Wind speed data processing method and device

Abstract

The invention provides a wind speed data processing method and device. According to the reanalysis data of the target area, meteorological characteristic data of a historical time period matched with the quarter of the to-be-predicted time period and wind speed data of a previous time period of the to-be-predicted time period are acquired; smoothing the wind speed data of the previous time period of the to-be-predicted time period to obtain a first wind speed sequence corresponding to the meteorological characteristic data in the to-be-predicted time period; performing feature combination extraction processing on the meteorological feature data by using a first module of a preset wind speed prediction model to obtain a combined meteorological feature; performing wind speed prediction processing on the combined meteorological characteristics and the first wind speed sequence by using a second module of a preset wind speed prediction model to obtain a predicted wind speed sequence; and determining predicted power corresponding to the predicted wind speed sequence, and generating a fan protection strategy for the fan in the target area according to the predicted power. Therefore, the accuracy of wind speed sequence prediction is improved, and the safety of fan equipment is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of wind and solar energy technology, and in particular to a method and device for processing wind speed data. Background Art

[0002] At present, wind and solar energy technologies, as an important part of clean energy, play a key role in the transformation of the global energy structure. In the application of wind and solar energy, wind speed prediction is a crucial key technology, especially in the prediction of wind turbine power in wind farms, which directly affects the power generation efficiency. However, the traditional wind speed prediction model has large errors under complex terrain and microclimate conditions, resulting in insufficient wind speed prediction accuracy.

[0003] To address the above problems, no effective solution has been proposed yet. Summary of the invention

[0004] The present specification provides a method and device for processing wind speed data, which solves the problems of large errors and low prediction accuracy in existing wind speed prediction models.

[0005] This specification provides a wind speed data processing method, including:

[0006] Acquire meteorological characteristic data of a historical period matching the quarter of the period to be predicted and wind speed data of a period before the period to be predicted based on the reanalysis data of the target area;

[0007] Smoothing the wind speed data of a period before the period to be predicted to obtain a first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data;

[0008] Using the first module of the preset wind speed prediction model, the meteorological characteristic data is subjected to feature combination extraction processing to obtain combined meteorological characteristics;

[0009] Using the second module of the preset wind speed prediction model, the combined meteorological characteristics and the first wind speed sequence are processed for wind speed prediction to obtain a predicted wind speed sequence;

[0010] Determining a predicted power corresponding to the predicted wind speed sequence, and generating a wind turbine protection strategy for wind turbines in the target area based on the predicted power;

[0011] Among them, the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained according to reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

[0012] In one embodiment, before the first module using the preset wind speed prediction model performs feature combination extraction processing on the meteorological feature data to obtain the combined meteorological features, it also includes:

[0013] According to the reanalysis data of each of the first regions, historical meteorological characteristic data of the first region in a historical period having the same seasonal characteristics as the period to be predicted, and wind speed data of a period before the historical period are obtained;

[0014] Smoothing the wind speed data of a period before the historical period to obtain a second wind speed sequence corresponding to the historical meteorological characteristic data;

[0015] Using the first module of the preset wind speed prediction model, performing feature combination extraction processing on the historical meteorological feature data to obtain a first combined meteorological feature;

[0016] Using the second module of the preset wind speed prediction model, performing wind speed prediction processing on the first combined meteorological feature and the second wind speed sequence to obtain a first predicted wind speed sequence;

[0017] Determine whether the preset wind speed prediction model converges based on the first combined meteorological feature, the first predicted wind speed sequence, and the third wind speed sequence. If it is determined that the preset wind speed prediction model has not converged, continue to train the preset wind speed prediction model based on the historical meteorological feature data, the second wind speed sequence, and the third wind speed sequence until the preset wind speed prediction model converges.

[0018] In one embodiment, determining whether the preset wind speed prediction model converges according to the first combined meteorological feature, the first predicted wind speed sequence, and the third wind speed sequence includes:

[0019] Determining a correlation between the first combined meteorological feature and the third wind speed sequence, and a difference between the first predicted wind speed sequence and the third wind speed sequence;

[0020] A loss value is determined according to a preset loss function, the correlation and the difference, and whether the preset wind speed prediction model converges is determined according to the loss value.

[0021] In one embodiment, the determining the predicted power corresponding to the predicted wind speed sequence includes:

[0022] Obtaining fan parameters of the fans in the target area;

[0023] The predicted power corresponding to the predicted wind speed sequence is determined based on the wind turbine parameters and the predicted wind speed sequence using a preset power determination model, wherein the preset power determination model is a model constructed based on a preset machine learning algorithm.

[0024] In one embodiment, the wind turbine protection strategy includes a diversion strategy and / or a voltage limiting strategy.

[0025] In one embodiment, generating a wind turbine protection strategy for wind turbines in the target area according to the predicted power includes:

[0026] When the predicted power is not lower than the preset rated power of the wind turbine, the electric energy generated by the power difference between the predicted power and the preset rated power is distributed to the preset energy storage device through the preset power adjustment device; and / or the blade angle of the wind turbine is controlled to perform voltage reduction processing to generate a wind turbine protection strategy for the wind turbine in the target area.

[0027] In one embodiment, the smoothing process of the wind speed data of the previous period of the period to be predicted to obtain a first wind speed sequence corresponding to the meteorological characteristic data includes:

[0028] Smoothing the wind speed data of a period before the period to be predicted to obtain a candidate wind speed sequence;

[0029] According to a preset sliding window, the candidate wind speed sequence is subjected to outlier detection processing, and when it is detected that the candidate wind speed sequence contains outliers, the outliers contained in the candidate wind speed sequence are processed using a moving average algorithm to obtain a first wind speed sequence corresponding to the meteorological characteristic data.

[0030] This specification provides a wind speed data processing device, comprising:

[0031] A data acquisition module, for acquiring meteorological characteristic data of a historical period matching the quarter of the period to be predicted, and wind speed data of a period before the period to be predicted, based on the reanalysis data of the target area;

[0032] A data processing module, used for smoothing the wind speed data of the previous period of the period to be predicted, to obtain a first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data;

[0033] A feature combination module, used to perform feature combination extraction processing on the meteorological feature data using the first module of the preset wind speed prediction model to obtain combined meteorological features;

[0034] A wind speed prediction module, configured to use the second module of the preset wind speed prediction model to perform wind speed prediction processing on the combined meteorological features and the first wind speed sequence to obtain a predicted wind speed sequence;

[0035] A strategy determination module, used to determine the predicted power corresponding to the predicted wind speed sequence, and generate a wind turbine protection strategy for the wind turbines in the target area according to the predicted power;

[0036] Among them, the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained according to reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

[0037] This specification also provides an electronic device, including a processor and a memory for storing instructions executable by the processor, wherein the processor implements a wind speed data processing method when executing the instructions.

[0038] This specification also provides a computer-readable storage medium on which computer instructions are stored, and when the instructions are executed, a wind speed data processing method is implemented.

[0039] A wind speed data processing method provided in the present specification is based on the re-analysis data of the target area, obtaining meteorological characteristic data of a historical period matching the quarter of the period to be predicted and wind speed data of a period before the period to be predicted; smoothing the wind speed data of the period before the period to be predicted to obtain a first wind speed sequence corresponding to the meteorological characteristic data of the period to be predicted; using a first module of a preset wind speed prediction model, performing feature combination extraction processing on the meteorological characteristic data to obtain combined meteorological features; using a second module of the preset wind speed prediction model, performing wind speed prediction processing on the combined meteorological features and the first wind speed sequence to obtain a predicted wind speed sequence; determining a predicted power corresponding to the predicted wind speed sequence, and generating a wind turbine protection strategy for the wind turbines in the target area according to the predicted power; wherein the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained from the re-analysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area. In this way, on the one hand, the meteorological characteristic data of the historical period matching the quarter of the period to be predicted obtained by the reanalysis data can be combined with the wind speed data of the previous period of the period to be predicted in the target area to accurately predict the predicted wind speed sequence corresponding to the predicted period of the target area. On the other hand, in the prediction process of the wind speed sequence, firstly, the first module of the preset wind speed prediction model can be used to perform feature combination extraction processing to reveal the potential nonlinear relationship and interaction between meteorological characteristics and reduce the interference of irrelevant features. The obtained combined meteorological characteristics can more comprehensively reflect the complex changes of meteorological conditions compared with single meteorological characteristic data. Secondly, the second module of the preset wind speed prediction model can be used to combine the combined meteorological characteristics and the first wind speed sequence for wind speed prediction processing, which can effectively characterize the temporal variation law of wind speed, and at the same time improve the ability of the preset wind speed prediction model to cope with complex meteorological conditions and the accuracy and stability of the prediction results. Finally, according to the predicted power corresponding to the obtained predicted wind speed sequence, a wind turbine protection strategy for the wind turbine in the target area can be generated in a targeted manner to ensure the equipment safety of the wind turbine in the target area through wind turbine protection measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of this specification, the drawings required for use in the embodiments will be briefly introduced below. The drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 is a flow chart of a wind speed data processing method provided by an embodiment of this specification;

[0042] Figure 2 It is a schematic diagram of the structure of an electronic device provided by an embodiment of this specification;

[0043] Figure 3 It is a schematic diagram of the structure of a wind speed data processing device provided by an embodiment of the present specification;

[0044] Figure 4 It is a flow chart of a local wind speed correction method provided by an embodiment of this specification;

[0045] Figure 5 It is a wind speed prediction result comparison schematic diagram provided by an embodiment of this specification;

[0046] Figure 6 is a correlation analysis schematic diagram provided by an embodiment of this specification;

[0047] Figure 7 It is a schematic diagram of an error analysis wind rose provided by an embodiment of this specification;

[0048] Figure 8 It is a schematic diagram of wind resource map analysis results provided by an embodiment of this specification. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0050] Wind speed prediction is crucial in wind energy resource assessment and weather forecasting, especially in wind farm site selection and power forecasting, which directly affects power generation efficiency. However, traditional wind speed prediction models have large errors under complex terrain and microclimate conditions, resulting in insufficient wind speed prediction accuracy. Existing wind speed prediction methods mainly rely on numerical weather forecast models (such as Weather Research and Forecasting Model, WRF) and reanalysis data (such as ECMWF Reanalysis 5th Generation, ERA5). Although the WRF model has a high spatial resolution, its accuracy decreases under microclimate conditions. Although the reanalysis data has a wide coverage, it is not timely enough and it is difficult to reflect local meteorological details.

[0051] Wind towers can provide local measured wind speeds, but due to the limited number of wind towers, spatial representation is insufficient. Traditional wind speed forecasting technology has shortcomings in single data source and model construction, and fails to effectively integrate multi-source data to improve forecast accuracy.

[0052] The existing technology has certain technical defects. The first is the limitation of a single data source. Traditional methods mostly rely on WRF model simulations. Although they can provide high-resolution forecast data, they are insufficient in capturing local wind speeds, especially under complex terrain and microclimate conditions. Second, when using the ERA5 reanalysis data, it is difficult to reflect the detailed characteristics of local wind speeds due to the low resolution. Existing wind speed corrections are mostly based on linear models, which cannot capture complex nonlinear relationships and have limited correction effects. Under dynamic meteorological conditions, the model's adaptability is weak, and real-time updates and long-term stable forecasts cannot be achieved.

[0053] In view of the root cause of the above problems, this specification considers that on the one hand, the meteorological characteristic data of the historical period matching the quarter of the period to be predicted obtained by the reanalysis data can be combined with the wind speed data of the previous period of the period to be predicted in the target area to accurately predict the predicted wind speed sequence corresponding to the predicted period of the target area. On the other hand, in the prediction process of the wind speed sequence, firstly, the first module of the preset wind speed prediction model can be used to perform feature combination extraction processing to reveal the potential nonlinear relationship and interaction between meteorological characteristics and reduce the interference of irrelevant features. The obtained combined meteorological characteristics can more comprehensively reflect the complex changes of meteorological conditions compared with single meteorological characteristic data. Secondly, the second module of the preset wind speed prediction model can be used to combine the combined meteorological characteristics and the first wind speed sequence for wind speed prediction processing, which can effectively characterize the temporal variation law of wind speed, and at the same time improve the ability of the preset wind speed prediction model to cope with complex meteorological conditions and the accuracy and stability of the prediction results. Finally, according to the predicted power corresponding to the obtained predicted wind speed sequence, a wind turbine protection strategy for the wind turbine in the target area can be generated in a targeted manner to ensure the equipment safety of the wind turbine in the target area through wind turbine protection measurement.

[0054] See also Figure 1 As shown, the embodiment of this specification provides a method for processing wind speed data, wherein the method is specifically applied to a server side. When implemented specifically, the method may include the following contents:

[0055] S101: Acquire meteorological characteristic data of a historical period matching the season of the period to be predicted, and wind speed data of a period before the period to be predicted, based on the reanalysis data of the target area;

[0056] S102: Smoothing the wind speed data of the previous period of the period to be predicted to obtain a first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data;

[0057] S103: using the first module of the preset wind speed prediction model, performing feature combination extraction processing on the meteorological feature data to obtain combined meteorological features;

[0058] S104: using the second module of the preset wind speed prediction model, performing wind speed prediction processing on the combined meteorological features and the first wind speed sequence to obtain a predicted wind speed sequence;

[0059] S105: Determine the predicted power corresponding to the predicted wind speed sequence, and generate a wind turbine protection strategy for the wind turbines in the target area based on the predicted power; wherein the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained based on reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

[0060] Among them, the above-mentioned reanalysis data can be ERA5 reanalysis data, which can provide historical weather data covering the whole world. For example, ERA5 reanalysis data can include temperature, precipitation, wind speed and other weather data recorded with high temporal and spatial resolution (hourly level, 31 kilometers horizontal resolution).

[0061] The above-mentioned meteorological characteristic data may refer to various observational or calculated data related to atmospheric conditions screened out from the reanalysis data. Specifically, the meteorological characteristic data may include characteristic data such as wind speed, wind direction, temperature, air pressure, humidity, precipitation, solar radiation, and cloud cover.

[0062] In some embodiments, taking the fourth quarter of this year as the period to be predicted and the period before the prediction period as the third quarter of this year as an example, the meteorological characteristic data of the historical period matching the quarter of the period to be predicted and the wind speed data of the period before the period to be predicted are obtained according to the reanalysis data of the target area. When implemented specifically, the following may be included:

[0063] Based on the reanalysis data of the target area, the meteorological characteristic data of the fourth quarter of last year and the wind speed data of the third quarter of this year were obtained.

[0064] In this way, the wind speed sequence of the time period to be predicted can be accurately predicted by using the meteorological characteristic data that can characterize the meteorological characteristics of the time period to be predicted and the wind speed data that can characterize the wind speed characteristics of the target area in the previous time period.

[0065] In some embodiments, the smoothing of the wind speed data of the previous period of the period to be predicted to obtain the first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data may include:

[0066] The wind speed data is smoothed using a preset smoothing algorithm to obtain a first wind speed sequence corresponding to the predicted time period and the meteorological characteristic data; wherein the wind speed data is collected during the previous time period at a preset fixed interval by a preset wind speed measuring device.

[0067] In some embodiments, the first module of the preset wind speed prediction model is used to perform feature combination extraction processing on the meteorological feature data to obtain combined meteorological features. When specifically implemented, it may include:

[0068] S1: using the first module of the preset wind speed prediction model, performing feature combination extraction processing on the meteorological feature data, and constructing secondary meteorological features and / or cross meteorological features of the meteorological feature data through a preset feature combination algorithm;

[0069] S2: Determine the combined meteorological feature according to the secondary meteorological feature and / or cross meteorological feature of the meteorological feature data.

[0070] Among them, the above-mentioned secondary meteorological characteristics can refer to new features obtained by square or other mathematical transformation of a single meteorological variable (such as wind speed, temperature, etc.), while the above-mentioned cross-meteorological characteristics are composite features formed by combining the interactions of multiple meteorological variables (such as the combination of wind speed and air pressure). These secondary meteorological characteristics and cross-meteorological characteristics can reveal the complex nonlinear relationship between meteorological variables, thereby providing more information dimensions for wind speed prediction and increasing the sensitivity of the model to meteorological changes. Through these enhanced meteorological characteristics, richer and more diverse combined meteorological characteristics can be constructed, providing more accurate and comprehensive input data for subsequent wind speed prediction.

[0071] In this way, by introducing secondary and cross-meteorological features, the model can capture the complex interactions between meteorological variables and improve the accuracy and flexibility of wind speed forecasting. These optimized meteorological features can help the model better understand the impact of meteorological conditions on wind speed changes and improve the adaptability and accuracy of wind speed forecasting under variable weather conditions.

[0072] In some embodiments, the first module of the preset wind speed prediction model can be constructed according to a feature extraction algorithm, and the second module can be constructed according to a preset machine learning algorithm. For example, the first module can be a module constructed according to feature extraction algorithms such as neural network algorithms and transformer algorithms, and the second module can be a module constructed according to machine learning algorithms such as time series neural network algorithms and LSTM algorithms.

[0073] In some embodiments, the target area and the first area may have the same terrain features. For example, the altitudes of the target area and the first area may be in the same altitude range, or the target area and the first area may both be plateau terrains.

[0074] In some embodiments, the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained from reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

[0075] In this way, on the one hand, by using historical meteorological characteristic data and the corresponding second wind speed sequence, a deep mapping relationship between wind speed and meteorological conditions can be established. On the other hand, by combining extensive historical data (i.e., reanalysis data) and data collected by wind towers, the correlation between wind speed time series changes and meteorological factors can be extracted, so that the preset wind speed prediction model has a certain generalization ability and better wind speed prediction performance.

[0076] In some embodiments, when applied to a target area, the preset wind speed prediction model can also be adjusted and optimized in combination with the meteorological characteristic data of the target area. Through the method of transfer learning, the learning results of the first area can be transferred to the target area, making full use of the existing training results and reducing the dependence on the amount of data in the target area.

[0077] Specifically, training is performed based on historical meteorological characteristic data obtained from reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area to obtain an intermediate model of the multiple first areas, and then based on the intermediate model, through transfer learning, the preset wind speed prediction model that can be applied to the target area is determined.

[0078] In some embodiments, after using the second module of the preset wind speed prediction model to perform wind speed prediction processing on the combined meteorological characteristics and the first wind speed sequence to obtain a predicted wind speed sequence, the specific implementation may also include: determining the location information of the wind farm in the target area based on the predicted wind speed sequence.

[0079] Specifically, the predicted wind speed sequence can be used to characterize the wind speed distribution at various locations in the target area. By performing regional identification processing on the predicted wind speed sequences at different locations in the target area, areas with higher wind speeds and greater stability can be identified. These areas may have relatively abundant wind energy resources and may have greater power generation potential. Therefore, the location information of these areas can be determined as the location information of the wind farm.

[0080] In this way, by performing regional identification processing on the predicted wind speed sequence, the exploratory survey and repeated evaluation process that may occur in the traditional wind farm site selection can be reduced, the site selection error can be reduced, time and cost can be saved, and the accuracy and efficiency of wind farm site selection can be improved. At the same time, wind resources can be used to the maximum extent to ensure that the wind farm can achieve higher power generation efficiency and economic benefits within the forecast period, thereby promoting the smooth implementation and sustainable development of renewable energy projects.

[0081] Based on the above embodiment, on the one hand, the meteorological characteristic data of the historical period matching the quarter of the period to be predicted obtained by the reanalysis data can be combined with the wind speed data of the previous period of the period to be predicted in the target area to accurately predict the predicted wind speed sequence corresponding to the predicted period of the target area. On the other hand, in the prediction process of the wind speed sequence, firstly, the first module of the preset wind speed prediction model can be used to perform feature combination extraction processing to reveal the potential nonlinear relationship and interaction between meteorological characteristics and reduce the interference of irrelevant features. The obtained combined meteorological characteristics can more comprehensively reflect the complex changes of meteorological conditions compared with single meteorological characteristic data. Secondly, the second module of the preset wind speed prediction model can be used to combine the combined meteorological characteristics and the first wind speed sequence for wind speed prediction processing, which can effectively characterize the temporal variation law of wind speed, and at the same time improve the ability of the preset wind speed prediction model to cope with complex meteorological conditions and the accuracy and stability of the prediction results. Finally, according to the predicted power corresponding to the obtained predicted wind speed sequence, a wind turbine protection strategy for the wind turbine in the target area can be generated in a targeted manner to ensure the equipment safety of the wind turbine in the target area through wind turbine protection measurement.

[0082] In some embodiments, before the first module of the preset wind speed prediction model is used to perform feature combination extraction processing on the meteorological feature data to obtain the combined meteorological features, the method may further include the following contents when it is specifically implemented:

[0083] S1: acquiring, based on the reanalysis data of each of the first regions, historical meteorological characteristic data of the first region in a historical period having the same seasonal characteristics as the period to be predicted, and wind speed data of a period before the historical period;

[0084] S2: Smoothing the wind speed data of the previous period of the historical period to obtain a second wind speed sequence corresponding to the historical meteorological characteristic data;

[0085] S3: using the first module of the preset wind speed prediction model, performing feature combination extraction processing on the historical meteorological feature data to obtain a first combined meteorological feature;

[0086] S4: using the second module of the preset wind speed prediction model, performing wind speed prediction processing on the first combined meteorological feature and the second wind speed sequence to obtain a first predicted wind speed sequence;

[0087] S5: Determine whether the preset wind speed prediction model converges based on the first combined meteorological feature, the first predicted wind speed sequence and the third wind speed sequence; if it is determined that the preset wind speed prediction model has not converged, continue to train the preset wind speed prediction model based on the historical meteorological feature data, the second wind speed sequence and the third wind speed sequence until the preset wind speed prediction model converges.

[0088] In some embodiments, the first module is used to train a module constructed by a preset feature extraction algorithm based on the historical meteorological feature data and the third wind speed sequence; the second module is used to train a module constructed by a preset machine learning algorithm based on the output data of the first module, the second wind speed sequence and the third wind speed sequence.

[0089] In some embodiments, the method of determining whether the preset wind speed prediction model converges according to the first combined meteorological feature, the first predicted wind speed sequence, and the third wind speed sequence may further include the following when implemented:

[0090] S1: determining a correlation between the first combined meteorological feature and the third wind speed sequence, and a difference between the first predicted wind speed sequence and the third wind speed sequence;

[0091] S2: Determine a loss value according to a preset loss function, the correlation and the difference, and determine whether the preset wind speed prediction model converges according to the loss value.

[0092] Among them, the above-mentioned correlation can be a measure of the strength of the relationship between meteorological characteristics and wind speed, and can be quantified using a correlation coefficient (such as the Pearson correlation coefficient), that is, it can be used to represent the intrinsic correlation between the first combined meteorological characteristics (such as wind speed, temperature, etc.) and the third wind speed sequence. For example, the stronger the correlation between the first combined meteorological characteristics and the third wind speed sequence, the more accurate the prediction of the preset wind speed prediction model.

[0093] The above difference may be a root mean square error and / or a mean absolute error between the first predicted wind speed sequence and the third wind speed sequence, and is used to evaluate the accuracy of the wind speed predicted by the preset wind speed prediction model.

[0094] In some embodiments, the method for determining the predicted power corresponding to the predicted wind speed sequence may further include the following when implemented:

[0095] S1: Obtaining fan parameters of the fan in the target area;

[0096] S2: Determine the predicted power corresponding to the predicted wind speed sequence based on the wind turbine parameters and the predicted wind speed sequence using a preset power determination model, wherein the preset power determination model is a model constructed based on a preset machine learning algorithm.

[0097] The above-mentioned fan parameters may include rated power, cut-in wind speed, cut-out wind speed, fan speed, fan efficiency, etc.

[0098] In some embodiments, the use of a preset power determination model to determine the predicted power corresponding to the predicted wind speed sequence based on the wind turbine parameters and the predicted wind speed sequence may include:

[0099] The predicted wind speed sequence is inputted by using a preset power determination model (such as a mathematical model based on a wind speed-power curve), and combined with the wind turbine parameters (such as rated power, cut-in wind speed, cut-out wind speed, wind turbine efficiency, etc.), to determine the theoretical power output at each moment. The predicted wind speed sequence represents the changing trend of the future wind speed. By matching the model with the wind turbine parameters, the predicted power under different wind speed conditions can be obtained. For example, when the predicted wind speed is between the cut-in wind speed and the rated wind speed, the preset power model will calculate the corresponding power value based on the wind speed and efficiency, and when the predicted wind speed exceeds the rated wind speed, the preset power model will output power according to the maximum power limit of the wind turbine.

[0100] Based on the above embodiment, by combining wind turbine parameters with predicted wind speed sequence to determine the predicted power of wind turbines, not only the accuracy of wind farm power prediction is improved, but also the resource scheduling and operation management of wind farms can be optimized. Accurate power prediction enables wind farms to effectively reduce wind energy waste, improve power generation efficiency, and reduce equipment loss and safety risks caused by overspeed or underspeed operation.

[0101] In some embodiments, when the method is implemented, it may also include the following contents: the wind turbine protection strategy includes a diversion strategy and / or a voltage limiting strategy.

[0102] In some embodiments, the method of generating a wind turbine protection strategy for wind turbines in the target area according to the predicted power may further include the following when the method is specifically implemented:

[0103] When the predicted power is not lower than the preset rated power of the wind turbine, the electric energy generated by the power difference between the predicted power and the preset rated power is distributed to the preset energy storage device through the preset power adjustment device; and / or the blade angle of the wind turbine is controlled to perform voltage reduction processing to generate a wind turbine protection strategy for the wind turbine in the target area.

[0104] Specifically, when the predicted power is not lower than the preset rated power of the wind turbine, in order to avoid overload damage to the wind turbine, the wind turbine protection strategy can realize dynamic adjustment of power through a preset power adjustment device. Specifically, when the predicted power exceeds the rated power of the wind turbine, the system can adjust the power difference and divert the excess power to a preset energy storage device, such as a battery or an energy storage system. The preset energy storage device can temporarily receive excess electrical energy to avoid grid fluctuations or excessive load on the wind turbine. At the same time, the preset energy storage device can also release the stored electrical energy when the wind speed is low or when the grid demand increases, providing additional support for the grid. This diversion and energy storage strategy effectively protects the wind turbine from high load damage while optimizing the utilization efficiency of wind power resources.

[0105] Furthermore, if the predicted power of the wind turbine continues to approach or exceed the rated power, blade angle control (also known as variable pitch control) can be used for voltage reduction. By adjusting the angle of the wind turbine blades, the efficiency of the wind turbine in capturing wind energy is reduced, thereby reducing the input wind energy and avoiding mechanical damage caused by excessive wind turbine output power. Specifically, increasing the windward angle of the blades can reduce the effective force area of ​​the wind turbine blades, thereby reducing the speed and power output of the wind turbine and ensuring that the wind turbine operates within a safe range. This voltage reduction process not only protects the wind turbine, but also can smoothly adjust the power output of the wind turbine when the wind speed is high, thereby improving the stability and reliability of the system. This series of wind turbine protection strategies provides important guarantees for the long-term operation of wind farms, ensuring the safe and efficient operation of wind turbines under different meteorological conditions.

[0106] In some embodiments, the wind speed data of the previous period of the period to be predicted is smoothed to obtain a first wind speed sequence corresponding to the meteorological characteristic data. When the method is specifically implemented, the following contents may also be included:

[0107] S1: Smoothing the wind speed data of the previous period of the period to be predicted to obtain a candidate wind speed sequence;

[0108] S2: According to a preset sliding window, the candidate wind speed sequence is subjected to outlier detection processing, and when it is detected that the candidate wind speed sequence contains outliers, the outliers contained in the candidate wind speed sequence are processed using a moving average algorithm to obtain a first wind speed sequence corresponding to the meteorological characteristic data.

[0109] Specifically, by applying a smoothing algorithm (such as a simple moving average or an exponentially weighted average) to eliminate short-term fluctuations or noise in the wind speed data, the data can be made more stable and consistent. In this way, the candidate wind speed sequence can better reflect the overall trend and change pattern of the wind speed, providing more stable input data for subsequent predictions. This helps to reduce the prediction error caused by instantaneous fluctuations or data noise and improve the prediction accuracy of the model.

[0110] Furthermore, a preset sliding window technology is used to perform outlier detection processing on the candidate wind speed sequence to identify the unconventional outliers in the sequence. The outliers are caused by measurement errors, meteorological interference and other factors. After the outliers are detected, the moving average algorithm is used to smooth the outliers, and these extremely fluctuating data values ​​are adjusted to reasonable values ​​that are more in line with the overall trend, ensuring that each data point in the candidate wind speed sequence reflects normal wind speed changes. Finally, the processed wind speed sequence obtained is the first wind speed sequence corresponding to the meteorological characteristic data, which can provide accurate and reliable data support for subsequent wind speed predictions.

[0111] Based on the above embodiments, by smoothing the wind speed data and detecting outliers, the quality and stability of the data can be significantly improved, and the interference of noise or abnormal fluctuations on the prediction results can be reduced. This processing helps to ensure the accuracy and consistency of wind speed data, provide more reliable input for the wind speed prediction model, and thus improve the prediction accuracy.

[0112] As can be seen from the above, a wind speed data processing method provided in an embodiment of the present specification obtains meteorological characteristic data of a historical period matching the quarter of the period to be predicted and wind speed data of a period before the period to be predicted based on re-analysis data of a target area; smoothes the wind speed data of the period before the period to be predicted to obtain a first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data; uses a first module of a preset wind speed prediction model to perform feature combination extraction processing on the meteorological characteristic data to obtain combined meteorological features; uses a second module of the preset wind speed prediction model to perform wind speed prediction processing on the combined meteorological features and the first wind speed sequence to obtain a predicted wind speed sequence; determines a predicted power corresponding to the predicted wind speed sequence, and generates a wind turbine protection strategy for the wind turbine in the target area based on the predicted power; wherein the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained from re-analysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area. In this way, on the one hand, the meteorological characteristic data of the historical period matching the quarter of the period to be predicted obtained by the reanalysis data can be combined with the wind speed data of the previous period of the period to be predicted in the target area to accurately predict the predicted wind speed sequence corresponding to the predicted period of the target area. On the other hand, in the prediction process of the wind speed sequence, firstly, the first module of the preset wind speed prediction model can be used to perform feature combination extraction processing to reveal the potential nonlinear relationship and interaction between meteorological characteristics and reduce the interference of irrelevant features. The obtained combined meteorological characteristics can more comprehensively reflect the complex changes of meteorological conditions compared with single meteorological characteristic data. Secondly, the second module of the preset wind speed prediction model can be used to combine the combined meteorological characteristics and the first wind speed sequence for wind speed prediction processing, which can effectively characterize the temporal variation law of wind speed, and at the same time improve the ability of the preset wind speed prediction model to cope with complex meteorological conditions and the accuracy and stability of the prediction results. Finally, according to the predicted power corresponding to the obtained predicted wind speed sequence, a wind turbine protection strategy for the wind turbine in the target area can be generated in a targeted manner to ensure the equipment safety of the wind turbine in the target area through wind turbine protection measurement.

[0113] See also Figure 2 As shown, the embodiment of this specification also provides a specific electronic device, wherein the electronic device includes a network communication port 201, a processor 202 and a memory 203, and the above structures are connected through internal cables so that each structure can perform specific data interaction.

[0114] The network communication port 201 can specifically obtain meteorological characteristic data of a historical period matching the season of the period to be predicted, and wind speed data of a period before the period to be predicted, based on the reanalysis data of the target area.

[0115] The processor 202 can be specifically used to smooth the wind speed data of the previous time period of the time period to be predicted to obtain a first wind speed sequence corresponding to the meteorological characteristic data of the time period to be predicted; use the first module of the preset wind speed prediction model to perform feature combination extraction processing on the meteorological characteristic data to obtain combined meteorological features; use the second module of the preset wind speed prediction model to perform wind speed prediction processing on the combined meteorological features and the first wind speed sequence to obtain a predicted wind speed sequence; determine the predicted power corresponding to the predicted wind speed sequence, and generate a wind turbine protection strategy for the wind turbine in the target area based on the predicted power; wherein the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained according to re-analysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

[0116] The memory 203 may be specifically used to store corresponding instruction programs.

[0117] Based on the above method, the relevant structural performance of the electronic equipment can be effectively utilized, the data processing speed of the electronic equipment can be improved, and the wind speed data processing method can be efficiently implemented.

[0118] In this embodiment, the network communication port 201 can be a virtual port that is bound to different communication protocols so that different data can be sent or received. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. In addition, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM, CDMA, etc.; it can also be a Wifi chip; it can also be a Bluetooth chip.

[0119] In this embodiment, the processor 202 may be implemented in any appropriate manner. For example, the processor may take the form of a microprocessor or a processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) executable by the (micro)processor, a logic gate, a switch, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, etc. This specification does not limit this.

[0120] In this embodiment, the memory 203 may include multiple levels. In a digital system, anything that can store binary data can be a memory; in an integrated circuit, a circuit with a storage function that has no physical form is also called a memory, such as RAM, FIFO, etc.; in a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0121] The embodiment of the present specification also provides a computer-readable storage medium based on the above-mentioned wind speed data processing method, wherein the computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed, the following are achieved: obtaining meteorological characteristic data of a historical period matching the quarter of the period to be predicted and wind speed data of a period before the period to be predicted according to the reanalysis data of the target area; smoothing the wind speed data of the period before the period to be predicted to obtain a first wind speed sequence corresponding to the meteorological characteristic data of the period to be predicted; performing feature combination extraction processing on the meteorological characteristic data using a first module of a preset wind speed prediction model to obtain a combined meteorological feature; performing wind speed prediction processing on the combined meteorological feature and the first wind speed sequence using a second module of the preset wind speed prediction model to obtain a predicted wind speed sequence; determining a predicted power corresponding to the predicted wind speed sequence, and generating a wind turbine protection strategy for the wind turbine in the target area according to the predicted power; wherein the preset wind speed prediction model is obtained by training the historical meteorological characteristic data obtained according to the reanalysis data of multiple first areas, the second wind speed sequence corresponding to the historical meteorological characteristic data, and the third wind speed sequence of the first area collected by a wind tower in the first area.

[0122] In this embodiment, the storage medium includes but is not limited to a random access memory (RAM), a read-only memory (ROM), a cache, a hard disk (HDD) or a memory card. The memory may be used to store computer program instructions. The network communication unit may be an interface for network connection communication set in accordance with the standard specified by the communication protocol.

[0123] In this embodiment, the functions and effects specifically implemented by the program instructions stored in the computer-readable storage medium can be explained in comparison with other implementations and will not be described in detail here.

[0124] See also Figure 3 At the software level, the embodiment of this specification also provides a wind speed data processing device, which may specifically include the following structural modules:

[0125] The data acquisition module 301 is used to acquire the meteorological characteristic data of the historical period matching the quarter of the period to be predicted and the wind speed data of the period before the period to be predicted according to the reanalysis data of the target area;

[0126] The data processing module 302 is used to perform smoothing processing on the wind speed data of the previous period of the period to be predicted to obtain a first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data;

[0127] A feature combination module 303 is used to perform feature combination extraction processing on the meteorological feature data using the first module of the preset wind speed prediction model to obtain combined meteorological features;

[0128] A wind speed prediction module 304, configured to use the second module of the preset wind speed prediction model to perform wind speed prediction processing on the combined meteorological features and the first wind speed sequence to obtain a predicted wind speed sequence;

[0129] A strategy determination module 305 is used to determine the predicted power corresponding to the predicted wind speed sequence, and generate a wind turbine protection strategy for the wind turbines in the target area according to the predicted power;

[0130] Among them, the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained according to reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

[0131] In some embodiments, before the above-mentioned feature combination module 303, during specific implementation, historical meteorological feature data of the historical period with the same seasonal characteristics as the period to be predicted in the first area and wind speed data of the previous period of the historical period are obtained according to the reanalysis data of each of the first areas; the wind speed data of the previous period of the historical period is smoothed to obtain a second wind speed sequence corresponding to the historical meteorological feature data; the historical meteorological feature data is subjected to feature combination extraction processing using the first module of the preset wind speed prediction model to obtain a first combined meteorological feature; the first combined meteorological feature and the second wind speed sequence are subjected to wind speed prediction processing using the second module of the preset wind speed prediction model to obtain a first predicted wind speed sequence; a convergence determination module is used to determine whether the preset wind speed prediction model converges according to the first combined meteorological feature, the first predicted wind speed sequence and the third wind speed sequence; when it is determined that the preset wind speed prediction model has not converged, the preset wind speed prediction model is continued to be trained according to the historical meteorological feature data, the second wind speed sequence, and the third wind speed sequence until the preset wind speed prediction model converges.

[0132] In some embodiments, the above-mentioned convergence determination module, when implemented, determines the correlation between the first combined meteorological feature and the third wind speed sequence, and the difference between the first predicted wind speed sequence and the third wind speed sequence; determines the loss value according to the preset loss function, the correlation and the difference, and determines whether the preset wind speed prediction model converges according to the loss value.

[0133] In some embodiments, during specific implementation, the wind turbine protection strategy includes a diversion strategy and / or a voltage limiting strategy.

[0134] In some embodiments, the above-mentioned strategy determination module 305, when implemented specifically, obtains the wind turbine parameters of the wind turbine in the target area; uses a preset power determination model to determine the predicted power corresponding to the predicted wind speed sequence based on the wind turbine parameters and the predicted wind speed sequence, and the preset power determination model is a model built based on a preset machine learning algorithm. In the case where the predicted power is not lower than the preset rated power of the wind turbine, the electric energy generated by the power difference between the predicted power and the preset rated power is allocated to a preset energy storage device through a preset power adjustment device; and / or the blade angle of the wind turbine is controlled to perform a voltage reduction process to generate a wind turbine protection strategy for the wind turbine in the target area.

[0135] In some embodiments, the above-mentioned data processing module 302, when implemented, performs smoothing processing on the wind speed data of the previous time period of the time period to be predicted to obtain a candidate wind speed sequence; performs outlier detection processing on the candidate wind speed sequence according to a preset sliding window, and when it is detected that the candidate wind speed sequence contains outliers, uses a moving average algorithm to process the outliers contained in the candidate wind speed sequence to obtain a first wind speed sequence corresponding to the meteorological characteristic data.

[0136] It should be noted that the units, devices or modules described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. For the convenience of description, the above devices are described separately by functions divided into various modules. Of course, when implementing this specification, the functions of each module can be implemented in the same or more software and / or hardware, or the modules that implement the same function can be implemented by a combination of multiple sub-modules or sub-units. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0137] As can be seen from the above, based on a wind speed data processing device provided by the embodiment of this specification, on the one hand, the meteorological characteristic data of the historical period matching the quarter of the period to be predicted obtained by the reanalysis data can be combined with the wind speed data of the previous period of the period to be predicted in the target area to accurately predict the predicted wind speed sequence corresponding to the predicted period of the target area. On the other hand, in the prediction process of the wind speed sequence, first, the first module of the preset wind speed prediction model can be used to perform feature combination extraction processing to reveal the potential nonlinear relationship and interaction between meteorological features, reduce the interference of irrelevant features, and the obtained combined meteorological features can more comprehensively reflect the complex changes of meteorological conditions compared to single meteorological characteristic data. Secondly, using the second module of the preset wind speed prediction model, the wind speed prediction processing can be combined with the combined meteorological features and the first wind speed sequence, which can effectively characterize the temporal variation law of wind speed, and at the same time improve the ability of the preset wind speed prediction model to cope with complex meteorological conditions and the accuracy and stability of the prediction results. Finally, according to the predicted power corresponding to the obtained predicted wind speed sequence, a wind turbine protection strategy for the wind turbine in the target area can be generated in a targeted manner to ensure the equipment safety of the wind turbine in the target area through wind turbine protection measurement.

[0138] In a specific scenario example, a wind speed data processing method and device provided in this specification can be applied to solve the problems of large error and low prediction accuracy in the existing wind speed prediction model. The specific implementation process may include the following contents.

[0139] S1: Use distributed database architecture to store massive data;

[0140] The NoSql database HBase distributed data storage cluster is used to store massive amounts of wind measurement data to support high-concurrency queries. Each data record mainly includes wind tower information, longitude and latitude, temperature, wind speed, wind direction, temperature, humidity and other fields, which are partitioned and stored according to region and time period to improve query processing efficiency.

[0141] S2: Data collection - actual wind tower data collection;

[0142] By obtaining the distributed database, wind towers are set up in the actual project target area to collect data such as wind speed, wind direction, temperature, humidity, etc. These wind towers are distributed in different terrains to cover the meteorological characteristics of the target area.

[0143] S3: Wind tower data cleaning and anomaly detection;

[0144] There may be missing values ​​in the massive wind tower data in the distributed database. In order to prevent the accuracy of the generated correction model, the present invention introduces sliding window and time series trend analysis, combines the box plot method (IQR) to detect outliers in the wind speed data of the wind tower, and uses rolling mean or model completion method to process the abnormal data.

[0145] S4: Obtain reanalysis data;

[0146] Obtain high-quality reanalysis data from ERA5, including regional wind speed, temperature, humidity and other meteorological elements, to provide historical meteorological background data for the model. The temporal resolution is hourly and the spatial resolution is about 31 km. Store the reanalysis data in a distributed data cluster. The reanalysis data may have a low resolution, but it can reflect long-term climate trends and enhance the applicability of the calibration model.

[0147] S5: Preprocessing of all data;

[0148] Linear interpolation was used to fill missing data to ensure data integrity. Standard methods were used to remove outliers to improve the reliability of subsequent corrections. For spatial discontinuities between data, inverse distance weighted (IDW) method was used for spatial interpolation.

[0149] S6: Time-align the wind resource data in the data storage cluster;

[0150] Ensure that the two types of data are synchronized for comparison and analysis. Smooth the wind resource data to eliminate the impact of extreme weather on the model and ensure that the model can accurately reflect the long-term wind speed trend. For time, use the python programming language through the pandas module to perform time index operations on the data.

[0151] S7: Feature extraction;

[0152] Extract multiple feature variables from wind tower and ERA5 data, including wind speed, wind direction, temperature, humidity, air pressure and other meteorological features to ensure the diversity of model input. Select the most relevant features through correlation analysis (such as Pearson correlation coefficient), and remove redundant or low-correlation features to reduce the computational complexity of the model and improve accuracy. Construct secondary features (such as wind speed squared) and cross features (such as wind speed × temperature), and optimize the fitting effect of the machine learning model by adding new feature combinations.

[0153] S8: Model error analysis;

[0154] Initial wind speed forecast data is generated by WRF and compared with the measured data of the wind tower. The root mean square error (RMSE) and mean absolute error (MAE) are calculated to identify the error characteristics under different meteorological conditions and time periods. In the error analysis process, the error values ​​are calculated for different time periods and weather conditions (such as sunny days, rainy days, and nights) to identify the differences in the adaptability of the model under various conditions.

[0155] S9: Dynamic update of the correction model;

[0156] Retrain and calibrate the model at regular intervals (e.g., one month) to ensure that the model parameters can be dynamically updated with meteorological conditions and seasonal changes. During retraining, the updated data set is the wind measurement data and WRF output data containing new data to ensure that the model maintains high evaluation accuracy in long-term operation.

[0157] S10: Build service interface.

[0158] Develop a flexible API interface to support the correction model called by external systems and realize the integrated service of correction estimation. Use a python-based framework to support high-performance and low-latency service requests. Use uvicorn to deploy the API and enable asynchronous mode. In high-concurrency scenarios, combine with Nginx to achieve load balancing.

[0159] In some embodiments, see Figure 4 As shown, a flow chart of the local wind speed correction method is shown, which specifically includes: obtaining wind tower data, collecting reanalysis data, data preprocessing, establishing a correction model (i.e., presetting a wind speed prediction model), correcting wind speed prediction (i.e., predicting wind speed sequence), and outputting wind speed prediction results.

[0160] Specifically, first, the real wind speed data is obtained through the wind tower (step 1), and the relevant reanalysis data is collected at the same time to conduct a more comprehensive wind speed analysis (step 2). Then, the collected data is preprocessed to ensure that it is suitable for model input (step 3). Then, a correction model is established based on the preprocessed data (step 4), and the corrected wind speed is calculated using the model (step 5). Finally, the corrected wind speed is output as the wind speed prediction result (step 6), providing more accurate data support for the management and decision-making of wind energy resources.

[0161] In some embodiments, see Figure 5 As shown, the measured wind speed (i.e., the third wind speed sequence collected by the wind tower), the wind speed prediction before simulation (i.e., the first wind speed sequence), and the wind speed prediction after correction (i.e., the predicted wind speed sequence) are displayed.

[0162] In some embodiments, see Figure 6 As shown, the correlation between the first combined meteorological feature and the third wind speed sequence is demonstrated.

[0163] In some embodiments, see Figure 7 As shown, the difference between the first predicted wind speed sequence and the third wind speed sequence is simulated to obtain an error analysis wind rose diagram.

[0164] In some embodiments, see Figure 8 As shown, the second module of the preset wind speed prediction model is used to perform wind speed prediction processing on the combined meteorological characteristics and the first wind speed sequence to obtain a resource spectrum analysis result diagram of the predicted wind speed sequence.

[0165] Based on the above embodiments, the accuracy and reliability of local wind speed prediction are significantly improved, especially for complex terrain and microclimate areas. Specific technical effects include: improving the accuracy of wind speed prediction: by integrating wind tower data, ERA5 reanalysis data and WRF model output, integrating big data analysis technology and machine learning model correction, reducing the systematic deviation of WRF model prediction, improving the spatial resolution and detail capture ability of wind speed prediction, and realizing the accurate integration of multi-source data. Enhance the stability and adaptability of the model: construct a machine learning correction model that can adapt to complex meteorological conditions. Through in-depth analysis of error characteristics, the correction model shows a high degree of stability under various meteorological conditions, and is suitable for stable prediction capabilities under long-term and changeable meteorological conditions. Real-time adjustment and dynamic update: based on big data real-time technology, through real-time data integration and model retraining, ensure that the model is automatically updated as meteorological conditions change. Through an automated model retraining mechanism, the model parameters are continuously optimized to adapt to changes in meteorological conditions and meet the needs of wind farms and other areas that require high-precision predictions. Reliability in dealing with data and model anomalies: A complete anomaly detection and correction mechanism is built to provide rapid response and repair capabilities for data anomalies (such as sudden changes in wind resource data) and model anomalies (such as prediction deviations from the confidence interval), ensuring the stable operation of the system under high data pressure and complex application scenarios. Comprehensive support is provided for scenarios such as resource assessment in microclimate environments with complex terrain, and the corrected wind resource data can provide guarantees for wind farm power generation analysis and optimized scheduling.

[0166] Further, the application of enhanced big data processing technology: Introduce a distributed computing framework to support efficient analysis and processing of wind tower data, reanalysis data, and model output data. Store wind tower data in regional shards, combine distributed storage technology, provide fast query retrieval efficiency and high-performance access efficiency, use Kafka to collect data streams, and support high-frequency updates of model input features. Rely on Nosql data storage to improve the timeliness of wind resource data correction. Feature engineering automation: Fusion of wind speed forecast data generated by WRF and wind tower measured data, application of automated feature extraction technology, extraction of complex cross-features from multi-source data. The final generated correction model is based on multi-source data fusion, using the common data foundation of ERA5, WRF, and wind towers to ensure that the wind speed correction model maintains high stability and high accuracy under the support of multi-source data. Data set abnormal situation processing: During the data processing process, a meteorological emergency detection model is constructed, and the abnormal frequency components are analyzed using fast Fourier transform to mark and remove extreme values. Combined with the data from neighboring wind towers, the spatiotemporal collaborative correction method of krige interpolation is used to ensure the geographical and temporal continuity of abnormal data. A cross-validation mechanism is introduced for reanalysis data and wind tower data, and the time matching degree (cosine similarity) is calculated to calibrate suspicious data. Handling of model anomalies: Cross-validation analysis methods are used during model establishment to ensure that the model complexity is consistent and to avoid overfitting specific data points. A breakpoint continuation training mechanism is introduced to automatically resume after training is interrupted. Confidence interval calculation is introduced for model prediction results. Special processing of application scenarios: Combined with the digital terrain model dem data, the elevation input of the wind speed correction model is dynamically adjusted to optimize the prediction accuracy of complex plant terrain areas. Extreme weather data sets (such as data under the influence of regional typhoons) are introduced in model correction to improve the robustness of wind resource prediction in extreme weather.

[0167] In this way, by integrating multi-source data, introducing a preset wind speed prediction model, and combining it with a dynamic update mechanism, the problems of single data source limitations, insufficient model correction capabilities, and poor dynamic adaptability in existing technologies are solved. The optimized correction model can significantly improve the accuracy of local wind speed prediction, especially for complex terrain and microclimate areas, providing reliable support for wind energy resource assessment and wind farm power prediction.

[0168] Although the present specification provides method operation steps as described in the embodiments or flow charts, more or fewer operation steps may be included based on conventional or non-creative means. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only order of execution. When the device or client product in practice is executed, it can be executed in sequence or in parallel according to the method shown in the embodiments or the drawings (for example, a parallel processor or a multi-threaded processing environment, or even a distributed data processing environment). The terms "include", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, product or device including a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also includes elements inherent to such a process, method, product or device. In the absence of more restrictions, it is not excluded that there are other identical or equivalent elements in the process, method, product or device including the elements. The words first, second, etc. are used to represent the name, and do not represent any particular order.

[0169] Those skilled in the art also know that, in addition to implementing the controller in a purely computer-readable program code, the controller can be made to implement the same function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered as a hardware component, and the devices for implementing various functions included therein can also be considered as structures within the hardware component. Or even, the devices for implementing various functions can be considered as both software modules for implementing the method and structures within the hardware component.

[0170] Through the description of the above embodiments, it can be known that those skilled in the art can clearly understand that the present specification can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present specification can essentially be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a mobile terminal, a server, or a network device, etc.) to execute the methods described in each embodiment of the present specification or some parts of the embodiments.

[0171] Although the present specification is described through embodiments, those skilled in the art will appreciate that there are many modifications and changes to the present specification without departing from the spirit of the present specification, and it is intended that the appended claims include these modifications and changes without departing from the spirit of the present specification.

Claims

1. A method for processing wind speed data, characterized in that: include: Acquire meteorological characteristic data of a historical period matching the quarter of the period to be predicted and wind speed data of a period before the period to be predicted based on the reanalysis data of the target area; Smoothing the wind speed data of a period before the period to be predicted to obtain a first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data; Using the first module of the preset wind speed prediction model, the meteorological characteristic data is subjected to feature combination extraction processing to obtain combined meteorological characteristics; Using the second module of the preset wind speed prediction model, performing wind speed prediction processing on the combined meteorological characteristics and the first wind speed sequence to obtain a predicted wind speed sequence; Determining a predicted power corresponding to the predicted wind speed sequence, and generating a wind turbine protection strategy for wind turbines in the target area based on the predicted power; Among them, the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained according to reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

2. The method according to claim 1, characterized in that: Before the first module using the preset wind speed prediction model performs feature combination extraction processing on the meteorological feature data to obtain the combined meteorological features, the method further includes: According to the reanalysis data of each of the first regions, historical meteorological characteristic data of the first region in a historical period having the same seasonal characteristics as the period to be predicted, and wind speed data of a period before the historical period are obtained; Smoothing the wind speed data of a period before the historical period to obtain a second wind speed sequence corresponding to the historical meteorological characteristic data; Using the first module of the preset wind speed prediction model, performing feature combination extraction processing on the historical meteorological feature data to obtain a first combined meteorological feature; Using the second module of the preset wind speed prediction model, performing wind speed prediction processing on the first combined meteorological feature and the second wind speed sequence to obtain a first predicted wind speed sequence; Determine whether the preset wind speed prediction model converges based on the first combined meteorological feature, the first predicted wind speed sequence, and the third wind speed sequence. If it is determined that the preset wind speed prediction model has not converged, continue to train the preset wind speed prediction model based on the historical meteorological feature data, the second wind speed sequence, and the third wind speed sequence until the preset wind speed prediction model converges.

3. The method according to claim 2, characterized in that The determining, according to the first combined meteorological feature, the first predicted wind speed sequence, and the third wind speed sequence, whether the preset wind speed prediction model converges includes: Determining a correlation between the first combined meteorological feature and the third wind speed sequence, and a difference between the first predicted wind speed sequence and the third wind speed sequence; A loss value is determined according to a preset loss function, the correlation and the difference, and whether the preset wind speed prediction model converges is determined according to the loss value.

4. The method according to claim 1, characterized in that: The determining of the predicted power corresponding to the predicted wind speed sequence comprises: Obtaining fan parameters of the fans in the target area; The predicted power corresponding to the predicted wind speed sequence is determined based on the wind turbine parameters and the predicted wind speed sequence using a preset power determination model, wherein the preset power determination model is a model constructed based on a preset machine learning algorithm.

5. The method according to claim 4, characterized in that The wind turbine protection strategy includes a diversion strategy and / or a voltage limiting strategy.

6. The method according to claim 5, characterized in that Generating a wind turbine protection strategy for wind turbines in the target area according to the predicted power includes: When the predicted power is not lower than the preset rated power of the wind turbine, the electric energy generated by the power difference between the predicted power and the preset rated power is distributed to the preset energy storage device through the preset power adjustment device; and / or the blade angle of the wind turbine is controlled to perform voltage reduction processing to generate a wind turbine protection strategy for the wind turbine in the target area.

7. The method according to claim 1, characterized in that The step of smoothing the wind speed data of a time period before the time period to be predicted to obtain a first wind speed sequence corresponding to the meteorological characteristic data includes: Smoothing the wind speed data of a period before the period to be predicted to obtain a candidate wind speed sequence; According to a preset sliding window, the candidate wind speed sequence is subjected to outlier detection processing, and when it is detected that the candidate wind speed sequence contains outliers, the outliers contained in the candidate wind speed sequence are processed using a moving average algorithm to obtain a first wind speed sequence corresponding to the meteorological characteristic data.

8. A wind speed data processing device, characterized in that: include: A data acquisition module, for acquiring meteorological characteristic data of a historical period matching the quarter of the period to be predicted, and wind speed data of a period before the period to be predicted, based on the reanalysis data of the target area; A data processing module, used for smoothing the wind speed data of the previous period of the period to be predicted, to obtain a first wind speed sequence corresponding to the period to be predicted and the meteorological characteristic data; A feature combination module, used to perform feature combination extraction processing on the meteorological feature data using the first module of the preset wind speed prediction model to obtain combined meteorological features; A wind speed prediction module, configured to use the second module of the preset wind speed prediction model to perform wind speed prediction processing on the combined meteorological features and the first wind speed sequence to obtain a predicted wind speed sequence; A strategy determination module, used to determine the predicted power corresponding to the predicted wind speed sequence, and generate a wind turbine protection strategy for the wind turbines in the target area according to the predicted power; Among them, the preset wind speed prediction model is trained based on historical meteorological characteristic data obtained according to reanalysis data of multiple first areas, a second wind speed sequence corresponding to the historical meteorological characteristic data, and a third wind speed sequence of the first area collected by a wind tower in the first area.

9. An electronic device, characterized in that: It comprises a processor and a memory for storing processor executable instructions, and when the processor executes the instructions, the steps of the wind speed data processing method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the wind speed data processing method according to any one of claims 1 to 7 are implemented.

Images