Generation method and device of wind power prediction model, equipment and storage medium
By generating questionnaires and using large language models to process data, building a wind power power prediction model is solved, and the problem of high technical threshold for traditional wind power plants is realized, and the automated modeling process is realized, which reduces costs and improves efficiency.
Patent Information
- Application Number
- CN202411439099.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-15
- Publication Date
- 2025-01-17
AI Technical Summary
Traditional wind power generation power prediction methods require customized special prediction algorithms for different wind power plants. The technical threshold is high and professional engineers are dependent on research and development and debugging, resulting in complex modeling process and high cost.
By obtaining model construction elements such as wind power plant type, prediction cycle type and operation data, use pre-programmed templates to generate question words and input them into the pre-trained large language model, generate training sets, verification sets and test sets, and according to the program code Build a wind power prediction model and determine the target model through training, optimization and testing.
The automatic generation of wind power power prediction models is realized, which reduces labor costs and technical thresholds, and improves model construction efficiency and accuracy.
Smart Images
Figure CN119323236A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the technical field of wind power prediction, and in particular to a method, device, equipment and storage medium for generating a wind power prediction model. Background Art
[0002] As a clean and renewable energy source, wind power plays a key role in the transformation of the global energy structure. However, due to the intermittent and uncertain output, it is easy to cause power waste or resource shortage. Therefore, it is necessary to accurately predict wind power generation to help grid operators plan in advance to avoid power waste and resource shortage.
[0003] However, the traditional wind power plant power generation prediction method requires customization of special prediction algorithms for different wind power plants, and needs to be developed or debugged by professional engineers before it can be put into use, which has a high technical threshold. Summary of the invention
[0004] In order to solve the above technical problems, the present disclosure provides a method, device, equipment and storage medium for generating a wind power prediction model.
[0005] In a first aspect, the present invention provides a method for generating a wind power prediction model, comprising:
[0006] Obtain model building elements, which include wind power plant type, forecast cycle type, operating data, a first requirement for model building and a second requirement for operating data; input the model building elements into a pre-compiled template to generate a question; input the question into a pre-trained large language model to obtain a training set, a validation set and a test set after the operating data is processed according to the second requirement, and obtain a program code that meets the wind power plant type, the forecast cycle type and the first requirement; build a wind power prediction model according to the program code; train the wind power prediction model using the training set, optimize the trained wind power prediction model using the validation set, and test the optimized wind power prediction model using the test set; when the wind power prediction model operates normally and the test results meet the preset test conditions, determine the wind power prediction model as the target model.
[0007] In a second aspect, the present invention provides a device for generating a wind power prediction model, comprising:
[0008] The first acquisition module is used to obtain model construction elements, which include wind power plant type, forecast cycle type, operating data, first requirements for model construction and second requirements for operating data; the first generation module is used to input the model construction elements into a pre-compiled template to generate a question; the second generation module is used to input the question into a pre-trained large language model to obtain a training set, a verification set and a test set obtained after the operating data is processed according to the second requirement, and obtain a program code that meets the wind power plant type, the forecast cycle type and the first requirement; the construction module is used to construct a wind power prediction model according to the program code; the training optimization test module is used to train the wind power prediction model using the training set, optimize the trained wind power prediction model using the verification set, and test the optimized wind power prediction model using the test set; the determination module is used to determine the wind power prediction model as the target model when the wind power prediction model operates normally and the test results meet the preset test conditions.
[0009] In a third aspect, the present invention provides a computer device, comprising:
[0010] The memory and the processor are communicatively connected with each other, the memory stores computer instructions, and the processor executes the method for generating a wind power prediction model of the first aspect of the invention and any embodiment thereof by executing the computer instructions.
[0011] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for generating a wind power prediction model of the first aspect of the invention and any embodiment thereof.
[0012] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:
[0013] The method for generating a wind power prediction model provided in the present embodiment obtains model construction elements, which include wind power plant type, prediction cycle type, operating data, a first requirement for the model to be constructed, and a second requirement for the operating data; the model construction elements are input into a pre-compiled template to generate a question; the question is input into a pre-trained large language model to obtain a training set, a verification set, and a test set obtained after the operating data is processed according to the second requirement, and a program code that meets the wind power plant type, the prediction cycle type, and the first requirement is obtained; a wind power prediction model is constructed according to the program code; the wind power prediction model is trained using the training set, the trained wind power prediction model is optimized using the verification set, and the optimized wind power prediction model is tested using the test set; when the wind power prediction model operates normally and the test results meet the preset test conditions, the wind power prediction model is determined as the target model; the present solution solves the problem of high technical threshold of the current modeling process, relies on the large language model, realizes an automated modeling process, reduces labor costs, and improves the efficiency of constructing a wind power prediction model. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.
[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0016] Figure 1 is a flow chart of a method for generating a wind power prediction model provided in accordance with an embodiment of the present invention;
[0017] Figure 2 is a structural connection diagram of a device for generating a wind power prediction model according to an embodiment of the present invention;
[0018] Figure 3 It is a structural connection diagram of a computer device provided according to an embodiment of the present invention. DETAILED DESCRIPTION
[0019] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.
[0020] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0021] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0022] As a clean and renewable energy source, wind power plays a key role in the transformation of the global energy structure. However, wind power generation is significantly affected by natural conditions, especially changes in wind speed and direction, resulting in intermittent and uncertain output. Accurate wind power generation forecasting can help grid operators plan the balance of power supply and demand in advance, reasonably arrange the peak-shaving and valley-filling strategies of traditional energy and energy storage equipment, reduce power waste or shortages caused by supply and demand mismatch, and thus improve the economy and reliability of the entire power grid. However, the traditional wind power generation prediction method requires customized prediction algorithms for different wind power plants, and requires research and development or debugging by professional engineers before it can be put into use, which has a high technical threshold.
[0023] To solve the above problems, an embodiment of the present invention provides a method for generating a wind power prediction model, which generates a question language according to the model construction elements selected by the user and the pre-compiled template, and then uses a large language model to generate a program code that meets the user's needs according to the question language, and constructs a wind power prediction model according to the program code. By training, optimizing and testing the wind power prediction model, a target model that can operate normally and whose test results meet the preset test conditions is obtained; thereby solving the problem of high technical threshold of the current modeling process, relying on the large language model, an automated modeling process is realized, labor costs are reduced, and the efficiency of building a wind power prediction model is improved.
[0024] According to an embodiment of the present invention, an embodiment of a method for generating a wind power prediction model is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0025] In this embodiment, a method for generating a wind power prediction model is provided. The method can be executed by a model generation device, which can be integrated into a PC, a data control device of a wind power plant, a server, and the like. Figure 1 is a flow chart of a method for generating a wind power prediction model according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:
[0026] S101, obtaining model building elements, where the model building elements include wind power plant type, forecast period type, operating data, a first requirement for model building, and a second requirement for operating data.
[0027] Among them, model building elements refer to the basic elements required in the model building process.
[0028] The wind farm type refers to the type of wind farm for which the model is to be constructed. The wind farm type is determined by the geographical location of the wind farm, such as offshore wind power, plain wind power, and hilly wind power.
[0029] The forecast period type refers to the type of forecast duration of the model to be built, such as ultra-short-term forecast (such as 0-6 am), short-term forecast (such as 6 am-3 days), medium-term forecast (such as 3 days-30 days), and long-term forecast (such as 30 days-360 days).
[0030] Operation data refers to the data corresponding to the factors that affect wind power within a preset historical period, such as wind speed, wind direction, temperature, air pressure, wind turbine hub height, number of wind turbines, wind turbine rated power generation and actual power generation.
[0031] The first requirement is the requirement for the model to be built. The first requirement at least includes the object-oriented requirements for the prediction model. The first requirement can also include the requirements for the basic model of the prediction model, the requirements for the prediction model training process and other personalized requirements of the user. The object-oriented requirements for the prediction model, such as overall modeling of the wind power plant and / or modeling of each wind turbine in the wind power plant. The requirements for the basic model of the prediction model, such as the prediction model gives priority to basic models such as artificial neural networks, support vector machines, gradient boosting decision trees and LSTM networks, and other algorithms can also be selected. The requirements for the prediction model training process, such as the maximum number of iterations in the prediction model training process is 10,000 times. Other personalized requirements of the user, such as the need to add Chinese comments line by line in the algorithm code, etc. It should be noted that the object-oriented requirements for the prediction model determine the number of models to be built. When the object-oriented requirement for the prediction model is to model the wind farm as a whole, there is only one model to be constructed; when the object-oriented requirement for the prediction model is to model each wind turbine in the wind farm, the number of models to be constructed is consistent with the number of wind turbines in the wind farm; when the object-oriented requirement for the prediction model is to model the wind farm as a whole and to model each wind turbine in the wind farm, the number of models to be constructed is the sum of the first two.
[0032] The second requirement is the requirement for operating data. The second requirement at least includes the requirements for the division of operating data, and the second requirement may also include the requirements for the preprocessing of operating data. Among them, the requirements for the division of operating data correspond to the first requirement. When the first requirement includes modeling for the entire wind farm, the requirements for the division of operating data are to divide the operating data into a training set, a validation set, and a test set according to a preset ratio. When the first requirement includes modeling for each wind turbine in the wind farm, the requirements for the division of operating data are to classify the operating data according to the wind turbine number, and divide the operating data corresponding to each wind turbine into a training set, a validation set, and a test set according to a preset ratio. When the first requirement includes modeling for the entire wind farm and modeling for each wind turbine in the wind farm, the requirements for the division of operating data include dividing the operating data into a training set, a validation set, and a test set according to a preset ratio, and classifying the operating data according to the wind turbine number, and dividing the operating data corresponding to each wind turbine into a training set, a validation set, and a test set according to a preset ratio. The preset ratio is set by the staff as needed, such as 70:15:15, and this embodiment does not specifically limit the preset ratio. In addition, the preprocessing requirements for the running data are to remove outliers and fill in missing values. The methods for removing outliers include but are not limited to the isolation forest or density-based clustering of applications with noise (DBSCAN), and the methods for filling in missing values include but are not limited to linear interpolation of previous and next data.
[0033] DBSCAN is a density-based spatial clustering algorithm that does not require specifying the number of clusters in advance. Instead, it relies on the neighborhood density of data points to determine the clustering structure. Therefore, it can discover clusters of any shape and effectively identify and exclude noise points. The DBSCAN algorithm mainly depends on two parameters: ε (neighborhood radius) and MinPts (minimum number of points). ε determines the neighborhood range of a point, that is, all points that can be reached within the given radius ε. MinPts specifies the minimum number of points required in the neighborhood for a point to become a core point. For any point P in the dataset, its ε-neighborhood N ε (P) is defined as the set of all points whose distance from P does not exceed ε, that is:
[0034] N ε (P) = {Q ∈ D | dist(P, Q) ≤ ε}
[0035]
[0036] where m is the dimension of the data points.
[0037] The calculation process is as follows: Select an unvisited point P and calculate the ε-neighborhood N ε (P). If |N ε (P)| < MinPts, then mark P as a noise point; otherwise, mark P as a core point and expand the cluster with it as the core. For each core point in N ε (P), recursively apply the above process until all points have been visited. If a point does not belong to a core point but is within the ε-neighborhood of a core point, it is defined as a border point.
[0038] Specifically, in response to the user's tick operation on the interaction interface, obtain the model construction elements ticked by the user.
[0039] S102, Input the model construction elements into a pre-compiled template to generate a question statement.
[0040] Among them, the pre-compiled template is a question template pre-written by the user before the implementation of this solution, and the question template includes model building elements. For example, the pre-compiled template is "If you are a wind power plant power generation prediction engineer, now provide you with a file, the file is the operation data of XX wind power plant (XX is the type of wind power plant) in the past period of time, the operation data includes YY and other data (YY is any one or more of wind speed, wind direction, temperature, air pressure, wind turbine hub height, number of wind turbines, wind turbine rated power and actual power). Now you need to establish the power generation prediction model of the wind power plant ZZ (ZZ is the prediction period type) based on the operation data. Please select the program code of the model to be built according to the first requirement, the first requirement is... Please process the operation data according to the second requirement, the second requirement is...
[0041] Specifically, in response to the user triggering the instruction of the function button corresponding to "generate prompt language", the model building elements obtained in S101 are input into the pre-compiled template to generate question language.
[0042] In an optional embodiment, the prompt is "If you are a wind farm power generation prediction engineer, now you are provided with a file. The file contains the operating data of a hilly wind farm in the past period of time. The operating data includes the wind speed, wind direction, temperature, air pressure, humidity, wind turbine hub height, number of wind turbines, wind turbine rated power and actual power generation of the wind farm. Now you need to build a short-term power generation prediction model for this wind farm based on these operating data. Please select the program code of the model to be built according to the first requirement. The first requirement is: (1) Modeling for the entire wind farm; (2) The prediction model is preferred The algorithms used are artificial neural network, support vector machine, gradient boosting decision tree and LSTM network, and other algorithms can also be selected; (3) The error in the prediction process uses the mean absolute percentage error and the coefficient of determination, and the maximum number of iterations of the algorithm training is 10,000 times; (4) The algorithm code is required to add Chinese comments line by line. Please process the running data according to the second requirement. The second requirement is: (1) Please sort the running data in chronological order, and remove outliers and fill in missing values for the sorted running data; (2) Divide the running data into training set, validation set and test set according to the preset ratio.
[0043] In another optional embodiment, the prompt is "If you are a wind power plant power generation prediction engineer, you are now provided with a file. The file contains the operating data of a hilly wind power plant in the past period of time. The operating data includes the wind speed, wind direction, temperature, air pressure, humidity, wind turbine hub height, number of wind turbines, wind turbine rated power and actual power generation of the wind power plant. Now you need to build a short-term power generation prediction model for this wind power plant based on these operating data. Please select the program code of the model to be built according to the first requirement. The first requirement is that (1) the model is oriented to each wind turbine in the wind power plant; (2) the algorithm preferred for the prediction model is the artificial neural network , support vector machine, gradient boosting decision tree and LSTM network, other algorithms can also be selected; (3) the error in the prediction process uses the mean absolute percentage error and the coefficient of determination, and the maximum number of iterations of the algorithm training is 10,000 times; (4) the algorithm code is required to add Chinese comments line by line. Please process the operating data according to the second requirement. The second requirement is: (1) please sort the operating data in chronological order, and remove outliers and fill in missing values for the sorted operating data; (2) classify the operating data according to the fan number, and divide the operating data corresponding to each fan into training set, validation set and test set according to the preset ratio.
[0044] In another optional embodiment, the prompt is "If you are a wind farm power generation prediction engineer, now you are provided with a file. The file is the operating data of a hilly wind farm in the past period of time. The operating data includes the wind speed, wind direction, temperature, air pressure, humidity, wind turbine hub height, number of wind turbines, wind turbine rated power and actual power generation of the wind farm. Now you need to build a short-term power generation prediction model for this wind farm based on these operating data. Please select the program code of the model to be built according to the first requirement. The first requirement is: (1) modeling for the entire wind farm and modeling for each wind turbine in the wind farm; (2) the algorithm preferred for the prediction model is artificial neural network, support vector machine, gradient boosting Decision tree and LSTM network, other algorithms can also be selected; (3) The error in the prediction process uses the mean absolute percentage error and the coefficient of determination, and the maximum number of iterations of the algorithm training is 10,000 times; (4) The algorithm code is required to add Chinese comments line by line. Please process the operation data according to the second requirement. The second requirement is: (1) Please sort the operation data in chronological order, and remove outliers and fill in missing values for the sorted operation data; (2) Directly divide the operation data into training set, validation set and test set according to the preset ratio; (3) Classify the operation data according to the fan number, and divide the operation data corresponding to each fan into training set, validation set and test set according to the preset ratio.
[0045] S103, input the question into the pre-trained large language model, obtain the training set, verification set and test set obtained after the operation data is processed according to the second requirement, and obtain the program code that meets the wind power plant type, prediction period type and the first requirement.
[0046] Specifically, the question generated in S102 is input into the large language model, and after being processed by the large language model, a training set, a validation set, and a test set processed according to the second requirement are obtained, and a program code that meets the type of wind power plant, the type of prediction cycle, and the first requirement is obtained. Since the program code is generated based on the question, it is ensured that the output program code meets the construction requirements of different users, which solves the problem of needing to customize special prediction algorithms for different wind power plants in the related technology, and directly obtains the program code through the large language model, which also solves the problem of high technical threshold in the current model construction process, reduces labor costs, and improves efficiency.
[0047] In an optional embodiment, the training set, validation set and test set obtained by processing the large language model are the results obtained by directly dividing the operating data according to a preset ratio. The program code obtained by processing the large language model corresponds to the entire wind power plant, that is, the program code is used to predict the wind power of the entire wind power plant in a certain time period in the future.
[0048] In another optional embodiment, the training set, validation set and test set obtained by the large language model processing are the result of classifying the operation data according to the wind turbine number and dividing the operation data corresponding to each wind turbine according to a preset ratio. The program code obtained by the large language model processing corresponds to each wind turbine in the wind farm, and each program code is only used to predict the wind power of the corresponding wind turbine in a certain time period in the future.
[0049] In another optional embodiment, the training set, validation set and test set obtained by processing the large language model include both the result obtained by directly dividing the operation data according to a preset ratio and the result obtained by classifying the operation data according to the wind turbine number and dividing the operation data corresponding to each wind turbine according to a preset ratio. The program code obtained by processing the large language model includes both the program code corresponding to the wind farm as a whole and the program code corresponding to each wind turbine in the wind farm.
[0050] S104, constructing a wind power prediction model according to the program code.
[0051] In an optional embodiment, the large language model outputs a program code for the entire wind farm, and a wind power prediction model for the entire wind farm can be directly constructed based on the program code. The wind power prediction model is used to predict the wind power of the wind farm within a certain period of time in the future.
[0052] In another optional embodiment, the large language model outputs the program code corresponding to each wind turbine in the wind farm, and then the wind power prediction model corresponding to each wind turbine can be constructed according to the program code corresponding to each wind turbine. Then, the wind power prediction models corresponding to all wind turbines in the wind farm are coupled to obtain a wind power prediction model for the entire wind farm.
[0053] In another optional embodiment, the program code output by the large language model includes the program code corresponding to the wind farm as a whole and the program code corresponding to each wind turbine in the wind farm. In this case, it is necessary to construct a first wind power prediction model for the wind farm as a whole based on the program code corresponding to the wind farm as a whole; it is also necessary to construct a wind power prediction model corresponding to the wind turbine based on the program code corresponding to each wind turbine, and couple the wind power prediction models corresponding to all wind turbines in the wind farm to obtain a second wind power prediction model for the wind farm as a whole.
[0054] S105, training the wind power prediction model using the training set, optimizing the trained wind power prediction model using the validation set, and testing the optimized wind power prediction model using the test set.
[0055] Specifically, after the wind power prediction model is obtained, the model is trained, optimized and tested to improve the accuracy of the wind power prediction model.
[0056] In an optional embodiment, when the wind power prediction model is constructed based on the program code corresponding to the entire wind power plant, the training set output by the large language model is directly used to train the wind power prediction model to obtain the trained wind power prediction model. Then, the trained wind power prediction model is optimized using the validation set to obtain the optimized wind power prediction model. Finally, the optimized wind power prediction model is tested using the test set to obtain the test results.
[0057] In another optional embodiment, when the wind power prediction model is obtained by coupling the wind power prediction models corresponding to each wind turbine, first, the wind power prediction model corresponding to the wind turbine is trained using the training set corresponding to each wind turbine to obtain the trained wind power prediction model corresponding to each wind turbine; then, the trained wind power prediction model corresponding to the wind turbine is optimized using the validation set corresponding to each wind turbine to obtain the optimized wind power prediction model corresponding to each wind turbine; finally, the optimized wind power prediction model corresponding to the wind turbine is tested using the test set corresponding to each wind turbine to obtain the test results corresponding to each optimized wind power prediction model. It should be noted that after obtaining the optimized wind power prediction model corresponding to each wind turbine, it is also necessary to obtain the test results of the wind power prediction model for the entire wind power plant after coupling based on the test results of the optimized wind power prediction models corresponding to all wind turbines.
[0058] In another optional embodiment, when the wind power prediction model includes a first wind power prediction model constructed based on the program code corresponding to the wind power plant as a whole, and a second wind power prediction model obtained by coupling the wind power prediction models corresponding to each wind turbine, the training, optimization and testing operations of the first wind power prediction model are completed in accordance with the processing method of the first embodiment corresponding to S105 to obtain a first test result corresponding to the first wind power prediction model; the training, optimization and testing operations of the second wind power prediction model are completed in accordance with the processing method of the second embodiment corresponding to S105 to obtain a second test result corresponding to the second wind power prediction model.
[0059] In another optional embodiment, when the wind power prediction model operates abnormally, abnormal information is obtained; the program code and the abnormal information are re-input into the large language model, and the program code is corrected according to the abnormal information using the large language model until the wind power prediction model constructed based on the corrected program code operates normally, and the correction is ended.
[0060] Specifically, since the wind power model constructed by S101-S104 may have errors in its operation stage (including training, optimization and testing stages), the process is forced to terminate. Therefore, when the wind power prediction model has an operational abnormality, the abnormal information can be obtained from the log file or the abnormal monitoring station, and the abnormal information and the program code corresponding to the wind power model are re-entered into the large language model, and the errors in the program code are corrected and debugged using the code correction function of the large language model itself to obtain the corrected program code. After obtaining the corrected program code, the wind power prediction model is re-constructed based on the program code, and the wind power prediction model is re-trained, optimized and tested until the wind power prediction model runs normally, and the code correction and debugging operations are terminated. In this embodiment, the abnormalities of the wind power prediction model are corrected and debugged by the large language model, thereby improving the efficiency of model construction.
[0061] S106, when the wind power prediction model operates normally and the test results meet the preset test conditions, the wind power prediction model is determined as the target model.
[0062] Specifically, when the wind power prediction model does not make any error during the training, optimization, and testing stages, it indicates that the wind power prediction model operates normally. At this point, it is necessary to determine whether the test results of the wind power prediction model meet the preset test conditions. If so, the wind power prediction model is determined as the target model. If not, the wind power prediction model needs to be retrained, optimized, and tested until the test results meet the preset test conditions. The preset test condition may be whether the accuracy of the test result reaches a preset accuracy threshold. If so, the preset test condition is met; if not, the preset test condition is not met.
[0063] In an optional embodiment, the wind power prediction model is constructed based on the program code corresponding to the entire wind power plant. In this case, when the wind power prediction model operates normally and the test results obtained according to S105 meet the preset test conditions, the wind power prediction model is directly determined as the target model for predicting the wind power of the wind power plant.
[0064] In another optional embodiment, the wind power prediction model is obtained by coupling the wind power prediction models corresponding to each wind turbine. In this case, when the wind power prediction model operates normally and the test results obtained according to S105 meet the preset test conditions, the coupled wind power prediction model is directly determined as the target model for predicting the wind power of the wind power plant.
[0065] In another optional embodiment, the wind power prediction model includes a first wind power prediction model constructed based on the program code corresponding to the wind power plant as a whole, and a second wind power prediction model obtained by coupling the wind power prediction models corresponding to each wind turbine. In this case, when the first wind power prediction model and the second wind power prediction model are both operating normally, first obtain the first prediction result of the first wind power prediction model and the second test result of the second wind power prediction model; then, select the test result with the highest accuracy from the first test result and the second test result; when the test result with the highest accuracy meets the preset test conditions, determine the wind power prediction model corresponding to the test result with the highest accuracy as the target model. This embodiment constructs two wind power prediction models, and determines the model with a higher test result accuracy and meets the preset test conditions as the target model to improve the accuracy of the prediction results.
[0066] The method for generating a wind power prediction model provided in the present embodiment obtains model construction elements, which include wind power plant type, prediction cycle type, operating data, a first requirement for the model to be constructed, and a second requirement for the operating data; the model construction elements are input into a pre-compiled template to generate a question; the question is input into a pre-trained large language model to obtain a training set, a verification set, and a test set obtained after the operating data is processed according to the second requirement, and a program code that meets the wind power plant type, the prediction cycle type, and the first requirement is obtained; a wind power prediction model is constructed according to the program code; the wind power prediction model is trained using the training set, the trained wind power prediction model is optimized using the verification set, and the optimized wind power prediction model is tested using the test set; when the wind power prediction model operates normally and the test results meet the preset test conditions, the wind power prediction model is determined as the target model; the present solution solves the problem of high technical threshold of the current modeling process, relies on the large language model, realizes an automated modeling process, reduces labor costs, and improves the efficiency of constructing a wind power prediction model.
[0067] In this embodiment, a device for generating a wind power prediction model is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0068] This embodiment provides a device for generating a wind power prediction model, such as Figure 2 As shown, including:
[0069] The first acquisition module 201 is used to acquire model building elements, which include wind power plant type, forecast period type, operating data, a first requirement for model building and a second requirement for operating data.
[0070] The first generating module 202 is used to input the model building elements into the pre-compiled template to generate the question language.
[0071] The second generation module 203 is used to input the question into the pre-trained large language model, obtain the training set, verification set and test set obtained after the operation data is processed according to the second requirement, and obtain the program code that meets the wind power plant type, prediction cycle type and the first requirement.
[0072] The construction module 204 is used to construct a wind power prediction model according to the program code.
[0073] The training optimization test module 205 is used to train the wind power prediction model using the training set, optimize the trained wind power prediction model using the validation set, and test the optimized wind power prediction model using the test set.
[0074] The determination module 206 is used to determine the wind power prediction model as the target model when the wind power prediction model operates normally and the test result meets the preset test conditions.
[0075] In some optional implementations, when the first requirement includes overall modeling of the wind power plant, the second requirement includes dividing the operating data into a training set, a validation set, and a test set according to a preset ratio.
[0076] In some optional embodiments, when the first requirement includes modeling each wind turbine in a wind power plant, the second requirement includes classifying the operating data according to the wind turbine number, and dividing the operating data corresponding to each wind turbine into a training set, a validation set, and a test set according to a preset ratio.
[0077] In some optional implementations, the training optimization test module 205 includes:
[0078] The training submodule is used to train the prediction model corresponding to each fan using the training set corresponding to each fan, so as to obtain the trained prediction model corresponding to each fan; the optimization submodule is used to optimize the trained prediction model corresponding to the fan using the validation set corresponding to each fan, so as to obtain the optimized prediction model corresponding to each fan; the testing submodule is used to test the optimized prediction model corresponding to the fan using the test set corresponding to each fan, so as to obtain the test result corresponding to each optimized prediction model.
[0079] In some optional implementations, the training optimization test module 205 further includes:
[0080] The generation submodule is used to test the optimized prediction model corresponding to each wind turbine using the test set corresponding to each wind turbine, and after obtaining the test result corresponding to each optimized prediction model, obtain the test result of the wind power prediction model according to the test result corresponding to each optimized prediction model.
[0081] In some optional embodiments, the device further comprises:
[0082] The second acquisition module is used to obtain abnormal information when the wind power prediction model operates abnormally; the debugging module is used to re-input the program code and the abnormal information into the large language model, and use the large language model to correct the program code according to the abnormal information until the wind power prediction model constructed based on the corrected program code operates normally, and then the correction is ended.
[0083] In some optional implementations, the determination module 206 includes:
[0084] An acquisition submodule is used to obtain a first prediction result of a wind power prediction model obtained by modeling the wind power plant as a whole and a second test result of a wind power prediction model obtained by modeling each wind turbine in the wind power plant; a selection submodule is used to select the test result with the highest accuracy from the first test result and the second test result; a determination submodule is used to determine the wind power prediction model corresponding to the test result with the highest accuracy as the target model when the test result with the highest accuracy meets the preset test conditions.
[0085] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.
[0086] The generating device of the wind power prediction model in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.
[0087] The embodiment of the present invention also provides a computer device having the above Figure 2 The generating device of the wind power prediction model is shown.
[0088] See also Figure 3 , Figure 3 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 3As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 3 A processor 10 is taken as an example.
[0089] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.
[0090] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.
[0091] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0092] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.
[0093] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.
[0094] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.
[0095] In addition to the above-mentioned computer device and computer-readable storage medium, the embodiments of the present application may also be a computer program product, which includes computer program instructions. When the computer program instructions are executed by a processor, the processor executes the steps of the sound source localization method provided by any embodiment of the present application.
[0096] The computer program product may be written in any combination of one or more programming languages to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages such as Java, C++, etc., and conventional procedural programming languages such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0097] The above are only specific embodiments of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to these embodiments herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for generating a wind power prediction model, characterized in that: include: Acquire model building elements, wherein the model building elements include a wind power plant type, a forecast period type, operating data, a first requirement for a model to be built, and a second requirement for the operating data; Inputting the model building elements into a pre-compiled template to generate a question language; Inputting the question into a pre-trained large language model, obtaining a training set, a validation set and a test set obtained after the operation data is processed according to the second requirement, and obtaining a program code that meets the wind power plant type, the forecast period type and the first requirement; Constructing a wind power prediction model according to the program code; Using the training set to train the wind power prediction model, using the validation set to optimize the trained wind power prediction model, and using the test set to test the optimized wind power prediction model; When the wind power prediction model operates normally and the test result meets the preset test conditions, the wind power prediction model is determined as the target model.
2. The method according to claim 1, characterized in that When the first requirement includes overall modeling of the wind power plant, the second requirement includes dividing the operating data into a training set, a validation set, and a test set according to a preset ratio.
3. The method according to claim 1, characterized in that When the first requirement includes modeling each wind turbine in the wind power plant, the second requirement includes classifying the operating data according to the wind turbine number, and dividing the operating data corresponding to each wind turbine into a training set, a validation set and a test set according to a preset ratio.
4. The method according to claim 3, characterized in that The step of training the wind power prediction model by using the training set, optimizing the trained wind power prediction model by using the validation set, and testing the optimized wind power prediction model by using the test set includes: Using the training set corresponding to each wind turbine respectively to train the prediction model corresponding to the wind turbine, to obtain the trained prediction model corresponding to each wind turbine; Optimizing the trained prediction model corresponding to each wind turbine using the validation set corresponding to each wind turbine to obtain an optimized prediction model corresponding to each wind turbine; The optimized prediction model corresponding to each wind turbine is tested using the test set corresponding to each wind turbine to obtain a test result corresponding to each optimized prediction model.
5. The method according to claim 4, characterized in that After the optimized prediction models corresponding to the wind turbines are tested using the test sets corresponding to each wind turbine to obtain the test results corresponding to each optimized prediction model, the method further includes: The test result of the wind power prediction model is obtained according to the test result corresponding to each of the optimized prediction models.
6. The method according to claim 1, characterized in that The method further comprises: When the wind power prediction model operates abnormally, obtaining abnormal information; The program code and the abnormal information are re-input into the large language model, and the program code is corrected according to the abnormal information using the large language model until the wind power prediction model constructed based on the corrected program code runs normally, and then the correction is ended.
7. The method according to claim 1, characterized in that When the first requirement includes modeling for the entire wind power plant and modeling for each wind turbine in the wind power plant, when the wind power prediction model operates normally and the test result meets the preset test conditions, determining the wind power prediction model as the target model includes: Obtaining a first prediction result of a wind power prediction model obtained by modeling the wind power plant as a whole and a second test result of a wind power prediction model obtained by modeling each wind turbine in the wind power plant; Selecting a test result with the highest accuracy from the first test result and the second test result; When the test result with the highest accuracy meets the preset test condition, the wind power prediction model corresponding to the test result with the highest accuracy is determined as the target model.
8. A device for generating a wind power prediction model, characterized in that: include: A first acquisition module is used to acquire model building elements, wherein the model building elements include a wind power plant type, a forecast period type, operating data, a first requirement for a model to be built, and a second requirement for the operating data; A first generating module, for inputting the model building elements into a pre-compiled template to generate a question language; A second generation module is used to input the question into a pre-trained large language model, obtain a training set, a validation set and a test set obtained after the operation data is processed according to the second requirement, and obtain a program code that meets the wind power plant type, the forecast period type and the first requirement; A construction module, used to construct a wind power prediction model according to the program code; A training optimization test module, used to train the wind power prediction model using the training set, optimize the trained wind power prediction model using the validation set, and test the optimized wind power prediction model using the test set; The determination module is used to determine the wind power prediction model as a target model when the wind power prediction model operates normally and the test result meets the preset test conditions.
9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for generating a wind power prediction model according to any one of claims 1 to 7 by executing the computer instructions.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for generating a wind power prediction model according to any one of claims 1 to 7.