W h h t-1 +W x x t +b h ) Formula (5)

[0094] In the above formula (5), x t The data input to the recurrent neural network for the current time step t, h t refers to the hidden state at the current time step t, h t-1 refers to the hidden state of the previous time step t-1, W h Refers to the current time step hidden layer weight matrix composed of the values ​​of the previous time step hidden layer, W x refers to the weight matrix from the input layer to the hidden layer, b h represents the bias term of the hidden state, and σ is the activation function (such as tanh, ReLU).

[0095] By using a recursive neural network to perform incremental learning on new power data, it is possible to learn the entities and / or relationships involved in each new power data. Then, for each new power data, the incrementally learned entities and / or relationships are searched in the graph database. When the corresponding entities and / or relationships are not found in the graph database, the incrementally learned entities and / or relationships are determined to be new entities and / or new relationships, and the new entities and / or new relationships are stored in the graph data. According to the new entities and / or new relationships, the graph structure corresponding to the original power knowledge graph is updated. Specifically, nodes corresponding to the new entities are added and / or edges corresponding to the new relationships are added to the graph structure corresponding to the original power knowledge graph.

[0096] Alternatively, when the corresponding entities and / or relationships are found in the graph database, the attribute information of the nodes and / or edges in the corresponding graph structure of the original power knowledge graph is updated according to the incremental learning results. For example, when the new power data involves the voltage change of a substation, the voltage value of the substation node is updated.

[0097] In addition, when new entities and / or new relationships are added to the graph structure, it is possible to detect whether the new entities and / or relationships have data conflicts (such as duplicate nodes, relationship conflicts) with the entities and / or relationships recorded in the original power knowledge graph. When data conflicts occur, corresponding conflict resolution strategies (such as: giving priority to retaining the latest data, merging information of duplicate nodes) can be used to process the data to ensure the accuracy and stability of the power knowledge graph.

[0098] In an optional embodiment, the above-mentioned method for constructing and updating the electric power knowledge graph further includes:

[0099] Step S251, using a web page ranking algorithm to evaluate the importance of any node in the graph structure to obtain a first evaluation result, wherein the first evaluation result is used to determine the importance of any node in the graph structure;

[0100] Step S252, clustering the nodes and edges in the graph structure using a community discovery algorithm to obtain a clustering result, wherein the clustering result is used to determine the nodes and edges in the graph structure whose correlation degree is higher than a preset value;

[0101] Step S253, using a graph neural network to analyze and decide on the first evaluation result and the clustering result to obtain a processing result, wherein the processing result is used to determine a predicted fault event of the power system;

[0102] Step S254, using the Bayesian network to perform probability evaluation on the predicted fault event to obtain a second evaluation result, wherein the second evaluation result is used to determine the probability of the predicted fault event occurring.

[0103] After constructing the power knowledge graph of the power system, deep knowledge mining can be performed on the power knowledge graph, and fault prediction and diagnosis can be performed, so that accident plans can be quickly formulated to reduce losses when power accidents occur. Specifically, the importance of each node in the graph structure can be evaluated by using the web page ranking algorithm (PageRank algorithm) shown in the following formula (6).

[0104]

[0105] In the above formula (6), PR(v i ) represents node v i PageRank value, d is the damping coefficient (usually 0.85), N is the total number of nodes, M(v i ) are all nodes v in the graph structure. i The node set, L(v j ) represents node v j The number of outgoing links.

[0106] Next, the community discovery algorithm Louvain shown in the following formula (7) can be used to identify the community structure in the graph structure, and cluster the nodes and edges involved in each community structure to discover closely related equipment, operating status, etc. in the power system.

[0107]

[0108] In the above formula (7), Q is the modularity, A is ij represents the edge weight between node i and node j, k i is the degree of node i, k j is the degree of node j, m is the total number of edges in the graph structure, δ is the Kronecker delta function, C i is the community structure to which node i belongs, C j is the community structure to which node j belongs.

[0109] Furthermore, the graph neural network (GNN) shown in the following formula (8) can be used to perform intelligent analysis and decision support on the first evaluation result of the web page ranking algorithm and the clustering result of the community discovery algorithm to predict possible fault events in the power system.

[0110]

[0111] In the above formula (8), represents the hidden state of node v at the k+1th time, φ(v) is the set of neighbor nodes of node v, W is the weight matrix representing the connection strength between neurons in the graph neural network, b is the bias term of the graph neural network, and σ is the activation function of the graph neural network (such as the Sigmoid function).

[0112] Next, the Bayesian network shown in the following formula (9) can be used to perform probability evaluation on the predicted fault events to obtain the probability of different predicted fault events occurring in the power system, thereby facilitating the preparation of accident plans and prevention of fault events based on the probability evaluation results.

[0113]

[0114] In the above formula (9), P(A|B) represents the probability of event A occurring under the condition that event B occurs, P(B|A) represents the probability of event B occurring under the condition that event A occurs, P(A) represents the prior probability of event A, and P(B) represents the prior probability of event B.

[0115] Through in-depth exploration of the above-mentioned power knowledge, it is possible to accurately predict possible fault events in the power system and accurately assess the probability of each fault event, thereby providing accurate guidance for fault management of the power system.

[0116] The method for constructing and updating the electric power knowledge graph provided by the above embodiment of the present invention achieves the following technical effects:

[0117] (1) Through comprehensive data collection and preprocessing, the integrity and consistency of various types of data in the power system are ensured, data standardization and standardized management are achieved, and data quality is improved. By using natural language processing technology and graph database technology, a power knowledge graph covering various types of power data in the power system is constructed, breaking down data silos, establishing the relationship between large-scale power data, and realizing the integration and sharing of power information.

[0118] (2) Through real-time data collection and stream processing technology, the dynamic changes of power data in the power system can be captured in a timely manner, and the power knowledge graph can be ensured to always reflect the latest power system status; the incremental update algorithm is adopted to make the update of the power knowledge graph more efficient, avoiding the high cost of full reconstruction; that is, the real-time update mechanism provided by the present invention not only ensures the timeliness of the power knowledge graph, but also improves the response capability of the power system to emergencies;

[0119] (3) The introduction of graph neural network and other technologies has achieved deep mining and intelligent analysis of the power knowledge graph; fault prediction and diagnosis based on the power knowledge graph has improved the fault prediction capability and emergency handling efficiency of the power system; the application of community discovery algorithm and web page ranking algorithm has made the identification of key equipment and important operating procedures more accurate, enhancing the decision support capability of the power system;

[0120] (4) The knowledge graph construction method provided by the present invention has good scalability and adaptability, can be applied to the fields of data management and knowledge mining of other complex systems, and can bring significant economic benefits.

[0121] In this embodiment, a device for constructing and updating an electric power knowledge graph is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and will not be repeated hereafter. As used below, a "module" is a combination of software and / or hardware that can implement a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0122] Figure 3 is a structural block diagram of a device for constructing and updating an electric power knowledge graph according to an embodiment of the present invention, such as Figure 3 As shown, the device comprises:

[0123] The collection module 301 is used to collect the original power data set in the power system, wherein the original power data set includes multiple categories of power data, and the sources of different categories of power data are different;

[0124] A preprocessing module 302 is used to preprocess the original power data set to obtain a target power data set;

[0125] A construction module 303 is used to construct an initial power knowledge graph according to a target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, the basic framework is used to characterize the architecture of power knowledge contained in the power system, and the graph database is used to store and query power knowledge based on a graph structure;

[0126] The updating module 304 is used to update the initial power knowledge graph based on the incremental learning algorithm to obtain the target power knowledge graph.

[0127] Optionally, the above-mentioned preprocessing module 302 is also used to: supplement the missing data in the original power data set based on the interpolation method to obtain a first data set; identify and remove abnormal data in the first data set based on the box plot to obtain a second data set; standardize the data of the same type in the second data set to obtain a third data set; convert the unstructured data in the third data set into structured data to obtain a target power data set.

[0128] Optionally, the above-mentioned construction module 303 is also used to: use natural language processing technology to extract data from the target power data set to obtain a fourth data set, wherein the fourth data set includes text data in the target power data set; perform word segmentation and part-of-speech tagging on the fourth data set to obtain a fifth data set; use a deep learning model to perform entity recognition on the fifth data set to obtain a sixth data set, wherein the sixth data set includes multiple categories of entities and attribute information of any entity; use a relationship extraction model to identify the relationship between any two entities in the sixth data set to obtain a seventh data set, wherein the seventh data set includes association information between any two entities; and construct an initial power knowledge graph based on the sixth data set and the seventh data set.

[0129] Optionally, the above-mentioned construction module 303 is also used to: use a word segmentation processing tool to perform word segmentation processing on any text data in the fourth data set to obtain an eighth data set, wherein the eighth data set includes a word sequence corresponding to any text data; use a part-of-speech tagging tool to perform part-of-speech tagging on any word sequence in the eighth data set to obtain a fifth data set.

[0130] Optionally, the above-mentioned construction module 303 is also used to: use the entities in the sixth data set as the nodes of the basic framework, and use the association information between any two entities in the seventh data set as the association information between the corresponding nodes of any two entities to construct a basic framework; store the node information in the basic framework in the nodes of the graph structure, and store the association information between any two nodes in the basic framework in the edges of the graph structure according to a preset storage method, to construct a graph database.

[0131] Optionally, the above-mentioned update module 304 is also used to: collect a new power data set of the power system based on the data stream processing architecture; use a recursive neural network to perform incremental learning on the new power data set to obtain a learning result, wherein the learning result is used to determine the entities and / or associated information involved in any new data in the new power data set; query the entity and / or associated information in the graph database to obtain a query result, wherein the query result is used to determine whether the entity is a newly added entity and / or whether the associated information is newly added associated information; update the initial power knowledge graph according to the query result and the learning result to obtain a target power knowledge graph.

[0132] Optionally, the above-mentioned update module 304 is also used to: in response to determining that the entity is a newly added entity and / or the associated information is newly added associated information according to the query results, store the newly added entity and / or the newly added associated information in the graph database, and add nodes corresponding to the newly added entity and / or edges corresponding to the newly added associated information in the graph structure according to the learning results; in response to determining that the entity is an existing entity and / or the associated information is existing associated information according to the query results, update the nodes corresponding to the existing entities and / or the edges corresponding to the existing associated information in the graph structure according to the learning results.

[0133] Optionally, the above-mentioned power knowledge graph construction and update device also includes: a prediction module 305, which is used to use a web page ranking algorithm to evaluate the importance of any node in the graph structure to obtain a first evaluation result, wherein the first evaluation result is used to determine the importance of any node in the graph structure; clustering the nodes and edges in the graph structure using a community discovery algorithm to obtain a clustering result, wherein the clustering result is used to determine the nodes and edges in the graph structure whose correlation degree is higher than a preset value; using a graph neural network to analyze and decide on the first evaluation result and the clustering result to obtain a processing result, wherein the processing result is used to determine the predicted fault event of the power system; using a Bayesian network to perform a probability assessment on the predicted fault event to obtain a second evaluation result, wherein the second evaluation result is used to determine the probability of the predicted fault event.

[0134] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0135] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, which, when executed by a processor, implements any one of the aforementioned methods for constructing and updating an electric power knowledge graph.

[0136] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the aforementioned methods for constructing and updating an electric power knowledge graph.

[0137] Optionally, in this embodiment, the computer-readable storage medium may be configured to store a computer program for performing the following steps:

[0138] Step S1, collecting an original power data set in the power system, wherein the original power data set includes multiple categories of power data, and different categories of power data have different sources;

[0139] Step S2, preprocessing the original power data set to obtain a target power data set;

[0140] Step S3, constructing an initial power knowledge graph according to the target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, the basic framework is used to characterize the architecture of power knowledge contained in the power system, and the graph database is used to store and query power knowledge based on the graph structure;

[0141] Step S4: update the initial power knowledge graph based on the incremental learning algorithm to obtain the target power knowledge graph.

[0142] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store a computer program for performing the following steps: supplementing the missing data in the original power data set based on the interpolation method to obtain a first data set; identifying and removing abnormal data in the first data set based on the box plot to obtain a second data set; standardizing the same type of data in the second data set to obtain a third data set; converting the unstructured data in the third data set into structured data to obtain a target power data set.

[0143] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store a computer program for performing the following steps: using natural language processing technology to extract data from the target power data set to obtain a fourth data set, wherein the fourth data set includes text data in the target power data set; performing word segmentation and part-of-speech tagging on the fourth data set to obtain a fifth data set; using a deep learning model to perform entity recognition on the fifth data set to obtain a sixth data set, wherein the sixth data set includes multiple categories of entities and attribute information of any entity; using a relationship extraction model to identify the relationship between any two entities in the sixth data set to obtain a seventh data set, wherein the seventh data set includes association information between any two entities; and constructing an initial power knowledge graph based on the sixth data set and the seventh data set.

[0144] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store a computer program for executing the following steps: using a word segmentation processing tool to perform word segmentation processing on any text data in the fourth data set to obtain an eighth data set, wherein the eighth data set includes a word sequence corresponding to any text data; using a part-of-speech tagging tool to perform part-of-speech tagging on any word sequence in the eighth data set to obtain a fifth data set.

[0145] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store a computer program for performing the following steps: using the entities in the sixth data set as the nodes of the basic framework, and using the association information between any two entities in the seventh data set as the association information between the corresponding nodes of any two entities, to construct a basic framework; according to a preset storage method, storing the node information in the basic framework in the nodes of the graph structure, and storing the association information between any two nodes in the basic framework in the edges of the graph structure, to construct a graph database.

[0146] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store a computer program for performing the following steps: collecting a new power data set of the power system based on a data stream processing architecture; using a recursive neural network to perform incremental learning on the new power data set to obtain a learning result, wherein the learning result is used to determine the entities and / or associated information involved in any new data in the new power data set; querying the entity and / or associated information in the graph database to obtain a query result, wherein the query result is used to determine whether the entity is a newly added entity and / or whether the associated information is newly added associated information; updating the initial power knowledge graph according to the query result and the learning result to obtain a target power knowledge graph.

[0147] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store a computer program for performing the following steps: in response to determining that the entity is a newly added entity and / or the associated information is newly added associated information based on the query results, storing the newly added entity and / or the newly added associated information in a graph database, and adding nodes corresponding to the newly added entity and / or edges corresponding to the newly added associated information in the graph structure based on the learning results; in response to determining that the entity is an existing entity and / or the associated information is existing associated information based on the query results, updating the nodes corresponding to the existing entities and / or the edges corresponding to the existing associated information in the graph structure based on the learning results.

[0148] Optionally, in this embodiment, the above-mentioned computer-readable storage medium can be configured to store a computer program for performing the following steps: using a web page ranking algorithm to perform an importance evaluation on any node in the graph structure to obtain a first evaluation result, wherein the first evaluation result is used to determine the importance of any node in the graph structure; using a community discovery algorithm to cluster the nodes and edges in the graph structure to obtain a clustering result, wherein the clustering result is used to determine the nodes and edges in the graph structure whose correlation degree is higher than a preset value; using a graph neural network to analyze and decide on the first evaluation result and the clustering result to obtain a processing result, wherein the processing result is used to determine a predicted fault event of the power system; using a Bayesian network to perform a probability evaluation on the predicted fault event to obtain a second evaluation result, wherein the second evaluation result is used to determine the probability of the predicted fault event occurring.

[0149] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0150] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiment and its optional implementation manners, and this embodiment will not be described in detail here.

[0151] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0152] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0153] In the several embodiments provided by the present invention, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0154] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0155] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0156] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk or optical disk, etc., various media that can store program codes.

[0157] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for constructing and updating an electric power knowledge graph, characterized in that: include: Collecting an original power data set in the power system, wherein the original power data set includes multiple categories of power data, and different categories of power data have different sources; Preprocessing the original power data set to obtain a target power data set; Constructing an initial power knowledge graph according to the target power data set, wherein the initial power knowledge graph includes a basic framework and a graph database, the basic framework is used to characterize the architecture of the power knowledge contained in the power system, and the graph database is used to store and query the power knowledge based on a graph structure; The initial power knowledge graph is updated based on an incremental learning algorithm to obtain a target power knowledge graph.

2. The method for constructing and updating the electric power knowledge graph according to claim 1, characterized in that: Preprocessing the original power data set to obtain the target power data set includes: Supplementing the missing data in the original power data set based on an interpolation method to obtain a first data set; Identify and remove abnormal data in the first data set based on the box plot to obtain a second data set; performing standardization processing on the data of the same type in the second data set to obtain a third data set; The unstructured data in the third data set is converted into structured data to obtain the target power data set.

3. The method for constructing and updating the electric power knowledge graph according to claim 1, characterized in that: Constructing the initial power knowledge graph according to the target power data set includes: Using natural language processing technology to extract data from the target power data set to obtain a fourth data set, wherein the fourth data set includes text data in the target power data set; Performing word segmentation and part-of-speech tagging on the fourth data set to obtain a fifth data set; Performing entity recognition on the fifth data set using a deep learning model to obtain a sixth data set, wherein the sixth data set includes entities of multiple categories and attribute information of any entity; Using a relationship extraction model to identify a relationship between any two entities in the sixth data set to obtain a seventh data set, wherein the seventh data set includes association information between the any two entities; The initial power knowledge graph is constructed according to the sixth data set and the seventh data set.

4. The method for constructing and updating the electric power knowledge graph according to claim 3 is characterized in that: The fourth data set is subjected to word segmentation and part-of-speech tagging to obtain the fifth data set including: Using a word segmentation processing tool to perform word segmentation processing on any one of the text data in the fourth data set to obtain an eighth data set, wherein the eighth data set includes a word sequence corresponding to the any one of the text data; Part-of-speech tagging tools are used to perform part-of-speech tagging on any word sequence in the eighth data set to obtain the fifth data set.

5. The method for constructing and updating the electric power knowledge graph according to claim 3 is characterized in that: Constructing the initial power knowledge graph according to the sixth data set and the seventh data set includes: The basic framework is constructed by using entities in the sixth data set as nodes of the basic framework and using association information between any two entities in the seventh data set as association information between nodes corresponding to the any two entities; According to a preset storage method, the node information in the basic framework is stored in the nodes of the graph structure, and the association information between any two nodes in the basic framework is stored in the edges of the graph structure, so as to construct the graph database.

6. The method for constructing and updating the electric power knowledge graph according to claim 1, characterized in that: The initial power knowledge graph is updated based on the incremental learning algorithm to obtain the target power knowledge graph, including: Collecting new power data sets of the power system based on a data stream processing architecture; Performing incremental learning on the new power data set using a recursive neural network to obtain a learning result, wherein the learning result is used to determine an entity and / or associated information involved in any new data in the new power data set; Querying the entity and / or the associated information in a graph database to obtain a query result, wherein the query result is used to determine whether the entity is a newly added entity and / or whether the associated information is newly added associated information; The initial power knowledge graph is updated according to the query result and the learning result to obtain the target power knowledge graph.

7. The method for constructing and updating the electric power knowledge graph according to claim 6, characterized in that: The initial power knowledge graph is updated according to the query result and the learning result to obtain the target power knowledge graph, including: In response to determining that the entity is a newly added entity and / or the associated information is newly added associated information according to the query result, storing the newly added entity and / or the newly added associated information in the graph database, and adding a node corresponding to the newly added entity and / or an edge corresponding to the newly added associated information in the graph structure according to the learning result; In response to determining that the entity is an existing entity and / or the associated information is existing associated information according to the query result, the nodes corresponding to the existing entity and / or the edges corresponding to the existing associated information in the graph structure are updated according to the learning result.

8. The method for constructing and updating the electric power knowledge graph according to claim 1, characterized in that: The method for constructing and updating the electric power knowledge graph further includes: Using a web page ranking algorithm to evaluate the importance of any node in the graph structure to obtain a first evaluation result, wherein the first evaluation result is used to determine the importance of the any node in the graph structure; Using a community discovery algorithm to cluster the nodes and edges in the graph structure to obtain a clustering result, wherein the clustering result is used to determine the nodes and edges in the graph structure whose correlation degree is higher than a preset value; Using a graph neural network to analyze and decide on the first evaluation result and the clustering result to obtain a processing result, wherein the processing result is used to determine a predicted fault event of the power system; A Bayesian network is used to perform a probability assessment on the predicted fault event to obtain a second assessment result, wherein the second assessment result is used to determine the probability of the predicted fault event occurring.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, it implements the method for constructing and updating the power knowledge graph of any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the method for constructing and updating the power knowledge graph as described in any one of claims 1 to 8.

Images