Mountain fire tripping data analysis system based on large language model

Abstract

The embodiment of the invention discloses a forest fire tripping data analysis system based on a large language model. Comprising the steps that a data acquisition module acquires historical forest fire tripping data associated with a forest fire tripping event in a historical time period of an area where a power transmission line is located, and sends the historical forest fire tripping data to a data processing module; the data processing module processes the received historical mountain fire tripping data to obtain standard mountain fire tripping data, and sends the standard mountain fire tripping data to the large language model module; and the big language model module determines an association relationship between the historical forest fire tripping event and a historical forest fire event corresponding to the historical forest fire tripping event for each historical forest fire tripping event in the received standard forest fire tripping data so as to predict the forest fire tripping probability after the forest fire event occurs based on the association relationship. According to the technical scheme provided by the embodiment of the invention, the historical mountain fire tripping data can be processed and analyzed, and information support is provided for operation and maintenance work.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of data processing, and in particular to a wildfire tripping data analysis system based on a large language model. Background Art

[0002] The power grid accumulates a large amount of historical wildfire tripping data every year, which summarizes various information of each wildfire tripping.

[0003] However, these large amounts of historical wildfire tripping data cannot be effectively processed and analyzed, making it difficult to play a role in operation and maintenance work, and an urgent problem needs to be solved. Summary of the invention

[0004] The embodiment of the present invention provides a wildfire tripping data analysis system based on a large language model to process and analyze historical wildfire tripping data and provide information support for operation and maintenance work.

[0005] According to one aspect of the present invention, a mountain fire tripping data analysis system based on a large language model is provided, which may include:

[0006] A data acquisition module, used to acquire historical wildfire tripping data associated with wildfire tripping events within a historical time period in the area where the transmission line is located, and send the historical wildfire tripping data to the data processing module;

[0007] A data processing module is used to process the received historical wildfire tripping data to obtain standard wildfire tripping data, and send the standard wildfire tripping data to the large language model module;

[0008] The large language model module is used to determine the correlation between the historical wildfire tripping event and the historical wildfire event corresponding to the historical wildfire tripping event for each historical wildfire tripping event in the received standard wildfire tripping data, so as to predict the wildfire tripping probability after the wildfire event occurs based on the correlation relationship.

[0009] The technical solution of the embodiment of the present invention includes: a data acquisition module, a data processing module and a large language model module; wherein the data acquisition module is used to acquire historical wildfire tripping data associated with wildfire tripping events within a historical time period of the area where the transmission line is located, and send the historical wildfire tripping data to the data processing module so that the data processing module can acquire comprehensive historical wildfire tripping data; the data processing module is used to process the received historical wildfire tripping data to obtain standard wildfire tripping data, and send the standard wildfire tripping data to the large language model module, and obtain standard wildfire tripping data that meets the input of the large language model module by processing the historical wildfire tripping data, so as to facilitate further processing by the large language model module; the large language model module is used to determine the association relationship between the historical wildfire tripping event and the historical wildfire event corresponding to the historical wildfire tripping event for each historical wildfire tripping event in the received standard wildfire tripping data, so as to predict the wildfire tripping probability after the wildfire event occurs based on the association relationship, and to obtain the association relationship by analyzing the standard wildfire tripping data, so as to achieve accurate prediction of the wildfire tripping probability. The above technical solution, by processing and analyzing historical wildfire tripping data, can predict the probability of wildfire tripping after a wildfire event occurs, and provide information support for operation and maintenance work.

[0010] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 It is a structural block diagram of a mountain fire tripping data analysis system based on a large language model provided according to an embodiment of the present invention;

[0013] Figure 2 It is a structural block diagram of another wildfire tripping data analysis system based on a large language model provided according to an embodiment of the present invention;

[0014] Figure 3 It is a structural block diagram of another wildfire tripping data analysis system based on a large language model provided according to an embodiment of the present invention;

[0015] Figure 4The present invention is another flowchart of wildfire tripping data processing in a wildfire tripping data analysis system based on a large language model provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0016] 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.

[0017] 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 interchangeable 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. The situations of "target", "original", etc. are similar and will not be repeated here. In addition, the terms "including" and "having" and any of their variations 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.

[0018] Figure 1 1 is a structural block diagram of a wildfire tripping data analysis system based on a large language model provided by an embodiment of the present invention. This embodiment is applicable to the case of processing historical wildfire tripping data.

[0019] See also Figure 1 The wildfire tripping data analysis system based on the large language model of the embodiment of the present invention includes: a data acquisition module 11, a data processing module 12 and a large language model module 13; wherein,

[0020] The data acquisition module 11 is used to acquire historical wildfire tripping data associated with wildfire tripping events within a historical time period in the area where the power transmission line is located, and send the historical wildfire tripping data to the data processing module 12;

[0021] The data processing module 12 is used to process the received historical wildfire tripping data to obtain standard wildfire tripping data, and send the standard wildfire tripping data to the large language model module 13;

[0022] The large language model module 13 is used to determine the correlation between the historical wildfire tripping event and the historical wildfire event corresponding to the historical wildfire tripping event for each historical wildfire tripping event in the received standard wildfire tripping data, so as to predict the wildfire tripping probability after the wildfire event occurs based on the correlation relationship.

[0023] Among them, historical wildfire tripping data can be understood as various data associated with wildfire tripping events during the historical time period in the area where the transmission line is located. Optionally, it can include wildfire record data, such as wildfire records that occurred in the area where the transmission line is located, including the time, location, cause of the wildfire, and the scope of influence of the wildfire; meteorological data, such as ambient temperature, humidity, wind speed and wind direction when the wildfire occurred; geographic information system data, such as topography, landforms and vegetation coverage in the area where the transmission line is located; transmission line operation data, such as the load, voltage and historical tripping records of the transmission line. The above data can reflect the specific impact of wildfire events on transmission lines.

[0024] The data acquisition module 11 can be understood as a module for acquiring and transmitting various data in the historical wildfire tripping data. Optionally, the data acquisition module 11 can acquire the historical wildfire tripping data through multiple sensor data sources, such as voltage sensors, temperature sensors, weather stations, and geographic information systems, and transmit the acquired historical wildfire tripping data to the data processing module 12.

[0025] The data processing module 12 can be understood as a module for processing historical wildfire tripping data. Since the historical wildfire tripping data contains unstructured data of various types and data formats, in order to enable it to be fully identified and utilized, it is necessary to process it to obtain standard wildfire tripping data that can be used by the large language model module 13. The optional data processing module 12 can pre-process the historical wildfire tripping data, such as data cleaning, noise filtering, data standardization and data annotation, to obtain standard wildfire tripping data. And send the standard wildfire tripping data to the large language model module 13 for processing.

[0026] The large language model module 13 can be understood as a natural language processing model based on a deep neural network, which is used to learn and analyze standard wildfire tripping data. Optionally, the large language model module 13 can be trained using each historical wildfire tripping event in the standard wildfire tripping data and the historical wildfire events corresponding to the historical wildfire tripping events, so that the large language model module 13 learns the correlation between wildfire events and wildfire tripping events, such as the impact of various wildfire conditions in wildfire events on the wildfire tripping of power transmission lines. In this way, the large language model module 13 can predict the probability of wildfire tripping after the wildfire event occurs based on the correlation.

[0027] The technical solution of the embodiment of the present invention includes: a data acquisition module, a data processing module and a large language model module; wherein the data acquisition module is used to acquire historical wildfire tripping data associated with wildfire tripping events within a historical time period of the area where the transmission line is located, and send the historical wildfire tripping data to the data processing module so that the data processing module can acquire comprehensive historical wildfire tripping data; the data processing module is used to process the received historical wildfire tripping data to obtain standard wildfire tripping data, and send the standard wildfire tripping data to the large language model module, and obtain standard wildfire tripping data that meets the input of the large language model module by processing the historical wildfire tripping data, so as to facilitate further processing by the large language model module; the large language model module is used to determine the association relationship between the historical wildfire tripping event and the historical wildfire event corresponding to the historical wildfire tripping event for each historical wildfire tripping event in the received standard wildfire tripping data, so as to predict the wildfire tripping probability after the wildfire event occurs based on the association relationship, and to obtain the association relationship by analyzing the standard wildfire tripping data, so as to achieve accurate prediction of the wildfire tripping probability. The above technical solution, by processing and analyzing historical wildfire tripping data, can predict the probability of wildfire tripping after a wildfire event occurs, and provide information support for operation and maintenance work.

[0028] An optional technical solution, the data processing module includes a data preprocessing submodule, a data standardization submodule and a data sending submodule; wherein the data preprocessing submodule is used to perform data preprocessing operations on the received historical wildfire tripping data to obtain preprocessed wildfire tripping data, wherein the data preprocessing operation includes at least one of data cleaning, noise filtering and data annotation; the data standardization submodule is used to perform standardization processing on the preprocessed wildfire tripping data to obtain standard wildfire tripping data; the data sending submodule is used to send the standard wildfire tripping data to the large language model module.

[0029] Among them, the data preprocessing submodule can be understood as a module for preprocessing historical wildfire tripping data. Optionally, preprocessing operations can include data cleaning, noise filtering and data annotation. Among them, data cleaning aims to remove missing values, inconsistencies and duplicate data in historical wildfire tripping data. These problems arise from format differences between different data sources or errors in the collection process of historical wildfire tripping data; the noise filtering step is used to eliminate irrelevant or erroneous data in historical wildfire tripping data that may interfere with model training, such as abnormal meteorological data generated under extreme weather conditions; data annotation is to label the key attributes and events in the historical wildfire tripping data so that they can be better recognized and learned by the large language model module in subsequent steps. The preprocessed wildfire tripping data can then be standardized by the data normalization submodule to obtain standard wildfire tripping data. The specific formula is as follows:

[0030]

[0031] Among them, X norm is the standardized data, X is the wildfire tripping data after preprocessing, μ is the mean of the wildfire tripping data after preprocessing, and σ is the standard deviation of the wildfire tripping data after preprocessing.

[0032] Through standardization, it can be ensured that different types of data in the pre-processed wildfire tripping data have a unified scale before entering the large language model, which is conducive to improving the training effect and prediction accuracy of the large language model module. Finally, the standard wildfire tripping data is sent to the large language model module through the data sending submodule.

[0033] The above technical solution can greatly improve the cleanliness of historical wildfire tripping data and the input quality of the large language model module by performing various preprocessing operations on the historical wildfire tripping data.

[0034] Figure 2 It is a structural block diagram of another wildfire trip data analysis system based on a large language model provided by an embodiment of the present invention. This embodiment is optimized based on the above-mentioned technical solutions. In this embodiment, optionally, the wildfire trip data analysis system based on a large language model also includes: a database construction module, which is used to construct a wildfire trip database based on the received standard wildfire trip data, so as to manage the historical wildfire trip data based on the wildfire trip database.

[0035] The explanations of the terms that are the same as or corresponding to the above embodiments are not repeated here.

[0036] For details, see Figure 2The mountain fire tripping data analysis system based on the large language model of this embodiment includes: a data acquisition module 21, a data processing module 22, a large language model module 23 and a database construction module 24; wherein,

[0037] The data acquisition module 21 is used to acquire historical wildfire tripping data associated with wildfire tripping events within a historical time period in the area where the power transmission line is located, and send the historical wildfire tripping data to the data processing module;

[0038] The data processing module 22 is used to process the received historical wildfire tripping data to obtain standard wildfire tripping data, and send the standard wildfire tripping data to the large language model module 23;

[0039] The large language model module 23 is used to determine the correlation between the historical wildfire tripping event and the historical wildfire event corresponding to the historical wildfire tripping event for each historical wildfire tripping event in the received standard wildfire tripping data, so as to predict the wildfire tripping probability after the wildfire event occurs based on the correlation relationship.

[0040] The database construction module 24 is used to construct a wildfire trip database based on the received standard wildfire trip data, so as to manage the historical wildfire trip data based on the wildfire trip database.

[0041] Among them, the database construction module 24 can be understood as a module for summarizing, storing and managing the received standard wildfire tripping data to construct a wildfire tripping database. Optionally, the database construction module 24 can use the standard wildfire tripping data to construct a database with a hierarchical structure, which can contain multiple levels of information, such as standard wildfire tripping data and professional annotations provided by domain experts, etc. The database can be constructed through an optimization algorithm to make the database construction module efficient and accurate. When managing historical wildfire tripping data based on the wildfire tripping database, it can quickly retrieve and output the most relevant data entries based on the received questions.

[0042] The above technical solution can manage the standard wildfire tripping data through the database construction module, making it convenient to view and call the standard wildfire tripping data.

[0043] Figure 3 It is a structural block diagram of another wildfire tripping data analysis system based on a large language model provided by an embodiment of the present invention. This embodiment is optimized based on the above-mentioned technical solutions. In this embodiment, optionally, the wildfire tripping data analysis system based on a large language model also includes: a question and answer generation module, which is used to generate a target answer corresponding to the wildfire tripping operation and maintenance problem based on the association relationship or the wildfire tripping database for the received wildfire tripping operation and maintenance problem.

[0044] The explanations of the terms that are the same as or corresponding to the above embodiments are not repeated here.

[0045] For details, see Figure 3 The wildfire tripping data analysis system based on the large language model of this embodiment includes: a data acquisition module 31, a data processing module 32, a large language model module 33, a database construction module 34 and a question and answer generation module 35; wherein,

[0046] The data acquisition module 31 is used to acquire historical wildfire tripping data associated with wildfire tripping events within a historical time period in the area where the power transmission line is located, and send the historical wildfire tripping data to the data processing module;

[0047] The data processing module 32 is used to process the received historical wildfire tripping data to obtain standard wildfire tripping data, and send the standard wildfire tripping data to the large language model module 33;

[0048] The large language model module 33 is used to determine the correlation between the historical wildfire tripping event and the historical wildfire event corresponding to the historical wildfire tripping event for each historical wildfire tripping event in the received standard wildfire tripping data, so as to predict the wildfire tripping probability after the wildfire event occurs based on the correlation relationship.

[0049] The database construction module 34 is used to construct a wildfire trip database based on the received standard wildfire trip data, so as to manage the historical wildfire trip data based on the wildfire trip database.

[0050] The question and answer generation module 35 is used to generate a target answer corresponding to the mountain fire tripping operation and maintenance problem based on the association relationship or the mountain fire tripping database for the received mountain fire tripping operation and maintenance problem.

[0051] Among them, the wildfire tripping operation and maintenance problem can be understood as the operation and maintenance problem of the transmission line related to the wildfire tripping event. For example, the wildfire tripping data at a certain moment in history can be retrieved and the probability of wildfire tripping can be predicted when a wildfire occurs.

[0052] The question and answer generation module 35 can be understood as issuing the wildfire operation and maintenance problem and generating a target answer corresponding to the wildfire tripping operation and maintenance problem based on the received return answer. The question and answer generation module 35 can generate a target answer corresponding to the wildfire tripping operation and maintenance problem based on the association relationship or the wildfire tripping database for the received wildfire tripping operation and maintenance problem.

[0053] The technical solution of the embodiment of the present invention, through the question and answer generation module, can quickly and accurately generate corresponding target answers according to the received wildfire tripping operation and maintenance questions based on the association relationship or the wildfire tripping database, thereby providing fast and reliable information support for operation and maintenance work.

[0054] An optional technical solution, a question and answer generation module, includes: a question parsing submodule, a question sending submodule and an answer generation submodule; wherein the question parsing submodule is used to parse the received wildfire tripping operation and maintenance problem and determine the problem type of the wildfire tripping operation and maintenance problem; the question sending submodule is used to send the wildfire tripping operation and maintenance problem to the large language model module when the problem type is a prediction type problem, so as to predict the wildfire tripping probability in the large language model module based on the association relationship, or send the wildfire tripping operation and maintenance problem to the database construction module when the problem type of the wildfire tripping operation and maintenance problem is a query type problem, so as to query the target wildfire tripping data corresponding to the wildfire tripping operation and maintenance problem from the wildfire tripping database based on the database construction module; the answer generation submodule is used to generate a target answer corresponding to the wildfire tripping operation and maintenance problem based on the wildfire tripping probability or the target wildfire tripping data.

[0055] Among them, the problem analysis submodule can be understood as a module that performs preliminary analysis on the wildfire tripping operation and maintenance problem, and can determine the problem type of the wildfire tripping operation and maintenance problem, such as a prediction problem or a query problem; the problem sending submodule can be understood as a module that sends the wildfire tripping operation and maintenance problem according to the problem type, and can send the wildfire tripping operation and maintenance problem to the large language model module when the problem type is a prediction problem, so as to predict the wildfire tripping probability based on the association relationship in the large language model module, or send the wildfire tripping operation and maintenance problem to the database construction module when the problem type of the wildfire tripping operation and maintenance problem is a query problem, so as to query the target wildfire tripping data corresponding to the wildfire tripping operation and maintenance problem from the wildfire tripping database based on the database construction module; the answer generation submodule can be understood as a module that generates a target answer based on the received question answer, and can generate a target answer corresponding to the wildfire tripping operation and maintenance problem based on the wildfire tripping probability or the target wildfire tripping data.

[0056] The above technical solution, through different sub-modules in the question-answer generation module, can send the question to different modules according to the type of the wildfire tripping operation and maintenance problem, and generate the target answer based on the returned answer.

[0057] On this basis, an optional answer generation submodule includes: a semantic extraction unit and a target answer generation unit; wherein the semantic extraction unit is used to extract semantic information in the wildfire tripping operation and maintenance problem; the target answer generation unit is used to integrate the semantic information into the wildfire tripping probability or target wildfire tripping data to generate a target answer corresponding to the wildfire tripping operation and maintenance problem.

[0058] Among them, the semantic extraction unit can be understood as a unit that extracts key semantic information from the mountain fire tripping operation and maintenance problem. Optionally, the semantic extraction unit can use the attention mechanism to extract the key semantic information of the mountain fire tripping operation and maintenance problem in the question and answer history to perceive the contextual information in the question and answer history. Then, the target answer generation unit integrates these key semantic information into the answer generation process of the current mountain fire tripping operation and maintenance problem to generate a target answer corresponding to the mountain fire tripping operation and maintenance problem.

[0059] The above technical solution can ensure the relevance of the target answers generated before and after by capturing the key semantic information in the question and answer history, thereby generating more coherent and relevant target answers.

[0060] Alternatively, the question and answer generation module also includes: an answer style adjustment submodule, which is used to identify the expected answer style for wildfire tripping operation and maintenance questions, and adjust the style of the target answer based on the expected answer style, wherein the expected answer style includes a brief answer or a detailed answer.

[0061] Among them, the answer style adjustment submodule can identify the expected answer style for wildfire tripping operation and maintenance issues, determine whether the answer style is a brief answer or a detailed answer, and adjust the target answer accordingly according to the answer style. For example, when the answer style is a brief answer, the answer style adjustment submodule will meet the needs by shortening the length of the target answer and simplifying the language expression; and when the answer style is a detailed answer, the answer style adjustment submodule will add more detailed information and explanation content. The specific formula is as follows:

[0062] R(Q, S) = R(Q) + F(S)

[0063] Among them, R(Q, S) is the answer that meets the user's expected style, R(Q) is the final answer, S is the style adjustment parameter, and F(S) is the style adjustment function.

[0064] The above technical solution can better meet the diverse needs of users by adjusting the answer style, thereby improving user satisfaction and saving user time.

[0065] Another optional technical solution, the large language model module, is also used to: extract data information from the received wildfire tripping operation and maintenance problem, and predict the wildfire tripping probability associated with the wildfire tripping operation and maintenance problem based on the data information and correlation relationship.

[0066] Among them, after learning the association relationship, the large language model module can also extract the data information in the received mountain fire tripping operation and maintenance problem, predict the data information through the association relationship, and obtain the mountain fire tripping probability associated with the fire tripping operation and maintenance problem. The specific formula is as follows:

[0067] A=argmax(P(A|Q))

[0068] Among them, A is the final generated wildfire tripping probability, and P(A|Q) is the probability distribution of the answer based on the wildfire tripping operation and maintenance problem Q.

[0069] The above technical solution predicts the probability of wildfire tripping through the association relationship learned by the large language model module, which can provide information support for operation and maintenance work.

[0070] Another optional technical solution also includes: an interface module for sending wildfire tripping operation and maintenance issues to the question and answer generation module, and receiving and displaying the target answers returned by the question and answer generation module.

[0071] Among them, the interface module can send the wildfire tripping operation and maintenance questions input by the user to the question and answer generation module, and receive and display the target answers returned by the question and answer generation module, so as to realize the transmission of wildfire tripping operation and maintenance questions and target answers between the user and the system.

[0072] On this basis, optionally, the interface module includes: an input identification submodule and an output display submodule; wherein the input identification submodule is used to convert the received wildfire tripping operation and maintenance problems into machine language, and send the machine language to the question and answer generation module; the output display submodule is used to receive the target answer generated by the question and answer generation module, and display the target answer based on a pre-set display format.

[0073] Among them, the interface module may include an input identification submodule and an output display submodule. The input identification submodule can convert the received wildfire tripping operation and maintenance problems into machine language, and send the machine language to the question and answer generation module to facilitate the question and answer generation module to identify the wildfire tripping operation and maintenance problems. Optionally, the input identification submodule can receive different user inputs, such as voice input, text input, and graphical interface operations, and convert them into machine language. The output display submodule can receive the target answer generated by the question and answer generation module, and display the target answer based on a pre-set display format. Optionally, the output display submodule displays the target answer in a pre-set display format, such as voice output, text output, and graphical output, so that the user can intuitively understand and use the target answer.

[0074] The above technical solution can provide flexible options for input and output through the interface module, thereby improving the user experience.

[0075] In order to better understand the above-mentioned technical solutions as a whole, the following is an exemplary description of them in combination with specific examples. In this specific example, the transmission line is a transmission corridor, and the wildfire tripping data processing flow chart is as follows: Figure 4 shown.

[0076] S100: Collect historical wildfire tripping data related to the transmission corridor from data sources to form a raw data pool.

[0077] The data sources include wildfire record database, meteorological data, geographic information system data and transmission line operation data. The historical wildfire tripping data related to the transmission corridor is collected from the data sources to form a raw data pool.

[0078] S200: Preprocess the original data pool (i.e. historical wildfire tripping data), including data cleaning, noise filtering, data standardization and data labeling

[0079] S300: Standardize the preprocessed data to convert the preprocessed data (i.e., standard wildfire tripping data) into data suitable for input into a large language model.

[0080] S400: Training and reasoning the processed historical wildfire tripping data (i.e., standard wildfire tripping data), and learning the association (i.e., correlation relationship) between wildfires and historical wildfire tripping events by training the historical data (i.e., standard wildfire tripping data).

[0081] By training on standard wildfire tripping data, the association between wildfire events and historical wildfire tripping events is learned, and prediction is made using the following formula:

[0082] P(A|Q)=softmax(W2·ReLU(W1·Q+b1)+b2)

[0083] Among them, Q is the vector representation of the question asked by the user, P(A|Q) is the probability distribution of the answer, W1 and W2 are the weight matrices of the model, and b1 and b2 are bias terms.

[0084] S500: Based on historical wildfire tripping data and professional knowledge, a knowledge base with a hierarchical structure (i.e., a wildfire tripping database) is constructed.

[0085] S600: Generate a response to the user's question (ie, a target answer) based on the large language model and the knowledge base.

[0086] Based on the large language model, the user input question (i.e., the wildfire tripping operation and maintenance question) is parsed, the data information in the question is extracted, and the wildfire tripping probability associated with the question is predicted based on the data information and the association relationship to generate a response to the user's question. Alternatively, the target wildfire tripping data corresponding to the question is queried based on the knowledge base to generate a response to the user's question.

[0087] In the above specific example, by processing and analyzing historical wildfire tripping data, information support can be provided for operation and maintenance work in the form of questions and answers.

[0088] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A wildfire tripping data analysis system based on a large language model, characterized in that: include: Data acquisition module, data processing module and large language model module; among them, The data acquisition module is used to acquire historical wildfire tripping data associated with wildfire tripping events within a historical time period in the area where the power transmission line is located, and send the historical wildfire tripping data to the data processing module; The data processing module is used to process the received historical wildfire tripping data to obtain standard wildfire tripping data, and send the standard wildfire tripping data to the large language model module; The large language model module is used to determine, for each historical wildfire tripping event in the received standard wildfire tripping data, an association relationship between the historical wildfire tripping event and the historical wildfire event corresponding to the historical wildfire tripping event, so as to predict the wildfire tripping probability after the wildfire event occurs based on the association relationship.

2. The system according to claim 1, characterized in that Also includes: The database construction module is used to construct a wildfire trip database based on the received standard wildfire trip data, so as to manage the historical wildfire trip data based on the wildfire trip database.

3. The system according to claim 2, characterized in that Also includes: The question and answer generation module is used to generate a target answer corresponding to the received wildfire tripping operation and maintenance problem based on the association relationship or the wildfire tripping database.

4. The system according to claim 3, characterized in that The question-answer generation module includes: a question parsing submodule, a question sending submodule and an answer generation submodule; wherein, The problem analysis submodule is used to analyze the received mountain fire tripping operation and maintenance problem and determine the problem type of the mountain fire tripping operation and maintenance problem; The problem sending submodule is used to send the mountain fire tripping operation and maintenance problem to the large language model module when the problem type is a prediction type problem, so as to predict the mountain fire tripping probability in the large language model module based on the association relationship, or send the mountain fire tripping operation and maintenance problem to the database construction module when the problem type of the mountain fire tripping operation and maintenance problem is a query type problem, so as to query the target mountain fire tripping data corresponding to the mountain fire tripping operation and maintenance problem from the mountain fire tripping database based on the database construction module; The answer generation submodule is used to generate a target answer corresponding to the wildfire tripping operation and maintenance problem based on the wildfire tripping probability or the target wildfire tripping data.

5. The system according to claim 4, characterized in that The answer generation submodule includes: a semantic extraction unit and a target answer generation unit; wherein, The semantic extraction unit is used to extract semantic information in the mountain fire tripping operation and maintenance problem; The target answer generation unit is used to integrate the semantic information into the wildfire trip probability or the target wildfire trip data to generate a target answer corresponding to the wildfire trip operation and maintenance problem.

6. The system according to claim 4, characterized in that The question-answer generation module also includes: The answer style adjustment submodule is used to identify the expected answer style of the wildfire tripping operation and maintenance problem, and adjust the style of the target answer based on the expected answer style, wherein the expected answer style includes a brief answer or a detailed answer.

7. The system according to claim 3, characterized in that The large language model module is also used for: For the received wildfire tripping operation and maintenance problem, data information in the wildfire tripping operation and maintenance problem is extracted, and the wildfire tripping probability associated with the wildfire tripping operation and maintenance problem is predicted based on the data information and the association relationship.

8. The system according to claim 3, characterized in that Also includes: The interface module is used to send the mountain fire tripping operation and maintenance problem to the question and answer generation module, and receive and display the target answer returned by the question and answer generation module.

9. The system according to claim 8, characterized in that The interface module includes: an input identification submodule and an output display submodule; wherein, The input recognition submodule is used to convert the received mountain fire tripping operation and maintenance question into machine language, and send the machine language to the question and answer generation module; The output display submodule is used to receive the target answer generated by the question and answer generation module, and display the target answer based on a preset display format.

10. The system according to claim 1, characterized in that The data processing module includes a data preprocessing submodule, a data standardization submodule and a data sending submodule; wherein, The data preprocessing submodule is used to perform a data preprocessing operation on the received historical wildfire tripping data to obtain preprocessed wildfire tripping data, wherein the data preprocessing operation includes at least one of data cleaning, noise filtering and data annotation; The data standardization submodule is used to perform standardization processing on the pre-processed wildfire tripping data to obtain standard wildfire tripping data; The data sending submodule is used to send the standard wildfire tripping data to the large language model module.

Images