CN114418210B - Methods and devices for predicting intraday wind power - Google Patents
Methods and devices for predicting intraday wind powerInfo
- Publication number
- CN114418210B CN114418210B CN202210051045.6A CN202210051045A CN114418210B CN 114418210 B CN114418210 B CN 114418210B CN 202210051045 A CN202210051045 A CN 202210051045A CN 114418210 B CN114418210 B CN 114418210B
- Authority
- CN
- China
- Prior art keywords
- data
- neural network
- prediction
- network model
- wind power
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/251—Fusion techniques of input or preprocessed data
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Tourism & Hospitality (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Public Health (AREA)
- Evolutionary Biology (AREA)
- Primary Health Care (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本发明公开了一种日内风功率的预测方法以及预测装置，该方法包括：获取多个初始大气预测数据，每个初始大气预测数据从不同数据源获取；对多个初始大气预测数据执行融合操作，获得融合后数据；对融合后数据执行经验模态分解，获得分解后数据；获取复合神经网络模型；基于复合神经网络模型对分解后数据进行分析，生成对应的日内风功率预测值。通过对多个数据源的天气数据进行精确性处理，提高天气预报数据的精确性，对天气数据中的风电场数据进行经验模态分解，将原始的基于时序的天气数据转换为能够输入智能学习模型进行处理的平稳数据，通过智能化学习模型对上述数据进行自动学习，输出精准的超短期风功率预测值，满足了技术人员的实际需求。
This invention discloses a method and device for predicting intraday wind power. The method includes: acquiring multiple initial atmospheric forecast data, each acquired from a different data source; performing a fusion operation on the multiple initial atmospheric forecast data to obtain fused data; performing empirical mode decomposition on the fused data to obtain decomposed data; acquiring a composite neural network model; and analyzing the decomposed data based on the composite neural network model to generate corresponding intraday wind power prediction values. By performing precision processing on weather data from multiple data sources, the accuracy of weather forecast data is improved. Empirical mode decomposition is performed on wind farm data within the weather data, converting the original time-series-based weather data into stationary data that can be input into an intelligent learning model for processing. The intelligent learning model automatically learns from the above data and outputs accurate ultra-short-term wind power prediction values, meeting the practical needs of technical personnel.
Description
Technical Field
The present invention relates to the field of prediction of weather data, and in particular, to a method for predicting solar wind power, a device for predicting solar wind power, and a computer-readable storage medium.
Background
With the great development of low-carbon economy, wind energy is used as a clean and environment-friendly renewable energy source, is rich in resources, does not need to be mined and transported, and becomes one of the most promising new energy sources, so that the wind power generation technology is widely applied and popularized on a large scale.
With the rapid development of wind power generation, the installed capacity of wind power also rapidly develops, the proportion of wind power in a power grid is continuously increased, but the inherent intermittence and fluctuation characteristics of wind power bring about serious threat to the electric energy quality and the safety and stability of a power system, so that the wind power is required to be effectively predicted in order to solve the problem of impact of large-scale wind power access to the power grid on the power system.
There are various techniques for forecasting weather in the prior art, such as a weather prediction method based on a physical model, a weather prediction method based on statistics, and a weather prediction method based on a learning algorithm. In the practical application process, however, the prediction method based on the physical model cannot take into consideration various complex factors actually influencing weather, so that the accuracy is low, the prediction method based on statistics can predict weather to a certain extent, the long-term prediction accuracy is still low, the weather prediction method based on the traditional neural network model has the problems of weak processing capacity on long-time sequences, gradient disappearance, overfitting and the like, and the prediction accuracy is low, so that the actual demands of technicians still cannot be met
Disclosure of Invention
In order to overcome the technical problems in the prior art, the embodiment of the invention provides a method for predicting the daily wind power, which is used for accurately processing data acquired by a plurality of data sources and analyzing the data by adopting an intelligent analysis model after empirical mode decomposition, so that accurate wind power prediction is realized and prediction accuracy is improved.
In order to achieve the above purpose, the embodiment of the invention provides a prediction method of solar wind power, which comprises the steps of obtaining a plurality of initial atmospheric prediction data, obtaining each initial atmospheric prediction data from a different data source, performing fusion operation on the plurality of initial atmospheric prediction data to obtain fused data, performing empirical mode decomposition on the fused data to obtain decomposed data, obtaining a composite neural network model, and analyzing the decomposed data based on the composite neural network model to generate a corresponding solar wind power predicted value.
Preferably, the method comprises the steps of performing fusion operation on the plurality of initial atmospheric prediction data to obtain fused data, wherein the fusion operation comprises the steps of determining real-time weight information of each initial atmospheric prediction data, and performing time sequence fusion processing on the plurality of initial atmospheric prediction data based on the real-time weight information to generate the fused data.
Preferably, the determining the real-time weight information of each initial atmosphere prediction data comprises the steps of obtaining real-time forecast quality evaluation information of each initial atmosphere prediction data, and processing the real-time forecast quality evaluation information based on a preset neural network to generate the real-time weight information of each initial atmosphere prediction data.
Preferably, the empirical mode decomposition is performed on the fused data to obtain decomposed data, the method comprises the steps of obtaining a first random time sequence signal from the fused data, determining a first local extremum of the first random time sequence signal, generating corresponding first envelope information based on the first local extremum, performing the empirical mode decomposition on the fused data based on the first envelope information to obtain a corresponding plurality of first intrinsic mode functions and first residual component data, wherein the fused data is the sum of the plurality of first intrinsic mode functions and the first residual component data, and taking the plurality of first intrinsic mode functions and the first residual component data as the decomposed data.
Preferably, the method comprises the steps of performing empirical mode decomposition on the fused data to obtain decomposed data, obtaining a second random time sequence signal in the fused data, obtaining a preset average order value and preset standard white noise, performing noise addition processing on the second random time sequence signal based on the preset average order value and the preset standard white noise to obtain an added signal, determining a second local extremum of the added signal, generating corresponding second envelope information based on the second local extremum, performing empirical mode decomposition on the added signal based on the second envelope information to obtain a corresponding plurality of second intrinsic mode functions and second residual component data, wherein the fused data is the sum of the plurality of second intrinsic mode functions and the second residual component data, and taking the plurality of second intrinsic mode functions and the second residual component data as decomposed data.
Preferably, the composite neural network model comprises a convolutional neural network model and a gating cyclic neural network model, the decomposed data is analyzed based on the composite neural network model to generate a corresponding daily wind power predicted value, the method comprises the steps of performing first data processing on the decomposed data based on the convolutional neural network model to obtain first processed data, and performing second data processing on the first processed data based on the gating cyclic neural network model to generate the daily wind power predicted value.
Preferably, the convolutional neural network model comprises a convolutional layer and a pooling layer, the first data processing is performed on the decomposed data based on the convolutional neural network model to obtain first processed data, the first processed data is obtained by performing feature extraction on the decomposed data based on the convolutional layer to obtain corresponding spatial features, and the first processed data is obtained by performing dimension reduction processing on the spatial features based on the pooling layer.
Preferably, the method for generating the daily wind power predicted value based on the gating cyclic neural network model performs second data processing on the first processed data, and comprises the steps of performing data analysis on the first processed data based on the gating cyclic neural network model, obtaining time compliance information between each data in the first processed data, and generating the daily wind power predicted value based on the time compliance information.
Correspondingly, the embodiment of the invention also provides a prediction device for the solar wind power, which comprises an initial data acquisition unit, a fusion unit, a decomposition unit, a model acquisition unit and a prediction unit, wherein the initial data acquisition unit is used for acquiring a plurality of initial atmospheric prediction data, each initial atmospheric prediction data is acquired from a different data source, the fusion unit is used for performing fusion operation on the plurality of initial atmospheric prediction data to obtain fused data, the decomposition unit is used for performing empirical mode decomposition on the fused data to obtain decomposed data, the model acquisition unit is used for acquiring a composite neural network model, and the prediction unit is used for analyzing the decomposed data based on the composite neural network model to generate a corresponding solar wind power prediction value.
Preferably, the fusion unit comprises a weight determination module and a fusion module, wherein the weight determination module is used for determining real-time weight information of each piece of initial atmospheric prediction data, and the fusion module is used for performing time sequence fusion processing on a plurality of pieces of initial atmospheric prediction data based on the real-time weight information to generate fused data.
Preferably, the weight determining module is specifically configured to obtain real-time forecast quality evaluation information of each piece of initial atmospheric forecast data, and process the real-time forecast quality evaluation information based on a preset neural network to generate real-time weight information of each piece of initial atmospheric forecast data.
Preferably, the decomposition unit comprises a first decomposition module, wherein the first decomposition module is specifically configured to obtain a first random time sequence signal from the fused data, determine a first local extremum of the first random time sequence signal, generate corresponding first envelope information based on the first local extremum, perform empirical mode decomposition on the fused data based on the first envelope information to obtain a corresponding plurality of first intrinsic mode functions and first residual component data, the fused data is a sum of the plurality of first intrinsic mode functions and the first residual component data, and take the plurality of first intrinsic mode functions and the first residual component data as decomposed data.
Preferably, the decomposition unit further comprises a second decomposition module, wherein the second decomposition module is specifically configured to acquire a second random time sequence signal from the fused data, acquire a preset average order value and a preset standard white noise, perform noise addition processing on the second random time sequence signal based on the preset average order value and the preset standard white noise to acquire an added signal, determine a second local extremum of the added signal, generate corresponding second envelope information based on the second local extremum, perform empirical mode decomposition on the added signal based on the second envelope information to acquire a corresponding plurality of second intrinsic mode functions and second residual component data, and the fused data is a sum of the plurality of second intrinsic mode functions and the second residual component data, and take the plurality of second intrinsic mode functions and the second residual component data as decomposed data.
Preferably, the composite neural network model comprises a convolutional neural network model and a gating cyclic neural network model, and the prediction unit comprises a first processing module and a prediction module, wherein the first processing module is used for performing first data processing on the decomposed data based on the convolutional neural network model to obtain first processed data, and the prediction module is used for performing second data processing on the first processed data based on the gating cyclic neural network model to generate a solar wind power predicted value.
Preferably, the convolutional neural network model comprises a convolutional layer and a pooling layer, and the first processing module is specifically configured to perform feature extraction on the decomposed data based on the convolutional layer to obtain corresponding spatial features, and perform dimension reduction processing on the spatial features based on the pooling layer to obtain first processed data.
Preferably, the prediction module is specifically configured to perform data analysis on the first processed data based on the gated recurrent neural network model, obtain time compliance information between each data in the first processed data, and perform wind power prediction based on the time compliance information, so as to generate the intra-day wind power prediction value.
In another aspect, the embodiment of the present invention further provides a computer readable storage medium, where a computer program is stored, where the program when executed by a processor implements the prediction method provided by the embodiment of the present invention.
Through the technical scheme provided by the invention, the invention has at least the following technical effects:
The weather data are acquired from a plurality of data sources, the weather data are accurately processed, so that the accuracy of weather forecast data is effectively improved, on the basis, the wind power plant data in the weather data are subjected to empirical mode decomposition, so that the original time sequence-based weather data are converted into stable data which can be input into an intelligent learning model for processing, the data are automatically learned through the intelligent learning model, and therefore accurate ultra-short-term wind power predicted values are output, and the actual demands of technicians are met.
Additional features and advantages of embodiments of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings are included to provide a further understanding of embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain, without limitation, the embodiments of the invention. In the drawings:
FIG. 1 is a flowchart of a specific implementation of a method for predicting solar wind power according to an embodiment of the present invention;
FIG. 2 is a flowchart of a specific implementation of obtaining fused data in the method for predicting solar wind power according to the embodiment of the present invention;
FIG. 3 is a flowchart of a specific implementation of obtaining decomposed data in the method for predicting solar wind power according to the embodiment of the present invention;
FIG. 4 is a flowchart of a specific implementation of obtaining decomposed data in a method for predicting solar wind power according to another embodiment of the present invention;
FIG. 5 is a schematic diagram of an empirical mode decomposition of fused data to obtain decomposed data in a method for predicting solar wind power according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a device for predicting solar wind power according to an embodiment of the present invention.
Detailed Description
The following describes the detailed implementation of the embodiments of the present invention with reference to the drawings. It should be understood that the detailed description and specific examples, while indicating and illustrating the invention, are not intended to limit the invention.
The terms "system" and "network" in embodiments of the invention may be used interchangeably. "plurality" means two or more, and "plurality" may also be understood as "at least two" in this embodiment of the present invention. "and/or" describes an association relationship of an association object, and indicates that there may be three relationships, for example, a and/or B, and may indicate that there are three cases of a alone, a and B together, and B alone. The character "/", unless otherwise specified, generally indicates that the associated object is an "or" relationship. In addition, it should be understood that in the description of embodiments of the present invention, the words "first," "second," and the like are used merely for distinguishing between the descriptions and not be construed as indicating or implying a relative importance or order.
Referring to fig. 1, an embodiment of the present invention provides a method for predicting intra-solar wind power, where the method includes:
s10) acquiring a plurality of initial atmosphere prediction data, wherein each initial atmosphere prediction data is acquired from a different data source;
S20) performing fusion operation on the plurality of initial atmosphere prediction data to obtain fused data;
S30) performing empirical mode decomposition on the fused data to obtain decomposed data;
S40) acquiring a composite neural network model;
S50) analyzing the decomposed data based on the composite neural network model to generate a corresponding predicted value of the daily wind power.
In one possible implementation, a plurality of initial atmospheric prediction data is first acquired. In the existing weather prediction field, weather prediction mode data of various data sources, such as a European middle weather forecast (European Centre for Medium-RANGE WEATHER Forecasts, abbreviated as EC mode), a eastern area middle scale numerical weather forecast (SMS-WARMS, abbreviated as Shanghai mode) and a rapid updating assimilation mode, can be correspondingly acquired, and the EC mode data, the Shanghai mode data and the rapid updating assimilation mode data can be acquired. Therefore, in the embodiment of the invention, the advantages of each data can be integrated by collecting the data of a plurality of data sources and performing comprehensive processing, and the defects of each data are eliminated, so that more accurate atmospheric data processing is realized.
After a plurality of initial atmospheric prediction data are acquired, a fusion operation is further performed on the plurality of initial atmospheric prediction data. Referring to fig. 2, in an embodiment of the present invention, the performing a fusion operation on the plurality of initial atmospheric prediction data to obtain fused data includes:
S21) determining real-time weight information of each initial atmospheric prediction data;
S22) performing time-series fusion processing on the plurality of initial atmospheric prediction data based on the real-time weight information, and generating fused data.
In the embodiment of the invention, the plurality of initial atmosphere prediction data are fused by determining the real-time weight information of each initial atmosphere prediction data. However, in the conventional data fusion method, the weight information of each data to be fused is determined mainly by past experience or working experience of a technician, and thus the actual accuracy requirement cannot be satisfied.
The method comprises the steps of obtaining real-time forecast quality evaluation information of each initial atmosphere forecast data, and processing the real-time forecast quality evaluation information based on a preset neural network to generate the real-time weight information of each initial atmosphere forecast data.
In one possible implementation, the real-time forecast quality evaluation information of each initial atmospheric forecast data is obtained first, for example, EC mode data, shanghai mode data and rapid-update assimilation mode data can be respectively scored according to historical real-time weather forecast data to evaluate the forecast quality under different time effects, and the forecast quality is analyzed by using a convolutional neural network at this time, so that weight information of each mode data is dynamically generated, that is, real-time weight information of each initial atmospheric forecast data is obtained.
For example, in the embodiment of the invention, the convolutional neural network comprises an input layer, a hidden layer and an output layer, wherein the hidden layer comprises a convolutional layer, a pooling layer and an activation layer, and in the process of analyzing the forecast quality by using the convolutional neural network, the characteristic information of the forecast quality is firstly extracted through the convolutional layer, and the connection between layers of the network can be effectively reduced by extracting the characteristic information of the forecast quality through the convolutional layer, and meanwhile, the risk of overfitting is reduced. And then, carrying out nonlinear mapping on the result of the convolution layer linear calculation through an activation layer, and increasing the nonlinear segmentation capability so as to improve the expression capability of the model. At this time, the data after nonlinear mapping is further processed by the pooling layer to compress the data and the parameter, so as to prevent overfitting and reduce the complexity of the network, for example, sub-sampling methods such as mean sub-sampling (mean pooling) and maximum sub-sampling (max pooling) can be adopted, and at this time, accurate real-time weight information can be obtained. And at this time, performing time sequence fusion processing on the acquired plurality of initial atmospheric prediction data according to the real-time weight information to generate fused data, for example, the fused data is wind farm data.
In the embodiment of the invention, the intelligent fusion of the plurality of mode data is realized by intelligently analyzing and determining the real-time weight information of each mode data, so that the accuracy of the generated atmospheric prediction data and the prediction timeliness are improved on the basis of the advantages of compatibility with various data.
At this time, the data can be analyzed, however, in the practical application process, because the wind speed and the power data have certain continuity, besides the accurate weather forecast data of the wind power plant are obtained, the live wind speed and the power data of the wind power plant at a near-period time are also required to be obtained as the input of the model. However, the actual wind speed and wind power curve is time sequence data with fluctuation, and the direct input of the time sequence data is unfavorable for learning and feature extraction of the intelligent model. Therefore, it is also necessary to perform data decomposition processing on the acquired fused data.
Referring to fig. 3, in an embodiment of the present invention, the performing empirical mode decomposition on the fused data to obtain decomposed data includes:
s311) acquiring a first random time sequence signal from the fused data;
S312) determining a local extremum of the first random timing signal;
s313) generating corresponding first envelope information based on the local extremum;
S314) performing empirical mode decomposition on the fused data based on the first envelope information to obtain a corresponding plurality of first intrinsic mode functions and first residual component data, wherein the fused data is the sum of the plurality of first intrinsic mode functions and the first residual component data;
S315) using the plurality of first intrinsic mode functions and the first residual component data as decomposed data.
In a possible implementation, first determining the first random timing signal x (t) in the fused data, then determining a first local extremum of the first random timing signal x (t), for example, the first local extremum is a local maximum point and a local minimum point of the first random timing signal x (t), then fitting the upper envelope and the lower envelope of the first random timing signal x (t) by using cubic spline interpolation as the first envelope information of the first random timing signal x (t), at this time, performing empirical mode decomposition on the fused data, specifically, first making the mean value of the upper envelope and the lower envelope be m (t), then calculating h (t) = (t) -m (t), and judging whether the calculated h (t) meets a preset decomposition condition, if not, continuing to perform the calculation with the h (t) as a new x (t) until the obtained h k (t) meets the preset decomposition condition, determining to obtain a first mode function c 34 (t), and performing a first partial decomposition on the first random component (t) such as a first partial component (t) of the first partial component (t) at this time, for example, until a residual component (37 r) such as a first partial component (t) is obtained by stopping the step (37 r) and a final decomposition (step (37 r) is performed, for example, until a residual component (37 r) is obtained, and a final component (step (37) is obtained, such as a final component (37) is obtained, in this case, n internal pattern functions c i (t) and a final residual component r n (t) are obtained after the first random timing signal x (t) is decomposed, and in the embodiment of the present invention, the first random timing signal x (t) is the sum of the multiple internal pattern functions c i (t) and the residual function r n (t), which can be expressed asAt this time, the plurality of first intrinsic mode functions c i (t) and the residual function r n (t) are taken as decomposed data.
In the embodiment of the invention, the non-stationary signal can be effectively decomposed into a plurality of stationary internal mode functions step by step according to the fluctuation or trend of different scales by adopting the data decomposition method based on the empirical mode decomposition, so that the stable decomposition of the unstable and dynamic weather forecast data is realized, the subsequent intelligent model is convenient for data analysis, and meanwhile, the empirical mode decomposition method is simpler, so that the calculation complexity is reduced, and the calculation accuracy and calculation efficiency are improved.
However, in the practical application process, since the simple empirical mode decomposition method has mode aliasing linearity and end effect, and has a certain influence on the final decomposition effect, in order to further improve the decomposition accuracy of weather forecast data, the embodiment of the invention suppresses the aliasing phenomenon by introducing white noise so as to realize a more accurate decomposition effect.
Referring to fig. 4, in an embodiment of the present invention, the performing empirical mode decomposition on the fused data to obtain decomposed data further includes:
s321) acquiring a second random time sequence signal from the fused data;
s322) obtaining a preset average number and preset standard white noise;
S323) performing noise addition processing on the second random timing signal based on the preset average number of times and the preset standard white noise, to obtain an added signal;
s324) determining a second local extremum of the added signal;
S325) generating corresponding second envelope information based on the second local extremum;
S326), performing empirical mode decomposition on the noise superimposed data based on the second envelope information to obtain a corresponding plurality of second intrinsic mode functions and second residual component data, wherein the fused data is the sum of the plurality of second intrinsic mode functions and the second residual component data;
s327) using the plurality of second intrinsic mode functions and the second residual component data as decomposed data.
In one possible implementation, a second random timing signal is first determined in the fused data, e.g., the second random timing signal is also denoted by x (t), where a preset average secondary value M and a preset standard white noise n i (t) are further obtained, and then the preset standard white noise n i (t) is added to the second random timing signal x (t) according to the preset average secondary value M to generate a new signal (i.e., an added signal), e.g., the new signal is characterized by x i(t)＝x(t)+ni (t), i=1, 2. At this time, empirical mode decomposition is performed on the obtained signal containing white noise, based on the same principle, firstly, a local maximum value and a local limit value of the added signal are determined, then, a corresponding upper envelope and a lower envelope are formed by means of cubic spline interpolation fitting, then, the average value of the upper envelope and the lower envelope is taken, gradual signal decomposition is performed, a plurality of intrinsic mode functions and one residual component data of the added signal are obtained, at this time, the principle that the statistical average value of an uncorrelated sequence is 0 can be utilized to perform average operation on the corresponding plurality of intrinsic mode functions, so as to obtain an intrinsic mode function after final decomposition, and please refer to fig. 5, which is a schematic diagram of the decomposed data obtained by performing empirical mode decomposition on the fused data provided by the embodiment of the invention.
In the embodiment of the invention, the improved empirical mode decomposition method is adopted, so that the mode aliasing phenomenon and the end effect existing in the traditional empirical mode decomposition method can be effectively inhibited, the accuracy of final decomposition data is further improved, the accuracy of the subsequent data analysis and the accuracy of the generated wind power prediction result are ensured, and the actual demands of technicians are met.
At this time, the decomposed data may be directly input into an intelligent analysis model to predict the wind power. In the prior art, various algorithms or models for intelligently analyzing data exist, however, due to the problems of low accuracy and difficult modeling caused by high uncertainty and noise interference of the data of the wind power plant, the traditional single intelligent model cannot meet the actual requirements.
In the embodiment of the invention, the composite neural network model comprises a convolutional neural network model and a gating cyclic neural network model, wherein the analysis is performed on the decomposed data based on the composite neural network model to generate a corresponding solar wind power predicted value, and the method comprises the steps of performing first data processing on the decomposed data based on the convolutional neural network model to obtain first processed data, and performing second data processing on the first processed data based on the gating cyclic neural network model to generate the solar wind power predicted value.
In one possible implementation, first a first data process is performed on the decomposed data based on a convolutional neural network to obtain first processed data. For example, in the embodiment of the invention, the convolutional neural network model comprises a convolutional layer and a pooling layer, the first data processing is performed on the decomposed data based on the convolutional neural network model to obtain first processed data, the first processed data is obtained by performing feature extraction on the decomposed data based on the convolutional layer to obtain corresponding spatial features, and the dimension reduction processing is performed on the spatial features based on the pooling layer to obtain first processed data.
In the embodiment of the invention, the convolution layer in the convolution neural network can automatically extract the data characteristics, and the high-dimensional data can be processed rapidly and efficiently by sharing the convolution kernel, and the dimension of the data can be reduced by the pooling layer, so that the technical problem of over-fitting can be effectively solved, the technical effect of rapidly and accurately extracting the spatial characteristics in the decomposed data can be realized, and the precision of the convolution neural network model can be effectively improved by continuous iterative training.
Further, in the embodiment of the invention, the step of performing second data processing on the first processed data based on the gating cyclic neural network model to generate a daily wind power prediction value comprises the steps of performing data analysis on the first processed data based on the gating cyclic neural network model to obtain time compliance information between each data in the first processed data, and performing wind power prediction based on the time compliance information to generate the daily wind power prediction value.
For example, after the first processed data is obtained, the first processed data is further analyzed through a gating cyclic neural network model to further mine time sequence features in the first processed data, for example, time compliance information between each data in the first processed data is calculated to determine, wind power is predicted according to the time compliance information, and an accurate daily wind power predicted value is obtained.
In the embodiment of the invention, the wind power plant data with time sequence characteristics is analyzed by adopting the gating circulating neural network model, so that gradient in the prediction process can be effectively counteracted, thereby realizing more efficient information processing capability, meeting the prediction requirement of technicians on ultra-short-term (intra-day) wind power information, effectively ensuring the accuracy of a processing result and improving the prediction accuracy of wind power values.
The following describes a device for predicting solar wind power according to an embodiment of the present invention with reference to the accompanying drawings.
Referring to fig. 6, based on the same inventive concept, an embodiment of the present invention provides a prediction apparatus for solar wind power, which includes an initial data acquisition unit configured to acquire a plurality of initial atmospheric prediction data, each of which is acquired from a different data source, a fusion unit configured to perform a fusion operation on the plurality of initial atmospheric prediction data to obtain fused data, a decomposition unit configured to perform empirical mode decomposition on the fused data to obtain decomposed data, a model acquisition unit configured to acquire a composite neural network model, and a prediction unit configured to analyze the decomposed data based on the composite neural network model to generate a corresponding solar wind power prediction value.
The fusion unit comprises a weight determining module and a fusion module, wherein the weight determining module is used for determining real-time weight information of each piece of initial atmospheric prediction data, and the fusion module is used for executing time sequence fusion processing on a plurality of pieces of initial atmospheric prediction data based on the real-time weight information to generate fused data.
In the embodiment of the invention, the weight determining module is specifically used for acquiring real-time forecast quality evaluation information of each piece of initial atmosphere forecast data, and processing the real-time forecast quality evaluation information based on a preset neural network to generate real-time weight information of each piece of initial atmosphere forecast data.
In the embodiment of the invention, the decomposition unit comprises a first decomposition module, wherein the first decomposition module is specifically used for acquiring a first random time sequence signal from the fused data, determining a first local extremum of the first random time sequence signal, generating corresponding first envelope information based on the first local extremum, executing empirical mode decomposition on the fused data based on the first envelope information to obtain a plurality of corresponding first intrinsic mode functions and first residual component data, the fused data is the sum of the plurality of first intrinsic mode functions and the first residual component data, and taking the plurality of first intrinsic mode functions and the first residual component data as decomposed data.
The decomposition unit further comprises a second decomposition module, wherein the second decomposition module is specifically configured to acquire a second random time sequence signal from the fused data, acquire a preset average order value and preset standard white noise, perform noise addition processing on the second random time sequence signal based on the preset average order value and the preset standard white noise to acquire an added signal, determine a second local extremum of the added signal, generate corresponding second envelope information based on the second local extremum, perform empirical mode decomposition on the added signal based on the second envelope information to acquire a corresponding plurality of second intrinsic mode functions and second residual component data, and the fused data is the sum of the plurality of second intrinsic mode functions and the second residual component data, and take the plurality of second intrinsic mode functions and the second residual component data as decomposed data.
In the embodiment of the invention, the composite neural network model comprises a convolutional neural network model and a gating cyclic neural network model, wherein the prediction unit comprises a first processing module and a prediction module, wherein the first processing module is used for performing first data processing on the decomposed data based on the convolutional neural network model to obtain first processed data, and the prediction module is used for performing second data processing on the first processed data based on the gating cyclic neural network model to generate a solar wind power predicted value.
In the embodiment of the invention, the convolutional neural network model comprises a convolutional layer and a pooling layer, and the first processing module is specifically used for extracting features of the decomposed data based on the convolutional layer to obtain corresponding spatial features, and performing dimension reduction processing on the spatial features based on the pooling layer to obtain first processed data.
In the embodiment of the invention, the prediction module is specifically used for carrying out data analysis on the first processed data based on the gating cyclic neural network model, obtaining time compliance information between each data in the first processed data, and carrying out wind power prediction based on the time compliance information to generate the daily wind power predicted value.
Further, the embodiment of the present invention also provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the method described in the embodiment of the present invention.
The foregoing details of the optional implementation of the embodiment of the present invention have been described in conjunction with the accompanying drawings, but the embodiment of the present invention is not limited to the specific details of the foregoing implementation, and various simple modifications may be made to the technical solution of the embodiment of the present invention within the scope of the technical concept of the embodiment of the present invention, where all the simple modifications belong to the protection scope of the embodiment of the present invention.
In addition, the specific features described in the above embodiments may be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, various possible combinations of embodiments of the present invention are not described in detail.
Those skilled in the art will appreciate that all or part of the steps in implementing the methods of the embodiments described above may be implemented by a program stored in a storage medium, including instructions for causing a single-chip microcomputer, chip or processor (processor) to perform all or part of the steps of the methods of the embodiments described herein. The storage medium includes a U disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, an optical disk, or other various media capable of storing program codes.
In addition, any combination of various embodiments of the present invention may be performed, so long as the concept of the embodiments of the present invention is not violated, and the disclosure of the embodiments of the present invention should also be considered.
Claims (11)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210051045.6A CN114418210B (en) | 2022-01-17 | 2022-01-17 | Methods and devices for predicting intraday wind power |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210051045.6A CN114418210B (en) | 2022-01-17 | 2022-01-17 | Methods and devices for predicting intraday wind power |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114418210A CN114418210A (en) | 2022-04-29 |
| CN114418210B true CN114418210B (en) | 2025-11-18 |
Family
ID=81273605
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210051045.6A Active CN114418210B (en) | 2022-01-17 | 2022-01-17 | Methods and devices for predicting intraday wind power |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114418210B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104102951A (en) * | 2014-05-05 | 2014-10-15 | 南方电网科学研究院有限责任公司 | Short-term wind power prediction method based on EMD historical data preprocessing |
| CN110555784A (en) * | 2019-07-23 | 2019-12-10 | 中国地质大学（武汉） | wind power prediction method based on gated cyclic neural network wind speed correction |
Family Cites Families (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB201201230D0 (en) * | 2012-01-25 | 2012-03-07 | Univ Delft Tech | Adaptive multi-dimensional data decomposition |
| CN110377984B (en) * | 2019-07-02 | 2023-08-18 | 中国科学院深圳先进技术研究院 | Industrial equipment residual effective life prediction method and system and electronic equipment |
| CN110782071A (en) * | 2019-09-25 | 2020-02-11 | 天津大学 | Wind power prediction method based on convolutional neural network fusion of spatiotemporal characteristics |
| CN111160621B (en) * | 2019-12-06 | 2022-06-14 | 江苏方天电力技术有限公司 | A short-term wind power prediction method integrating multi-source information |
| CN111476402A (en) * | 2020-03-16 | 2020-07-31 | 云南电网有限责任公司 | Wind power generation capacity prediction method coupling meteorological information and EMD technology |
| CN111562358B (en) * | 2020-05-06 | 2021-03-16 | 武汉大学 | Prediction method and system of gas content in transformer oil based on joint model |
| CN111695724B (en) * | 2020-06-01 | 2023-05-05 | 浙江大学 | A Wind Speed Prediction Method Based on Hybrid Neural Network Model |
| CN111832825B (en) * | 2020-07-14 | 2024-04-16 | 山东大学 | Wind power prediction method and system integrating long-term memory network and extreme learning machine |
| CN112329987A (en) * | 2020-10-15 | 2021-02-05 | 南京理工大学 | Short-term wind power plant power prediction method based on Adaboost-EMD-SVM |
| CN112287606A (en) * | 2020-11-09 | 2021-01-29 | 山西大学 | Exhaust gas mixing prediction model method based on self-organizing features |
| CN112529282A (en) * | 2020-12-03 | 2021-03-19 | 清华大学 | Wind power plant cluster short-term power prediction method based on space-time graph convolutional neural network |
| CN112580899A (en) * | 2021-02-23 | 2021-03-30 | 国能日新科技股份有限公司 | Medium-and-long-term wind power generation prediction method and system fused with machine learning model |
-
2022
- 2022-01-17 CN CN202210051045.6A patent/CN114418210B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104102951A (en) * | 2014-05-05 | 2014-10-15 | 南方电网科学研究院有限责任公司 | Short-term wind power prediction method based on EMD historical data preprocessing |
| CN110555784A (en) * | 2019-07-23 | 2019-12-10 | 中国地质大学（武汉） | wind power prediction method based on gated cyclic neural network wind speed correction |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114418210A (en) | 2022-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114330097A (en) | Heterogeneous data and deep learning wind power cluster power prediction method and device | |
| CN111582551B (en) | Wind power plant short-term wind speed prediction method and system and electronic equipment | |
| CN110443417A (en) | Multi-model integrated load prediction method based on wavelet transformation | |
| CN114819374B (en) | Regional new energy ultra-short term power prediction method and system | |
| CN119382103B (en) | TCN-LSTM wind power generation prediction method combining time-frequency analysis and attention mechanism | |
| CN116629448A (en) | Photovoltaic power prediction method based on combination of empirical mode decomposition and neural network | |
| CN118033784A (en) | Meteorological prediction method and system for micro-topography overhead transmission line | |
| CN119009942A (en) | Wind-solar integrated power prediction method and system | |
| CN114418210B (en) | Methods and devices for predicting intraday wind power | |
| CN119271971A (en) | Ultra-short-term wind power forecasting method based on quadratic decomposition and TCN-ECA-BiGRU | |
| CN114118499A (en) | Method and device for short-term wind power prediction | |
| CN108427742B (en) | A distribution network reliability data repair method and system based on low-rank matrix | |
| Zhang et al. | Load prediction based on depthwise separable convolution model | |
| CN118429133A (en) | A method and system for predicting the full cycle of wide-area distributed photovoltaic power generation | |
| CN117810984A (en) | A distributed photovoltaic power generation power prediction method, system, equipment and medium | |
| CN117833216A (en) | Photovoltaic power station generated power prediction method and device based on hybrid neural network | |
| CN117009752A (en) | Power consumption demand prediction method based on POA-VMD-SSA-KELM model | |
| CN118069717B (en) | Time sequence data characteristic prediction method based on cyclic neural network | |
| Zhang et al. | A Hybrid Model based on Deep Learning and Cross-attention for Short-term Wind Power Prediction | |
| CN119622182B (en) | A conventional power output data fitting method and system based on cloud transformation fitting | |
| CN119944615A (en) | A microgrid energy management method and system based on dual sequence prediction model | |
| CN121076813A (en) | A method and system for identifying low-voltage distribution area loads based on density-adaptive clustering | |
| CN119863334A (en) | Quick climbing and scene switching point extraction method and device for electric power system | |
| Zhuo et al. | The evaluation of wind power time series multi-scale modeling methods | |
| Wang et al. | A Comprehensive Power Consumption Model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |