A power grid dispatching knowledge extraction method and system

Abstract

The application discloses a power grid dispatching knowledge extraction method and system, and comprises the following steps: acquiring a first parameter of a target power dispatching system, and performing first preprocessing on the first parameter to obtain a second parameter; establishing a first entity recognition model according to the second parameter; establishing a second joint model according to the first entity recognition model and in combination with preset integration rules; and performing target power grid dispatching knowledge extraction according to the second joint model. Through the construction of the efficient and intelligent power grid dispatching knowledge extraction method, the intelligent level of the power grid dispatching system is improved, and the safety and stability of power grid operation are ensured.

Description

Technical Field

[0001] This invention relates to the field of power grid dispatching knowledge extraction technology, and in particular to a method and system for extracting power grid dispatching knowledge. Background Technology

[0002] Power grid dispatching is a crucial aspect of power system operation, with its core task being to ensure the safe, stable, and efficient operation of the power system. During power grid dispatching, dispatchers rely on extensive professional knowledge, including the status of power grid equipment, fault handling procedures, and dispatching operation processes. However, traditional power grid dispatching knowledge is largely stored in document form, resulting in low efficiency in knowledge acquisition and updating, making it difficult to meet the real-time and accuracy requirements of modern smart grids.

[0003] Existing technologies such as forward maximum matching and fuzzy matching may involve significant computation, especially when processing long texts and large-scale data, potentially leading to slow processing speeds and impacting overall system performance. Furthermore, fixed multi-label dictionaries and pre-defined algorithms may lack sufficient adaptability and flexibility when facing varied natural language expressions and complex contexts, making it difficult to handle the diversity and complexity of language use. Some existing technologies integrating multiple high-level models and complex training processes may require substantial computational resources, particularly when processing large amounts of data, potentially resulting in slow processing speeds and high resource consumption. Moreover, in the multi-level processing flow from dialogue state tracking to keyword extraction and summary generation, errors in previous steps may be amplified and propagated to subsequent steps, affecting the quality of the final summary.

[0004] In existing technologies, power grid dispatching knowledge is mostly in document form, resulting in low efficiency in knowledge acquisition and updating, which makes it difficult to meet the real-time and accuracy requirements of modern smart grids. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] Therefore, the present invention provides a method and system for extracting power grid dispatching knowledge, which can solve the problems mentioned in the background art.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0009] In a first aspect, the present invention provides a method for extracting power grid dispatching knowledge, comprising:

[0010] Obtain the first parameter of the target power dispatching system, and perform a first preprocessing on the first parameter to obtain the second parameter;

[0011] A first entity recognition model is established based on the second parameter;

[0012] Based on the first entity recognition model and combined with preset integration rules, a second joint model is established;

[0013] Target power grid scheduling knowledge is extracted based on the second joint model.

[0014] As a preferred embodiment of the power grid dispatching knowledge extraction method described in this invention, the first entity recognition model includes:

[0015] The first entity recognition model is any type of model whose input is the second parameter and whose output is a key entity or key entity-related data;

[0016] The first entity recognition model is used at least to capture the context information of the second parameter and the relationship between the labels in the second parameter.

[0017] As a preferred embodiment of the power grid dispatching knowledge extraction method described in this invention, the second joint model includes an entity recognition task and a relation extraction task.

[0018] The entity recognition task is performed using the first entity recognition model;

[0019] The relation extraction task is any operation performed on the result of the recognition of the first entity recognition model to obtain relations.

[0020] As a preferred embodiment of the power grid dispatching knowledge extraction method of the present invention, the preset integration rules include:

[0021] Construct a shared BERT encoding layer, as well as task-specific layers for entity recognition and relation extraction tasks, respectively;

[0022] This approach integrates entity recognition and relation extraction tasks into a unified framework. By sharing BERT's encoding layer, it leverages feedback information from the entity recognition task to improve the performance of the relation extraction task.

[0023] In a preferred embodiment of the power grid dispatching knowledge extraction method described in this invention, the second parameter is a concatenation of word vectors and character embeddings.

[0024] As a preferred embodiment of the power grid dispatching knowledge extraction method described in this invention, the first entity recognition model further includes:

[0025] We choose a bidirectional long short-term memory network and a conditional random field to jointly construct the first entity recognition model;

[0026] In the first entity recognition model architecture, a bidirectional long short-term memory network is used to capture the contextual information of the second parameter, and a conditional random field is used to capture the relationship between the labels in the second parameter.

[0027] As a preferred embodiment of the power grid dispatching knowledge extraction method described in this invention, the second joint model includes adding a self-attention layer on top of the BERT encoding layer.

[0028] Secondly, the present invention provides a power grid dispatching knowledge extraction system, comprising:

[0029] The data acquisition and processing module is used to acquire the first parameters of the target power dispatching system and perform a first preprocessing on the first parameters to obtain the second parameters;

[0030] The first model building module is used to build a first entity recognition model based on the second parameter;

[0031] The second model building module is used to build a second joint model based on the first entity recognition model and a preset integration rule.

[0032] The knowledge extraction module is used to extract target power grid scheduling knowledge based on the second joint model.

[0033] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0034] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0035] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention proposes a method and system for extracting power grid dispatch knowledge. It obtains first parameters of a target power dispatch system, performs a first preprocessing on these parameters to obtain second parameters, establishes a first entity recognition model based on the second parameters, establishes a second joint model based on the first entity recognition model and pre-defined integration rules, and extracts target power grid dispatch knowledge based on the second joint model. By constructing a joint model architecture including a BERT encoding layer, a self-attention layer, and a relation classifier, and employing a joint training method of NER and RE, the accuracy of named entity recognition and relation extraction is significantly improved. The adoption of the iLSTM-CRF entity recognition model enables the model to effectively capture contextual information and label dependencies, improving the accuracy and robustness of named entity recognition. The introduction of the self-attention mechanism allows the model to focus on important characters and words, enhancing its ability to capture fine-grained features, thereby improving the overall model performance. Through the joint training method of NER and RE, the model can simultaneously learn entity and relation information, improving the model's generalization ability and training efficiency. By constructing an intelligent power grid dispatching knowledge base, precise decision support can be provided to power grid dispatchers, improving the accuracy and efficiency of power grid dispatching operations. Furthermore, by developing efficient and intelligent power grid dispatching knowledge extraction methods, the intelligence level of the power grid dispatching system can be enhanced, ensuring the safety and stability of power grid operation. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0037] Figure 1 A flowchart of a power grid dispatching knowledge extraction method and system provided in one embodiment of the present invention;

[0038] Figure 2 This is an internal structure diagram of a computer device for a power grid dispatching knowledge extraction method and system provided in one embodiment of the present invention. Detailed Implementation

[0039] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0040] Example 1

[0041] Reference Figures 1-2 This is the first embodiment of the present invention, which provides a method and system for extracting power grid dispatch knowledge, including:

[0042] Existing technologies have several limitations, such as low accuracy in entity recognition and relation extraction, insufficient model generalization ability, and a lack of effective power grid dispatch knowledge base support.

[0043] This application provides a method that can effectively solve the problems mentioned above. The following will describe in detail how to implement the power grid dispatch knowledge extraction method with reference to several embodiments.

[0044] Figure 1 A flowchart illustrating a method and system for extracting knowledge for power grid dispatching is shown, including:

[0045] S101, Obtain the first parameter of the target power dispatching system, and perform the first preprocessing on the first parameter to obtain the second parameter;

[0046] In an optional embodiment, the first parameter of the target power dispatching system may include text data such as daily operation records, dispatching logs, and accident reports from the power dispatching system;

[0047] In an optional embodiment, the first parameter of the target power dispatching system may also include relevant technical documents, operation manuals, rules and regulations, etc., in the power company's internal database;

[0048] In an optional embodiment, the first parameter of the target power dispatch system may also include publicly available data such as publicly available power industry reports, research papers, and standards and specifications.

[0049] In this embodiment of the application, in order to design an intelligent power grid dispatch knowledge base, the required data needs to be very comprehensive. Therefore, all the relevant data mentioned above are selected for processing. This application sacrifices time in exchange for high accuracy.

[0050] In an optional embodiment, the first preprocessing may include steps such as text cleaning, word segmentation, part-of-speech tagging, and entity tagging. Text cleaning is used to remove irrelevant characters and noisy data, word segmentation divides the text into individual words, part-of-speech tagging assigns a corresponding part-of-speech label to each word, and entity tagging identifies specific entities in the text, such as device names, time, and location.

[0051] In an optional embodiment, the first preprocessing is to improve the data quality and ensure the accuracy and efficiency of subsequent model training. The preprocessed data will provide more accurate input for the entity recognition model, thereby improving the performance of the entire system. During the preprocessing, natural language processing techniques can be adopted, such as using specific algorithms to identify and process proper nouns, terms, and abbreviations in the text, ensuring that these key information will not be lost or misinterpreted in subsequent processing. In addition, the preprocessing steps can also include removing stop words, correcting spelling mistakes, etc., to further optimize the data quality.

[0052] In the embodiment of the present application, the first preprocessing specifically includes:

[0053] Uniformly format the obtained various data into plain text files and remove the format information in the data.

[0054] Furthermore, use the text processing technology of regular expressions to remove the noise in the text, that is, remove the extra spaces, irrelevant symbols, and specific format characters.

[0055] Furthermore, use the natural language processing (NLP) tool NLTK to tokenize and perform part-of-speech tagging on the text. Tokenization divides the continuous text into independent words, and part-of-speech tagging tags the part of speech of each word.

[0056] Furthermore, remove the common meaningless words such as "of", "is", "in", etc. that are not helpful for knowledge extraction.

[0057] Furthermore, use the NLP tool to perform named entity recognition on the text, label the key entities in the text, such as device names, fault types, operation actions, etc., and then normalize the terms of the labeled entities to ensure the consistency of the same entities in different texts.

[0058] Furthermore, store the processed and labeled data in a structured database.

[0059] It should be noted that the data stored in the structured database is the second parameter.

[0060] In the embodiment of the present application, the second parameter is the concatenation of word vectors and character embeddings.

[0061] It should be noted that obtaining the first parameter of the target power dispatching system and performing a first preprocessing on it to obtain the second parameter provides more accurate and richer information for subsequent entity recognition and relation extraction. This second parameter, as input data, will be used to train and optimize the power grid dispatching knowledge extraction model. In this way, the model can better understand the contextual information of the power dispatching system, thereby achieving more accurate entity recognition and relation extraction in practical applications. Furthermore, by constructing an intelligent power grid dispatching knowledge base, dynamic updates and maintenance of power grid dispatching knowledge can be achieved, ensuring the timeliness and accuracy of the knowledge base content. This not only helps improve the efficiency and security of power grid dispatching but also provides more precise decision support for power grid dispatchers, thereby improving the overall operational efficiency and reliability of the power grid system.

[0062] S102, Establish the first entity recognition model based on the second parameter;

[0063] In this embodiment of the application, the first entity recognition model includes:

[0064] The first entity recognition model is any type of model whose input is the second parameter and whose output is a key entity or key entity-related data.

[0065] The first entity recognition model is used at least to capture the contextual information of the second parameter and the relationship between the labels in the second parameter.

[0066] In an optional embodiment, key entities include equipment name, fault type, and operation action. Text: Transformer T1 failed due to overload at substation S1. Annotations: {Transformer T1: [Equipment]}, {Substation S1: [Location]}, {Overload: [Fault Type]};

[0067] In an optional embodiment, key entity-related data may include the equipment model, the specific time of the fault occurrence, the operator's name, etc. For example, for the text "Transformer T1 failed due to overload at substation S1", the annotation can be expanded to: {Transformer T1: [Equipment, model T1]}, {Substation S1: [Location, specific location S1]}, {Overload: [Fault type, fault occurrence time is specific time]}. Such detailed annotations facilitate subsequent fault analysis and processing, as well as enriching and updating the power grid dispatch knowledge base.

[0068] In an alternative embodiment, the first entity recognition model can be designed using various methods, such as deep learning-based natural language processing techniques like recurrent neural networks (RNNs), long short-term memory networks (LSTMs), or Transformer models. These models are capable of processing sequential data and capturing long-range dependencies, which is crucial for understanding complex power dispatch text. For example, LSTM models can effectively learn time-series information in text through their gating mechanism, while Transformer models utilize self-attention mechanisms to capture the relationships between any two positions in the text, which is particularly effective for entity recognition and relation extraction.

[0069] In this embodiment of the application, the first entity recognition model further includes:

[0070] We choose a bidirectional long short-term memory network and a conditional random field to jointly construct the first entity recognition model;

[0071] In the first entity recognition model architecture, a bidirectional long short-term memory network is used to capture the contextual information of the second parameter, and a conditional random field is used to capture the relationship between the labels in the second parameter.

[0072] Specifically, a Bidirectional Long Short-Term Memory (BiLSTM) network and a Conditional Random Field (CRF) are chosen to jointly construct the Named Entity Recognition (ENR) model. In the BiLSTM-CRF model architecture, BiLSTM is used to capture contextual information, and CRF is used to capture the dependencies between labels.

[0073] In this embodiment of the application, the model training steps of BiLSTM-CRF are as follows:

[0074] 1) Feature extraction and input representation: The pre-trained word vector model GloVe is used to convert the text into word vectors for initialization, and then the training set, validation set and test set are formed; the input representation is the concatenation of word vectors and character embeddings.

[0075] 2) Model initialization

[0076] The embedding layer is initialized with pre-trained word vectors; the parameters of the BiLSTM layer are initialized with a Gaussian distribution; and the parameters of the CRF layer, including the state score and transition score, are randomly initialized.

[0077] 3) Model Training: The BiLSTM and CRF layers are trained using an alternating optimization approach, and the parameters are updated using a loss function. The loss function is defined as:

[0078]

[0079] Where θ is the model parameter, P(y i|x i ;θ) is a given input sequence x i and label sequence y i The conditional probability.

[0080] 4) Model optimization and validation

[0081] The Adam algorithm is used to optimize the model, its performance is evaluated on the validation set, and the model's hyperparameters are adjusted until the best results are achieved.

[0082] It should be noted that establishing the first entity recognition model based on the second parameter can significantly improve the accuracy of entity recognition. The BiLSTM-CRF model can effectively process sequential data in power dispatch text while capturing complex relationships between entities. During model training, alternately optimizing the BiLSTM and CRF layers allows the model to better understand the contextual information of the text and the dependencies between labels during learning. Furthermore, by evaluating model performance on the validation set, the hyperparameters can be adjusted in a timely manner to ensure the model achieves optimal recognition results in practical applications. Ultimately, this optimization and validation process helps to build a more accurate and efficient power grid dispatch knowledge extraction model, providing more accurate decision support for power grid dispatch, thereby improving the overall operational efficiency and reliability of the power grid system.

[0083] S103, Based on the first entity recognition model and combined with the preset integration rules, establish a second joint model;

[0084] In this embodiment of the application, the second joint model includes an entity recognition task and a relation extraction task;

[0085] The entity recognition task uses the first entity recognition model for identification;

[0086] The relation extraction task involves performing arbitrary operations to obtain relations from the results of the first entity recognition model.

[0087] In this embodiment of the application, the preset integration rules include:

[0088] Construct a shared BERT encoding layer, as well as task-specific layers for entity recognition and relation extraction tasks, respectively;

[0089] This approach integrates entity recognition and relation extraction tasks into a unified framework. By sharing BERT's encoding layer, it leverages feedback information from the entity recognition task to improve the performance of the relation extraction task.

[0090] In this embodiment of the application, the second joint model includes adding a self-attention layer on top of the BERT encoding layer.

[0091] In an optional embodiment, the relation extraction task involves any operation on the results of the first entity recognition model to obtain relations. For example, the relation extraction task may include, but is not limited to, the following steps: First, using the first entity recognition model to identify key entities in the text; second, determining the type of relationship between entities based on semantic relationships and contextual information; then, extracting the relationships between entities through predefined relation extraction rules or pattern matching; finally, integrating the extracted relation information into the power grid dispatch knowledge base to enrich its content. For example, for the text "Transformer T1 failed due to overload at substation S1," after entity recognition, the causal relationship between "transformer T1" and "overload" can be further identified, and this relationship information can be stored in the knowledge base. Such relation extraction not only helps to understand the background of the fault but also provides data support for future fault prediction and prevention. In this way, the power grid dispatch knowledge base can continuously update and optimize itself, thereby providing more intelligent and automated decision support for power grid dispatch.

[0092] In an alternative embodiment, the relation extraction task can be implemented by designing a relation extraction model or employing machine learning algorithms. For example, methods such as Support Vector Machines (SVM), Random Forests, or deep learning models can be used to identify relationships between entities. These models can predict the types of relationships that may exist between entities by learning from a large amount of labeled data based on the results of entity identification.

[0093] In an optional embodiment, the relation extraction task can also improve accuracy through ensemble learning. Ensemble learning enhances overall performance by combining the predictions of multiple models. For example, a voting mechanism can be used, where each model votes on its prediction of relationships between entities, and the final decision is based on the consensus of the majority of models. This approach reduces the potential bias of individual models, thereby improving the robustness and accuracy of relation extraction. Furthermore, ensemble learning can further optimize performance by adjusting the weights of different models, ensuring that the relation information stored in the power grid dispatch knowledge base is more reliable and useful. Through these optimization and validation steps, the power grid dispatch knowledge extraction system can more accurately identify and extract key information, providing more intelligent and efficient decision support for power grid dispatch.

[0094] In this embodiment of the application, a BERT multi-task relation extraction learning model is established as the second joint model;

[0095] In the embodiments of this application, the second joint model structure includes a shared BERT encoding layer and task-specific layers for NER and RE, respectively.

[0096] This paper integrates Named Entity Recognition (NER) and Relation Extraction (RE) tasks into a unified framework. By sharing a BERT encoding layer, it leverages feedback from the NER task to improve RE performance. The model architecture includes a shared BERT encoding layer and task-specific layers for NER and RE, respectively. Through this shared architecture, the model can simultaneously learn features for entity recognition and relation extraction, thereby enhancing its overall performance. During training, a self-attention mechanism is introduced to dynamically assign different attention weights to each entity pair, enhancing the model's ability to understand and process contextual information.

[0097] In this embodiment of the application, all possible entity pairs are selected from the identified entities as candidates for relation extraction, and then input features are constructed for each entity pair, including the entity's context information and location encoding.

[0098] 1) Entity recognition

[0099] The BiLSTM-CRF entity recognition model is used to identify all entities, and the identified entities are classified by category, such as device, location, time, etc.

[0100] 2) Selection of intelligent physical pairs

[0101] Entity pair selection is performed using a hybrid strategy of rule-based and machine learning, filtering entity pairs according to predefined rules. Only entities in adjacent sentences are likely to have a relationship. A classification model is then trained using the classic machine learning method of Random Forest to predict whether a relationship exists between entity pairs based on their contextual features.

[0102] 3) Contextual Feature Construction

[0103] For each entity pair, a contextual feature is constructed, including the entity's contextual information, location encoding, and category information. For example, Entity 1: Power Plant A; Entity 2: Transformer B; Contextual information: Power Plant A is connected to Transformer B. Location encoding: The location of Power Plant A and the location of Transformer B.

[0104] In this embodiment, an input sequence is constructed by concatenating entity pairs and their contexts. For example: "[CLS] Power Station A [SEP] Power Station A is connected to Transformer B [SEP] Transformer B [SEP]". Then, special markers [CLS] and [SEP] are added to the input sequence for processing by the BERT model.

[0105] Furthermore, the input sequence is fed into the BERT model to generate a contextual representation h. ij

[0106]

[0107] Among them, h i and h j Let i and j represent the context representations of entity i and entity j, respectively. The symbol represents element-wise multiplication, the hyphen (-) represents element-wise subtraction, and [·; ·] represents vector concatenation.

[0108] Furthermore, a self-attention layer is added on top of the BERT encoding layer to further process the context vector h. ij This improves the accuracy of relation extraction. The combination representation h... ij The input is taken from the attention layer to generate a weighted context representation. The attention weights are represented as follows:

[0109]

[0110] Among them, e ij Let represent the attention score for entity pairs i and j, and n represent the total number of entity pairs. Weighted context. This is represented as:

[0111]

[0112] Furthermore, this application constructs a Softmax classifier as the implementation model for the relation extraction task, classifies the relations on the weighted context representation, and then assigns a relation label to each entity pair to indicate the type of relationship between the entity pairs.

[0113] In this embodiment of the application, a training set containing entity labels and relation labels is constructed, and then the model is trained on a constructed model architecture containing a BERT encoding layer, a self-attention layer and a relation classifier.

[0114] In an optional embodiment, assume the training dataset contains N samples, each containing T words. The true label for each word t is y. t The model predicts the label probability distribution as P(y t |x t ), where x t Let represent the input features of the t-th word. Then, the NER loss function is specifically defined as follows:

[0115]

[0116] In an optional embodiment, assume the training dataset contains M entity pairs, each entity pair has a true relation label r, and the model predicts the relation label probability distribution as follows: in Let the contextual representation of entity pair (i,j) be given. Then, the RE loss function is specifically defined as follows:

[0117]

[0118] Therefore, the joint loss function for NER and RE tasks can be set as follows:

[0119]

[0120] in, This represents the loss function of the NER task. The loss function λ for the RE task represents the weight parameters.

[0121] It should be noted that the model training process includes joint training of two tasks: NER and RE. Through joint training, the model can simultaneously learn features for entity recognition and relation extraction, thereby enhancing the overall performance of the model.

[0122] S104, Extract target power grid scheduling knowledge based on the second joint model.

[0123] In an optional embodiment, new information from various sources is collected, including new scheduling procedures, fault handling cases, and equipment maintenance records. The collected information is then validated to ensure its accuracy and reliability. Using a pre-built named entity recognition and relation extraction model, key knowledge points are extracted from the new information, updating the entities and relations in the knowledge base. The new knowledge is then integrated with the existing knowledge base to eliminate redundancy and conflicts, ensuring the consistency and integrity of the knowledge base. Evaluation metrics for knowledge updates are set, such as accuracy, completeness, and usability.

[0124] In an optional embodiment, target power grid dispatching knowledge is extracted based on the second joint model, and an intelligent power grid dispatching knowledge base is constructed. This intelligent knowledge base provides precise decision support for power grid dispatchers, improving the accuracy and efficiency of power grid dispatching operations. By constructing an efficient and intelligent power grid dispatching knowledge extraction method, the intelligence level of the power grid dispatching system is enhanced, ensuring the safety and stability of power grid operation.

[0125] In summary, this invention proposes a method for extracting power grid dispatch knowledge. The method involves obtaining first parameters of the target power dispatch system, performing a first preprocessing step on these parameters to obtain second parameters, establishing a first entity recognition model based on the second parameters, establishing a second joint model based on the first entity recognition model and pre-defined integration rules, and extracting target power grid dispatch knowledge based on the second joint model. By constructing a joint model architecture including a BERT encoding layer, a self-attention layer, and a relation classifier, and employing a joint training method of NER and RE, the accuracy of named entity recognition and relation extraction is significantly improved. The adoption of the iLSTM-CRF entity recognition model enables the model to effectively capture contextual information and label dependencies, improving the accuracy and robustness of named entity recognition. The introduction of the self-attention mechanism allows the model to focus on important characters and words, enhancing its ability to capture fine-grained features and thus improving the overall model performance. Through the joint training method of NER and RE, the model can simultaneously learn entity and relation information, improving the model's generalization ability and training efficiency. By constructing an intelligent power grid dispatch knowledge base, precise decision support can be provided to power grid dispatchers, improving the accuracy and efficiency of power grid dispatch operations. By constructing an efficient and intelligent method for extracting knowledge from power grid dispatch, the intelligence level of the power grid dispatch system can be improved, ensuring the safety and stability of power grid operation.

[0126] This embodiment also provides a power grid dispatching knowledge extraction system, including:

[0127] The data acquisition and processing module is used to acquire the first parameters of the target power dispatching system and perform a first preprocessing on the first parameters to obtain the second parameters;

[0128] The first model building module is used to build a first entity recognition model based on the second parameter;

[0129] The second model building module is used to build a second joint model based on the first entity recognition model and a preset integration rule.

[0130] The knowledge extraction module is used to extract target power grid scheduling knowledge based on the second joint model.

[0131] The above-mentioned unit modules can be embedded in the processor of the computer device in hardware form or independent of it, or they can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of the above modules.

[0132] This embodiment also provides a computer device, which may be a terminal, and its internal structure diagram may be as follows. Figure 2As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a power grid dispatching knowledge extraction method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0133] This embodiment also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, it performs the following steps:

[0134] Obtain the first parameters of the target power dispatching system, and perform a first preprocessing on the first parameters to obtain the second parameters;

[0135] A first entity recognition model is established based on the second parameter;

[0136] Based on the first entity recognition model and combined with the preset integration rules, a second joint model is established;

[0137] Knowledge extraction for target power grid scheduling is performed based on the second joint model.

[0138] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0139] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0140] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0141] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0142] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0143] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0144] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for extracting power grid dispatching knowledge, characterized in that, include: Obtain the first parameter of the target power dispatching system, and perform a first preprocessing on the first parameter to obtain the second parameter; The first parameter includes: Text data of daily operation records, dispatch logs, and accident reports in the power dispatching system; Relevant technical documents, operation manuals, and regulations in the power company's internal database; Publicly available power industry reports, research papers, and standard specifications; A first entity recognition model is established based on the second parameter; Based on the first entity recognition model and combined with preset integration rules, a second joint model is established; Target power grid scheduling knowledge is extracted based on the second joint model; The first entity recognition model includes: The first entity recognition model is any type of model whose input is the second parameter and whose output is a key entity or key entity-related data; The first entity recognition model is used at least to capture the context information of the second parameter and the relationship between the labels in the second parameter; The second joint model includes entity recognition tasks and relation extraction tasks; The entity recognition task is performed using the first entity recognition model; The relation extraction task is any operation that extracts relations from the results of the recognition of the first entity recognition model. The preset integration rules include: Construct a shared BERT encoding layer, as well as task-specific layers for entity recognition and relation extraction tasks, respectively; The entity recognition task and the relation extraction task are integrated into a unified framework. By sharing the encoding layer of BERT, the performance of the relation extraction task is improved by utilizing the feedback information from the entity recognition task. The second parameter is the concatenation of word vectors and character embeddings; The first entity recognition model also includes: We choose a bidirectional long short-term memory network and a conditional random field to jointly construct the first entity recognition model; In the first entity recognition model architecture, a bidirectional long short-term memory network is used to capture the context information of the second parameter, and a conditional random field is used to capture the relationship between the labels in the second parameter; The second joint model includes adding a self-attention layer on top of the BERT encoding layer.

2. A power grid dispatching knowledge extraction system applying the method described in claim 1, characterized in that, include: The data acquisition and processing module is used to acquire the first parameters of the target power dispatching system and perform a first preprocessing on the first parameters to obtain the second parameters; The first model building module is used to build a first entity recognition model based on the second parameter; The second model building module is used to build a second joint model based on the first entity recognition model and a preset integration rule. The knowledge extraction module is used to extract target power grid scheduling knowledge based on the second joint model.

3. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method described in claim 1.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method described in claim 1.

Images