Generated power prediction method and system based on CFS meteorology

Abstract

The invention relates to the technical field of new energy power generation, in particular to a generated power prediction method and system based on CFS meteorology, and the method or system mainly comprises the steps: carrying out the preprocessing of measured data and CFS meteorological data, carrying out the feature selection and feature extraction after preprocessing, and constructing a data set. Dividing the data set into a training set and a test set, verifying the prediction model through the test set, outputting an optimal prediction model, integrating obtained prediction results, and outputting a prediction result. Through the method, the meteorological condition and the electric field data are fused, medium and long term power prediction is realized, and the accuracy of medium and long term power prediction is improved.

Description

Technical Field

[0001] The present invention relates to the field of renewable energy power generation technology, and in particular to a method and system for predicting power generation based on CFS meteorology. Background Art

[0002] In the prior art, when predicting medium- and long-term power, only the historical data of electric field measurements are generally used to predict the power in the future. However, for renewable energy power generation, meteorological factors have a great impact on the accuracy of power prediction. The prior art lacks a method to integrate meteorological data and electric field data to achieve medium- and long-term power prediction. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for predicting power generation based on CFS meteorology to solve the above problems in the prior art.

[0004] The present invention is achieved through the following technical solutions:

[0005] First, the power generation prediction method based on CFS meteorology includes:

[0006] Obtaining the measured power generation data of the power station and CFS meteorological data, wherein the CFS meteorological data includes data of four dimensions, and preprocessing the measured data and the CFS meteorological data, and performing feature selection and feature extraction after the preprocessing;

[0007] The data after the physical sign extraction is normalized to obtain data with the same scale and a data set is constructed;

[0008] Divide the data set into a training set and a test set, establish a prediction model, train the prediction model with the training set, and when the set number of iterations is reached, verify the prediction model with the test set and output the optimal prediction model;

[0009] The prediction model is used to combine the CFS meteorological data of four dimensions to predict the power generation, and the prediction results of the power generation are obtained. The obtained prediction results are integrated and output.

[0010] Preferably, the pretreatment comprises:

[0011] Identify missing values, duplicate data, and outliers, interpolate missing values, remove outliers, and then interpolate and remove duplicate data.

[0012] Preferably, the interpolation includes:

[0013] Determine the data interval to be interpolated and select the order of B-spline;

[0014] Calculate the B-spline basis function and obtain the spline curve through the basis function and the data interval;

[0015] Select the interpolated value range and randomly obtain the data of a point on the spline curve as the interpolated value.

[0016] Preferably, the normalization process includes:

[0017]

[0018] In the formula, X is the original feature variable, X′ is the normalized feature variable, and X max is the maximum value of the characteristic variable, X min is the minimum value of the characteristic variable.

[0019] Preferably, the establishing of the prediction model comprises:

[0020] Calculate the distance between data points, select the K nearest adjacent data points, and determine their categories;

[0021] The data with the largest number of categories is selected as the prediction result.

[0022] Preferably, calculating the distance between data points includes:

[0023]

[0024] In the formula, x i and x j are two data points respectively, n is the feature dimension, x k is the feature vector of the target data point, x jk is the feature vector of the jth data point in the training set and k is the number of features.

[0025] Preferably, the categories described and determined include:

[0026]

[0027] In the formula, Class new is the category of the data point to be classified, Class i is the category of the i-th neighboring data point, c is a traversal of all possible categories, I() is the indicator function, and K is the selected value.

[0028] Preferably, integrating the obtained prediction results and outputting the prediction results comprises:

[0029]

[0030] In the formula, E avg is the prediction result, N is the number of prediction results, S j is the prediction result of the jth dimension.

[0031] In a second aspect, the present invention further provides a power generation prediction system based on CFS meteorology, comprising:

[0032] A data acquisition module is configured to acquire the measured power generation data of the power station and CFS meteorological data, wherein the CFS meteorological data includes data of four dimensions, and pre-process the measured data and the CFS meteorological data, and perform feature selection and feature extraction after the pre-processing;

[0033] The model building module is configured to normalize the data after the vital signs are extracted to obtain data with the same scale and construct a data set; divide the data set into a training set and a test set, establish a prediction model, train the prediction model through the training set, and when the set number of iterations is reached, verify the prediction model through the test set, and output the optimal prediction model;

[0034] The output module is configured to predict the power generation by combining the CFS meteorological data of four dimensions through the prediction model, obtain the prediction result of the power generation, integrate the obtained prediction results, and output the prediction results;

[0035] The main control device is connected with the data acquisition module, the model building module and the output module, and is used to execute the above-mentioned CFS meteorological-based power generation prediction method.

[0036] The technical solution of the embodiment of the present invention has at least the following advantages and beneficial effects:

[0037] The method or system provided by the present invention mainly includes preprocessing the measured data and CFS meteorological data, performing feature selection and feature extraction after preprocessing, and constructing a data set; dividing the data set into a training set and a test set, verifying the prediction model through the test set, outputting the optimal prediction model, integrating the obtained prediction results, and outputting the prediction results. By integrating meteorological conditions with electric field data through the above method, medium- and long-term power prediction is realized, and the accuracy of medium- and long-term power prediction is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0041] The division of modules in this application is a logical division. There may be other division methods when implemented in actual applications. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0042] The modules or submodules described independently may be physically separated or not: they may be implemented by software or hardware, and some modules or submodules may be implemented by software, and the processor may call the software to implement the functions of these modules or submodules, and other modules or submodules may be implemented by hardware, such as by hardware circuits. In addition, some or all of the modules may be selected according to actual needs to achieve the purpose of the present application.

[0043] Please refer to Figure 1 The present invention provides a method for predicting power generation based on CFS meteorology, comprising:

[0044] S101: obtaining the measured power generation data of the power station and CFS meteorological data, wherein the CFS meteorological data includes data of four dimensions, and preprocessing the measured data and the CFS meteorological data, and performing feature selection and feature extraction after the preprocessing;

[0045] Specifically, the weather forecast of the CFS climate prediction model includes 4 members, corresponding to four dimensions, each member has 5 points, of which point 001 is the longitude and latitude of the electric field center. The weather forecast duration can meet the needs of the next 45 days. The model is modeled using multiple regression algorithms and the results of different models are integrated.

[0046] The data of the four dimensions include

[0047] Time Dimension:

[0048] The time dimension refers to the changes and development of data over time, usually including different time steps, such as hours, days, months and years.

[0049] Spatial Dimensions:

[0050] The spatial dimension involves data about geographic locations, usually represented as a grid of longitude and latitude (such as a global grid or a regional grid). Each grid point corresponds to a specific geographic location, covering both land and ocean.

[0051] Meteorological variable dimensions:

[0052] The meteorological variable dimension refers to the various meteorological variables measured, such as temperature, humidity, precipitation, wind speed, air pressure, etc. Each variable has its own unique value in time and space.

[0053] Depth dimension (or layer dimension):

[0054] For three-dimensional spatial data, such as atmospheric or ocean models, the depth dimension represents different levels of height or depth. For example, meteorological data in the atmosphere may include different altitude layers from the ground to the upper atmosphere (such as troposphere, stratosphere, etc.), while ocean data may involve different depths from the sea surface to the seabed.

[0055] S102: normalizing the data after the physical sign extraction to obtain data with the same scale, and constructing a data set;

[0056] Select or construct features that have a significant impact on power output, such as the five features selected, such as historical power, temperature, humidity, wind speed, and precipitation; then standardize the data to have the same scale and construct a data set.

[0057] S103: Divide the data set into a training set and a test set, establish a prediction model, train the prediction model through the training set, and when the set number of iterations is reached, verify the prediction model through the test set, and output the optimal prediction model;

[0058] S104: The power generation is predicted by combining the CFS meteorological data of the four dimensions with the prediction model to obtain the prediction result of the power generation, and the obtained prediction results are integrated to output the prediction results.

[0059] The method or system provided by the present invention mainly includes preprocessing the measured data and CFS meteorological data, performing feature selection and feature extraction after preprocessing, and constructing a data set; dividing the data set into a training set and a test set, verifying the prediction model through the test set, outputting the optimal prediction model, integrating the obtained prediction results, and outputting the prediction results. The method of fusing meteorological conditions with electric field data by the above method lacks the meteorological prediction dimension in the medium and long-term power prediction, which affects the accuracy of the medium and long-term power prediction.

[0060] In an exemplary embodiment of the present invention, the pretreatment comprises:

[0061] Identify missing values, duplicate data, and outliers, interpolate missing values, remove outliers, and then interpolate and remove duplicate data.

[0062] Specifically, interpolation includes:

[0063] S201: Determine the data interval to be interpolated and select the order of the B-spline;

[0064] For example: linear splines are of order 1, quadratic splines are of order 2, cubic splines are of order 3, etc.). The order determines the polynomial degree of each spline segment.

[0065] S202: Calculate the B-spline basis function, and obtain a spline curve through the basis function and the data interval;

[0066]

[0067] Where N i，0 (u) is the basis function, i is the index value of the basis function, k is the order, and u is the parameter value.

[0068]

[0069] Where C(u) is the generated spline curve, P i is the control point and n is a constant.

[0070] S203: Selecting a value range to be interpolated, and randomly obtaining data of a point on the spline curve as the interpolated value.

[0071] In an exemplary embodiment of the present invention, the normalization process includes:

[0072]

[0073] X=X′(X max-Xmin +X min

[0074] In the formula, X is the original feature variable, X′ is the normalized feature variable, and X max is the maximum value of the characteristic variable, X min is the minimum value of the characteristic variable.

[0075] An exemplary implementation of the present invention uses KNN to establish a prediction model. KNN, i.e., the K-nearest neighbor algorithm (k-NearestNeighbor), is the K nearest neighbors. When an unknown sample needs to be predicted, it is determined by the K neighbors closest to the sample. KNN can be used for both classification problems and regression problems. When performing classification prediction, the one with the largest number of categories (or the most weighted) among the K neighbors is used as the prediction result; when performing regression prediction, the mean (or weighted mean) of the K neighbors is used as the prediction result.

[0076] In the KNN algorithm, choosing a suitable k value is crucial because it directly affects the performance and effect of the algorithm. A smaller k value means that the model is more dependent on local features in the training set, which will make the model too sensitive to noise data and easily lead to overfitting; a larger k value will cause the algorithm to consider more neighboring points when making decisions. Although it can reduce the impact of noise data, it will cause the model to be unable to capture the local features of the data, thus causing underfitting. Therefore, choosing a moderate k value is the key to achieving the best generalization ability of the model. In practical applications, the selection of k value is usually determined by cross-validation and other techniques, with the aim of finding an optimal balance between bias and variance, thereby improving the model's prediction accuracy for new data.

[0077] The decision-making rules in the KNN algorithm are usually based on the "majority voting" principle. For each sample to be classified, the KNN algorithm first determines the k nearest neighbors in the training set, and then determines the category of the sample to be classified based on the categories of these neighbors. If the number of samples in a certain category among the k nearest neighbors is the largest, then the sample to be classified is assigned to this category. In some cases, in order to improve the accuracy of classification, a weighted voting mechanism can also be used, that is, neighbors closer to the sample to be classified have a greater influence on the final voting result. This method ensures that samples with closer distances have higher weights in the decision-making process, so that the classification results can more accurately reflect the local characteristics of the data.

[0078] Specifically, they include:

[0079] Calculate the distance between data points, select the K nearest adjacent data points, and determine their categories;

[0080] The data with the largest number of categories is selected as the prediction result.

[0081] Preferably, calculating the distance between data points includes:

[0082]

[0083] In the formula, x i and xj are two data points respectively, n is the feature dimension, x k is the feature vector of the target data point, x jk is the feature vector of the jth data point in the training set and k is the number of features.

[0084] Preferably, the categories described and determined include:

[0085]

[0086] In the formula, Class new is the category of the data point to be classified, Class i is the category of the i-th neighboring data point, c is a traversal of all possible categories, I() is the indicator function, and K is the selected value.

[0087] Specifically, the obtained prediction results are integrated and outputted, including:

[0088]

[0089] In the formula, E avg is the prediction result, N is the number of prediction results, S j is the prediction result of the jth dimension.

[0090] In another embodiment of the present invention, the prediction model can also be constructed by using the SVR algorithm, which is the full name of the SVR algorithm, support vector regression, and is an application of SVM (support vector machine) to regression problems. The SVR model can be simply understood as creating an "interval band" on both sides of the linear function, and the spacing of this "interval band" is ∈ (this value is often given based on experience), and no loss is calculated for all samples falling into the interval band, that is, only the support vector will affect its function model, and finally the optimized model is obtained by minimizing the total loss and maximizing the interval.

[0091] Specifically, when the electric field power is predicted by the Support Vector Regression (SVR) algorithm using the measured power generation data of the power station and the CFS (Climate Forecast System) meteorological data, the data set is constructed in the same way as the data set construction process in the knn algorithm, and then the training set data is used to train the SVR model. During the training process, the kernel function is selected through cross-validation, and the optimal hyperplane is obtained, so that the sum of the distances from all training samples to this hyperplane is minimized.

[0092] In another embodiment of the present invention, ElasticNet can also be used to construct a prediction model. The ElasticNet algorithm is a linear regression model, which is a hybrid of Lasso regression and Ridge regression. The ElasticNet algorithm can control the complexity of the model by adjusting two parameters λ1 and λ2, so that the model performs better on the training set. Among them, when λ1=0, the ElasticNet algorithm degenerates into Ridge regression; when λ2=0, the ElasticNet algorithm degenerates into Lasso regression; when λ1=λ2, the ElasticNet algorithm degenerates into an ordinary linear regression model.

[0093] In a second aspect, the present invention further provides a power generation prediction system based on CFS meteorology, comprising:

[0094] A data acquisition module is configured to acquire the measured power generation data of the power station and CFS meteorological data, wherein the CFS meteorological data includes data of four dimensions, and pre-process the measured data and the CFS meteorological data, and perform feature selection and feature extraction after the pre-processing;

[0095] The model building module is configured to normalize the data after the vital signs are extracted to obtain data with the same scale and construct a data set; divide the data set into a training set and a test set, establish a prediction model, train the prediction model through the training set, and when the set number of iterations is reached, verify the prediction model through the test set, and output the optimal prediction model;

[0096] The output module is configured to respectively predict the CFS meteorological data of four dimensions through the prediction model, integrate the obtained prediction results, and output the prediction results;

[0097] The main control device is connected with the data acquisition module, the model building module and the output module, and is used to execute the above-mentioned CFS meteorological-based power generation prediction method.

[0098] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0099] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0100] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The power generation prediction method based on CFS meteorology is characterized by: include: Obtaining the measured power generation data of the power station and CFS meteorological data, wherein the CFS meteorological data includes data of four dimensions, and preprocessing the measured data and the CFS meteorological data, and performing feature selection and feature extraction after the preprocessing; The data after the physical sign extraction is normalized to obtain data with the same scale and a data set is constructed; Divide the data set into a training set and a test set, establish a prediction model, train the prediction model with the training set, and when the set number of iterations is reached, verify the prediction model with the test set and output the optimal prediction model; The prediction model is used to combine the CFS meteorological data of four dimensions to predict the power generation, and the prediction results of the power generation are obtained. The obtained prediction results are integrated and output.

2. The power generation prediction method based on CFS meteorology according to claim 1 is characterized in that: The pre-processing comprises: Identify missing values, duplicate data, and outliers, interpolate missing values, remove outliers, and then interpolate and remove duplicate data.

3. The power generation prediction method based on CFS meteorology according to claim 2 is characterized in that: The interpolation includes: Determine the data interval to be interpolated and select the order of B-spline; Calculate the B-spline basis function and obtain the spline curve through the basis function and the data interval; Select the interpolated value range and randomly obtain the data of a point on the spline curve as the interpolated value.

4. The power generation prediction method based on CFS meteorology according to claim 3 is characterized in that: The normalization process includes: X=X′(X max -X min )+X min In the formula, X is the original feature variable, X′ is the normalized feature variable, and X max is the maximum value of the characteristic variable, X min is the minimum value of the characteristic variable.

5. The method for predicting power generation based on CFS meteorology according to claim 4, characterized in that: The establishment of the prediction model comprises: Calculate the distance between data points, select the K nearest adjacent data points, and determine their categories; The data with the largest number of categories is selected as the prediction result.

6. The method for predicting power generation based on CFS meteorology according to claim 5, characterized in that: The calculation of the distance between data points includes: In the formula, x i and x j are two data points respectively, n is the feature dimension, x k is the feature vector of the target data point, x jk is the feature vector of the jth data point in the training set and k is the number of features.

7. The method for predicting power generation based on CFS meteorology according to claim 6, characterized in that: Described and identified as including: In the formula, Class new is the category of the data point to be classified, Class i is the category of the i-th neighboring data point, c is a traversal of all possible categories, I() is the indicator function, and K is the selected value.

8. The method for power generation prediction based on CFS meteorology according to claim 7, characterized in that: The obtained prediction results are integrated and the output prediction results include: In the formula, E avg is the prediction result, N is the number of prediction results, S j is the prediction result of the jth dimension.

9. The power generation prediction system based on CFS meteorology is characterized by: include: A data acquisition module is configured to acquire the measured power generation data of the power station and CFS meteorological data, wherein the CFS meteorological data includes data of four dimensions, and pre-process the measured data and the CFS meteorological data, and perform feature selection and feature extraction after the pre-processing; The model building module is configured to normalize the data after the vital signs are extracted to obtain data with the same scale and construct a data set; divide the data set into a training set and a test set, establish a prediction model, train the prediction model through the training set, and when the set number of iterations is reached, verify the prediction model through the test set, and output the optimal prediction model; The output module predicts the power generation by combining the CFS meteorological data of four dimensions through the prediction model, obtains the prediction results of the power generation, integrates the obtained prediction results, and outputs the prediction results; A main control device is connected to the data acquisition module, the model building module and the output module, and is used to execute the power generation prediction method based on CFS meteorology as described in any one of claims 1-8.

Images