Distributed power supply power generation prediction method, system, equipment and medium

Abstract

The invention relates to the technical field of distributed power supply power generation prediction, and provides a distributed power supply power generation prediction method, system and device and a medium, and the method comprises the steps: collecting the power generation data of a distributed power supply, and carrying out the data cleaning and per-unit processing; performing decomposition and dimension reduction on the standard distributed power supply power generation data through a variational mode decomposition algorithm; performing correlation analysis on each modal component through a correlation analysis method; obtaining a new modal component according to the correlation coefficient between the modal components; inputting the new modal components and the residual components into a distributed power generation prediction model to obtain predicted values of the modal components and predicted values of the residual components; and superposing each mode component prediction value and the residual component prediction value to obtain a distributed power supply power generation prediction value. The method effectively reduces the interference of external factors on the power generation prediction of the distributed power supply, improves the prediction precision, reduces the calculation complexity of the model, and improves the prediction efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of distributed power generation prediction, and in particular to a distributed power generation prediction method, system, device and medium. Background Art

[0002] The balance of power supply and demand in the distribution network system is of great significance to ensuring energy security, and the prediction of distributed power generation is the basis for ensuring the safe and stable operation of the regional distribution network system. Its prediction accuracy directly affects the reliability, safety and economy of the distribution network operation. Accurate prediction results can not only effectively ensure the stable operation of the power system, but also reduce the cost of power generation equipment and improve the efficiency of power grid dispatching. In addition, with the continuous expansion of the distribution network scale, the efficiency of distributed power generation prediction in processing large amounts of data has become particularly important. In recent years, with the popularization of smart grid technology and the large-scale investment of smart detection equipment such as smart meters, rich and high-quality data support has been provided for distributed power generation prediction based on deep learning. At the same time, the rapid development of modern power measurement technology and communication technology has made it possible to monitor and analyze distributed power generation data. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the related art. To this end, the present invention provides a distributed power generation prediction method, system, device and medium, which fully considers the cleaning process of the original distributed power generation data, fully cleans the abnormal data, and performs standardization on the original data, thereby improving the accuracy of subsequent data analysis and facilitating the prediction accuracy of the distributed power generation prediction model. Taking full account of the dimensionality reduction of the original data, a CNN-LSTM prediction model based on the VMD algorithm is constructed, which effectively reduces the interference of external factors on the distributed power generation prediction, improves the prediction accuracy, reduces the computational complexity of the model, and improves the prediction efficiency.

[0004] The present invention provides a distributed power generation prediction method, comprising: S1: Collect distributed power generation data, clean and normalize the distributed power generation data, and obtain standard distributed power generation data; S2: Decompose and reduce the dimensionality of the standard distributed power generation data through the variational mode decomposition algorithm to obtain multiple modal components and residual components; S3: Perform correlation analysis on each modal component by using the correlation analysis method to obtain the correlation coefficient between the modal components; S4: classify each modal component according to the correlation coefficient between the modal components to obtain a new modal component; S5: input each new modal component into the distributed power generation prediction model to obtain a prediction value of each modal component; input the residual component into the distributed power generation prediction model to obtain a prediction value of the residual component; S6: Superimpose the predicted values ​​of each modal component and the predicted value of the residual component to obtain the predicted value of distributed generation power generation.

[0005] Furthermore, step S1 includes: S11: Identify abnormal data and extract basic information of normal data on both sides of the abnormal data point, the basic information includes the frequency, amplitude and phase of the signal; S12: construct a standard distributed power generation data waveform based on basic information; S13: replacing abnormal data point data with a standard distributed power generation data waveform to obtain cleaned and completed data; S14: normalize the cleaned and completed data to obtain standard distributed power generation data.

[0006] Furthermore, in step S2, The parameters of the quadratic penalty term are obtained by an alternating direction multiplier algorithm, and the parameters of the quadratic penalty term control the bandwidth of the modal component of the distributed generation data; The standard distributed generation data is decomposed into multiple modal components and residual components through the quadratic penalty parameter, and each modal component has a central frequency and bandwidth signal.

[0007] Furthermore, the correlation analysis method includes a Pearson correlation coefficient analysis method.

[0008] Furthermore, in step S4, a correlation coefficient threshold is set. If the absolute value of the correlation coefficient between the modal components is greater than or equal to the correlation coefficient threshold, the modal components are superimposed to obtain new modal components; If the absolute value of the correlation coefficient between the modal components is less than the correlation coefficient threshold, the modal component is taken as a new modal component.

[0009] Furthermore, the distributed power generation prediction model is a CNN-LSTM prediction model based on the VMD algorithm, including an input layer, a CNN network layer, an LSTM network layer and an output layer; The input layer includes a fully connected layer, the CNN network layer includes two one-dimensional convolutional layers and a maximum pooling layer, and both convolutional layers use the ReLU activation function; the LSTM network layer includes two LSTM layers, a fully connected layer and a random inactivation layer, and the output layer includes a fully connected layer.

[0010] Furthermore, the distributed power generation data is collected, transmitted, processed and analyzed based on the intelligent power data measurement device.

[0011] The present invention also provides a distributed power generation prediction system, which is used to execute any of the above-mentioned distributed power generation prediction methods, comprising: A standardization module, wherein the standardization module collects distributed power generation data, performs data cleaning and standardization processing on the distributed power generation data, and obtains standard distributed power generation data; A decomposition and dimensionality reduction module, wherein the decomposition and dimensionality reduction module decomposes and reduces the standard distributed power generation data through a variational modal decomposition algorithm to obtain modal components and residual components; A correlation analysis module, wherein the correlation analysis module performs correlation analysis on each modal component by a correlation analysis method to obtain correlation coefficients between modal components; A classification module, wherein the classification module classifies each modal component according to the correlation coefficient between the modal components to obtain a new modal component; A prediction module, wherein the prediction module inputs each new modal component into a distributed power generation prediction model to obtain a prediction value of each modal component; and inputs a residual component into the distributed power generation prediction model to obtain a prediction value of the residual component; A superposition module is used to superimpose the predicted values ​​of each modal component and the predicted value of the residual component to obtain the predicted value of distributed power generation.

[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of a distributed power generation prediction method as described in any one of the above are implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of a distributed power generation prediction method as described in any one of the above are implemented.

[0014] The above one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: The present invention provides a distributed power generation prediction method, system, device and medium, which fully considers the cleaning process of the original distributed power generation data, fully cleans the abnormal data, and performs standardization on the original data, thereby improving the accuracy of subsequent data analysis and the prediction accuracy of the distributed power generation prediction model. Taking full consideration of the dimensionality reduction of the original data, a CNN-LSTM prediction model based on the VMD algorithm is constructed, which effectively reduces the interference of external factors on the distributed power generation prediction, improves the prediction accuracy, reduces the calculation complexity of the model, and improves the prediction efficiency.

[0015] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a flow chart of a distributed power generation prediction method provided by the present invention.

[0018] Figure 2 It is a structural schematic diagram of a distributed power generation prediction system provided by the present invention.

[0019] Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention.

[0020] Reference numerals: 101. Standardization module; 102. Decomposition and dimensionality reduction module; 103. Correlation analysis module; 104. Classification module; 105. Prediction module; 106. Superposition module; 201. Processor; 202. Communication bus; 203. Communication interface; 204. Memory. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme in the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0022] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the embodiment of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0023] Combine the following Figures 1 to 3 A distributed power generation prediction method, system, device and medium of the present invention are described.

[0024] like Figure 1 As shown, a distributed power generation prediction method includes: S1: Collect distributed power generation data, clean and normalize the distributed power generation data, and obtain standard distributed power generation data; During the data collection, transmission and storage process of distributed power sources, data collection equipment and storage devices may be affected by factors such as electromagnetic interference and extreme natural conditions, resulting in abnormalities in the distributed power generation data, such as data loss. Such abnormal data will not only affect the accuracy of data analysis, but also interfere with the operation monitoring of distributed power sources; S11: Identify abnormal data and extract basic information of normal data on both sides of the abnormal data point, the basic information includes the frequency, amplitude and phase of the signal; S12: construct a standard distributed power generation data waveform based on basic information; S13: replacing abnormal data point data with a standard distributed power generation data waveform to obtain cleaned and completed data; S14: normalize the cleaned and completed data to obtain standard distributed power generation data.

[0025] In some specific embodiments of the present invention, by analyzing the normal data on both sides of the abnormal data point, the frequency, amplitude and phase information of the data signals on both sides are extracted to construct a standard distributed power generation data waveform. This standard signal waveform is used to supplement the missing data, realize data cleaning and completion, and thus ensure the integrity and consistency of the data.

[0026] After the original distributed power generation data is cleaned, the distributed power generation data is further normalized to ensure that the data can be standardized in the subsequent distributed power generation forecast.

[0027] Since different distributed power supply devices have differences in models and specifications, the present invention performs per-unit processing on the data based on the voltage and current ratings of each distributed power supply, thereby ensuring the comparability of data between different devices. The calculation expression for voltage per-unit processing is: in, is the voltage per unit value, is the actual value of the distributed power generation voltage, It is the standard value of the distributed power generation voltage; The calculation expression for current normalization is: in, is the per unit current value, is the actual value of the distributed power generation current, It is the standard value of the distributed power generation current.

[0028] S2: Decompose and reduce the dimensionality of the standard distributed power generation data through the variational modal decomposition algorithm to obtain the modal components and residual components; The parameters of the quadratic penalty term are obtained by an alternating direction multiplier algorithm, and the parameters of the quadratic penalty term control the bandwidth of the modal component of the distributed generation data; The standard distributed generation data is decomposed into multiple modal components and residual components through the quadratic penalty parameter, and each modal component has a central frequency and a bandwidth.

[0029] In some specific embodiments of the present invention, the distributed power generation is subject to the influence of temperature, humidity, light intensity, wind speed and other factors, and exhibits strong randomness. The variational mode decomposition (VMD) algorithm can decompose the data into a series of modal components and residual components according to the time scale of the distributed power generation data itself, and different modal components represent the characteristic components of the distributed power generation data at different frequencies.

[0030] The VMD algorithm is used to decompose and reduce the dimension of distributed power generation data to improve the prediction accuracy of distributed power generation. The VMD algorithm can adaptively decompose the distributed power generation data into multiple narrowband modes and simultaneously obtain the center frequency of each mode, so as to more accurately analyze the characteristics of each modal component of the distributed power generation data. Multi-component distributed power generation data It can be expressed as: in, for The first modal component of the distributed generation data at time, for The second modal component of the distributed generation data at time, for Distributed power generation data at the moment modal components, , is the total number of modal components, Has a center frequency and a bandwidth.

[0031] The goal of VMD is to transform the variational problem of the modal components into a modal function that minimizes the sum of the estimated bandwidths of each modal component. It can be expressed as an unconstrained optimization problem of the extended Lagrangian expression L. The calculation expression is: in, For the modal components, For the The center frequency of the modal components, is the Lagrangian operator, is the quadratic penalty parameter, For time Find the derivative, is the Hilbert transform kernel, is the Dirac function, is the frequency adjustment factor, is an imaginary unit, is the convolution operator, is the frequency offset, is the square of the norm, for The time Lagrangian operator, is the set of modal components, is the center frequency set.

[0032] To limit the modal components bandwidth; To represent the modal components Changes in signal frequency; To be used Convert to analytical signal; Ensure that the modal components Focused on its center frequency nearby; The original signal With all components The difference between the sum; is the Lagrange multiplier term, which is included in the objective function as a constraint.

[0033] The alternating direction multiplier algorithm (ADMM) is used, that is, by alternatingly updating the The modal component The estimated value of the iteration , No. The center frequency of the modal component The estimated value of the iteration and the Lagrangian operator The estimated value of the iteration , solve the above variational problem, the specific steps of the solution are as follows: S21: Initialize the modal component set for the first iteration , the center frequency set of the first iteration , the Lagrangian operator of the first iteration , and assign an initial value of zero; S22: Iterative update according to the following formula , the real part of its inverse Fourier transform is the required time domain component signal , the calculation expression is: in, For the modal components versus frequency No. The estimated value of the iteration, is the original signal frequency The Fourier transform of For the modal components versus frequency No. The estimated value of the iteration, For the modal components versus frequency No. The estimated value of the iteration, is the Lagrangian operator for the frequency No. The estimated value of the iteration, For the The center frequency of the modal component The estimated value of the iteration; S23: Iterative update according to the following formula ; S24: Iterative update according to the following formula ; in, is the penalty parameter, is the Lagrangian operator The estimated value of the iteration, is the Fourier transform of the original signal; S25: loop from step S22 to step S25 until the iterative convergence condition shown in the following formula is met: in, is the convergence allowable value, For the The modal component The estimated value of the iteration, the present invention will The value is .

[0034] Parameters in VMD algorithm The bandwidth of each modal component can be controlled. The larger the value, the narrower the bandwidth of each decomposed mode; conversely, The smaller the value, the wider the bandwidth of each decomposed mode. =500, which can effectively separate various modal components and residual components.

[0035] S3: Perform correlation analysis on each modal component by using the correlation analysis method; obtain the correlation coefficient between the modal components; Correlation analysis methods include Pearson correlation coefficient analysis; When using the VMD algorithm to decompose and reduce the dimension of distributed power generation data, similar modes may be decomposed independently. In order to further improve the prediction efficiency, the present invention uses the Pearson correlation analysis method to perform correlation analysis on each modal component. Pearson correlation coefficient The calculation formula is as follows: in, For data series No. individual data; For data series The average value of is the length of the data sequence; For data series No. individual data; For data series The formula is used to calculate the correlation coefficients between the distributed generation historical data series, the user load historical data series and their corresponding characteristic parameter data series.

[0036] S4: classify each modal component according to the correlation coefficient between the modal components to obtain a new modal component; Set the correlation coefficient threshold. If the absolute value of the correlation coefficient between the modal components is greater than or equal to the correlation coefficient threshold, the modal components are superimposed to obtain new modal components; If the absolute value of the correlation coefficient between the modal components is less than the correlation coefficient threshold, the modal component is taken as a new modal component.

[0037] In some specific embodiments of the present invention, the correlation coefficient threshold is set to 0.9. When the Pearson correlation coefficient between the modal components When , the modal components are superimposed to obtain new modal components, If the absolute value of the correlation coefficient between the modal components is less than , then the modal component is taken as the new modal component; The new modal component participates in subsequent calculations and analysis as an independent modal component.

[0038] S5: input each new modal component into the distributed power generation prediction model to obtain a prediction value of each modal component; input the residual component into the distributed power generation prediction model to obtain a prediction value of the residual component; The distributed power generation prediction model is a CNN-LSTM prediction model based on the VMD algorithm, including an input layer, a CNN network layer, a LSTM network layer and an output layer; The input layer includes a fully connected layer, the CNN network layer includes two one-dimensional convolutional layers and a maximum pooling layer, and both convolutional layers use the ReLU activation function; the LSTM network layer includes two LSTM layers, a fully connected layer and a random loss layer, and the output layer includes a fully connected layer; The main function of the input layer is to pass the new modal component as input data into the CNN network layer, and deeply mine the input data through the advantages of the CNN structure to extract the main features of the input sequence.

[0039] The function of the CNN layer is to extract features from the input data. After the convolution layer, the maximum pooling operation is performed to downsample the features extracted by the convolution layer to reduce the number of parameters of the model. The features extracted by the CNN network are passed to the LSTM network for further processing. By learning these features, LSTM reveals the intrinsic relationship and periodicity between data, thereby predicting data.

[0040] LSTM is an improved recurrent neural network (RNN). By adding three gated units to control information transmission in the RNN structure, it effectively overcomes the gradient vanishing or gradient exploding problems that are prone to occur when RNN processes long sequences. The three gated units in LSTM are the forget gate, input gate, and output gate. The main function of the forget gate is to filter information, retaining useful information while avoiding the continued transmission of invalid information at the previous moment; the input gate is used to read new data at the current moment; and the output gate is responsible for transmitting the processed information to the next moment.

[0041] The LSTM network consists of two LSTM layers. The role of the LSTM layer is to learn the features extracted by the CNN layer, retain useful information and forget invalid information. Since most of the components processed by the VMD algorithm have strong regularity, two LSTM layers are selected to ensure better prediction results in a shorter training time. A fully connected layer is added after the two LSTM layers to compress the data before the output layer, reduce the volume of the data flow and the network parameters of the output layer. The addition of the Dropout layer helps to reduce overfitting. The output layer is a fully connected layer, whose main function is to process the output of the LSTM layer through the fully connected layer, calculate the final prediction value and output it.

[0042] Since the power generation of distributed power sources has large fluctuations and high randomness, the present invention first uses the VMD algorithm to decompose the distributed power data into multiple modal components and residual components. Then, according to the results of the Pearson correlation analysis, the components with strong correlation are superimposed to form new modal components, and then each new modal component is input into the neural network for prediction. The patent of this invention uses a convolutional neural network (CNN) to extract features from the data processed by VMD, and then passes the extracted features into the LSTM network for prediction.

[0043] S6: Superimpose the predicted values ​​of each modal component and the predicted value of the residual component to obtain the predicted value of distributed generation power generation.

[0044] The distributed power generation data is collected, transmitted, processed and analyzed through intelligent measurement devices based on electric energy data.

[0045] In order to improve the efficiency of distributed power generation prediction, the present invention combines modern intelligent measurement technology and modern control technology to design a distributed power data acquisition, transmission, processing and analysis system based on electric energy data intelligent measurement device. Through this system, the historical power generation data and real-time power generation data of distributed power sources can be comprehensively collected, thereby providing reliable data support for efficient prediction.

[0046] All collected data is first transmitted to the distributed power station controller for data analysis and processing. The processing results are then transmitted to the master station for comparison and comprehensive analysis with the upper-level power grid data obtained by the master station, thereby generating optimized control signals and realizing remote control of distributed power sources. In addition, this system supports energy interaction between energy storage units and upper-level power grids, further ensuring the stability and reliability of system operation.

[0047] After completing the collection of historical power generation data and real-time power generation data, the system conducts preliminary analysis through the distributed power station controller, and transmits the analysis results to the main station for higher-level data processing and analysis. In order to ensure the security and privacy of data transmission, the present invention introduces an intelligent gateway and intranet cloud architecture, which strictly limits data transmission and storage devices to the intranet, and only authorized users can access them, thereby effectively avoiding external threats. In addition, the system's data and control signal transmission is based on a variety of dedicated network channels, including 1.8GHz power line wireless private network, 230MHz power wireless private network and optical fiber private network, and the communication protocol strictly follows DL / T645, DL / T698 and other industry standards. This design effectively ensures the stability, security and efficiency of data during transmission, and meets the actual needs of modern distributed power systems.

[0048] like Figure 2 As shown, a distributed power generation prediction system is used to execute the above-mentioned distributed power generation prediction method, including: The standardization module 101 collects the distributed power generation data, and performs data cleaning and standardization on the distributed power generation data to obtain standard distributed power generation data; The decomposition and dimensionality reduction module 102 decomposes and reduces the dimensionality of the standard distributed power generation data by using a variational modal decomposition algorithm to obtain modal components and residual components; The correlation analysis module 103 performs correlation analysis on each modal component by a correlation analysis method to obtain the correlation coefficient between the modal components; The classification module 104 classifies each modal component according to the correlation coefficient between the modal components to obtain a new modal component; The prediction module 105 inputs each new modal component into the distributed power generation prediction model to obtain the prediction value of each modal component; inputs the residual component into the distributed power generation prediction model to obtain the prediction value of the residual component; The superposition module 106 superimposes the predicted values ​​of each modal component and the predicted value of the residual component to obtain the predicted value of distributed power generation.

[0049] Through the collaborative work of the above modules, by fully considering the cleaning process of the original distributed power generation data, fully cleaning the abnormal data, and standardizing the original data, the accuracy of subsequent data analysis is improved, which is conducive to improving the prediction accuracy of the distributed power generation prediction model. Taking full account of the dimensionality reduction of the original data, a CNN-LSTM prediction model based on the VMD algorithm is constructed, which effectively reduces the interference of external factors on the distributed power generation prediction, improves the prediction accuracy, reduces the calculation complexity of the model, and improves the prediction efficiency.

[0050] Figure 3 A block diagram of an electronic device is shown as an example. Figure 3 As shown, the electronic device may include: a processor 201 (processor), a communication interface 203 (Communications Interface), a memory 204 (memory) and a communication bus 202, wherein the processor 201, the communication interface 203, and the memory 204 communicate with each other through the communication bus 202. The processor 201 may call the logic instructions in the memory 204 to execute a distributed power generation prediction method.

[0051] In addition, the logic instructions in the above-mentioned memory 204 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and 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 method described in each embodiment 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, etc. Various media that can store program codes.

[0052] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute a distributed power generation prediction method provided by the above-mentioned methods.

[0053] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a distributed power generation prediction method provided by the above methods.

[0054] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0055] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A distributed power generation prediction method, characterized in that: include: S1: Collect distributed power generation data, clean and normalize the distributed power generation data, and obtain standard distributed power generation data; S2: Decompose and reduce the dimensionality of the standard distributed power generation data through the variational modal decomposition algorithm to obtain the modal components and residual components; S3: Perform correlation analysis on each modal component by using the correlation analysis method to obtain the correlation coefficient between the modal components; S4: classify each modal component according to the correlation coefficient between the modal components to obtain a new modal component; S5: input each new modal component into the distributed power generation prediction model to obtain a prediction value of each modal component; input the residual component into the distributed power generation prediction model to obtain a prediction value of the residual component; S6: Superimpose the predicted values ​​of each modal component and the predicted value of the residual component to obtain the predicted value of distributed generation power generation.

2. A distributed power generation prediction method according to claim 1, characterized in that: The S1 step includes: S11: Identify abnormal data and extract basic information of normal data on both sides of the abnormal data point, the basic information includes the frequency, amplitude and phase of the signal; S12: construct a standard distributed power generation data waveform based on basic information; S13: replacing abnormal data point data with a standard distributed power generation data waveform to obtain cleaned and completed data; S14: normalize the cleaned and completed data to obtain standard distributed power generation data.

3. A distributed power generation prediction method according to claim 1, characterized in that: In step S2, The parameters of the quadratic penalty term are obtained by an alternating direction multiplier algorithm, and the parameters of the quadratic penalty term control the bandwidth of the modal component of the distributed generation data; The standard distributed generation data is decomposed into multiple modal components and residual components through the quadratic penalty parameter, and each modal component has a central frequency and bandwidth signal.

4. A distributed power generation prediction method according to claim 1, characterized in that: The correlation analysis method includes the Pearson correlation coefficient analysis method.

5. A distributed power generation prediction method according to claim 1, characterized in that: In step S4, the correlation coefficient threshold is set. If the absolute value of the correlation coefficient between the modal components is greater than or equal to the correlation coefficient threshold, the modal components are superimposed to obtain new modal components; If the absolute value of the correlation coefficient between the modal components is less than the correlation coefficient threshold, the modal component is taken as a new modal component.

6. A distributed power generation prediction method according to claim 1, characterized in that: The distributed power generation prediction model is a CNN-LSTM prediction model based on the VMD algorithm, including an input layer, a CNN network layer, a LSTM network layer and an output layer; The input layer includes a fully connected layer, the CNN network layer includes two one-dimensional convolutional layers and a maximum pooling layer, and both convolutional layers use the ReLU activation function; the LSTM network layer includes two LSTM layers, a fully connected layer and a random inactivation layer, and the output layer includes a fully connected layer.

7. A distributed power generation prediction method according to claim 1, characterized in that: The distributed power generation data is collected, transmitted, processed and analyzed by an intelligent power data measurement device.

8. A distributed power generation prediction system, characterized in that: A method for predicting distributed power generation according to any one of claims 1 to 7, comprising: A standardization module, wherein the standardization module collects distributed power generation data, performs data cleaning and standardization processing on the distributed power generation data, and obtains standard distributed power generation data; A decomposition and dimensionality reduction module, wherein the decomposition and dimensionality reduction module decomposes and reduces the standard distributed power generation data through a variational modal decomposition algorithm to obtain modal components and residual components; A correlation analysis module, wherein the correlation analysis module performs correlation analysis on each modal component by a correlation analysis method to obtain correlation coefficients between modal components; A classification module, wherein the classification module classifies each modal component according to the correlation coefficient between the modal components to obtain a new modal component; A prediction module, wherein the prediction module inputs each new modal component into a distributed power generation prediction model to obtain a prediction value of each modal component; and inputs a residual component into the distributed power generation prediction model to obtain a prediction value of the residual component; A superposition module is used to superimpose the predicted values ​​of each modal component and the predicted value of the residual component to obtain the predicted value of distributed power generation.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of a distributed power generation prediction method as described in any one of claims 1 to 7 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a distributed power generation prediction method as described in any one of claims 1 to 7 are implemented.

Images