Construction method, system and equipment of wind power prediction data set and medium

Abstract

The invention discloses a method, a system and equipment for constructing a wind power prediction data set and a medium. The method comprises the following steps: acquiring historical meteorological data and historical power data of a target wind power station as an original data set; dividing the original data set according to a first division rule to obtain an extreme weather event data set; wherein the extreme weather event data set comprises sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events; in the extreme weather event data set, performing related feature analysis based on environmental factors contained in the sub-historical meteorological data to obtain a correlation analysis result of each environmental factor and the historical power data; and screening the extreme weather event data set based on a correlation analysis result to obtain a target wind power prediction data set. The accuracy of wind power prediction under the extreme weather condition is remarkably improved, the problem that a traditional prediction method is insufficient in adaptability to small-probability extreme weather events is solved, and more reliable data support is provided for power grid dispatching.

Description

Technical Field

[0001] The present invention relates to the technical field of power engineering data processing, and in particular to a method, system, equipment and medium for constructing a wind power prediction data set. Background Art

[0002] Wind power generation systems are extremely sensitive to extreme weather conditions, which can lead to large-scale equipment outages and sudden drops in power generation. In extreme cases, they may even pose a threat to the safe and stable operation of the power grid. Extreme weather generally refers to catastrophic weather conditions in which meteorological factors deviate significantly from normal levels and bring serious consequences. From the perspective of wind power generation, any meteorological factor that causes power generation fluctuations, such as wind speed and direction, and any weather phenomenon that undergoes drastic changes in a short period of time can be defined as extreme weather. These typically include severe weather conditions such as strong winds, low temperatures and freezing temperatures, cold waves, and typhoons. As global warming exacerbates the frequency of extreme weather events and the penetration rate of wind power generation continues to increase, the impact of extreme weather on power grid operations is becoming increasingly significant. Therefore, accurately predicting and warning of wind power output in complex extreme weather conditions and enhancing the resilience of the power system are key areas for future research.

[0003] However, current mainstream wind power forecasting technologies rely primarily on statistical models such as machine learning and deep learning that draw on large amounts of data. Since extreme weather events are low-probability events and sample data is scarce, wind power forecasting models based on small samples of complex extreme weather events are prone to overfitting during training, seriously impacting the accuracy and reliability of power forecasts. Furthermore, the dramatic dynamic changes in meteorological data under complex extreme weather conditions significantly increase the difficulty of data processing and sample extraction. Furthermore, the causes of dramatic changes in wind power under abnormal weather conditions are complex, making it difficult to effectively capture the inherent connection between meteorological fluctuations and power fluctuations, making accurate wind power forecasting under complex extreme weather conditions extremely difficult.

[0004] Therefore, how to build an accurate wind power prediction dataset for extreme weather events with low probability has become an urgent problem to be solved. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a method, system, device and medium for constructing a wind power prediction dataset to solve the problem of inaccurate dataset construction for extreme weather events with low probability.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a method for constructing a wind power prediction dataset, comprising:

[0009] Obtain the historical meteorological data and historical power data of the target wind farm as the original data set;

[0010] Divide the original data set according to the first division rule to obtain an extreme weather event data set; wherein the extreme weather event data set includes sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events;

[0011] In the extreme weather event data set, perform correlation analysis based on the environmental factors included in the sub-historical meteorological data to obtain the correlation analysis results between each environmental factor and the historical power data;

[0012] Based on the correlation analysis results, screen the extreme weather event data set to obtain a target wind power prediction data set.

[0013] As a preferred solution of the method for constructing the wind power prediction data set according to the present invention, wherein: the dividing the original data set according to the first division rule to obtain an extreme weather event data set includes:

[0014] Obtain each meteorological factor type and each meteorological factor threshold corresponding to the preset extreme weather;

[0015] Based on the meteorological factor type and the meteorological factor threshold, divide and screen the historical meteorological data to obtain sub-historical meteorological data;

[0016] Extract the historical power data corresponding to the time series of the sub-historical meteorological data as sub-historical power data;

[0017] Use the sub-historical meteorological data and the sub-historical power data as the extreme weather event data set.

[0018] The beneficial effect of this preferred technical solution is: accurately dividing the historical meteorological data and historical power data can effectively identify the operation characteristics of the wind farm under various extreme weather events.

[0019] As a preferred solution of the method for constructing the wind power prediction data set according to the present invention, wherein: the obtaining of the correlation analysis results between each environmental factor and the historical power data includes:

[0020] Perform correlation analysis on the environmental factors based on at least two correlation analysis methods to obtain corresponding sub-correlation analysis results; wherein the sub-correlation analysis results include the environmental factors and the correlation influence factors corresponding to each of the environmental factors;

[0021] Based on the correlation influence factors, sort the environmental factors to obtain a sorting result;

[0022] In each of the sub-correlation analysis results, obtain a target environmental factor based on the sorting result;

[0023] Determine the correlation influence weight of the target environmental factor based on the influence factors corresponding to each of the target environmental factors; wherein, the correlation influence weight is positively correlated with the influence factor;

[0024] Based on the correlation influence weight and by fusing the target environmental factors included in each of the sub-correlation analysis results, obtain the correlation analysis result of each environmental factor with respect to the historical power data.

[0025] As a preferred embodiment of the method for constructing a wind power prediction data set according to the present invention, wherein: the acquisition of the target wind power prediction data set includes:

[0026] Obtain a preset environmental factor based on the correlation analysis result;

[0027] Match the sub-historical meteorological data in the extreme weather event data set and the sub-historical meteorological data corresponding to the time series of the sub-historical meteorological data based on the preset environmental factor, as the target wind power prediction data set.

[0028] The beneficial effects of this preferred technical solution are: significantly improving the accuracy of wind power prediction under extreme weather conditions, solving the problem of insufficient adaptability of traditional prediction methods to small-probability extreme weather events, and providing more reliable data support for power grid dispatching.

[0029] As a preferred embodiment of the method for constructing a wind power prediction data set according to the present invention, wherein: the actual correlation between the environmental factor corresponding to each target meteorological data in the target wind power prediction data set and the historical power data is greater than a preset correlation.

[0030] As a preferred embodiment of the method for constructing a wind power prediction data set according to the present invention, wherein: after obtaining the target wind power prediction data set, it further includes:

[0031] Based on a pre-constructed WGAN-GP model, perform at least one round of expansion on the target wind power prediction data set to obtain an expanded wind power prediction data.

[0032] As a preferred embodiment of the method for constructing a wind power prediction data set according to the present invention, wherein: after obtaining the target wind power prediction data set, it further includes:

[0033] In each round of expansion, determine the actual data quality standard of the output result of the WGAN-GP model;

[0034] When the actual data quality standard meets the first data quality standard, add the augmented wind power prediction data obtained by augmentation to the target wind power prediction data set.

[0035] The beneficial effects of this preferred technical solution are as follows: By means of the gradient penalty mechanism of the generative adversarial network, the stability of the data augmentation process is ensured. The generated high-quality augmented data not only retains the typical characteristics of the original extreme weather events but also significantly enriches the diversity of training samples.

[0036] In a second aspect, the present invention provides a system for constructing a wind power prediction data set, including:

[0037] A data acquisition module, configured to acquire historical meteorological data and historical power data of a target wind farm as an original data set;

[0038] A division module, configured to divide the original data set according to a first division rule to obtain an extreme weather event data set; wherein the extreme weather event data set includes sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events;

[0039] A correlation analysis module, configured to perform correlation feature analysis based on environmental factors included in the sub-historical meteorological data in the extreme weather event data set to obtain a correlation analysis result of each environmental factor and the historical power data;

[0040] A screening module, configured to screen the extreme weather event data set based on the correlation analysis result to obtain a target wind power prediction data set; wherein, for each target meteorological data in the target wind power prediction data set, the actual correlation between the corresponding environmental factor and the historical power data is greater than a preset correlation.

[0041] In a third aspect, the present invention provides an electronic device, including a memory and a processor; the memory is used to store computer-executable instructions, and when the processor executes the computer-executable instructions, the steps of the method for constructing a wind power prediction data set are implemented.

[0042] In a fourth aspect, the present invention provides a computer-readable storage medium, which stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, the steps of the method for constructing a wind power prediction data set are implemented.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a method, system, equipment and medium for constructing a wind power prediction data set, and by establishing extreme weather classification rules based on meteorological factor types and meteorological factor thresholds, accurately divides historical meteorological data and historical power data, and effectively identifies the operating characteristics of wind farms under various extreme weather events; in each extreme weather event data set, a correlation feature analysis is performed based on the environmental factors contained in the sub-historical meteorological data to obtain the correlation analysis results of each environmental factor and the historical power data; based on the correlation analysis results, the extreme weather event data set is screened to obtain a target wind power prediction data set, which significantly improves the accuracy of wind power prediction under extreme weather conditions, solves the problem of insufficient adaptability of traditional prediction methods to low-probability extreme weather events, provides more reliable data support for power grid dispatching, and enhances the operating safety and stability of the power system under severe meteorological conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 The figure is a schematic diagram of the overall process logic of a method for constructing a wind power prediction dataset according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0047] Example 1, with reference to Figure 1 As an embodiment of the present invention, a method for constructing a wind power prediction data set is provided, such as Figure 1 The specific steps shown include:

[0048] S100: Acquire historical meteorological data and historical power data of a target wind farm as an original data set;

[0049] S200: Divide the original data set according to a first division rule to obtain an extreme weather event data set; wherein the extreme weather event data set includes sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events;

[0050] S300: In the extreme weather event dataset, performing correlation feature analysis based on environmental factors included in the sub-historical meteorological data to obtain correlation analysis results between each environmental factor and the historical power data;

[0051] S400: Filtering the extreme weather event data set based on the correlation analysis result to obtain a target wind power prediction data set; wherein, the actual correlation between the environmental factors corresponding to each target meteorological data in the target wind power prediction data set and the historical power data is greater than a preset correlation.

[0052] It should be noted that since extreme weather is a low-probability event and sample data is scarce, wind power prediction models based on small samples of complex extreme weather are prone to overfitting during training, seriously affecting the accuracy and reliability of power predictions. At the same time, the causes of drastic changes in wind power under abnormal weather are complex, and it is difficult to effectively capture the intrinsic connection between meteorological fluctuations and power fluctuations, making it very difficult to accurately predict wind power under complex extreme weather conditions.

[0053] Therefore, in order to solve the problem of inaccurate data sets constructed for extreme weather events with low probability, the above steps S100 to S400 establish extreme weather classification rules based on meteorological factor types and meteorological factor thresholds, accurately divide historical meteorological data and historical power data, and effectively identify the operating characteristics of wind farms under various extreme weather events; in each extreme weather event data set, relevant feature analysis is performed based on the environmental factors contained in the sub-historical meteorological data to obtain the correlation analysis results of each environmental factor and the historical power data; based on the correlation analysis results, the extreme weather event data set is screened to obtain the target wind power prediction data set, which significantly improves the accuracy of wind power prediction under extreme weather conditions, solves the problem of insufficient adaptability of traditional prediction methods to low-probability extreme weather events, provides more reliable data support for power grid dispatching, and enhances the operation safety and stability of the power system under severe meteorological conditions.

[0054] Example 2: Based on the previous example, this example introduces a specific implementation method of a method for constructing a wind power prediction dataset, in order to illustrate the technical solution of this method.

[0055] S100: Acquire historical meteorological data and historical power data of a target wind farm as an original data set;

[0056] In an embodiment of the present application, the historical meteorological data and historical power data of the target wind farm are obtained by acquiring recorded data of the target wind farm, wherein the historical meteorological data are obtained by acquiring historical numerical weather forecast data through a meteorological data acquisition device provided at the target wind farm.

[0057] Specifically, the historical meteorological data may include the collected temperature, humidity, air pressure, wind speed, wind direction, predicted temperature, humidity, air pressure, wind speed, and wind direction, etc.

[0058] Specifically, after obtaining the historical meteorological data and historical power data, preliminary cleaning and reconstruction are performed on the outliers in the historical data; for the historical data, data with a wind speed less than the cut-in wind speed but a wind power data greater than zero, a wind power data greater than zero when the wind speed is greater than the cut-out wind speed, and a wind power data equal to zero when the wind speed is greater than the cut-in wind speed and less than the cut-out wind speed are excluded.

[0059] It should be noted that the multi-dimensional and long-time-span original operation data obtained in the above step S100 not only completely records the actual response characteristics of the wind farm under various meteorological conditions, but more importantly, provides a true and reliable data source for the accurate identification and feature analysis of extreme weather events, effectively solving the problem of insufficient training data for the prediction model caused by the lack of extreme weather samples in traditional methods, and ensuring the accuracy and reliability of subsequent data analysis from the source.

[0060] S200: Divide the original data set according to the first division rule to obtain an extreme weather event data set; the extreme weather event data set includes sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events;

[0061] In an alternative embodiment, the first division rule can be based on dynamic clustering division based on meteorological pattern recognition. An unsupervised clustering algorithm is used to perform multi-dimensional feature clustering on the historical meteorological data (such as wind speed mutation rate, duration of continuous low temperature, air pressure gradient, etc.), and automatically identify extreme weather event categories with similar meteorological characteristics without presetting fixed thresholds, adapting to the climate characteristics of different regions.

[0062] In another alternative embodiment, the first division rule can also be based on spatio-temporal association division triggered by events. Define the triggering logic of extreme weather events (such as strong winds lasting for more than 6 hours within a radius of 50 km affected by the typhoon path), and combine spatio-temporal association analysis to couple and match the meteorological data with the wind farm location and terrain data to divide an event data set with clear spatio-temporal boundaries, enhancing the causal relationship between extreme events and power fluctuations.

[0063] In the embodiments of the present application, for wind farms, common extreme weather events usually include cold wave events, low temperature freezing events, and strong wind events. Therefore, the first division rule may include a cold wave event division rule, a low temperature freezing event division rule, and a strong wind division rule.

[0064] Specifically, the meteorological factor type corresponding to the cold wave event is the daily minimum temperature. The division rule for the cold wave event is that the daily minimum temperature is not greater than the first preset temperature value, and the temperature drop amplitude within the first preset duration is not less than the second preset temperature value, the temperature drop amplitude within the second preset duration is not less than the third preset temperature value, or the temperature drop amplitude within the third preset duration is not less than the fourth preset temperature value. Among them, the numerical values of the first preset temperature value, the second preset temperature value, the third preset temperature value, and the fourth preset temperature value increase in sequence, and the numerical values of the first preset duration, the second preset duration, and the third preset duration increase in sequence.

[0065] Exemplarily, the division rule for the cold wave event is that the daily minimum temperature is not greater than 4°C, and the temperature drop amplitude within 24 hours is not less than 8°C, the temperature drop amplitude within 48 hours is not less than 10°C, or the temperature drop amplitude within the third preset duration is not less than 12°C.

[0066] Specifically, the meteorological factor types corresponding to the low-temperature freezing are temperature and humidity. The division rule for the cold wave event is that the temperature is less than the fifth preset temperature value, the humidity is not less than the preset humidity value, and the duration is not less than the first preset duration.

[0067] Exemplarily, the division rule for the cold wave event is that the temperature is less than 3°C, the humidity is not less than 70%, and the duration is not less than 24 hours.

[0068] Specifically, the meteorological factor type corresponding to the strong wind event is the wind speed at the height of the wind turbine rotor. The division rule for the strong wind event is that the wind speed at the height of the wind turbine rotor is greater than the cut-out wind speed of the wind turbine, and the duration is not less than the second preset duration. Among them, the cut-out wind speed of the wind turbine and the second preset duration can be determined according to the wind turbine parameters.

[0069] In the embodiment of the present application, on the basis of the above first division work, the steps of dividing the original data set to obtain the extreme weather event data set specifically include:

[0070] Obtain each meteorological factor type and each meteorological factor threshold corresponding to the preset extreme weather;

[0071] Based on the meteorological factor type and the meteorological factor threshold, divide and screen the historical meteorological data to obtain sub-historical meteorological data;

[0072] Extract the historical power data corresponding to the time series of the sub-historical meteorological data as sub-historical power data;

[0073] Use the sub-historical meteorological data and the sub-historical power data as the extreme weather event data set.

[0074] Specifically, each type of meteorological factor is the daily minimum temperature, temperature and humidity, and wind speed at the wind turbine hub height, and each meteorological factor threshold is the corresponding value of the minimum temperature, temperature and humidity, and wind speed at the wind turbine hub height.

[0075] It should be noted that the above step S200 accurately classifies various extreme weather events based on scientifically defined meteorological factor types and thresholds, not only realizing the targeted identification of different extreme weather scenarios, but also establishing an association framework between meteorological data and power fluctuations, significantly improving the accuracy and interpretability of the extraction of wind farm operation characteristics under extreme weather conditions.

[0076] It should be noted that after dividing the original data set according to the first division rule to obtain the extreme weather event data set, since the sub-historical meteorological data includes multiple environmental factors such as the collected temperature and humidity, and the correlations of each environmental factor with the power performance under extreme weather are different, it is necessary to perform a correlation analysis on the meteorological data in each extreme weather event data set.

[0077] S300: In the extreme weather event data set, perform a correlation feature analysis based on the environmental factors included in the sub-historical meteorological data to obtain the correlation analysis results of each environmental factor and the historical power data;

[0078] In an optional embodiment, a correlation analysis method can be used to perform a correlation feature analysis based on the environmental factors included in the sub-historical meteorological data to obtain the correlation analysis results of each environmental factor and the historical power data.

[0079] Although using a single correlation analysis method alone can achieve the correlation feature analysis of the environmental factors included in the sub-historical meteorological data, the correlation analysis results are limited by the inherent properties of the correlation analysis method, and the correlation analysis results are not fully considered.

[0080] To solve this problem, in the embodiments of the present application, at least two correlation analysis methods are used to perform a correlation feature analysis on the environmental factors included in the sub-historical meteorological data to obtain the sub-correlation analysis results corresponding to each correlation analysis method;

[0081] In the embodiments of the present application, the acquisition of the correlation analysis results of each environmental factor and the historical power data includes:

[0082] Perform a correlation analysis on the environmental factors based on at least two correlation analysis methods to obtain the corresponding sub-correlation analysis results; wherein, the sub-correlation analysis results include the environmental factors and the correlation influence factors corresponding to each environmental factor;

[0083] Sort the environmental factors based on the correlation influence factors to obtain the sorting result;

[0084] In the results of each sub-correlation analysis, the top N environmental factors are extracted as target environmental factors based on the sorting results;

[0085] Based on the impact factors corresponding to each target environmental factor, the correlation impact weight of the target environmental factor is determined; among them, the correlation impact weight is positively correlated with the impact factor;

[0086] Based on the correlation impact weight and integrating the target environmental factors included in each sub-correlation analysis result, the correlation analysis result of each environmental factor relative to the historical power data is obtained.

[0087] Specifically, the correlation analysis methods include the mRMR method, the grey relational degree method, the random forest method, and the Gaussian distance method.

[0088] Specifically, N is a positive integer greater than zero.

[0089] In an optional embodiment, N is taken as 10.

[0090] Exemplarily, taking the preset correlation analysis methods of the mRMR method, the grey relational degree method, the random forest method, and the Gaussian distance method as examples, the solution of this embodiment is described:

[0091] After obtaining the original data set, four methods of the maximum relevance minimum redundancy method, the grey relational degree method, the random forest method, and the Gaussian distance method are respectively used to perform correlation analysis between the environmental factors related to the power under extreme weather events and the historical power, and the correlation impact factors of each environmental factor relative to the historical power data corresponding to each correlation analysis method are obtained;

[0092] Based on the correlation impact factors, the environmental factors are sorted from high to low to obtain the sorting result;

[0093] In the results of each sub-correlation analysis, the top 10 environmental factors are selected as target environmental factors based on the sorting results, and the weight assignment from 10 to 1 is sequentially performed on the target environmental factors as the correlation impact weight;

[0094] The correlation impact weights of each target environmental factor in each sub-correlation analysis result are summed to obtain the comprehensive sorting of the environmental factor and power correlation as the correlation analysis result of each environmental factor relative to the historical power data.

[0095] It should be noted that in the above step S300, by deeply analyzing the correlation between environmental factors and historical power data in various extreme weather events, the internal mechanism of wind power fluctuations under different meteorological conditions is effectively revealed; a quantitative analysis method is used to explore the correlation law between key environmental factors and power output, which not only overcomes the defect of insufficient analysis of the causes of abnormal power fluctuations under extreme weather by traditional methods, enables the constructed prediction model to accurately capture the non-linear relationship between extreme meteorological parameters and wind power output, and significantly improves the prediction sensitivity to power mutations under complex weather conditions.

[0096] S400: Screen the extreme weather event data set based on the correlation analysis results to obtain a target wind power prediction data set; wherein, the actual correlation between the environmental factors corresponding to each target meteorological data in the target wind power prediction data set and the historical power data is greater than the preset correlation.

[0097] In the embodiment of the present application, the acquisition of the target wind power prediction data set includes:

[0098] Select the top N target environmental factors as preset environmental factors based on the correlation analysis results;

[0099] Match the sub-historical meteorological data in the extreme weather event data set and the sub-historical meteorological data corresponding to the time series of the sub-historical meteorological data based on the preset environmental factors as the target wind power prediction data set.

[0100] It should be noted that after obtaining the correlation analysis results, the extreme weather event data set is screened based on the correlation analysis results to obtain a target wind power prediction data set; the actual correlation between the environmental factors corresponding to each historical target meteorological data included in the target wind power prediction data set and the historical power data is greater than the preset correlation; therefore, the finally constructed target wind power prediction data set contains multiple historical target meteorological data with an actual correlation greater than the preset correlation with extreme weather events, and has the advantage of higher accuracy when constructing a wind power prediction data set for extreme weather of small probability events.

[0101] It should be noted that the above step S400 accurately screens the extreme weather data set based on strict correlation analysis results, effectively eliminates noise data and weakly associated samples by setting a preset correlation threshold, and significantly improves the signal-to-noise ratio of the data set; it not only solves the problem of the decline in the accuracy of the prediction model caused by traditional methods retaining low-correlation data, but also constructs a high-quality target wind power prediction data set, providing a more reliable extreme weather power prediction result for power grid dispatching decisions.

[0102] In the method for constructing the wind power prediction data set in the above embodiments, the finally constructed target wind power prediction data set contains multiple historical target meteorological data whose actual correlation with extreme weather events is greater than the preset correlation. The method has the advantage of being more accurate for extreme weather with low-probability events when constructing the wind power prediction data set. Moreover, the method of determining the correlation influence weight of the target environmental factor according to the influence factor corresponding to the target environmental factor and further fusing by sorting considering the correlation between the environmental factor and the power can consider multiple correlation analysis strategies and combine the application of multiple feature selection methods, and can accurately identify and screen out the key features closely related to wind power output in complex extreme weather situations. However, since extreme weather is a low-probability event and the sample data is scarce, the wind power prediction model based on small samples of complex extreme weather is prone to overfitting during training, seriously affecting the accuracy and reliability of power prediction.

[0103] To solve this problem, in the embodiments of the present application, after screening the extreme weather event data set based on the correlation analysis results and obtaining the target wind power prediction data set, it further includes:

[0104] Based on the pre-constructed WGAN-GP model, the target wind power prediction data set is expanded at least once to obtain the expanded wind power prediction data after expansion.

[0105] Specifically, the WGAN-GP model is used to expand the target wind power prediction data set at least once; the Wasserstein Generative Adversarial Network (WGAN) has achieved significant improvements on the basis of the original generative adversarial network by introducing the Wasserstein distance loss function with gradient penalty, not only effectively avoiding the problem of mode collapse, but also significantly improving the stability during the network training process. And WGAN-GP (i.e., WassersteinGAN with gradient penalty) combines the advantages of the Wasserstein distance metric and combines the gradient penalty term, thus ensuring that the model can maintain a more robust and efficient performance during the training stage.

[0106] Specifically, after obtaining the expanded wind power prediction data after expansion, the wind power prediction sample set under complex extreme weather conditions after expansion is detected for abnormal data again. Cluster analysis is used to distinguish normal power data and abnormal power data, and more accurate abnormal detection results are obtained by appropriately adjusting the model parameters.

[0107] Specifically, in order to ensure that the sample quality level of the augmented wind power prediction data can still meet the preset quality level, as an exemplary embodiment, the method further includes: in each round of augmentation, determining the actual data quality standard of the output result of the WGAN-GP model based on T-distributed Stochastic Neighbor Embedding (t-SNE), Principal Component Analysis (PCA), and Maximum Mean Discrepancy (MMD); when the actual data quality standard meets the preset data quality standard, using the augmented wind power prediction data as the target wind power prediction data set.

[0108] Further, after obtaining the target wind power prediction data set, perform normalization processing on the target wind power prediction data set.

[0109] It should be noted that, on the one hand, for the construction method of the target wind power prediction data set, for the screened extreme weather event data sets under cold snaps, low-temperature freezing, and strong wind conditions, four different methods, namely mRMR (Minimum Redundancy Maximum Relevance), Grey Relational Analysis, Random Forest Algorithm, and Gaussian Distance Metric, are used to comprehensively evaluate and analyze the correlation between all environmental factors associated with wind power output and it, and explore the target environmental factors that have the most significant impact on wind power generation; finally, based on the comprehensive ranking results, select the environmental factors with the highest correlation with wind power, and construct a feature data set based on this, ensuring the accuracy and reliability of the features. On the other hand, use the WGAN-GP model to perform data augmentation on the prediction samples, effectively increasing the number of samples and improving the generalization ability of the model. A variety of methods are used to conduct quality inspection and preprocessing on the augmented sample data, ensuring the accuracy and reliability of the data. These innovative measures together constitute the core advantages of this solution, providing strong data support for the accurate prediction of wind power under complex extreme weather conditions.

[0110] Embodiment 3 provides a system for constructing a wind power prediction data set, including a data acquisition module, a division module, a correlation analysis module, and a screening module;

[0111] Specifically, the data acquisition module is used to obtain the historical meteorological data and historical power data of the target wind farm as the original data set;

[0112] Specifically, the division module is used to divide the original data set according to the first division rule to obtain an extreme weather event data set; where the extreme weather event data set includes sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events;

[0113] Specifically, the correlation analysis module is used to perform correlation feature analysis on the environmental factors included in the sub-historical meteorological data in the extreme weather event data set to obtain the correlation analysis results between each environmental factor and the historical power data;

[0114] Specifically, the screening module is used to screen the extreme weather event dataset based on the correlation analysis result to obtain the target wind power prediction dataset; wherein, the environmental factors corresponding to the target meteorological data in the target wind power prediction dataset have an actual correlation with the historical power data greater than the preset correlation.

[0115] It should be noted that the technical solution of the system for constructing the wind power prediction dataset and the technical solution of the method for constructing the wind power prediction dataset described above belong to the same concept. For the details not described in detail in the technical solution of the system for constructing the wind power prediction dataset in this embodiment, reference can be made to the description of the technical solution of the method for constructing the wind power prediction dataset described above.

[0116] The above-mentioned unit modules can be embedded in the processor in the electronic device in hardware form or be independent of the processor, or can be stored in the memory in the electronic device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0117] This embodiment also provides an electronic device, which includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic 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 communication interface of the electronic device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a carrier network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it realizes the method for constructing the wind power prediction dataset. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device can be a touch layer covered on the display screen, or a button, a trackball, or a touchpad set on the shell of the electronic device, or an external keyboard, a touchpad, or a mouse, etc.

[0118] This embodiment also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by the processor, it realizes the method proposed in the above embodiment.

[0119] The storage medium proposed in this embodiment and the method proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0120] From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general-purpose hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disc of a computer, and includes several instructions for causing an electronic device (which can be a personal computer, server, or network device, etc.) to execute the method of the embodiments of the present invention.

[0121] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

[0122] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can be implemented in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be implemented in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present application can be implemented in various computer languages.

[0123] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0124] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to work in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device that implements the functions specified in one or more of the processes and / or blocks Figure 1 in one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0125] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, such that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the processes and / or blocks Figure 1 in one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0126] Although the preferred embodiments of the present application have been described, additional changes and modifications can be made to these embodiments by those skilled in the art once they learn of the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present application.

[0127] It is obvious that those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application is also intended to include these modifications and variations.

Claims

1. A method for constructing a wind power prediction data set, characterized in that Including: Obtain the historical meteorological data and historical power data of the target wind farm as the original data set; Divide the original data set according to the first division rule to obtain an extreme weather event data set; Wherein the extreme weather event data set includes sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events; In the extreme weather event data set, perform correlation analysis based on the environmental factors included in the sub-historical meteorological data to obtain the correlation analysis results between each environmental factor and the historical power data; Filter the extreme weather event data set based on the correlation analysis results to obtain a target wind power prediction data set.

2. The method for constructing a wind power prediction data set according to claim 1, wherein The step of dividing the original data set according to the first division rule to obtain an extreme weather event data set includes: Obtain various meteorological factor types and meteorological factor thresholds corresponding to preset extreme weather; Divide and filter the historical meteorological data based on the meteorological factor types and the meteorological factor thresholds to obtain sub-historical meteorological data; Extract the historical power data corresponding to the time series of the sub-historical meteorological data as sub-historical power data; Use the sub-historical meteorological data and the sub-historical power data as the extreme weather event data set.

3. The construction method of the wind power prediction data set according to claim 2, characterized in that The acquisition of the correlation analysis results between each environmental factor and the historical power data includes: Perform correlation analysis on the environmental factors based on at least two correlation analysis methods to obtain corresponding sub-correlation analysis results; wherein, the sub-correlation analysis results include the environmental factors and the correlation influence factors corresponding to each of the environmental factors; Sort the environmental factors based on the correlation influence factors to obtain a sorting result; In each of the sub-correlation analysis results, obtain the target environmental factors based on the sorting result; Determine the correlation influence weight of the target environmental factors based on the influence factors corresponding to each of the target environmental factors; wherein, the correlation influence weight is positively correlated with the influence factor; Based on the correlation influence weight and integrating the target environmental factors included in each of the sub-correlation analysis results, obtain the correlation analysis results of each environmental factor relative to the historical power data.

4. The method for constructing a wind power prediction data set according to claim 3, wherein, The acquisition of the target wind power prediction data set includes: Obtain preset environmental factors based on the correlation analysis results; Match the sub-historical meteorological data in the extreme weather event data set and the sub-historical meteorological data corresponding to the time series of the sub-historical meteorological data based on the preset environmental factors as the target wind power prediction data set.

5. The method for constructing a wind power prediction data set according to claim 4, characterized in that, The actual correlation between the environmental factors corresponding to each target meteorological data in the target wind power prediction data set and the historical power data is greater than the preset correlation.

6. The construction method of the wind power prediction data set according to claim 5, characterized in that, After obtaining the target wind power prediction data set, it further includes: Based on the pre-constructed WGAN-GP model, perform at least one round of expansion on the target wind power prediction data set to obtain the expanded wind power prediction data after expansion.

7. The method for constructing a wind power prediction data set according to claim 6, characterized in that, After obtaining the target wind power prediction data set, it further includes: In each round of expansion, determine the actual data quality standard of the output result of the WGAN-GP model; When the actual data quality standard meets the first data quality standard, add the augmented wind power prediction data obtained by augmentation to the target wind power prediction data set.

8. A construction system for a wind power prediction data set, applying the construction method of the wind power prediction data set according to any one of claims 1 to 7, characterized in that, Including: A data acquisition module, configured to acquire historical meteorological data and historical power data of a target wind farm as an original data set; A division module, configured to divide the original data set according to a first division rule to obtain an extreme weather event data set; wherein the extreme weather event data set includes sub-historical meteorological data and sub-historical power data corresponding to multiple types of extreme weather events; A correlation analysis module, configured to perform correlation feature analysis based on environmental factors included in the sub-historical meteorological data in the extreme weather event data set to obtain a correlation analysis result of each environmental factor and the historical power data; A screening module, configured to screen the extreme weather event data set based on the correlation analysis result to obtain a target wind power prediction data set; wherein, for each target meteorological data in the target wind power prediction data set, the actual correlation between the corresponding environmental factor and the historical power data is greater than a preset correlation.

9. An electronic device, comprising a memory and a processor, characterized in that: The memory is used to store computer-executable instructions, and when the processor executes the computer-executable instructions, the steps of the method for constructing a wind power prediction data set according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: When the computer-executable instructions are executed by the processor, the steps of the method for constructing a wind power prediction data set according to any one of claims 1 to 7 are implemented.

Images