Photovoltaic generating capacity prediction method and system based on stability correction

Abstract

The invention discloses a photovoltaic generating capacity prediction method and system based on stationarity correction, and the method comprises the steps: carrying out the partitioning of preprocessed photovoltaic generating capacity data according to the period length, and obtaining a photovoltaic generating capacity data block sequence composed of a plurality of data blocks; inputting the photovoltaic generating capacity data block sequence into a deep feature extraction model, respectively performing deep time feature extraction and deep space feature extraction to obtain time features and space features, and fusing the time features and the space features to obtain deep features; introducing an autocorrelation matrix to perform stability correction on the deep features to obtain corrected features; and inputting the correction characteristics into a generating capacity prediction model for prediction to obtain a prediction result. According to the method, the deep feature extraction model is designed to extract the deep features, and the extracted features are corrected, so that the prediction precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and in particular to a photovoltaic power generation prediction method and system based on stationarity correction. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

[0003] As the core carrier of clean energy, accurate prediction of photovoltaic power generation has become a key technical foundation for optimizing power system scheduling, energy storage capacity configuration and market-based transactions. In order to ensure efficient operation and stable power generation of photovoltaic power generation systems, accurate prediction of photovoltaic power generation has become an urgent problem to be solved.

[0004] The prediction of photovoltaic power generation is affected by many factors, including not only long-term time characteristics such as monthly cycles and annual cycles in historical data, but also short-term local time patterns such as daily cycles. At the same time, external environmental factors such as sunshine intensity, ambient temperature, and wind speed also have an important impact on power generation. The existing photovoltaic power generation prediction algorithms have some obvious shortcomings: (1) Many common photovoltaic power generation prediction methods rely on shallow feature extraction models and fail to fully explore the deep temporal characteristics and the complex relationships between variables, resulting in low prediction accuracy.

[0005] (2) Traditional methods fail to effectively deal with non-stationary problems caused by factors such as trends, seasonal changes or irregular fluctuations, and ignore the time dependence within the data, which limits the accurate modeling and prediction of PV power generation characteristics.

[0006] (3) Existing time series modeling methods usually treat the entire historical time series as a whole and fail to effectively capture local temporal patterns at different levels, resulting in insufficient forecasting flexibility and accuracy. Summary of the invention

[0007] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a photovoltaic power generation prediction method and system based on stationarity correction, designs a deep feature extraction model to extract deep features, and corrects the extracted features to improve the prediction accuracy.

[0008] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: In a first aspect, the present invention provides a photovoltaic power generation prediction method based on stationarity correction, comprising: Acquire photovoltaic power generation data and perform preprocessing; The pre-processed photovoltaic power generation data is divided into blocks according to the cycle length to obtain a photovoltaic power generation data block sequence consisting of multiple data blocks; Inputting the photovoltaic power generation data block sequence into a deep feature extraction model, performing deep time feature extraction and deep space feature extraction respectively to obtain time features and space features, and fusing the time features and space features to obtain deep features; An autocorrelation matrix is ​​introduced to perform stationarity correction on the deep features to obtain correction features; The correction feature is input into a power generation prediction model to perform prediction and obtain a prediction result.

[0009] A further technical solution is to divide the preprocessed photovoltaic power generation data into blocks according to the cycle length: the last data value of the photovoltaic power generation data is repeated multiple times and filled to the end of the data, and the cycle length is used as the block length to divide it to obtain multiple data blocks.

[0010] According to a further technical solution, the deep feature extraction model is composed of multi-layer perceptrons connected in sequence.

[0011] A further technical solution is to obtain the time feature by first extracting local features within each data block, and then extracting global features between different data blocks to obtain the time feature.

[0012] A further technical solution to obtain spatial features is as follows: combining the meteorological data block sequence and the photovoltaic power generation data block sequence into a multivariate time series, inputting the multivariate time series into a deep feature extraction model for modeling, performing feature extraction on the variable dimension, and obtaining spatial features.

[0013] A further technical solution is to obtain the correction characteristics as follows: Calculate the mean, variance and autocorrelation matrix of the PV power generation data block sequence respectively; Calculate the stationary correction coefficient so that the autocorrelation matrix after multiplying the stationary correction coefficient with the deep features is close to the autocorrelation matrix of the photovoltaic power generation data block sequence, and use fast Fourier transform to accelerate the calculation of the stationary correction coefficient; The deep features are affine transformed using stationarity correction coefficients to obtain correction features.

[0014] Further technical solution, the expression of the stability correction coefficient is expressed as:

[0015]

[0016] in, represents the stability correction coefficient, Indicates the length of input data. Represents a sequence of photovoltaic power generation data blocks No. The lagged series and The correlation coefficient between the lagged series is Representing deep features No. The lagged series and The correlation coefficient between the lagged series.

[0017] In a second aspect, the present invention provides a photovoltaic power generation prediction system based on stationarity correction, comprising: A data acquisition module is configured to: acquire photovoltaic power generation data and perform preprocessing; A data block module is configured to: block the preprocessed photovoltaic power generation data according to the cycle length to obtain a photovoltaic power generation data block sequence consisting of multiple data blocks; A deep feature extraction module is configured to: input the photovoltaic power generation data block sequence into a deep feature extraction model, perform deep time feature extraction and deep space feature extraction respectively to obtain time features and space features, and fuse the time features and space features to obtain deep features; A stationarity correction module is configured to: introduce an autocorrelation matrix to perform stationarity correction on the deep features to obtain corrected features; The prediction output module is configured to: input the correction feature into the power generation prediction model for prediction to obtain a prediction result.

[0018] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in a method for predicting photovoltaic power generation based on stationarity correction as described in the first aspect.

[0019] In a fourth aspect, the present invention provides a computer 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 in a method for predicting photovoltaic power generation based on stationarity correction as described in the first aspect are implemented.

[0020] One or more of the above technical solutions have the following beneficial effects: The present invention divides photovoltaic power generation data into periods through sequence blocks, which can effectively capture local time characteristics and improve the adaptability of the power generation prediction model to periodic changes; a deep-level time feature extractor is used to fully explore the local time pattern and global time pattern of photovoltaic power generation data, and a spatial feature extractor is used to fully explore the complex relationship between photovoltaic power generation data and meteorological data, thereby enhancing the feature learning ability of the power generation prediction model; an autocorrelation matrix is ​​introduced to perform stationarity correction on the extracted features, and the data dependency and model stability are further guaranteed by recovering the lost non-stationary information; finally, the final prediction is performed through the LSTM long short-term memory neural network, which significantly improves the accuracy and robustness of the photovoltaic power generation prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0022] Figure 1 is a sequence block flow chart of a photovoltaic power generation prediction method according to an embodiment of the present invention; Figure 2 is a structural diagram of a deep feature extraction model according to an embodiment of the present invention; Figure 3 is a flow chart of time feature extraction according to an embodiment of the present invention; Figure 4 is a flow chart of spatial feature extraction according to an embodiment of the present invention; Figure 5 is a flow chart of deep feature stability correction according to an embodiment of the present invention; Figure 6 It is a module diagram of a photovoltaic power generation prediction system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0024] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0025] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0026] Embodiment 1 This embodiment discloses a photovoltaic power generation prediction method based on stationarity correction, which includes the following steps: S1: Acquire photovoltaic power generation data and perform preprocessing; In this embodiment, the photovoltaic power generation data is obtained and defined as , Indicates the length of the data. In order to enhance the stability and performance of the prediction method, the photovoltaic power generation data is first preprocessed, that is, normalized, to adjust the data to a more appropriate range. The following formula is used for normalization: (1) in, Representation data The mean of Representation data The variance of .

[0027] S2: Divide the preprocessed photovoltaic power generation data into blocks according to the cycle length to obtain a photovoltaic power generation data block sequence consisting of multiple data blocks; In this embodiment, if Figure 1 As shown in the figure, the steps of data sequence segmentation are as follows: Repeat the last data value of the photovoltaic power generation data multiple times and fill it to the end of the data; The cycle length is used as the block length to divide the data into multiple data blocks; A statistic of each data block is obtained based on the plurality of data blocks.

[0028] Specifically, after normalization, repeat The last data times and pad it to the end of the original data sequence before chunking.

[0029] In order to capture the local features of short periods, the block length Defined as the length of the daily cycle, that is, assuming the time step between two data points is 1 hour, the block length is 24.

[0030] The step size between two consecutive data blocks is defined as , the number of blocks is obtained according to the following formula : (2) in, represents the lower limit function, Indicates the length of the input data.

[0031] Therefore, the result after block division is obtained, that is, multiple data blocks constitute a data block sequence , called the photovoltaic power generation data block sequence. Since subsequent operations will cause the data distribution in the data block to change, it is also necessary to record The mean ,variance and the autocorrelation matrix (First calculate the covariance based on the data, and then get the autocorrelation matrix based on the covariance) for subsequent stationarity correction operations.

[0032] The preprocessed photovoltaic power generation data is divided into blocks to better capture the local time characteristics in the photovoltaic power generation data.

[0033] S3: Inputting the photovoltaic power generation data block sequence into a deep feature extraction model, performing deep time feature extraction and deep space feature extraction respectively to obtain time features and space features, and fusing the time features and space features to obtain deep features; In this embodiment, if Figure 3 As shown, after the data is divided into blocks according to the period, the time features and spatial features of the data are extracted respectively using a deep feature extraction model, and the model structure is composed of multiple layers of MLP (multi-layer perceptron) connected in sequence. In this embodiment, Figure 2 As shown in Figure 1, the deep feature extraction model uses 3 layers of MLP (6 MLPs in total), which can be flexibly set according to the actual application.

[0034] The deep temporal feature extraction is specifically as follows: firstly, local feature extraction is performed inside each data block, and then global feature extraction is performed between different data blocks to obtain the temporal feature.

[0035] Specifically, in the deep temporal feature extraction process, local features are first extracted within each data block, and then global features are extracted between different data blocks. of The data of each block are input into the deep feature extraction model to model the features in each daily cycle, that is, The short-term time features are extracted from the data values ​​and defined as local features. Then, the local features are transposed and input into the deep feature extraction model. The global features between blocks can be modeled to capture long-term time features such as monthly cycles and annual cycles, and the long-term time features are defined as global features. Finally, the data after global feature modeling is transposed to the original dimension to obtain .

[0036] The deep spatial feature extraction is specifically as follows: combining the meteorological data block sequence and the photovoltaic power generation data block sequence into a multivariate time series, inputting the multivariate time series into the deep feature extraction model for modeling, performing feature extraction on the variable dimension, and obtaining spatial features.

[0037] like Figure 4 As shown, specifically, for deep spatial feature extraction, since the photovoltaic power generation prediction is strongly correlated with meteorological variables such as light intensity, the meteorological data (meteorological data block sequence) and the photovoltaic power generation data (photovoltaic power generation data block sequence) are first combined into multivariate time series data, and then the multivariate time series data is input into the deep feature extraction model for modeling and extracting spatial features. The deep feature extraction model extracts features from the multivariate time series data in the variable dimension to obtain spatial features. The variable dimension here refers to multiple time series, such as photovoltaic power generation data, light intensity data, temperature data, rainfall data, etc., which are all independent variables (time series). Feature extraction in the variable dimension refers to modeling the association between these variables, and spatial features refer to the features after the photovoltaic power generation data is modeled and associated with other meteorological variables.

[0038] Specifically, define meteorological data , here Represents the number of meteorological variables, including light intensity (Light), rainfall (Rain), temperature (Temp), etc., as shown in the following formula: . (3) The meteorological data is divided into blocks, and the result after the block division is that multiple data blocks constitute a data block sequence , called the meteorological data block sequence. and Combine to get multivariate time series data , then Transpose and get After the data combination is completed, Input to the deep feature extraction model, the size is The variable dimension is modeled, the data of the dimension where the photovoltaic power generation is located is taken out, and the process of fusion of meteorological data features is completed.

[0039] Finally, the temporal features and spatial features are weighted and fused to obtain the final deep feature extraction result. , that is, deep features.

[0040] S4: introducing an autocorrelation matrix to perform stationarity correction on the deep features to obtain corrected features; Since the feature extraction process after block segmentation will cause significant changes in the distribution of data, existing methods for recovering data distribution mainly focus on statistical measures such as mean and variance, without considering the time dependency within the data. Therefore, basic features such as trend and seasonality in the original data may be affected. To this end, the present invention introduces a novel stationarity correction method, which constrains the relationship between the stationarity of the time series before and after model processing to correct the data distribution while preserving data dependency.

[0041] The stationarity of a time series is mainly reflected in two aspects: mean and covariance. The mean is mainly responsible for constraining the distribution from a statistical perspective, while the covariance constrains the distribution from a time-dependent perspective. Changes in the mean of data distribution are usually achieved through global addition or subtraction operations and will not affect the covariance. However, adjusting the covariance may cause changes in the mean of data distribution. Therefore, the present invention first adjusts the covariance of the data. Can represent a sequence of photovoltaic power generation data blocks With its The dependence between the lagged series is not sufficient to fully describe the temporal dependence of the entire time series. Figure 5 As shown, the present invention introduces the autocorrelation matrix to provide a more comprehensive constraint on the time series dependency, defining for The autocorrelation matrix is ​​expressed as: (4) (5) in, express No. The lagged series and The correlation coefficient between the lagged series is express No. A lagged series, express The variance of express No. A lagged series, express The variance of .

[0042] Calculating the Matrix (i.e., the stability correction coefficient) to achieve its deep features after feature extraction The autocorrelation matrix after multiplication Approach , It is expressed as: (6) (7) in, Indicates the length of input data. express No. The lagged series and The correlation coefficient between the lagged series.

[0043] and The constraints between are expressed as: (8) in, express and The constraints between Represents the square of the Frobenius norm, which is used to measure the size of a matrix. The calculation formula is the square root of the sum of the squares of all elements in the matrix.

[0044] is a non-stationary time series. Multiply will result in The change between the data points in Zoom, so It is equivalent to the following formula: . (9) According to the Wiener-Khinchin theorem, it can be accelerated by fast Fourier transform (FFT) The calculation process is as follows: (10) (11) (12) in, express The form after removing the mean is Subtract the mean from all values , in order to The mean of is 0 to satisfy the conditions of the Wiener-Khinchin theorem; express The form after removing the mean; Express Take the complex conjugate after Fourier transformation; Express Take the complex conjugate after Fourier transformation; yes The mean of yes The mean of .

[0045] Therefore, deep features Can be updated to: (13) (14) in, represents the calibration characteristic, express The mean vector of express The mean vector of ; Represents the mean correction term, used to adjust and The mean difference between .

[0046] The present invention restores the non-stationary information that may be lost in the normalization process by introducing an autocorrelation matrix, corrects data distribution, and simultaneously maintains the dependency relationship between data.

[0047] S5: Inputting the correction feature into a power generation prediction model for prediction to obtain a prediction result.

[0048] Finally, the results of temporal feature modeling, spatial feature modeling and stationarity correction will be completed. Flatten it into a continuous sequence and input it into the long short-term memory neural network (LSTM) to get the prediction result. Flattening the data into a continuous sequence is to merge the block sequences into a one-dimensional continuous sequence, which is convenient for inputting into the LSTM model to output the prediction result.

[0049] The modeling and output process of LSTM is shown in Equations (15) to (20): (15) (16) (17) (18) (19) (20) in, represents the input gate, represents the sigmoid activation function, , , , Represent the weight matrices of the input gate, forget gate, candidate memory unit and output gate respectively, Indicates the hidden state at the current moment, Represents the input at the current moment, , , , Represent the bias items of the input gate, forget gate, candidate memory unit and output gate respectively, represents the forget gate, represents the candidate memory unit, represents the output gate, Represents the memory unit at the current moment.

[0050] The final prediction process is expressed as: (twenty one) in, Represents the prediction result.

[0051] Photovoltaic power generation prediction, as a time series prediction task, is a typical regression problem. Therefore, the present invention uses several common regression problem evaluation indicators to better demonstrate the effect of the proposed method. These indicators include mean square error (MSE), mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE). Formulas (23) to (26) show the calculation formulas of these evaluation indicators. The smaller the values ​​of these indicators, the better the prediction effect.

[0052] (twenty two) (twenty three) (twenty four) (25) in, Indicates the photovoltaic power generation at time The true value of The method of the present invention is The predicted value of photovoltaic power generation, is the prediction time length.

[0053] The method of the present invention was compared with four advanced methods: Transformer model, bidirectional long short-term memory neural network (Bi-LSTM), CNN-LSTM neural network and temporal convolutional network (TCN). The values ​​of the four evaluation indicators of the method of the present invention were the smallest, indicating the effectiveness of the proposed method.

[0054] Embodiment 2 like Figure 6As shown, this embodiment discloses a photovoltaic power generation prediction system based on stationarity correction, including: A data acquisition module is configured to: acquire photovoltaic power generation data and perform preprocessing; A sequence block module is configured to: block the pre-processed photovoltaic power generation data according to the cycle length to obtain multiple data blocks; A deep feature extraction module is configured to: input the multiple data blocks into a deep feature extraction model, perform deep temporal feature extraction and deep spatial feature extraction respectively to obtain temporal features and spatial features, and fuse the temporal features and spatial features to obtain deep features; A stationarity correction module is configured to: introduce an autocorrelation matrix to perform stationarity correction on the deep features to obtain corrected features; The prediction output module is configured to: input the correction feature into the power generation prediction model for prediction to obtain a prediction result.

[0055] Embodiment 3 The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method of embodiment 1 when executing the program.

[0056] Embodiment 4 The purpose of this embodiment is to provide a computer-readable storage medium, a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the steps of the method of embodiment 1 are performed.

[0057] The steps involved in the apparatus of the above embodiments 3 and 4 correspond to the method embodiment 1, and the specific implementation method can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0058] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0060] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A photovoltaic power generation prediction method based on stationarity correction, characterized in that: include: Acquire photovoltaic power generation data and perform preprocessing; The pre-processed photovoltaic power generation data is divided into blocks according to the cycle length to obtain a photovoltaic power generation data block sequence consisting of multiple data blocks; Inputting the photovoltaic power generation data block sequence into a deep feature extraction model, performing deep time feature extraction and deep space feature extraction respectively to obtain time features and space features, and fusing the time features and space features to obtain deep features; An autocorrelation matrix is ​​introduced to perform stationarity correction on the deep features to obtain correction features; The correction feature is input into a power generation prediction model to perform prediction and obtain a prediction result.

2. A photovoltaic power generation prediction method based on stationarity correction as claimed in claim 1, characterized in that: The preprocessed photovoltaic power generation data is divided into blocks according to the cycle length. Specifically, the last data value of the photovoltaic power generation data is repeated multiple times and filled to the end of the data, and the cycle length is used as the block length to obtain multiple data blocks.

3. A photovoltaic power generation prediction method based on stationarity correction as claimed in claim 1, characterized in that: The deep feature extraction model is composed of multi-layer perceptrons connected in sequence.

4. A photovoltaic power generation prediction method based on stationarity correction as claimed in claim 1, characterized in that: The specific steps of obtaining the time feature are as follows: firstly, local features are extracted within each data block, and then global features are extracted between different data blocks to obtain the time feature.

5. The photovoltaic power generation prediction method based on stationarity correction according to claim 1, characterized in that: The spatial features are obtained specifically as follows: the meteorological data block sequence and the photovoltaic power generation data block sequence are combined into a multivariate time series, the multivariate time series is input into a deep feature extraction model for modeling, and feature extraction is performed on the variable dimension to obtain the spatial features.

6. A photovoltaic power generation prediction method based on stationarity correction as claimed in claim 1, characterized in that: The correction characteristics are as follows: Calculate the mean, variance and autocorrelation matrix of the PV power generation data block sequence respectively; Calculate the stationary correction coefficient so that the autocorrelation matrix after multiplying the stationary correction coefficient with the deep features is close to the autocorrelation matrix of the photovoltaic power generation data block sequence, and use fast Fourier transform to accelerate the calculation of the stationary correction coefficient; The deep features are affine transformed using stationarity correction coefficients to obtain correction features.

7. A photovoltaic power generation prediction method based on stationarity correction as claimed in claim 6, characterized in that: The expression of the stability correction coefficient is expressed as: in, represents the stability correction coefficient, Indicates the length of input data. Represents a sequence of photovoltaic power generation data blocks No. The lagged series and The correlation coefficient between the lagged series is Representing deep features No. The lagged series and The correlation coefficient between the lagged series.

8. A photovoltaic power generation prediction system based on stationarity correction, characterized in that: include: A data acquisition module is configured to: acquire photovoltaic power generation data and perform preprocessing; A data block module is configured to: block the preprocessed photovoltaic power generation data according to the cycle length to obtain a photovoltaic power generation data block sequence consisting of multiple data blocks; A deep feature extraction module is configured to: input the photovoltaic power generation data block sequence into a deep feature extraction model, perform deep time feature extraction and deep space feature extraction respectively to obtain time features and space features, and fuse the time features and space features to obtain deep features; A stationarity correction module is configured to: introduce an autocorrelation matrix to perform stationarity correction on the deep features to obtain corrected features; The prediction output module is configured to: input the correction feature into the power generation prediction model for prediction to obtain a prediction result.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a photovoltaic power generation prediction method based on stationarity correction as described in any one of claims 1 to 7 are implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the photovoltaic power generation prediction method based on stationarity correction as described in any one of claims 1 to 7 are implemented.

Images