Power data augmentation method, device, equipment, storage medium and program product
By inputting historical power data into pre-trained large language model, generating power augmentation data and adjusting model parameters, the challenges of traditional models in power data augmentation are solved, and efficient power data augmentation and model performance are achieved. optimization.
Patent Information
- Application Number
- CN202411438995.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-15
- Publication Date
- 2025-02-07
AI Technical Summary
Traditional large language models face challenges in processing data augmentation of power data, especially in generating augmented data that is highly relevant to actual operating conditions.
By obtaining the historical power data of the target power system, inputting it into a pre-trained large language model, generating power augmentation data, and optimizing model performance by adjusting model parameters.
The power augmentation data is realized in a simple and efficient way to increase the power augmentation data. The generated power augmentation data can facilitate system detection and system upgrade, and improve the robustness and prediction accuracy of the model.
Smart Images
Figure CN119397272A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, equipment, storage medium and program product for power data augmentation. Background Art
[0002] With the rapid development of deep learning and natural language processing technology, large language models, such as Generative Pre-Trained Transformer (GPT), have shown excellent capabilities in text generation and understanding. These large language models are pre-trained on massive text data to generate text data that is highly relevant to the input prompts, and have demonstrated their powerful generation capabilities in applications in multiple fields.
[0003] However, power data is highly diverse and complex, and usually includes sensor data, time series data, and operating records. These data are not only highly time-dependent, but also have high requirements for real-time and accuracy. Based on this, traditional large language models face many challenges in data augmentation of power data. Summary of the invention
[0004] Based on this, it is necessary to provide a power data augmentation method, device, equipment, storage medium and program product that can perform data augmentation on power data in response to the above technical problems.
[0005] In a first aspect, the present application provides a method for augmenting power data, comprising:
[0006] Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data;
[0007] Input historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system;
[0008] Among them, the large language model is trained in the following way:
[0009] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0010] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0011] In one embodiment, according to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted, including:
[0012] Determine target power data according to the sample power data and the sample augmented data;
[0013] According to the target power data, the model parameters of the large language model to be trained are adjusted.
[0014] In one embodiment, the sample power data includes power time series data; accordingly, determining the target power data according to the sample power data and the sample augmented data includes:
[0015] Mapping the sample augmented data to the time series corresponding to the sample power data;
[0016] According to the mapping result of the sample augmented data and the sample power data, the target power data is determined.
[0017] In one embodiment, determining the target power data according to the mapping result of the sample augmented data and the sample power data includes:
[0018] The mapping result of the sample augmented data and the sample power data are used as candidate power data;
[0019] Target power data is screened from candidate power data.
[0020] In one embodiment, screening target power data from candidate power data includes:
[0021] Conduct legality verification on candidate power data, and conduct rationality verification on candidate power data that have passed legality verification;
[0022] The candidate power data that has passed the rationality verification is used as the target power data.
[0023] In one embodiment, obtaining historical power data generated when the target power system is operated in a historical period includes:
[0024] Obtaining status monitoring data of the operating status of the target power system from a monitoring terminal corresponding to the target power system;
[0025] When the status monitoring data indicates that there is an operation fault in the target power system, historical power data generated when the target power system is operated in a historical period is obtained.
[0026] In a second aspect, the present application also provides a power data augmentation device, including:
[0027] A data acquisition module is used to acquire historical power data generated when the target power system is running in a historical period; wherein the historical power data includes power time series data;
[0028] The data augmentation module is used to input historical power data into a pre-trained large language model to obtain power augmentation data corresponding to the target power system;
[0029] Among them, the large language model is trained in the following way:
[0030] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0031] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0032] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:
[0033] Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data;
[0034] Input historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system;
[0035] Among them, the large language model is trained in the following way:
[0036] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0037] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0038] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the following steps are implemented:
[0039] Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data;
[0040] Input historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system;
[0041] Among them, the large language model is trained in the following way:
[0042] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0043] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0044] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:
[0045] Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data;
[0046] Input historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system;
[0047] Among them, the large language model is trained in the following way:
[0048] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0049] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0050] The above-mentioned power data augmentation method, device, equipment, storage medium and program product, through a pre-trained large language model, performs data augmentation processing on the historical power data generated when the target power system is running in the historical period, making the method of data augmentation processing on the power data simpler. Since the power augmented data has a large amount of data and can represent standard power data, the power augmented data can facilitate system detection or system upgrade of the target power system. In other words, the above process can perform data augmentation on the power data, which is convenient for subsequent processing of the target power system. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0052] Figure 1 is a schematic diagram of a flow chart of a power data augmentation method in one embodiment;
[0053] Figure 2 A schematic diagram of a flow chart of a model parameter adjustment step in one embodiment;
[0054] Figure 3 A schematic diagram of a flow chart of a data acquisition step in an embodiment;
[0055] Figure 4Ais a flow chart of a power data augmentation method according to another embodiment;
[0056] Figure 4B A flowchart of a large language model training step in one embodiment;
[0057] Figure 5 is a structural block diagram of a power data augmentation device in one embodiment;
[0058] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0059] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0060] Before introducing the power data augmentation method provided by this application, it should be noted that with the increasing complexity of power production systems, the diversity and large-scale nature of power data have become challenges. Traditional data augmentation methods are mostly used for image or signal processing, and are less capable of augmenting power data such as the combination of text data and time series data.
[0061] In recent years, with the rapid development of deep learning and natural language processing technologies, large language models, such as GPT, have demonstrated outstanding capabilities in text generation and understanding. These large language models are able to generate text data that is highly relevant to the input prompts by pre-training on massive text data, and have demonstrated their powerful generation capabilities in applications in multiple fields. For example, in the fields of medicine, law, finance, etc., large language models can generate text content that meets domain-specific requirements for tasks such as automated report generation, document summarization, and dialogue systems. However, in data augmentation tasks, the potential of these large language models has not yet been fully explored. Traditional data augmentation methods are mostly used for image processing, such as random cropping, rotation, flipping, and other techniques, which can increase the diversity of training data and thus improve the generalization ability of the model. However, the application of these methods on text data and time series data is still relatively small.
[0062] Power data has significant diversity and complexity, and usually includes sensor data, time series data, and operating condition records. These data not only have high time-series dependence, but also have high requirements for real-time and accuracy. Traditional data augmentation techniques face many challenges in processing such data, especially in how to generate augmented data that is highly relevant to actual operating conditions. Although the current large language model performs well in text generation, its application in power production scenarios still needs further exploration. Specifically, how to effectively combine the text data generated by the large language model with the actual operating condition data of power production to generate augmented data that is beneficial to model training is an urgent problem to be solved. In addition, the data in power production (i.e., power data) often has complex structures and background knowledge, requiring data augmentation technology to not only generate diverse text data, but also ensure that these data are highly relevant and reasonable to the actual scenarios of power production. Therefore, using the powerful generation ability of large language models to enhance data sets in power production scenarios is not only a challenging research topic, but also a key link in promoting the intelligent development of power systems.
[0063] In one embodiment, Figure 1 As shown, a method for augmenting power data is provided. This embodiment uses the method applied to a server as an example for illustration. It can be understood that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:
[0064] S110, obtaining historical power data generated when the target power system is operated within a historical period.
[0065] The historical power data includes power time series data, which is the data with time series in the power data. The target power system is the power system that needs to be augmented with power data.
[0066] The length of the historical period can be determined based on manual experience or through a large number of experiments, and the present application does not impose any limitation on the length of the historical period. For example, the length of the historical period can be one week or one month.
[0067] Exemplarily, in this embodiment, all power data generated during the operation of the target power system can be pre-stored in a preset storage location. When it is detected that there is a need for data augmentation in the target power system, or when a data augmentation instruction for the target power system is received, the historical power data generated when the target power system is operating during a historical period is retrieved from the preset storage location.
[0068] Furthermore, in this embodiment, the power data may include multiple types of data such as sensor data, equipment operation data, equipment operation data, fault data, and technical text data. In order to facilitate data management, different types of data may be stored in different locations during data storage. Accordingly, when acquiring historical power data, the historical power data of the corresponding type is acquired from the storage location where the corresponding type of data is stored.
[0069] S120, inputting historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system.
[0070] Among them, the large language model is trained in the following way: the sample power data is input into the large language model to be trained to obtain sample augmented data; according to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0071] In the above-mentioned power data augmentation method, the historical power data generated when the target power system is running in the historical period is augmented by a pre-trained large language model, making the method of data augmentation processing for power data simpler. Since the power augmented data has a large amount of data and can represent standard power data, the power augmented data can facilitate system detection or system upgrade of the target power system. In other words, the above process can perform data augmentation on power data, which is convenient for subsequent processing of the target power system.
[0072] Furthermore, the power augmented data can be used to train and optimize the machine learning models of the target power system. These models can include fault prediction models, equipment health management models, and energy efficiency optimization models. By introducing multimodal data after power augmentation data, the robustness and prediction accuracy of the model can be improved.
[0073] Based on the technical solutions of the above embodiments, the present application also provides an optional embodiment. In this optional embodiment, the process of adjusting the model parameters of the large language model to be trained according to the sample power data and the sample augmented data is refined.
[0074] See also Figure 2 The model parameter adjustment steps shown include:
[0075] S210, determining target power data according to the sample power data and the sample augmented data.
[0076] The sample power data may be obtained from a database of the target power system, such as an online document library, an equipment management system, etc. In order to improve the data accuracy of the sample power data, after obtaining the sample power data, the sample power data may be preprocessed, including data cleaning, format conversion, denoising, etc., to ensure the consistency and accuracy of the sample power data. Furthermore, the sample power data may be segmented into data segments suitable for input into a large language model, and deduplication, word segmentation, and part-of-speech tagging may be performed, while the time series data and sensor data may be standardized.
[0077] In an optional embodiment, the sample power data and the sample augmented data may be combined, and the combined data may be used as the target power data.
[0078] In another optional embodiment, the sample power data includes power time series data, and accordingly, the sample augmented data can be mapped to the time series corresponding to the sample power data; the target power data is determined according to the mapping result of the sample augmented data and the sample power data. It can be understood that mapping the sample augmented data to the time series corresponding to the sample power data can ensure that the time series of the sample augmented data and the sample power data are consistent (i.e. aligned), which is convenient for subsequent training of the large language model.
[0079] Exemplarily, this embodiment may directly use the mapping result of the sample augmented data and the sample power data as the target power data.
[0080] Furthermore, in order to ensure the data accuracy of the sample augmented data, to improve the rationality of the sample augmented data, and thus to improve the training effect of the large language model. In the process of determining the target power data, the power data can be screened and verified. Exemplarily, the mapping result of the sample augmented data and the sample power data are used as candidate power data; the target power data is screened from the candidate power data.
[0081] Optionally, the candidate power data can be verified for legitimacy, and the candidate power data that has passed the legitimacy verification can be verified for rationality; the candidate power data that has passed the rationality verification can be used as the target power data. For example, the generated data can be preliminarily screened using preset rules and models to remove false data that does not conform to the actual scenario (i.e., legitimacy verification). Furthermore, the screened augmented data can be verified for rationality using an existing verification data set or expert knowledge base to ensure that it meets the actual needs and characteristics of the power production scenario.
[0082] S220: Adjust model parameters of the large language model to be trained according to the target power data.
[0083] Specifically, the target power data is input into the large language model to be trained to adjust the model parameters of the large language model to be trained, so as to obtain a trained large language model.
[0084] In the above embodiment, the process of adjusting the model parameters of the large language model to be trained according to the sample power data and the sample augmented data is refined. By processing the sample power data and the sample augmented data, the target power data is obtained, and then the model parameters of the large language model to be trained are adjusted according to the target power data, which can improve the model accuracy of the large language model.
[0085] Based on the technical solutions of the above embodiments, the present application also provides an optional embodiment. In this optional embodiment, the process of obtaining the historical power data generated when the target power system is running in the historical period is described in detail.
[0086] See also Figure 3 The data acquisition steps shown include:
[0087] S310, obtaining status monitoring data of the operating status of the target power system from a monitoring terminal corresponding to the target power system.
[0088] Among them, the monitoring terminal is used to monitor the operating status of the target power system in real time and analyze the monitored power data to determine the health status of each equipment in the target power system, potential fault warnings, energy efficiency optimization suggestions, etc.
[0089] Specifically, the monitoring terminal corresponding to the target power system will feed back the status monitoring data to the server.
[0090] S320, when the status monitoring data indicates that the target power system has an operation fault, obtaining historical power data generated when the target power system is operating in a historical period.
[0091] Specifically, after receiving the status monitoring data, the server analyzes the status monitoring data, and when it is determined that the status monitoring data indicates that there is an operating fault in the target power system, obtains historical power data generated when the target power system was operating in a historical period.
[0092] In the above embodiment, the process of obtaining the historical power data generated when the target power system is running in the historical period is described in detail, so that the server can further optimize the large language model according to the state monitoring data.
[0093] Based on the technical solutions of the above embodiments, the present application also provides an optional embodiment. In this optional embodiment, the power data augmentation method provided by the present application is introduced in detail.
[0094] See also Figure 4A The power data augmentation method shown includes:
[0095] S410A, obtaining sample power data, and adjusting model parameters of a large language model to be trained according to the sample power data to obtain a trained large language model;
[0096] For example, the trained large language master model has the ability to generate text data that is highly relevant to the production scenario of the target power system, while retaining the model's original natural language understanding and generation capabilities. The formula is as follows:
[0097] ;
[0098] in, represents the trained large language model, Represents the large language model to be trained.
[0099] S420A, obtaining status monitoring data of the operating status of the target power system from a monitoring terminal corresponding to the target power system;
[0100] S430A, when the state monitoring data indicates that the target power system has an operation fault, obtaining historical power data generated when the target power system was operated in a historical period;
[0101] S440A, inputs the historical power data into the trained large language model to obtain the power augmented data corresponding to the target power system.
[0102] Further, the process of S410A is described in detail, see Figure 4B The steps for training a large language model include:
[0103] S410B, acquiring initial power data, and performing data preprocessing on the initial power data to obtain sample power data;
[0104] S420B, inputting the sample power data into the large language model to be trained to obtain sample augmented data corresponding to the sample power data;
[0105] S430B, using the mapping result of the sample augmented data and the sample power data as candidate power data;
[0106] Exemplarily, the formula of candidate power data is as follows:
[0107] ;
[0108] ;
[0109] in, represents sample augmented data, Represents sensor data in power data; Represents the power time series data in the power data, Indicates candidate power data.
[0110] S440B, filtering and verifying the candidate power data to screen the candidate power data and obtain target power data;
[0111] Exemplarily, the formula of the target power data is as follows:
[0112] ;
[0113] in, Indicates the target power data, Represents a dataset used to validate candidate power data.
[0114] S450B, adjusting model parameters of the large language model to be trained based on the target power data.
[0115] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0116] Based on the same inventive concept, the embodiment of the present application also provides a power data augmentation device for implementing the power data augmentation method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more power data augmentation device embodiments provided below can refer to the limitations of the power data augmentation method above, and will not be repeated here.
[0117] In an exemplary embodiment, Figure 5 As shown, a power data augmentation device is provided, including: a data acquisition module 510 and a data augmentation module 520, wherein:
[0118] The data acquisition module 510 is used to acquire historical power data generated when the target power system is operated within a historical period.
[0119] Among them, historical power data includes power time series data.
[0120] The data augmentation module 520 is used to input the historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system.
[0121] Among them, the large language model is trained in the following way: the sample power data is input into the large language model to be trained to obtain sample augmented data; according to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0122] In one embodiment, the power data augmentation device includes a model training module, including a target power data determination unit, which is used to determine the target power data based on the sample power data and the sample augmented data; and a model parameter adjustment unit, which is used to adjust the model parameters of the large language model to be trained according to the target power data.
[0123] In one embodiment, the target power data determination unit includes a data mapping subunit for mapping the sample augmented data to a time series corresponding to the sample power data; and a data determination subunit for determining the target power data based on the mapping result of the sample augmented data and the sample power data.
[0124] In one embodiment, the data mapping subunit is specifically used to use the mapping result of the sample augmented data and the sample power data as candidate power data; and to filter the target power data from the candidate power data.
[0125] In one embodiment, the data mapping subunit is specifically used to perform legality verification on the candidate power data, and perform rationality verification on the candidate power data that has passed the legality verification; and use the candidate power data that has passed the rationality verification as the target power data.
[0126] In one embodiment, the data acquisition module 510 includes a status acquisition unit, which is used to obtain status monitoring data of the operating status of the target power system from the monitoring terminal corresponding to the target power system; the data acquisition unit is used to obtain historical power data generated when the target power system is operating within a historical period when the status monitoring data indicates that there is an operating fault in the target power system.
[0127] Each module in the above-mentioned power data augmentation device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.
[0128] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a method for augmenting power data is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.
[0129] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0130] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:
[0131] Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data;
[0132] Input historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system;
[0133] Among them, the large language model is trained in the following way:
[0134] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0135] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0136] In one embodiment, when the processor executes the computer program, the following steps are also implemented:
[0137] Determine target power data according to the sample power data and the sample augmented data;
[0138] According to the target power data, the model parameters of the large language model to be trained are adjusted.
[0139] In one embodiment, when the processor executes the computer program, the following steps are also implemented:
[0140] Mapping the sample augmented data to the time series corresponding to the sample power data;
[0141] According to the mapping result of the sample augmented data and the sample power data, the target power data is determined.
[0142] In one embodiment, when the processor executes the computer program, the following steps are also implemented:
[0143] The mapping result of the sample augmented data and the sample power data are used as candidate power data;
[0144] Target power data is screened from candidate power data.
[0145] In one embodiment, when the processor executes the computer program, the following steps are also implemented:
[0146] Conduct legality verification on candidate power data, and conduct rationality verification on candidate power data that have passed legality verification;
[0147] The candidate power data that has passed the rationality verification is used as the target power data.
[0148] In one embodiment, when the processor executes the computer program, the following steps are also implemented:
[0149] Obtaining status monitoring data of the operating status of the target power system from a monitoring terminal corresponding to the target power system;
[0150] When the status monitoring data indicates that there is an operation fault in the target power system, historical power data generated when the target power system is operated in a historical period is obtained.
[0151] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
[0152] Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data;
[0153] Input historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system;
[0154] Among them, the large language model is trained in the following way:
[0155] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0156] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0157] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0158] Determine target power data according to the sample power data and the sample augmented data;
[0159] According to the target power data, the model parameters of the large language model to be trained are adjusted.
[0160] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0161] Mapping the sample augmented data to the time series corresponding to the sample power data;
[0162] According to the mapping result of the sample augmented data and the sample power data, the target power data is determined.
[0163] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0164] The mapping result of the sample augmented data and the sample power data are used as candidate power data;
[0165] Target power data is screened from candidate power data.
[0166] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0167] Conduct legality verification on candidate power data, and conduct rationality verification on candidate power data that have passed legality verification;
[0168] The candidate power data that has passed the rationality verification is used as the target power data.
[0169] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0170] Obtaining status monitoring data of the operating status of the target power system from a monitoring terminal corresponding to the target power system;
[0171] When the status monitoring data indicates that there is an operation fault in the target power system, historical power data generated when the target power system is operated in a historical period is obtained.
[0172] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:
[0173] Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data;
[0174] Input historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system;
[0175] Among them, the large language model is trained in the following way:
[0176] Input the sample power data into the large language model to be trained to obtain sample augmented data;
[0177] According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
[0178] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0179] Determine target power data according to the sample power data and the sample augmented data;
[0180] According to the target power data, the model parameters of the large language model to be trained are adjusted.
[0181] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0182] Mapping the sample augmented data to the time series corresponding to the sample power data;
[0183] According to the mapping result of the sample augmented data and the sample power data, the target power data is determined.
[0184] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0185] The mapping result of the sample augmented data and the sample power data are used as candidate power data;
[0186] Target power data is screened from candidate power data.
[0187] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0188] Conduct legality verification on candidate power data, and conduct rationality verification on candidate power data that have passed legality verification;
[0189] The candidate power data that has passed the rationality verification is used as the target power data.
[0190] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:
[0191] Obtaining status monitoring data of the operating status of the target power system from a monitoring terminal corresponding to the target power system;
[0192] When the status monitoring data indicates that there is an operation fault in the target power system, historical power data generated when the target power system is operated in a historical period is obtained.
[0193] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.
[0194] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0195] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for augmenting power data, characterized in that: The method comprises: Acquire historical power data generated when the target power system is operated in a historical period; wherein the historical power data includes power time series data; Inputting the historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system; The large language model is trained in the following way: Input the sample power data into the large language model to be trained to obtain sample augmented data; According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
2. The method according to claim 1, characterized in that The adjusting the model parameters of the large language model to be trained according to the sample power data and the sample augmented data includes: Determining target power data according to the sample power data and the sample augmented data; According to the target power data, model parameters of the large language model to be trained are adjusted.
3. The method according to claim 2, characterized in that The sample power data includes power time series data; accordingly, determining the target power data according to the sample power data and the sample augmented data includes: Mapping the sample augmented data to the time series corresponding to the sample power data; The target power data is determined according to the mapping result of the sample augmented data and the sample power data.
4. The method according to claim 3, characterized in that The determining target power data according to the mapping result of the sample augmented data and the sample power data comprises: Using the mapping result of the sample augmented data and the sample power data as candidate power data; Target power data is screened from the candidate power data.
5. The method according to claim 4, characterized in that The screening of target power data from the candidate power data includes: Performing a legality verification on the candidate power data, and performing a rationality verification on the candidate power data that has passed the legality verification; The candidate power data that has passed the rationality verification is used as the target power data.
6. The method according to any one of claims 1 to 5, characterized in that The obtaining of historical power data generated when the target power system is operated in a historical period includes: Acquire status monitoring data of the operating status of the target power system by a monitoring terminal corresponding to the target power system; When the state monitoring data indicates that the target power system has an operation fault, historical power data generated when the target power system is operated within a historical period is obtained.
7. A power data augmentation device, characterized in that: The device comprises: A data acquisition module, used to acquire historical power data generated when the target power system is running in a historical period; wherein the historical power data includes power time series data; A data augmentation module, used for inputting the historical power data into a pre-trained large language model to obtain power augmented data corresponding to the target power system; The large language model is trained in the following way: Input the sample power data into the large language model to be trained to obtain sample augmented data; According to the sample power data and the sample augmented data, the model parameters of the large language model to be trained are adjusted to train the large language model.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.