Meteorological characteristic data processing method, device and equipment and wind power plant generating capacity prediction method, device and equipment
By calculating the deviation rate of meteorological forecast data and optimizing meteorological characteristic data, the power generation prediction model was corrected. By using machine learning and reinforcement learning to adjust parameters, the problem of low accuracy in wind farm power generation prediction was solved, the stability and adaptability of the model were improved, and more accurate power generation prediction was achieved.
Patent Information
- Application Number
- CN202511074772.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-01
- Publication Date
- 2025-11-28
AI Technical Summary
The accuracy of existing wind farm power generation forecasts is limited by errors in meteorological forecast data, resulting in low accuracy in power generation forecasts.
By calculating the deviation rate of meteorological forecast data, optimizing meteorological characteristic data, and correcting the power generation forecast model based on the optimized meteorological characteristic data, the model parameters are adjusted using machine learning and reinforcement learning until the verification results meet the conditions, thus obtaining the target power generation forecast model.
It significantly improves the accuracy of wind farm power generation prediction, enhances the stability and robustness of the model in complex and variable environments, and avoids the limitations and subjectivity of manual parameter tuning.
Smart Images
Figure CN121031850A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of computer technology, and in particular to meteorological feature data processing, wind farm power generation prediction methods, devices and equipment. Background Technology
[0002] With the increasing global demand for clean energy, wind power has been widely adopted as an important renewable energy source. Accurate wind farm power generation forecasting is crucial for the stable operation of the power system, power dispatch, and the operation and management of wind farms. Accurate power generation forecasting can help the power grid rationally plan power generation, reduce operational risks of the power system, improve wind energy utilization, and provide a basis for decision-making regarding the economic operation of wind farms.
[0003] However, the accuracy of wind farm power generation forecasting is currently limited by a variety of factors. Among them, the error in meteorological forecast data is one of the key factors affecting the accuracy of power generation forecasting. Meteorological forecast data often has a certain degree of deviation. For example, there may be significant differences between the predicted and actual values of meteorological elements such as wind speed, wind direction, temperature, and air pressure. These deviations directly lead to large errors in wind farm power generation forecasting models based on meteorological data, resulting in low accuracy of power generation forecasting. Summary of the Invention
[0004] Therefore, it is necessary to provide a meteorological characteristic data processing method, apparatus, computer equipment, computer-readable storage medium, and computer program product, as well as a wind farm power generation prediction device, computer equipment, computer-readable storage medium, and computer program product, to address the above-mentioned technical problems and improve the accuracy of power generation prediction.
[0005] Firstly, this application provides a method for processing meteorological characteristic data, including:
[0006] Based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database for the first time period, the deviation rate of meteorological prediction data is calculated.
[0007] The meteorological characteristic data for the second time period are optimized based on the deviation rate of meteorological forecast data. The second time period is later than the first time period.
[0008] Based on the correlation between meteorological characteristic data and power generation data, and the optimized meteorological characteristic data, the power generation prediction model is revised.
[0009] Based on the optimized meteorological characteristic data, the wind farm power generation in the second time period is predicted using the modified power generation prediction model.
[0010] The predicted wind farm power generation data is validated. If the validation results do not meet the validation conditions, the model parameters are adjusted during the optimization of meteorological characteristic data and the correction of the power generation prediction model until the validation results meet the validation conditions, thus obtaining the target power generation prediction model for power generation prediction.
[0011] In some embodiments, calculating the meteorological forecast data deviation rate based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data within a first time period stored in the associated database includes:
[0012] For each meteorological element, the meteorological element deviation between the predicted value and the observed value is calculated based on the correspondence between the historical actual meteorological characteristic data and the historical predicted meteorological characteristic data stored in the associated database within the first time period.
[0013] Based on the deviation of the meteorological elements, the deviation rate of the meteorological forecast data corresponding to the meteorological elements is determined.
[0014] In some embodiments, the optimization of meteorological characteristic data for the second time period based on the meteorological forecast data deviation rate includes:
[0015] Identify the factors influencing the deviation rate of meteorological forecast data;
[0016] A relationship model between the deviation rate of meteorological forecast data and influencing factors is constructed. The relationship model is trained and optimized using historical data to obtain a deviation rate prediction model.
[0017] Obtain the initial meteorological characteristic prediction data for the second time period, and determine the values of the influencing factors within the second time period;
[0018] Based on the influencing factors during the second time period, the deviation rate correction value is calculated using the deviation rate prediction model.
[0019] The initial meteorological feature prediction data are corrected based on the calculated deviation rate correction value to obtain optimized meteorological feature data.
[0020] In some embodiments, the step of correcting the power generation prediction model based on the correlation between meteorological characteristic data and power generation data, and the optimized meteorological characteristic data, includes:
[0021] Select the appropriate machine learning model based on the degree of linear correlation between different meteorological characteristic data and power generation data;
[0022] Perform initialization settings for the selected machine learning model;
[0023] When the machine learning model is a linear regression model, the regression coefficients are re-estimated based on the optimized meteorological feature data; when the machine learning model is a neural network model, the model is trained using the optimized meteorological feature data and the corresponding power generation data.
[0024] During training, the error between the predicted power generation and the actual power generation is calculated based on the set loss function. The error is then backpropagated to each layer of the network to update the weights and biases, thus obtaining the corrected power generation prediction model.
[0025] In some embodiments, adjusting the model parameters during the optimization of meteorological characteristic data and the correction of the power generation prediction model includes:
[0026] Define the reinforcement learning environment, including the state space, action space, and reward function;
[0027] Select a reinforcement learning algorithm and initialize the agent parameters;
[0028] Reinforcement learning training is conducted, including environment reset, action selection, environment interaction and state transition, reward calculation and storage, and agent learning;
[0029] After training, the trained agent strategy and the adjusted power generation prediction model are obtained. The trained agent strategy is used to optimize the meteorological feature data, and the target power generation prediction model is constructed based on the adjusted model parameters.
[0030] Secondly, this application also provides a meteorological characteristic data processing device, comprising:
[0031] The deviation rate calculation module is used to calculate the deviation rate of meteorological forecast data based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database within the first time period.
[0032] The meteorological feature data optimization module is used to optimize the meteorological feature data for a second time period based on the deviation rate of meteorological forecast data. The second time period is later than the first time period.
[0033] The model correction module is used to correct the power generation prediction model based on the correlation between meteorological feature data and power generation data, as well as the optimized meteorological feature data.
[0034] The power generation prediction module is used to predict the power generation of the wind farm in the second time period based on the optimized meteorological characteristic data and the modified power generation prediction model.
[0035] The model validation and adjustment module is used to validate the predicted wind farm power generation data. If the validation results do not meet the validation conditions, the model parameters are adjusted during the optimization of meteorological characteristic data and the correction of the power generation prediction model until the validation results meet the validation conditions, thus obtaining the target power generation prediction model for power generation prediction.
[0036] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described meteorological feature data processing method.
[0037] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described meteorological feature data processing method.
[0038] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described meteorological characteristic data processing method.
[0039] The aforementioned meteorological characteristic data processing methods, devices, computer equipment, computer-readable storage media, and computer program products calculate the meteorological forecast data deviation rate based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database within the first time period. This allows for a deeper understanding of the error characteristics of the meteorological forecast model. Based on this, the power generation forecast model is corrected to effectively compensate for the power generation forecast error caused by meteorological forecast deviation, thereby significantly improving the accuracy of power generation forecast. After optimizing the meteorological characteristic data and correcting the power generation forecast model based on the meteorological forecast data deviation rate, the optimized forecast results are continuously verified. When the forecast results are not satisfactory, the model parameters are adjusted during the optimization of meteorological characteristic data and the correction of the power generation forecast model. The optimal model parameter adjustment strategy is automatically learned, improving the performance and adaptability of the power generation forecast model to better cope with complex and variable meteorological conditions and improve the accuracy of wind farm power generation forecast. This avoids the limitations and subjectivity of manual parameter adjustment, improves the stability and robustness of the power generation forecast model in complex and variable environments, and thus improves the accuracy of power generation forecast.
[0040] Sixthly, this application provides a method for predicting the power generation of a wind farm, including:
[0041] Obtain initial meteorological characteristic data for the period to be predicted;
[0042] The initial meteorological feature data is optimized to obtain optimized meteorological feature data;
[0043] The optimized meteorological characteristic data is input into the target power generation prediction model to obtain the predicted power generation value of the wind farm during the prediction period.
[0044] The target power generation prediction model is trained using the meteorological characteristic data processing method described above.
[0045] Seventhly, this application provides a wind farm power generation prediction device, comprising:
[0046] The meteorological characteristic data acquisition module is used to acquire the initial meteorological characteristic data for the period to be predicted.
[0047] The meteorological feature data optimization module is used to optimize the initial meteorological feature data to obtain optimized meteorological feature data.
[0048] The wind farm power generation prediction module is used to input the optimized meteorological characteristic data into the target power generation prediction model to obtain the predicted value of wind farm power generation for the period to be predicted; wherein, the target power generation prediction model is trained using the above-mentioned meteorological characteristic data processing method.
[0049] Eighthly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described meteorological feature data processing method.
[0050] Ninthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described meteorological feature data processing method.
[0051] In a tenth aspect, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described meteorological feature data processing method.
[0052] The aforementioned wind farm power generation prediction method, device, computer equipment, computer-readable storage medium, and computer program product calculate the meteorological prediction data deviation rate based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database for the first time period. This allows for a deeper understanding of the error characteristics of the meteorological prediction model, and based on this, the power generation prediction model is corrected to effectively compensate for the power generation prediction error caused by meteorological prediction deviation, thereby significantly improving the accuracy of power generation prediction. After optimizing the meteorological characteristic data and correcting the power generation prediction model based on the meteorological prediction data deviation rate, the optimized prediction results are continuously verified. When the prediction results are not satisfactory, the model parameters are adjusted during the meteorological characteristic data optimization and power generation prediction model correction process. The optimal model parameter adjustment strategy is automatically learned to improve the performance and adaptability of the power generation prediction model, better cope with complex and variable meteorological conditions, and improve the accuracy of wind farm power generation prediction. This avoids the limitations and subjectivity of manual parameter adjustment, improves the stability and robustness of the power generation prediction model in complex and variable environments, and thus improves the accuracy of power generation prediction. Attached Figure Description
[0053] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0054] Figure 1 This is an application environment diagram of the meteorological feature data processing method and the wind farm power generation prediction method in one embodiment;
[0055] Figure 2 This is a flowchart illustrating a meteorological feature data processing method in one embodiment;
[0056] Figure 3 This is a flowchart illustrating the steps for correcting a power generation prediction model in one embodiment.
[0057] Figure 4 This is a flowchart illustrating the steps for adjusting model parameters in one embodiment;
[0058] Figure 5 This is a block diagram illustrating the principle of the reinforcement learning algorithm used in this application in one embodiment;
[0059] Figure 6 This is a flowchart illustrating a wind farm power generation prediction method in one embodiment;
[0060] Figure 7This is a structural block diagram of a meteorological feature data processing device in one embodiment;
[0061] Figure 8 This is a structural block diagram of a wind farm power generation prediction device in one embodiment;
[0062] Figure 9 This is an internal structural diagram of a computer device in one embodiment;
[0063] Figure 10 This is a diagram of the internal structure of a computer device in another embodiment. Detailed Implementation
[0064] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0065] The meteorological feature data processing method and wind farm power generation prediction provided in this application can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on the cloud or other network servers. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.
[0066] For example, server 104 can train a target power generation prediction model using the meteorological feature data processing method of this application. Server 104 can further send the target power generation prediction model to terminal 102. After terminal 102 deploys the target power generation prediction model, it can realize wind farm power generation prediction.
[0067] In one exemplary embodiment, such as Figure 2 As shown, a meteorological feature data processing method is provided. This method can be executed by a server or a terminal. In this embodiment, the method is applied to... Figure 1Taking the server in the example, the following steps are included:
[0068] Step 202: Calculate the meteorological forecast data deviation rate based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database for the first time period.
[0069] Specifically, historical actual meteorological characteristic data and corresponding historical predicted meteorological characteristic data can be collected to establish a correlation database between actual meteorological characteristic data and predicted meteorological characteristic data. Then, the server can calculate the deviation of meteorological data elements and the deviation rate of meteorological prediction data based on the correlation database between actual meteorological characteristic data and predicted meteorological characteristic data.
[0070] For example, historical actual meteorological characteristic data covers multiple meteorological elements such as wind speed, wind direction, temperature, air pressure, and humidity. For instance, the actual meteorological characteristic data for the previous year from June to December, and the corresponding historical predicted meteorological characteristic data are meteorological characteristic data predicted earlier based on neural network models. The associated database includes actual meteorological characteristic data, predicted meteorological characteristic data, and the mapping relationship between the two. For instance, the actual meteorological characteristic data, predicted meteorological characteristic data, and the mapping relationship between the two for the corresponding time period and region.
[0071] For example, the meteorological forecast data deviation rate is calculated based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database for the first time period. This includes: for each meteorological element, calculating the meteorological element deviation between the predicted value and the observed value of the meteorological element based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database for the first time period; and determining the meteorological forecast data deviation rate corresponding to the meteorological element based on the meteorological element deviation.
[0072] For example, the specific calculation method for the deviation rate of meteorological forecast data is as follows:
[0073] First, for each meteorological element (such as wind speed, temperature, and air pressure), calculate the deviation between the predicted and observed values. Let the predicted value of a certain meteorological element be P. i The observed value is O i Then the deviation D of this element i =P i -O i Here, 'i' represents different time points or data samples. First, determine the time interval and data points for calculating the bias. Based on the research objective and data characteristics, select an appropriate time interval to calculate the bias. For example, the bias of meteorological elements at different time scales can be calculated by hour, day, or month. Then, determine the range of data points for which the bias will be calculated—whether it's for the entire historical data period or a specific season or year.
[0074] Taking wind speed as an example, let the actual observed wind speed sequence be V. a (t), the predicted wind speed sequence is V p (t), where t represents a time point, at each selected time point t j The formula for calculating wind speed deviation is as follows:
[0075] ΔV(t j ) = V p (t j )-V a (t j )
[0076] For other meteorological elements such as temperature T and air pressure P, let the actual observed temperature series be T. a (t), the predicted temperature sequence is T p (t), the actual observed pressure sequence is P a (t), the predicted pressure sequence is P p (t), the deviation calculation expression is: temperature deviation ΔT(t) j ) = T p (t j )-T a (t j ), air pressure deviation ΔP(t) j ) = P p (t j )-P a (t j ).
[0077] Then, for each meteorological element, calculate the statistical index of the meteorological element deviation corresponding to that meteorological element, and at the same time calculate the statistical index of the observed value corresponding to that meteorological element. The statistical index may specifically include the average deviation. Deviation standard deviation σ D At least one of them, the formula for calculating the average deviation is: The standard deviation of the deviation is n is the number of data samples;
[0078] Therefore, for each meteorological element, the formula for calculating the deviation rate of meteorological forecast data can be:
[0079] R = Statistical index of meteorological element deviation / Statistical index of observed values, where the statistical index is a statistical value of the same dimension. For example, the deviation rate of meteorological forecast data for a certain meteorological element can be: R = Average deviation / The average of the observed values.
[0080] In some other embodiments, the formula for calculating the meteorological forecast data deviation rate may also be:
[0081]
[0082] Where i represents different sampling time points of meteorological elements, and m is the number of sampling time points of meteorological elements.
[0083] Step 204: Optimize the meteorological characteristic data for the second time period based on the deviation rate of the meteorological forecast data. The second time period is later than the first time period.
[0084] Specifically, the server can determine the influencing factors of the meteorological forecast data deviation rate, construct a relationship model between the meteorological forecast data deviation rate and the influencing factors, train and optimize the relationship model using historical data to obtain the deviation rate prediction model, acquire the initial meteorological characteristic prediction data for the second time period, determine the corresponding values of the influencing factors in the second time period, calculate the deviation rate correction value using the deviation rate prediction model based on the influencing factors in the second time period, and correct the initial meteorological characteristic prediction data based on the calculated deviation rate correction value to obtain the optimized meteorological characteristic data.
[0085] For example, optimizing meteorological characteristic data for the forecast period based on the meteorological forecast data deviation rate is specifically as follows:
[0086] First, identify the factors that affect the deviation rate: consider factors that may affect the meteorological forecast deviation rate, such as the forecast lead time (e.g., the forecast deviation rate may be different when the forecast is 1 hour, 6 hours, or 24 hours in advance), the current value of the meteorological characteristics themselves (e.g., the forecast deviation may be different when the current wind speed is high compared to when the wind speed is low), geographical location (the terrain and climate conditions of different regions may lead to different deviation patterns), seasonal factors, etc.
[0087] Furthermore, model construction can be employed, using methods such as multiple linear regression, decision trees, and neural networks to build a model of the relationship between the deviation rate and influencing factors. For example, a multiple linear regression model can be constructed with wind speed deviation rate as the dependent variable and forecast lead time, current wind speed, and season as independent variables.
[0088] r V =β0 + β1 × T + β2 × V c +β3×S+…+ε(where T is the prediction time lead, V…) c (where S is the current wind speed, β0, β1, β2, β3, ... are regression coefficients, and ε is the error term).
[0089] Furthermore, by training and optimizing the model using historical data, the model parameters can be determined so that the model can better fit the relationship between the bias rate and influencing factors. The performance of the model can be evaluated using methods such as cross-validation, and metrics such as mean squared error and coefficient of determination can be calculated.
[0090] Next, the initial meteorological characteristic prediction data for the period to be predicted are obtained, and the existing meteorological prediction model is used to predict the meteorological characteristics of the period to be predicted (such as the next 24 hours) to obtain the initial predicted values of wind speed, wind direction, temperature, air pressure, etc.
[0091] Next, determine the relevant influencing factors for the period to be predicted: calculate the forecast lead time for the period to be predicted (e.g., the time interval from the current moment to the start of the period to be predicted), the meteorological characteristic value at the current moment (as a reference for the initial forecast), the values of the current season, and other factors.
[0092] Next, the deviation rate correction value is calculated using the deviation rate model: The relevant influencing factor values for the period to be predicted are substituted into the established deviation rate relationship model to calculate the prediction deviation rate correction value for each meteorological characteristic. For example, the prediction deviation rate correction value for wind speed is calculated based on the model.
[0093] Finally, the meteorological feature data is optimized: the initial meteorological feature prediction data is corrected based on the calculated deviation rate correction value to obtain optimized meteorological feature data. For example, for wind speed, the corrected prediction value... Similarly, other meteorological characteristics (wind direction, temperature, air pressure, etc.) are corrected to obtain optimized meteorological characteristic data for the forecast period.
[0094] Step 206: Based on the correlation between meteorological characteristic data and power generation data, and the optimized meteorological characteristic data, the power generation prediction model is corrected.
[0095] Specifically, the correlation between meteorological characteristic data and power generation data can be measured using the Pearson correlation coefficient to assess the degree of linear correlation between the corresponding meteorological characteristic data and power generation. The Pearson correlation coefficient is as follows:
[0096] Let the meteorological characteristic data sequence be X = {x1, x2, ..., x...} n The power generation sequence is Y = {y1, y2, ..., y}. n}, then the Pearson correlation coefficient r XY The calculation formula is:
[0097]
[0098] in, These are the average values of meteorological characteristic data and power generation, respectively.
[0099] In some embodiments, reference Figure 3 Based on the correlation between meteorological characteristic data and power generation data, and the optimized meteorological characteristic data, the power generation prediction model is revised, including the following steps:
[0100] Step 302: Select the corresponding machine learning model based on the degree of linear correlation between different meteorological characteristic data and power generation data.
[0101] Step 304: Initialize the selected machine learning model.
[0102] Step 306: If the machine learning model is a linear regression model, re-estimate the regression coefficients based on the optimized meteorological feature data; if the machine learning model is a neural network model, train the model using the optimized meteorological feature data and the corresponding power generation data.
[0103] Step 308: During the training process, the error between the predicted power generation and the actual power generation is calculated according to the set loss function, and the error is backpropagated to each layer of the network to update the weights and biases, thereby obtaining the corrected power generation prediction model.
[0104] Specifically, based on the degree of linear correlation between different meteorological characteristic data and power generation, the corresponding machine learning model is selected. For example, when the Pearson coefficient indicates a high correlation between a certain meteorological characteristic data and power generation, a linear model can be selected; when the Pearson coefficient indicates a low correlation between a certain meteorological characteristic data and power generation, a statistical model can be selected.
[0105] For the selected machine learning model, initial settings are performed, determining initial values for parameters such as the number of network layers, the number of neurons in each layer, the activation function, and the learning rate. Further adjustments can be made to the model parameters: if the machine learning model is a linear regression model, the regression coefficients are re-estimated based on the optimized meteorological feature data; if the machine learning model is a neural network model, the model is trained using the optimized meteorological feature data and corresponding power generation data, with the optimized meteorological feature data as the input layer and the power generation data as the output layer, adjusting the weights and biases of the neural network through a backpropagation algorithm. During training, the error between the predicted power generation and the actual power generation is calculated according to the set loss function, and the error is backpropagated to each layer of the network to update the weights and biases. The aforementioned neural network models include, but are not limited to, BP neural networks, PSO-BP neural network models, and LSTM neural network models.
[0106] Step 208: Based on the optimized meteorological characteristic data, the wind farm power generation in the second time period is predicted using the modified power generation prediction model.
[0107] Specifically, the optimized meteorological characteristic data for the forecast period can be input into the modified power generation prediction model, ensuring that the data format and order are consistent with those used during model training. For example, if wind speed data was in the first position of the input vector during model training, then the optimized wind speed data should also be placed in the corresponding position during prediction. The modified power generation prediction model calculates the predicted power generation value for the forecast period based on the input optimized meteorological characteristic data, through its internal mathematical operations and model structure. Furthermore, some post-processing can be performed on the prediction results, such as checking the reasonableness of predictions based on the historical power generation fluctuation range of the wind farm and removing obviously unreasonable prediction values; or visualizing the prediction results to allow wind farm operation and management personnel to intuitively understand the predicted power generation situation, providing a basis for decisions regarding wind farm scheduling, operation and maintenance, and electricity market transactions.
[0108] Step 210: Validate the predicted wind farm power generation data. If the validation results do not meet the validation conditions, adjust the model parameters in the process of optimizing the meteorological characteristic data and correcting the power generation prediction model until the validation results meet the validation conditions, and obtain the target power generation prediction model for power generation prediction.
[0109] Specifically, the server can verify the predicted wind farm power generation data and obtain the verification results. If the verification results do not meet the verification conditions, the model parameters in the process of optimizing the meteorological characteristic data and correcting the power generation prediction model will be adjusted and continuously iterated until the verification conditions are met, and finally the target power generation prediction model is obtained. This target power generation prediction model can be used to predict the power generation of wind farms.
[0110] For example, to verify the predicted wind farm power generation data, specifically, the predicted wind farm power generation data can be compared with the actual wind farm power generation data for the corresponding period, and the root mean square error and mean absolute error can be calculated. If the root mean square error and mean absolute error are large, it indicates that the predicted wind farm power generation data is not accurate enough, and the model parameters in the process of optimizing meteorological characteristic data and correcting the power generation prediction model need to be adjusted.
[0111] In some embodiments, reference Figure 4 The model parameters are adjusted during the optimization of meteorological characteristic data and the correction of the power generation prediction model, including:
[0112] Step 402: Define the reinforcement learning environment, including the state space, action space, and reward function.
[0113] Step 404: Select a reinforcement learning algorithm and initialize the agent parameters.
[0114] Step 406 involves reinforcement learning training, including environment reset, action selection, environment interaction and state transition, reward calculation and storage, and agent learning.
[0115] Step 408: After training is completed, the trained agent strategy and the adjusted power generation prediction model are obtained. The trained agent strategy is used to optimize the meteorological feature data, and the target power generation prediction model is constructed based on the adjusted model parameters.
[0116] Specifically, first, a reinforcement learning environment is defined, including a state space, action space, and reward function. Then, a reinforcement learning algorithm is selected and the agent parameters are initialized. Finally, reinforcement learning training is performed, including environment reset, action selection, environment interaction and state transition, reward calculation and storage, and agent learning. After training, a trained agent policy and an adjusted power generation prediction model are obtained. The trained agent policy is used to optimize meteorological feature data, and a target power generation prediction model is constructed based on the adjusted model parameters. Figure 5 Specifically, it includes the following steps:
[0117] I. Defining the Reinforcement Learning Environment
[0118] (1) State space:
[0119] The state representation includes current meteorological characteristic data (such as wind speed, wind direction, temperature, air pressure, etc.), historical meteorological forecast deviation rate information (including deviation rate statistics for different time periods and under different meteorological conditions), and performance indicators of the current power generation forecast model (such as mean square error, accuracy, etc.). For example, the state can be represented as a vector: s = [V c D ω D t ,…,MSE prev ,], where V c This is the current wind speed, D ω This is related to wind speed deviation rate, D t This is related to temperature deviation rate, MSE prev It is the mean square error of the previous power generation forecast.
[0120] (2) Action space:
[0121] Actions correspond to adjustment strategies in the optimization process of meteorological feature data (such as adjusting the correction magnitude of wind speed forecasts, selecting the direction of wind direction deviation correction, etc.) and adjustments to the parameters of power generation prediction models (such as adjusting the update step size of weights and the learning rate in neural network models, etc.). For example, an action can be a discrete or continuous value vector representing the adjustment amount of different parameters: a = [ΔV a ,Δ θ ,…,Δ ω ], where ΔVa It is the adjustment amount for the wind speed forecast, Δ θ It is the adjustment amount of the step size, Δ ω It is the adjustment amount of the learning rate.
[0122] (3) Reward function:
[0123] The design of the reward function is a crucial step. The reward should reflect the contribution of the current action to the ultimate goal (improving the accuracy of power generation forecasting). For example, if the deviation between the predicted and actual power generation decreases after the action adjustment, a positive reward is given; conversely, if the deviation increases, a negative reward is given. Simultaneously, rewards can be given for the stability and rationality of the action; for instance, a smaller, rational adjustment receives a higher reward than a large, random adjustment. The reward function can be expressed as: r = f(ΔE, Δ a ), where ΔE is the change in the power generation forecast deviation, Δ a It refers to the range of motion.
[0124] II. Initialization of Reinforcement Learning Agent
[0125] (1) Selecting a reinforcement learning algorithm:
[0126] Deep Q-Networks (DQN), Proximal Policy Optimization (PPO), or other reinforcement learning algorithms suitable for continuous or discrete action spaces can be used. Taking DQN as an example, a neural network with a suitable structure needs to be initialized as a Q-network, whose input is the state space and whose output is the Q-value estimate for each action.
[0127] (2) Initialize agent parameters:
[0128] This includes parameters such as the weights of the neural network, the exploration rate (used to balance exploring new actions and utilizing existing experience), and the learning rate (used to update the neural network weights). For example, the weights of the neural network are initialized to small random values, the exploration rate is set to a high value (such as 0.9) to encourage exploration in the initial stage, and the learning rate is set to an appropriate value (such as 0.001) to facilitate stable learning.
[0129] III. Training Cycle
[0130] (1) Environment reset:
[0131] At the beginning of each training round, the environmental state is reset to the initial state, and the initial meteorological characteristic data, historical deviation rate information and power generation prediction model performance indicators are obtained as the initial input state s0 of the agent.
[0132] (2) Action selection:
[0133] Based on the current state, the agent selects an action a according to its policy.t .
[0134] (3) Environment interaction and state transition:
[0135] Execute the selected action a t During the optimization of meteorological feature data, adjustments are made to the meteorological feature data according to action instructions (such as correcting wind speed forecasts). During the correction of the power generation prediction model, model parameters are adjusted (such as updating neural network weights). Then, the new state s for the next time step is obtained. t+1 This includes updated meteorological characteristic data, new deviation rate information, and performance indicators of the power generation prediction model with adjusted parameters.
[0136] (4) Reward calculation and storage:
[0137] Calculate the action a based on the reward function. t The reward r obtained later t Transition the current state (s) t a t r t s t+1 The data is stored in the experience replay buffer for later use in training the agent.
[0138] (5) Agent learning:
[0139] Periodically sample a batch of state transition data from the experience replay buffer and use this data to train the agent's neural network (e.g., in DQN, calculate the target Q-value and update the neural network weights by minimizing the loss function between the target Q-value and the estimated Q-value). Adjust the agent's parameters according to the requirements of the learning algorithm (e.g., reduce the exploration rate, update the learning rate, etc.).
[0140] (6) Determining training termination conditions:
[0141] Continue training cycles until the preset training termination conditions are met, such as reaching the maximum number of training rounds or the performance indicators (such as mean square error) of the power generation prediction model converging to an acceptable range.
[0142] After training, a trained agent strategy and an adjusted power generation prediction model are obtained. The trained agent strategy can then be used to optimize meteorological feature data, and a target power generation prediction model can be constructed based on the adjusted model parameters.
[0143] By employing reinforcement learning algorithms, the optimal model parameter adjustment strategy can be automatically learned during the optimization of meteorological feature data and the correction of power generation prediction models, thereby improving the performance and adaptability of the entire prediction system, so as to better cope with complex and ever-changing meteorological conditions and improve the accuracy of wind farm power generation prediction.
[0144] In the aforementioned meteorological feature data processing method, the meteorological prediction data deviation rate is calculated based on the correspondence between historical actual meteorological feature data and historical predicted meteorological feature data stored in the associated database within the first time period. This allows for a deeper understanding of the error characteristics of the meteorological prediction model. Based on this, the power generation prediction model is corrected to effectively compensate for the power generation prediction error caused by meteorological prediction deviation, thereby significantly improving the accuracy of power generation prediction. After optimizing the meteorological feature data and correcting the power generation prediction model based on the meteorological prediction data deviation rate, the optimized prediction results are continuously verified. When the prediction results are not satisfactory, the model parameters are adjusted during the optimization of meteorological feature data and the correction of the power generation prediction model. The optimal model parameter adjustment strategy is automatically learned to improve the performance and adaptability of the power generation prediction model, so as to better cope with complex and variable meteorological conditions and improve the accuracy of wind farm power generation prediction. This avoids the limitations and subjectivity of manual parameter tuning and improves the stability and robustness of the power generation prediction model in complex and variable environments, thereby improving the accuracy of power generation prediction.
[0145] In some exemplary embodiments, such as Figure 6 As shown, this application also provides a method for predicting the power generation of a wind farm. This method can be executed by a server or by a terminal. In the embodiments of this application, this method is applied to... Figure 1 Taking the terminal in the example, the explanation includes the following steps:
[0146] Step 602: Obtain initial meteorological characteristic data for the period to be predicted.
[0147] Step 604: Optimize the initial meteorological characteristic data to obtain optimized meteorological characteristic data.
[0148] Step 606: Input the optimized meteorological characteristic data into the target power generation prediction model to obtain the predicted power generation value of the wind farm for the period to be predicted.
[0149] The target power generation prediction model can be trained using the meteorological feature data processing method provided in any of the above embodiments.
[0150] Specifically, the terminal can deploy a target power generation prediction model. When it is necessary to predict the power generation of a wind farm, the terminal can obtain the initial meteorological characteristic data for the period to be predicted, optimize the initial meteorological characteristic data to obtain optimized meteorological characteristic data, and input the optimized meteorological characteristic data into the target power generation prediction model to obtain the predicted value of the wind farm power generation for the period to be predicted.
[0151] For example, after acquiring the initial meteorological characteristic data for the period to be predicted, the terminal can optimize it using a trained intelligent agent strategy, and then input the optimized meteorological characteristic data into the target power generation prediction model to obtain the predicted value of wind farm power generation for the period to be predicted.
[0152] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.
[0153] Based on the same inventive concept, this application also provides a meteorological feature data processing device for implementing the meteorological feature data processing method described above, and a wind farm power generation prediction device for implementing the wind farm power generation prediction method described above. The solutions provided by these devices are similar to those described in the methods above. Therefore, the specific limitations of the one or more meteorological feature data processing devices and wind farm power generation prediction devices provided below can be found in the limitations of the corresponding methods above, and will not be repeated here.
[0154] In one exemplary embodiment, such as Figure 7 As shown, a meteorological feature data processing device 700 is provided, comprising:
[0155] The deviation rate calculation module 702 is used to calculate the deviation rate of meteorological forecast data based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database within the first time period.
[0156] The meteorological feature data optimization module 704 is used to optimize the meteorological feature data for the second time period based on the deviation rate of meteorological forecast data. The second time period is later than the first time period.
[0157] The model correction module 706 is used to correct the power generation prediction model based on the correlation between meteorological feature data and power generation data and the optimized meteorological feature data.
[0158] The power generation prediction module 708 is used to predict the power generation of the wind farm in the second time period based on the optimized meteorological characteristic data and the modified power generation prediction model.
[0159] The model verification and adjustment module 710 is used to verify the predicted wind farm power generation data. If the verification results do not meet the verification conditions, the model parameters are adjusted in the process of optimizing the meteorological characteristic data and correcting the power generation prediction model until the verification results meet the verification conditions, and then the target power generation prediction model for power generation prediction is obtained.
[0160] In some embodiments, the deviation rate calculation module 702 is further configured to, for each meteorological element, calculate the meteorological element deviation between the predicted value and the observed value of the meteorological element based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database within a first time period; and determine the meteorological prediction data deviation rate corresponding to the meteorological element based on the meteorological element deviation.
[0161] In some embodiments, the meteorological feature data optimization module 704 is further configured to: determine the influencing factors of the meteorological forecast data deviation rate; construct a relationship model between the meteorological forecast data deviation rate and the influencing factors; train and optimize the relationship model using historical data to obtain a deviation rate prediction model; acquire the initial meteorological feature prediction data for the second time period and determine the corresponding values of the influencing factors within the second time period; calculate the deviation rate correction value using the deviation rate prediction model based on the influencing factors within the second time period; and correct the initial meteorological feature prediction data based on the calculated deviation rate correction value to obtain optimized meteorological feature data.
[0162] In some embodiments, the model correction module 706 is further configured to select a corresponding machine learning model based on the degree of linear correlation between different meteorological feature data and power generation data; perform initialization settings for the selected machine learning model; if the machine learning model is a linear regression model, re-estimate the regression coefficients based on the optimized meteorological feature data; if the machine learning model is a neural network model, train the model using the optimized meteorological feature data and the corresponding power generation data; during the training process, calculate the error between the predicted power generation and the actual power generation based on the set loss function, and backpropagate the error to each layer of the network to update the weights and biases, thereby obtaining the corrected power generation prediction model.
[0163] In some embodiments, the model verification and adjustment module 710 is further used to define a reinforcement learning environment, including a state space, an action space, and a reward function; select a reinforcement learning algorithm and initialize agent parameters; perform reinforcement learning training, including environment reset, action selection, environment interaction and state transition, reward calculation and storage, and agent learning; after training, the trained agent policy and the adjusted power generation prediction model are obtained, the trained agent policy is used to optimize meteorological feature data, and the target power generation prediction model is constructed based on the adjusted model parameters.
[0164] In some exemplary embodiments, reference is made to Figure 8 This application provides a wind farm power generation prediction device 800, comprising:
[0165] The meteorological characteristic data acquisition module 802 is used to acquire the initial meteorological characteristic data for the period to be predicted;
[0166] The meteorological feature data optimization module 804 is used to optimize the initial meteorological feature data to obtain optimized meteorological feature data;
[0167] The wind farm power generation prediction module 806 is used to input the optimized meteorological characteristic data into the target power generation prediction model to obtain the predicted value of wind farm power generation for the period to be predicted; wherein, the target power generation prediction model is trained using the above-mentioned meteorological characteristic data processing method.
[0168] Each module in the above-mentioned device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.
[0169] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores XX data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements a meteorological characteristic data processing method.
[0170] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 10 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for predicting wind farm power generation. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.
[0171] Those skilled in the art will understand that Figure 9 , Figure 10The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0172] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the meteorological feature data processing method described above.
[0173] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described meteorological feature data processing method.
[0174] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the meteorological feature data processing method described above.
[0175] In one exemplary embodiment, another computer device is provided, including a memory and a processor, the memory storing a computer program that the processor executes to implement the steps of the wind farm power generation prediction method described above.
[0176] In one embodiment, another computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the wind farm power generation prediction method described above.
[0177] In one embodiment, another computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the wind farm power generation prediction method described above.
[0178] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0179] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.
[0180] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0181] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for processing meteorological characteristic data, characterized in that, The method includes: Based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database for the first time period, the deviation rate of meteorological prediction data is calculated. The meteorological characteristic data for the second time period are optimized based on the deviation rate of meteorological forecast data. The second time period is later than the first time period. Based on the correlation between meteorological characteristic data and power generation data, and the optimized meteorological characteristic data, the power generation prediction model is revised. Based on the optimized meteorological characteristic data, the wind farm power generation in the second time period is predicted using the modified power generation prediction model. The predicted wind farm power generation data is validated. If the validation results do not meet the validation conditions, the model parameters are adjusted during the optimization of meteorological characteristic data and the correction of the power generation prediction model until the validation results meet the validation conditions, thus obtaining the target power generation prediction model for power generation prediction.
2. The method according to claim 1, characterized in that, The step of calculating the meteorological forecast data deviation rate based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data within the first time period stored in the associated database includes: For each meteorological element, the meteorological element deviation between the predicted value and the observed value is calculated based on the correspondence between the historical actual meteorological characteristic data and the historical predicted meteorological characteristic data stored in the associated database within the first time period. Based on the deviation of the meteorological elements, the deviation rate of the meteorological forecast data corresponding to the meteorological elements is determined.
3. The method according to claim 2, characterized in that, The meteorological characteristic data for the second time period optimized based on the meteorological forecast data deviation rate includes: Identify the factors influencing the deviation rate of meteorological forecast data; A relationship model between the deviation rate of meteorological forecast data and influencing factors is constructed. The relationship model is trained and optimized using historical data to obtain a deviation rate prediction model. Obtain the initial meteorological characteristic prediction data for the second time period, and determine the values of the influencing factors within the second time period; Based on the influencing factors during the second time period, the deviation rate correction value is calculated using the deviation rate prediction model. The initial meteorological feature prediction data are corrected based on the calculated deviation rate correction value to obtain optimized meteorological feature data.
4. The method according to claim 1, characterized in that, The step of revising the power generation prediction model based on the correlation between meteorological characteristic data and power generation data, and the optimized meteorological characteristic data, includes: Select the appropriate machine learning model based on the degree of linear correlation between different meteorological characteristic data and power generation data; Perform initialization settings for the selected machine learning model; When the machine learning model is a linear regression model, the regression coefficients are re-estimated based on the optimized meteorological feature data; when the machine learning model is a neural network model, the model is trained using the optimized meteorological feature data and the corresponding power generation data. During training, the error between the predicted power generation and the actual power generation is calculated based on the set loss function. The error is then backpropagated to each layer of the network to update the weights and biases, thus obtaining the corrected power generation prediction model.
5. The method according to any one of claims 1 to 4, characterized in that, The adjustment of model parameters during the optimization of meteorological characteristic data and the correction of the power generation prediction model includes: Define the reinforcement learning environment, including the state space, action space, and reward function; Select a reinforcement learning algorithm and initialize the agent parameters; Reinforcement learning training is conducted, including environment reset, action selection, environment interaction and state transition, reward calculation and storage, and agent learning; After training, the trained agent strategy and the adjusted power generation prediction model are obtained. The trained agent strategy is used to optimize the meteorological feature data, and the target power generation prediction model is constructed based on the adjusted model parameters.
6. A method for predicting wind farm power generation, characterized in that, The method includes: Obtain initial meteorological characteristic data for the period to be predicted; The initial meteorological feature data is optimized to obtain optimized meteorological feature data; The optimized meteorological characteristic data is input into the target power generation prediction model to obtain the predicted power generation value of the wind farm during the prediction period. The target power generation prediction model is trained using the meteorological characteristic data processing method as described in any one of claims 1 to 5.
7. A meteorological characteristic data processing device, characterized in that, The device includes: The deviation rate calculation module is used to calculate the deviation rate of meteorological forecast data based on the correspondence between historical actual meteorological characteristic data and historical predicted meteorological characteristic data stored in the associated database within the first time period. The meteorological feature data optimization module is used to optimize the meteorological feature data for a second time period based on the deviation rate of meteorological forecast data. The second time period is later than the first time period. The model correction module is used to correct the power generation prediction model based on the correlation between meteorological feature data and power generation data, as well as the optimized meteorological feature data. The power generation prediction module is used to predict the power generation of the wind farm in the second time period based on the optimized meteorological characteristic data and the modified power generation prediction model. The model validation and adjustment module is used to validate the predicted wind farm power generation data. If the validation results do not meet the validation conditions, the model parameters are adjusted during the optimization of meteorological characteristic data and the correction of the power generation prediction model until the validation results meet the validation conditions, thus obtaining the target power generation prediction model for power generation prediction.
8. A wind farm power generation prediction device, characterized in that, The device includes: The meteorological characteristic data acquisition module is used to acquire the initial meteorological characteristic data for the period to be predicted. The meteorological feature data optimization module is used to optimize the initial meteorological feature data to obtain optimized meteorological feature data. The wind farm power generation prediction module is used to input the optimized meteorological characteristic data into the target power generation prediction model to obtain the predicted value of wind farm power generation for the period to be predicted; wherein, the target power generation prediction model is trained using the meteorological characteristic data processing method as described in any one of claims 1 to 5.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5 or claim 6.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5 or claim 6.