Wind power prediction method and system, computer equipment and storage medium

Abstract

The invention provides a wind power prediction method and system, computer equipment and a storage medium, and belongs to the technical field of wind power prediction.The method comprises the steps that firstly, a missing data filling method is proposed to be combined with cross multiple interpolation (CMI) to obtain a complete environment variable data set; secondly, feature variables are selected based on the maximum information coefficient, so that a data set is simplified, and needed feature vectors are obtained; and finally, training a time convolutional network (TCN) based on the data of the characteristic variable set to obtain a wind power prediction model, inputting the operation data of the to-be-detected fan into the wind power prediction model, and outputting a wind power prediction value. Compared with other models, the method has better prediction performance, and provides data support for subsequent total wind power analysis and prediction.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power prediction, and in particular relates to a wind power prediction method, system, computer equipment and storage medium. Background Art

[0002] Today, artificial intelligence and deep learning models are widely used. Existing technologies usually use machine learning models to model wind farms and make predictions based on physical models and historical data. Although wind power prediction methods have made progress, incomplete data and poor quality still affect the accuracy of predictions. In order to ensure the accuracy of data analysis, missing values ​​must be reasonably filled. Common methods for filling missing values ​​in wind power data include mean filling, median filling, mode filling, and interpolation filling. However, the above data filling methods still have problems such as a single reconstruction method and limited universality. As a result, the filling results are not accurate enough and cannot fully reflect the true distribution or characteristics of the data, affecting the subsequent analysis and prediction of the total amount of wind power. Summary of the invention

[0003] In order to solve the problems of data quality and model accuracy in wind power prediction, the present invention provides a wind power prediction method, system, computer equipment and storage medium.

[0004] In order to achieve the above object, the present invention provides the following technical solutions: A wind power prediction method comprises the following steps: Acquire the original data set of wind turbine operation; use the cross multiple interpolation method to fill the original data set of wind turbine operation; normalize the data set of wind turbine operation after filling the missing values ​​to obtain different operation variables; calculate the maximum information coefficient of the different operation variables, and select the variable whose maximum information coefficient with the wind target variable is greater than a set threshold as the characteristic variable according to the maximum information coefficient obtained by different variables; Taking the characteristic variables as input, training the time series prediction TCN model to obtain a wind power prediction model; The wind turbine operation data to be detected is input into the wind power prediction model, and the wind power prediction value is output.

[0005] Preferably, the cross multiple interpolation method is used to fill the original data set of the wind turbine operation, specifically using multiple interpolation models in the sklearn module to fill the data, and cross-validating the filling results of the multiple interpolation models to select the optimal interpolation result.

[0006] Preferably, the maximum information coefficient MIC of the different operating variables is calculated, specifically by the following formula: ; in, Operation variables for fans and The mutual information of yes and At the same time get and The joint probability of is a variable Get the value The marginal probability of is a variable Get the value The marginal probability of .

[0007] Preferably, the temporal convolution network (TCN) model uses causal convolution and dilated convolution and introduces a residual module.

[0008] Preferably, before the cross multiple interpolation method is used to fill the original data set of the fan operation, the method also includes clearing abnormal data, which specifically includes the following steps: setting an abnormal value standard; screening out data that meets the abnormal value standard; and replacing the cleared abnormal data with the mean.

[0009] Preferably, the characteristic variables input into the time network convolution TCN model specifically include: generator speed, converter torque, data availability status, average wind speed, power generation time and environmental time.

[0010] The present invention also provides a wind power prediction system, which specifically includes: A data module is used to obtain an original data set of wind turbine operation; fill in the original data set of wind turbine operation using a cross multiple interpolation method; normalize the data set of wind turbine operation after filling in missing values ​​to obtain different operating variables; calculate the maximum information coefficient of the different operating variables, and select a variable whose maximum information coefficient with a wind target variable is greater than a set threshold as a characteristic variable according to the maximum information coefficient obtained for different variables.

[0011] The variable module is used to take the characteristic variables as input to train the time series prediction TCN model to obtain a wind power prediction model.

[0012] The prediction module is used to input the wind turbine operation data to be detected into the wind power prediction model and output the wind power prediction value.

[0013] The present invention also provides a computer device, comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps described in the wind power prediction method.

[0014] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is loaded by a processor, the steps described in the wind power prediction method can be executed.

[0015] The wind power prediction method provided by the present invention has the following beneficial effects: The present invention obtains the original data set of wind turbine operation, and uses the cross multiple interpolation (CMI) method to fill the data to obtain a complete data set, providing reliable data support for model processing, while improving data quality and the accuracy of subsequent predictions. Based on the characteristic variable of the maximum information coefficient, the variable strongly correlated with the wind target variable is selected as the characteristic variable, and the characteristic vector is input into the wind power prediction model established based on the time convolution network (TCN), and the wind power prediction value is output, which is conducive to observing the situation of the wind farm. Compared with other models, by inputting a more complete data set, the effect of subsequent analysis and prediction of the total amount of wind power is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiment of the present invention and its design scheme, the following briefly introduces the drawings required for this embodiment. The drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 The present invention provides a flow chart of a wind power prediction method.

[0018] Figure 2 is a comparison diagram of wind power prediction results in an embodiment of the present invention, wherein: Figure 2 (a) is the wind power prediction result diagram of the multi-layer perceptron MLP model; Figure 2 (b) is the wind power prediction result diagram of the long short-term memory network LSTM model; Figure 2 (c) is the wind power prediction result diagram of the time series prediction TCN model.

[0019] Figure 3 This is an error histogram of wind power prediction using MSE (mean square error) in an embodiment of the present invention.

[0020] Figure 4 This is an error histogram of wind power prediction using MAE (mean absolute error) in an embodiment of the present invention.

[0021] Figure 5 The figure is an error histogram of wind power prediction using MAPE (mean absolute percentage error) in an embodiment of the present invention.

[0022] Figure 6The figure is an error histogram of wind power prediction using RMSE (root mean square error) in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the technical solution of the present invention and implement it, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the scope of protection of the present invention.

[0024] Example The present invention provides a wind power prediction method, such as Figure 1 As shown, the specific steps include: S1. Obtain the original data set, which includes generator set parameters, environmental variables, and wind power target variables.

[0025] S2. Process the original data. First, by defining the outlier standard, filter out the outliers that meet the standard and replace them with the mean to remove the abnormal data to ensure that the data set is suitable for further analysis. Then, the CMI filling method is used for data filling, and multiple interpolation models in the python external modules "sklearn.impute" and "sklearn.ensemble" are used for data filling. The interpolation modules specifically include: linear regression, logistic regression, decision tree and random forest. The filling results of multiple interpolation models are cross-validated, the complete data is artificially emptied, and the error is calculated based on the filling results of multiple interpolation models and the data that is not emptied, and the optimal interpolation result is selected. Involving a variety of interpolation techniques to fill missing data, including mean filling, which is used to obtain the mean of existing data to replace missing values; forward filling, using the previous valid data point to fill missing values; multiple interpolation filling, using statistical models to estimate and fill missing values; cross-multiple interpolation filling, on the basis of multiple interpolation filling, adding mutual references between data of different heights. Finally, normalize the data set: after filling the missing values, the data is normalized to ensure consistency between different variables.

[0026] S3. Calculate the maximum information coefficient MIC of different operating variables, and select the variable whose maximum information coefficient with the wind target variable is greater than a set threshold as the characteristic variable according to the maximum information coefficient obtained for different variables.

[0027] Input variables: The variables considered include average wind speed, wind direction, wind angle, ambient temperature and generator speed, etc. Calculated by the following MIC formula: ; in, is to run the variable under a certain grid partition and The mutual information of the grid refers to the gridding of the coordinate axes according to the set values ​​in two-dimensional space; yes and At the same time get and The joint probability of is a variable Get the value The marginal probability of is a variable Get the value The marginal probability of .

[0028] The six characteristic variables with the strongest correlation with the wind target variable were selected according to the maximum information coefficient (MIC) to filter and simplify the input data. These characteristic variables were standardized and then used as input data for the model.

[0029] S4. Build the CMI-TCN network, input the feature variables into the time network convolution TCN model, and output the wind power prediction value. Construct the TCN model for wind power prediction. The TCN model can effectively capture the time series characteristics of long sequences through causal convolution and dilated convolution, with low computational complexity and high prediction accuracy. Introduce the residual module to alleviate the gradient vanishing problem in the deep network and ensure the effective training of the model.

[0030] In order to verify the effectiveness and rationality of the model adopted by the present invention, the output of the model of the present invention is compared with that of the multi-layer perceptron MLP model and the long short-term memory network LSTM on the acquired complete data set, as shown in FIG. Figure 2 , where the vertical axis power is a normalized dimensionless unit. Display results: Visualize the prediction results of the model (for example, a comparison chart of the predicted value and the actual value).

[0031] Compute and compare evaluation metrics such as Figure 3 Shown is MAPE (mean absolute percentage error), Figure 4 Shown is RMSE (root mean square error), Figure 5 Shown is MAE (mean absolute error), Figure 6 Shown is the MSE (mean squared error) to assess the accuracy of the model.

[0032] The present invention also provides a wind power prediction system, which specifically includes: The data module is used to obtain the original data set of wind turbine operation; use the cross multiple interpolation method to fill the original data set of wind turbine operation; normalize the data set of wind turbine operation after filling the missing values ​​to obtain different operation variables; calculate the maximum information coefficient of different operation variables, and select the variables that are strongly correlated with the wind target variables as feature variables according to the maximum information coefficients obtained for different variables.

[0033] The variable module is used to take the characteristic variables as input to train the time series prediction TCN model to obtain the wind power prediction model.

[0034] The prediction module is used to input the wind turbine operation data to be detected into the wind power prediction model and output the wind power prediction value.

[0035] Each module in the above-mentioned wind power prediction system can be implemented in whole or in part by software, hardware and a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module above.

[0036] The present invention also provides a computer device, including a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps in a wind power prediction method embodiment. The specific implementation method can be found in the method embodiment, which will not be described in detail here.

[0037] Furthermore, the present invention also provides a non-temporary computer-readable storage medium containing instructions, and a computer program is stored on the storage medium. For example, a memory containing instructions, the above instructions can be executed by a processor of a computer device to complete the above method. For example, the non-temporary computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device. When the computer program is executed by the processor, the steps in an embodiment of a wind power prediction method can be implemented. The specific implementation method can be found in the method embodiment, which will not be repeated here.

[0038] It should be understood by those skilled in the art that embodiments of the present invention may provide methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0039] The present invention is described with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as a combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0040] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0041] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0042] It should be pointed out that the specific implementation methods described above can enable those skilled in the art to understand the invention more comprehensively, but do not limit the invention in any way. Therefore, although the invention has been described in detail in this specification and embodiments, those skilled in the art should understand that the invention can still be modified or replaced by equivalents; and all technical solutions and improvements that do not deviate from the spirit and scope of the invention are included in the protection scope of the patent for the invention. Any figure mark in the claims should not be regarded as limiting the claims involved. Any simple change or equivalent replacement of the technical solution that can be obviously obtained by any technician familiar with the field within the technical scope disclosed in the present invention belongs to the protection scope of the present invention.

Claims

1. A wind power prediction method, characterized in that: The following steps are involved: Acquire an original data set of the fan operation; use a cross multiple interpolation method to fill in the original data set of the fan operation; normalize the data set of the fan operation after the missing values ​​are filled to obtain different operation variables; Calculating the maximum information coefficients of the different operating variables, and selecting a variable whose maximum information coefficient with the wind power target variable is greater than a set threshold as a characteristic variable according to the maximum information coefficients obtained for the different variables; Taking the characteristic variables as input, training the time series prediction TCN model to obtain a wind power prediction model; The wind turbine operation data to be detected is input into the wind power prediction model, and the wind power prediction value is output.

2. A wind power prediction method according to claim 1, characterized in that: The cross multiple interpolation method is used to fill the original data set of the fan operation, specifically using multiple interpolation models in the sklearn module to fill the data, and cross-validating the filling results of multiple interpolation models, leaving the complete data blank, and calculating the error based on the filling results of multiple interpolation models and the data that are not blank, and selecting the optimal interpolation result.

3. A wind power prediction method according to claim 1, characterized in that: The maximum information coefficient MIC of the different operating variables is calculated, specifically by the following formula: ； in, Operation variables for fans and The mutual information of yes and At the same time get and The joint probability of is a variable Get the value The marginal probability of is a variable Get the value The marginal probability of .

4. A wind power prediction method according to claim 1, characterized in that: The temporal convolution network (TCN) model uses causal convolution and hole convolution and introduces a residual module.

5. A wind power prediction method according to claim 1, characterized in that: Before using the cross multiple interpolation method to fill the original data set of the fan operation, it also includes clearing abnormal data, which specifically includes the following steps: setting an abnormal value standard; screening out data that meets the abnormal value standard; replacing the cleared abnormal data with the mean.

6. A wind power prediction method according to claim 1, characterized in that: The characteristic variables input into the time network convolution TCN model specifically include: generator speed, converter torque, data availability status, average wind speed, power generation time and environment time.

7. A wind power prediction system, characterized in that: include: A data module is used to obtain an original data set of the fan operation; and fill the original data set of the fan operation with data by using a cross-multiple interpolation method; The wind turbine operation data set after filling the missing values ​​is normalized to obtain different operation variables; Calculating the maximum information coefficients of the different operating variables, and selecting a variable whose maximum information coefficient with the wind power target variable is greater than a set threshold as a characteristic variable according to the maximum information coefficients obtained for the different variables; A variable module is used to take the characteristic variables as input to train the time series prediction TCN model to obtain a wind power prediction model; The prediction module is used to input the wind turbine operation data to be detected into the wind power prediction model and output the wind power prediction value.

8. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is loaded into a processor, it can execute the steps of the method according to any one of claims 1 to 6.

Images