Wind speed forecast correction method combining prediction and error correction and related device

Abstract

The invention discloses a prediction and error correction combined wind speed forecast correction method and a related device, and belongs to the technical field of meteorological numerical forecast post-processing, and the method comprises the steps: obtaining actually measured wind speed data of a target meteorological station and numerical weather forecast wind speed prediction data of a corresponding position of the target meteorological station; calculating to obtain a wind speed prediction error sequence; performing prediction based on the historical meteorological data sequence and the wind speed prediction error sequence by adopting an Attention-BiLSTM model to obtain a preliminary error value; performing residual error correction on the initial error value by adopting a KELM model based on non-time-sequence environment characteristic data; calculating a total error predicted value based on the initial error value and the residual predicted value; and the corrected wind speed is obtained in combination with numerical weather forecast wind speed prediction data. According to the method, multi-source heterogeneous data can be deeply fused, staged modeling is realized, time sequence dynamics and environment complexity are considered, and efficiency and wind speed forecasting precision are improved.

Description

Technical Field

[0001] This invention belongs to the field of meteorological numerical forecast post-processing technology, specifically relating to a wind speed forecast correction method and related apparatus that combines forecasting and error correction. Background Technology

[0002] With the increasing proportion of renewable energy (especially wind power) in the power system, high-precision wind speed forecasting has become crucial for ensuring the safe and stable operation of the power grid. However, existing numerical weather prediction (NWP) models often exhibit significant systematic biases in their wind speed forecasts due to initial field errors, simplification of physical parameterization schemes, and insufficient terrain resolution. Traditional statistical post-processing methods struggle to fully capture the nonlinear, non-stationary, and spatiotemporal heterogeneous characteristics of wind speed errors.

[0003] Statistical methods in wind speed forecast correction, such as multiple linear regression (MOS) and Kalman filtering, assume that the error follows a linear or Gaussian distribution, making it difficult to handle the strong nonlinearity and non-stationarity of wind speed errors, resulting in limited correction effectiveness. Currently, single machine learning models are also used, such as support vector machines (SVM), random forests (RF), and single LSTM. While these can handle nonlinear relationships, they are prone to overfitting or feature interference when the input features include strong temporal variables and complex static variables, making it difficult to balance temporal modeling and multi-source information fusion. Furthermore, existing technologies may employ ensemble learning methods, such as gradient boosting trees (GBDT) and XGBoost, which perform well when feature engineering is sufficient, but have weak capabilities for modeling long-term dependencies and struggle to capture the long-term evolution patterns of wind speed errors. Therefore, in summary, current wind speed forecast correction methods struggle to effectively separate and collaboratively process temporal dynamic features and static / semi-static environmental features, lack adaptive attention to key historical moments in the error sequence, and have relatively singular models, making it difficult to balance prediction accuracy and computational efficiency. Summary of the Invention

[0004] This invention provides a wind speed forecast correction method and related apparatus that combines prediction and error correction. The purpose is to solve the problems in current wind speed forecast correction methods, such as the difficulty in effectively separating and coordinating the processing of temporal dynamic features and static / semi-static environmental features, the lack of adaptive attention to key historical moments in the error sequence, and the relatively simple model, which makes it difficult to balance prediction accuracy and computational efficiency.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: This invention provides a wind speed forecast correction method that combines prediction and error correction, comprising the following steps: S1. Obtain the measured wind speed data of the target meteorological station and the numerical weather forecast wind speed prediction data of the corresponding location of the target meteorological station; and calculate the wind speed prediction error sequence based on the measured wind speed data and the numerical weather forecast wind speed prediction data. S2. Using the Attention-BiLSTM model, predictions are made based on historical meteorological data sequences and the calculated wind speed prediction error sequence to obtain preliminary error values. S3. Using the KELM model, based on non-time-series environmental characteristic data, residual correction is performed on the preliminary error value to obtain the residual prediction value; S4. Based on the preliminary error value and the residual prediction value, the total error prediction value is calculated; based on the numerical weather prediction wind speed prediction data and the total error prediction value, the corrected wind speed is obtained, and the wind speed forecast correction combining prediction and error correction is completed. Among them, the Attention-BiLSTM model and the KELM model are pre-trained using historical meteorological data, wind speed prediction error sequences, and non-temporal environmental feature data; the Attention-BiLSTM model is a model that combines a bidirectional long short-term memory network with an attention mechanism, and the KELM model is a kernel extreme learning machine model.

[0006] In some implementations, the wind speed prediction error sequence is calculated in S1 using the following formula: ; in, express Wind speed prediction error at any given time express Numerical weather forecast wind speed prediction data at any given time. express Real-time measured wind speed data.

[0007] In some implementations, in S2, the historical meteorological data sequence includes historical measured wind speed, historical temperature, historical air pressure, and historical humidity sequences; the input of the Attention-BiLSTM model also includes the historical wind speed prediction error sequence.

[0008] In some implementations, in S3, residual correction is performed on the initial error value. First, the master model residuals are calculated, and then the KELM model is used as the target for prediction; wherein: The master model residuals are calculated using the following formula: ; in, express The master model residuals at time 10:00. express Wind speed prediction error at any given time express The initial error value at time.

[0009] In some implementations, in S3, the non-time-series environmental feature data includes at least one of terrain parameters, atmospheric stability index, time features, weather type labels, and numerical weather forecast set dispersion. The terrain parameters include elevation and surface roughness, the atmospheric stability index includes Richardson number, and the temporal characteristics include at least one of month, day and night, and season.

[0010] In some implementations, in S4, the total error prediction value is calculated using the following formula: ; in, express Total error prediction value at time 10:00 express The initial error value at time, express The predicted residual value at time step.

[0011] In some implementations, in S4, the corrected wind speed is calculated using the following formula: ; in, This indicates the corrected wind speed. This represents numerical weather forecast wind speed prediction data. This represents the total predicted error.

[0012] This invention also provides a wind speed forecast correction system combining prediction and error correction to implement the aforementioned wind speed forecast correction method combining prediction and error correction. The system includes a wind speed data acquisition module, an Attention-BiLSTM error calculation module, a KELM residual prediction module, and a correction and wind speed fusion output module, wherein: Wind speed data acquisition module: used to acquire measured wind speed data of the target meteorological station and numerical weather forecast wind speed prediction data of the corresponding location of the target meteorological station; and to calculate the wind speed prediction error sequence based on the measured wind speed data and numerical weather forecast wind speed prediction data. Attention-BiLSTM Error Calculation Module: Used to make predictions based on historical meteorological data sequences and calculated wind speed prediction error sequences using the Attention-BiLSTM model, and obtain preliminary error values; KELM Residual Prediction Module: Used to apply the KELM model to correct the residuals of the initial error values ​​based on non-time-series environmental feature data, and obtain the predicted residual values. Correction and wind speed fusion output module: used to calculate the total error prediction value based on the preliminary error value and the residual prediction value; and to obtain the corrected wind speed based on the numerical weather forecast wind speed prediction data and the total error prediction value, thus completing the wind speed forecast correction that combines prediction and error correction. Among them, the Attention-BiLSTM model and the KELM model are pre-trained using historical meteorological data, wind speed prediction error sequences, and non-temporal environmental feature data; the Attention-BiLSTM model is a model that combines a bidirectional long short-term memory network with an attention mechanism, and the KELM model is a kernel extreme learning machine model.

[0013] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the wind speed forecast correction method combining prediction and error correction as described above.

[0014] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the wind speed forecast correction method combining prediction and error correction as described above.

[0015] Compared with the prior art, the wind speed forecast correction method and related apparatus of the present invention, which combines prediction and error correction, has the following beneficial effects: This invention presents a wind speed forecast correction method combining prediction and error correction. It employs a two-stage hybrid model framework of main prediction and error correction, utilizing high-quality measured meteorological data as ground truth. An Attention-BiLSTM model captures the temporal dynamics of wind speed errors, while a Kernel Extreme Learning Machine (KELM) model integrates complex non-temporal factors such as topography, atmospheric stability, and temporal characteristics to refine the residuals of the main model, ultimately achieving high-precision correction of non-wind speed (NWP) forecasts. This invention significantly improves correction accuracy by capturing both the temporal dynamics of errors and the nonlinear effects of the environment through two-stage modeling. Attention-BiLSTM focuses on temporal modeling, while KELM focuses on multi-factor fusion, avoiding feature interference and improving model robustness. The Attention mechanism can output weights for key historical moments to aid in error analysis. KELM has fast training speed and is suitable for online applications. It is applicable to NWP correction for other meteorological elements such as temperature, humidity, and irradiance, demonstrating significant practical value.

[0016] On the other hand, the computer device of the present invention, through the execution of a specific computer program by a processor, can efficiently implement the steps of the method of the present invention. When performing data processing tasks, the computer device can accurately perform numerical calculations and logical judgments, ensuring the accuracy and consistency of the data processing results. The computer-readable storage medium of the present invention, by programming the steps of the method of the present invention into a computer program and storing it on the computer-readable storage medium, allows users to easily load and execute these programs on any compatible computer device without rewriting or converting the code, thus improving the convenience and flexibility of program execution. Attached Figure Description

[0017] The accompanying drawings are provided to further understand the invention and constitute a part of this invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0018] Figure 1 This is a flowchart illustrating a wind speed forecast correction method that combines prediction and error correction according to the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0021] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0022] It should be noted that the apparatus and methods disclosed in the embodiments herein can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments herein. In this regard, each block in a flowchart or block diagram may represent a module, program, or part of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system to perform the specified function or action, or can be implemented using a combination of dedicated hardware and computer instructions.

[0023] In addition, the functional modules in the various embodiments of this article can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0024] How can we provide a wind speed forecast correction method that can deeply integrate multi-source heterogeneous data, perform phased modeling, and take into account both temporal dynamics and environmental complexity? This method can be used to systematically correct the wind speed output of numerical weather prediction (NWP) models, thereby improving the accuracy of wind speed forecasts. It is particularly suitable for scenarios that require high-precision wind speed forecasts, such as wind power prediction, power grid dispatching, and meteorological services.

[0025] like Figure 1 As shown, the present invention provides a wind speed forecast correction method combining prediction and error correction, comprising the following steps: S1. Obtain the measured wind speed data of the target meteorological station and the numerical weather forecast wind speed prediction data of the corresponding location of the target meteorological station; and calculate the wind speed prediction error sequence based on the measured wind speed data and the numerical weather forecast wind speed prediction data. S2. Using the Attention-BiLSTM model, predictions are made based on historical meteorological data sequences and the calculated wind speed prediction error sequence to obtain preliminary error values. S3. Using the KELM model, based on non-time-series environmental characteristic data, residual correction is performed on the preliminary error value to obtain the residual prediction value; S4. Based on the preliminary error value and the residual prediction value, the total error prediction value is calculated; based on the numerical weather prediction wind speed prediction data and the total error prediction value, the corrected wind speed is obtained, and the wind speed forecast correction combining prediction and error correction is completed. Among them, the Attention-BiLSTM model and the KELM model are pre-trained using historical meteorological data, wind speed prediction error sequences, and non-temporal environmental feature data; the Attention-BiLSTM model is a model that combines a bidirectional long short-term memory network with an attention mechanism, and the KELM model is a kernel extreme learning machine model.

[0026] This invention combines prediction and error correction in a wind speed forecast correction method. It utilizes an Attention-BiLSTM model to specifically handle historical meteorological data and error sequences with strong time dependencies, capturing forward and backward temporal dependencies. The attention mechanism adaptively focuses on the most critical historical moments for the current prediction, effectively improving the model's ability to capture nonlinear and non-stationary temporal patterns in wind speed error sequences. By introducing a KELM model, it handles non-temporally dependent or weakly temporally dependent environmental features, such as topography and atmospheric stability. It models and corrects the nonlinear relationship between these complex factors and the main model residuals, avoiding feature interference and model overfitting problems that may occur when simply mixing temporally and non-temporally features and inputting them into a single model. This makes the invention more robust to multi-source heterogeneous data. Furthermore, by constructing and training the aforementioned models, this invention demonstrates high efficiency in practical applications, requiring no retraining during the online prediction phase and exhibiting fast response speed. By decomposing the complex wind speed correction problem into two sub-tasks handled by specialized models, it achieves higher accuracy correction of wind speed in numerical weather forecasts.

[0027] Furthermore, the historical meteorological data series includes historical measured wind speed, historical temperature, historical air pressure, and historical humidity sequences; the input to the Attention-BiLSTM model also includes the historical wind speed prediction error sequence. The multi-dimensional historical meteorological data sequences (wind speed, temperature, air pressure, and humidity) collectively reflect the evolution of atmospheric conditions, while the historical error sequence directly contains the evolutionary patterns of forecast bias. Using these features as input allows the model to learn the temporal dynamic mechanisms affecting wind speed errors from a more comprehensive perspective, rather than relying solely on a single wind speed or error sequence, thus making the initial error prediction more accurate and reliable.

[0028] This invention performs residual correction on the initial error value. First, the master model residual is calculated. Then, using the master model residual as the target, a KELM model is used for prediction. The master model residual represents the portion of error that the Attention-BiLSTM master model could not fully explain. By calculating and using it as the learning target of the KELM model, task transfer and error complementarity between the two model stages are achieved.

[0029] Furthermore, the non-time-series environmental feature data of this invention includes at least one of the following: topographic parameters, atmospheric stability indices, temporal features, weather type labels, and numerical weather prediction ensemble dispersion. Topographic parameters include elevation and surface roughness; atmospheric stability indices include Richardson number; and temporal features include at least one of month, day / night, and season. Topographic parameters directly affect the distribution of near-surface wind fields; atmospheric stability indices reflect the influence of atmospheric stratification on turbulence and wind speed; temporal features introduce climatic periodicity; and ensemble weather types and NWP ensemble dispersion, by introducing diverse, time-independent but physically meaningful features into the model, enable KELM to perform more refined and accurate corrections to the master model residuals.

[0030] Specifically, this invention provides a wind speed forecast correction method based on a hybrid model of Attention-BiLSTM and KELM, constructing a two-stage framework of main prediction and error correction, which is carried out according to the following steps: I. Data Preparation and Preprocessing 1) Obtain high-precision measured wind speed data of the target meteorological station and corresponding numerical weather prediction (NWP) wind speed forecast data; 2) Calculate the difference between the NWP wind speed and the measured wind speed, and construct a wind speed prediction error sequence: ; in, express Wind speed prediction error at any given time express Numerical weather forecast wind speed prediction data at any given time. express Real-time measured wind speed data.

[0031] II. Constructing the Master Prediction Model (Attention-BiLSTM) 1) Input features: Historical measured wind speed, temperature, air pressure, and humidity sequences over the past N hours, as well as historical error sequences e t 1, e t 2, ..., e t N ; 2) Model Structure: A Bidirectional Long Short-Term Memory (BiLSTM) network is used to extract the forward and backward dependencies of the historical sequence; an attention mechanism is introduced after the BiLSTM output layer to automatically learn the importance weights of different historical moments for the current error prediction; Output: Preliminary error prediction value. .

[0032] III. Constructing the Error Correction Model (KELM) 1) Calculate the master model residuals: ; in, express The master model residuals at time 10:00. express Wind speed prediction error at any given time express The initial error value at time.

[0033] 2) Input features: terrain parameters (elevation, roughness), atmospheric stability indices (such as Richardson number), time features (month, day / night, season), weather type labels, NWP set dispersion, etc. 3) Model Structure: The residual prediction model is trained using Kernel Extreme Learning Machine (KELM) to output the predicted residual values. .

[0034] IV. Final Wind Speed ​​Correction 1) Calculate the total error prediction value: ; in, express Total error prediction value at time 10:00 express The initial error value at time, express The predicted residual value at time step; 2) Output corrected wind speed: ; in, This indicates the corrected wind speed. This represents numerical weather forecast wind speed prediction data. This represents the total predicted error.

[0035] This invention also provides a wind speed forecast correction system combining prediction and error correction to implement the aforementioned wind speed forecast correction method combining prediction and error correction. The system includes a wind speed data acquisition module, an Attention-BiLSTM error calculation module, a KELM residual prediction module, and a correction and wind speed fusion output module, wherein: Wind speed data acquisition module: used to acquire measured wind speed data of the target meteorological station and numerical weather forecast wind speed prediction data of the corresponding location of the target meteorological station; and to calculate the wind speed prediction error sequence based on the measured wind speed data and numerical weather forecast wind speed prediction data. Attention-BiLSTM Error Calculation Module: Used to make predictions based on historical meteorological data sequences and calculated wind speed prediction error sequences using the Attention-BiLSTM model, and obtain preliminary error values; KELM Residual Prediction Module: Used to apply the KELM model to correct the residuals of the initial error values ​​based on non-time-series environmental feature data, and obtain the predicted residual values. Correction and wind speed fusion output module: used to calculate the total error prediction value based on the preliminary error value and the residual prediction value; and to obtain the corrected wind speed based on the numerical weather forecast wind speed prediction data and the total error prediction value, thus completing the wind speed forecast correction that combines prediction and error correction. Among them, the Attention-BiLSTM model and the KELM model are pre-trained using historical meteorological data, wind speed prediction error sequences, and non-temporal environmental feature data; the Attention-BiLSTM model is a model that combines a bidirectional long short-term memory network with an attention mechanism, and the KELM model is a kernel extreme learning machine model.

[0036] The following detailed description of a wind speed forecast correction method and related apparatus that combines prediction and error correction according to the present invention will be provided through specific embodiments.

[0037] Taking the correction of wind speed forecasts for a certain wind farm as an example, the specific implementation steps are as follows: I. Data Collection: 1) Obtain measured wind speed, temperature, and air pressure data once per minute from the wind farm's wind measurement tower; 2) Obtain hourly wind speed forecast data for the coordinate grid points of the wind farm using the CMA, ECMWF, or GFS meteorological forecasting model; 3) Collect wind farm topographic data (DEM elevation, surface roughness) and historical weather type labels (windy days, frontal passage, etc.).

[0038] II. Model Training: 1) Select data from the past year as the training set and construct the error sequence e. t ; 2) Use historical weather data from the past 6 hours as input to the Attention-BiLSTM to train the main model; 3) Calculate the training set residuals The model is trained and corrected by using terrain, season, atmospheric stability, and other factors as inputs to KELM.

[0039] III. Online Prediction: 1) Get the latest NWP wind speed forecast in real time; 2) Input the latest 6 hours of measured data into Attention-BiLSTM to obtain... ; 3) Input the current environment characteristics into KELM to obtain... ; 4) Calculate the final corrected wind speed, which will be used as input for the wind power prediction model.

[0040] In some practical applications, this invention can be deployed in wind farm SCADA systems, regional meteorological centers, or cloud platforms to support batch or real-time wind speed correction. It has a certain degree of applicability.

[0041] The present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the wind speed forecast correction method combining prediction and error correction as described above.

[0042] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the wind speed forecast correction method combining prediction and error correction as described above.

[0043] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions may also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means implemented in the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0044] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0045] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Anyone skilled in the art can readily implement the present invention according to the description and above. Any modifications, alterations, or equivalent variations made using the technical content disclosed above are equivalent embodiments of the present invention. Furthermore, any modifications, alterations, or variations made to the above embodiments based on the essential technology of the present invention are still within the protection scope of the present invention.

Claims

1. A wind speed forecast correction method combining prediction and error correction, characterized in that, Includes the following steps: S1. Obtain the measured wind speed data of the target meteorological station and the numerical weather forecast wind speed prediction data of the corresponding location of the target meteorological station; and calculate the wind speed prediction error sequence based on the measured wind speed data and the numerical weather forecast wind speed prediction data. S2. Using the Attention-BiLSTM model, predictions are made based on historical meteorological data sequences and the calculated wind speed prediction error sequence to obtain preliminary error values. S3. Using the KELM model, based on non-time-series environmental characteristic data, residual correction is performed on the preliminary error value to obtain the residual prediction value; S4. Based on the preliminary error value and the residual prediction value, the total error prediction value is calculated; Based on numerical weather prediction wind speed forecast data and total error prediction value, the corrected wind speed is obtained, and the wind speed forecast correction combining prediction and error correction is completed. Among them, the Attention-BiLSTM model and the KELM model are pre-trained using historical meteorological data, wind speed prediction error sequences, and non-temporal environmental feature data; the Attention-BiLSTM model is a model that combines a bidirectional long short-term memory network with an attention mechanism, and the KELM model is a kernel extreme learning machine model.

2. The wind speed forecast correction method combining prediction and error correction according to claim 1, characterized in that, In step S1, the wind speed prediction error sequence is calculated using the following formula: ； in, express Wind speed prediction error at any given time express Numerical weather forecast wind speed prediction data at any given time. express Real-time measured wind speed data.

3. The wind speed forecast correction method combining prediction and error correction according to claim 1, characterized in that, In S2, the historical meteorological data sequence includes historical measured wind speed, historical temperature, historical air pressure, and historical humidity sequences; the input of the Attention-BiLSTM model also includes the historical wind speed prediction error sequence.

4. The wind speed forecast correction method combining prediction and error correction according to claim 1, characterized in that, In step S3, residual correction is performed on the initial error value. First, the master model residuals are calculated. Then, using the master model residuals as the target, the KELM model is used for prediction. Wherein: The master model residuals are calculated using the following formula: ； in, express The master model residuals at time 10:

00. express Wind speed prediction error at any given time express The initial error value at time.

5. The wind speed forecast correction method combining prediction and error correction according to claim 1, characterized in that, In S3, the non-time-series environmental feature data includes at least one of terrain parameters, atmospheric stability index, time features, weather type labels, and numerical weather forecast set dispersion. The terrain parameters include elevation and surface roughness, the atmospheric stability index includes Richardson number, and the temporal characteristics include at least one of month, day and night, and season.

6. The wind speed forecast correction method combining prediction and error correction according to claim 1, characterized in that, In step S4, the total error prediction value is calculated using the following formula: ； in, express Total error prediction value at time 10:00 express The initial error value at time, express The predicted residual value at time step.

7. The wind speed forecast correction method combining prediction and error correction according to claim 1, characterized in that, In S4, the corrected wind speed is calculated using the following formula: ； in, This indicates the corrected wind speed. This represents numerical weather forecast wind speed prediction data. This represents the total predicted error.

8. A wind speed forecast correction system combining prediction and error correction, for implementing the wind speed forecast correction method combining prediction and error correction as described in any one of claims 1-7, characterized in that, It includes a wind speed data acquisition module, an Attention-BiLSTM error calculation module, a KELM residual prediction module, and a correction and wind speed fusion output module, wherein: Wind speed data acquisition module: used to acquire measured wind speed data of the target meteorological station and numerical weather forecast wind speed prediction data of the corresponding location of the target meteorological station; and to calculate the wind speed prediction error sequence based on the measured wind speed data and numerical weather forecast wind speed prediction data. Attention-BiLSTM Error Calculation Module: Used to make predictions based on historical meteorological data sequences and calculated wind speed prediction error sequences using the Attention-BiLSTM model, and obtain preliminary error values; KELM Residual Prediction Module: Used to apply the KELM model to correct the residuals of the initial error values ​​based on non-time-series environmental feature data, and obtain the predicted residual values. Correction and wind speed fusion output module: used to calculate the total error prediction value based on the preliminary error value and the residual prediction value; and to obtain the corrected wind speed based on the numerical weather forecast wind speed prediction data and the total error prediction value, thus completing the wind speed forecast correction that combines prediction and error correction. Among them, the Attention-BiLSTM model and the KELM model are pre-trained using historical meteorological data, wind speed prediction error sequences, and non-temporal environmental feature data; the Attention-BiLSTM model is a model that combines a bidirectional long short-term memory network with an attention mechanism, and the KELM model is a kernel extreme learning machine model.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the wind speed forecast correction method combining prediction and error correction as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the wind speed forecast correction method combining prediction and error correction as described in any one of claims 1-7.

Images