Method for predicting different meteorological power generation based on photovoltaic power generation prediction model

Abstract

The invention discloses a method for predicting different meteorological power generation based on a photovoltaic power generation prediction model. The method comprises the following steps: carrying out meteorological factor photovoltaic power generation correlation analysis processing on photovoltaic power station meteorological data information and photovoltaic power generation power data information based on a Spearman algorithm; based on an FCM fuzzy clustering algorithm, meteorological factor photovoltaic power generation correlation analysis processing result is carried out, and meteorological type power generation working condition classification processing is carried out; based on a VDM variational mode decomposition algorithm, carrying out generation power sub-sequence decomposition processing on the meteorological type generation working conditions, and extracting generation power fluctuation characteristics; based on a photovoltaic power generation prediction model, photovoltaic power generation capacity prediction processing is carried out on the power generation power fluctuation characteristics of the power generation power subsequences of all meteorological type working conditions; the generating capacity under different meteorological conditions can be accurately predicted based on the photovoltaic power generation prediction model.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation, and in particular to a method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model. Background Art

[0002] The main disadvantage of the photovoltaic power generation prediction method in the prior art is the poor prediction accuracy under different weather conditions. For example, the application number is CN202410109711.6, the invention patent name is: A photovoltaic power generation prediction method based on clustering algorithm fusion. The specific technical disadvantages are as follows: Existing prediction methods are sensitive to the selection of initial cluster centers, such as the K-means algorithm. Different initial cluster centers may lead to different clustering results, which affects the stability and reliability of the model.

[0003] Current forecasting models have difficulty adapting to rapidly changing weather conditions when dealing with photovoltaic power generation data that has obvious intermittency, randomness and volatility, thus affecting the accuracy of the forecast.

[0004] Traditional machine learning algorithms may not be as effective as deep learning methods when dealing with nonlinear relationships and large-scale datasets, which limits the accuracy and applicability of predictive models.

[0005] Most existing technologies use historical data as training sets, and the data types are fixed, usually irradiance, temperature, humidity, air pressure, wind speed, and power generation. The lack of parameters that directly reflect the operating conditions of photovoltaic panels leads to poor prediction accuracy under complex meteorological conditions such as cloudy or rainy days. Summary of the invention

[0006] In order to overcome the deficiencies of the prior art, the present invention provides a method for predicting power generation under different meteorological conditions based on a photovoltaic power generation prediction model, thereby realizing accurate prediction of power generation under different meteorological conditions based on the photovoltaic power generation prediction model.

[0007] In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical solutions: The first aspect of the present application provides a method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model, comprising the following steps: S101, performing meteorological factor photovoltaic power generation correlation analysis on the photovoltaic power station meteorological data information and photovoltaic power generation data information based on the Spearman algorithm; S102, classifying power generation conditions by meteorological type based on the results of meteorological factor photovoltaic power generation correlation analysis and processing based on a fuzzy clustering algorithm; S103, based on the variational mode decomposition algorithm, the meteorological type power generation condition classification processing results are processed, and the power generation subsequences of each condition are decomposed and the power generation fluctuation characteristics are extracted; S104, based on the photovoltaic power generation prediction model, the photovoltaic power generation prediction processing is performed on the power generation power fluctuation characteristics of the power generation power subsequences under each meteorological type working condition.

[0008] Further, the meteorological factor photovoltaic power generation correlation analysis processing based on the Spearman algorithm for the photovoltaic power station meteorological data information and photovoltaic power generation data information includes the following steps: Preprocess the meteorological data information and photovoltaic power generation data information of the photovoltaic power station; Based on the Spearman algorithm, the correlation coefficient of the pre-processed photovoltaic power station meteorological data and photovoltaic power generation data is calculated and processed; Based on the results of correlation coefficient calculation, the correlation analysis of meteorological factors and photovoltaic power generation is carried out to select parameters with a greater correlation with power generation as characteristic parameters of the prediction model; the parameters with a greater correlation with power generation are total radiation, temperature, humidity, wind speed and backplane temperature.

[0009] Furthermore, the correlation coefficient is calculated based on the Spearman algorithm for the pre-processed photovoltaic power station meteorological data information and photovoltaic power generation data information, and the expression is as follows:

[0010] in, X i is the magnitude of the influencing factors of the i-th sample point, including total radiation, temperature, humidity, atmospheric pressure, wind speed, wind direction, and backplane temperature. Y i is the power generation size of the i-th sample point, X and Y are the influencing factors and the power generation N The average value among the sample points.

[0011] Furthermore, based on the fuzzy clustering algorithm, the meteorological factor photovoltaic power generation correlation analysis results are processed to classify the meteorological type power generation conditions, including the following steps: Taking irradiance as the clustering index, the results of the meteorological factor photovoltaic power generation correlation analysis are preliminarily divided into different meteorological types of power generation conditions; The membership degree and distance calculation are performed on the preliminary classification results of meteorological type power generation conditions and the cluster centers of different meteorological type power generation conditions; Based on the membership degree and distance calculation processing results, the meteorological type power generation conditions are reclassified to iteratively optimize the preliminary classification processing results of the meteorological type power generation conditions.

[0012] Furthermore, the meteorological type power generation conditions are reclassified based on the membership degree and distance calculation processing results to iteratively optimize the preliminary classification processing results of the meteorological type power generation conditions. The iterative process is as follows: 1) Optimize the objective function through iteration J ( U , V ) to minimize the weighted sum of square distances between data points and cluster centers. The objective function is: ; 2) In order to ensure that the sum of the membership of each data point is 1, the constraint function is: ; 3) Randomly initialize the membership matrix U and cluster center V, and update the membership matrix u according to the following expression j and cluster center u ij : ; 4) Repeat step 3) until the membership matrix U converges or the maximum number of iterations is reached; In the formula, u ij is the sample point x i With cluster center v j The membership degree of is in the range of 0-1; m is the fuzzy index, and m>1, which determines the fuzziness of the clustering, and its value range is 1.5-2.5; d ij is the sample point x i With cluster center v j The distance between n is the sample size; k is the number of cluster centers.

[0013] Furthermore, based on the variational mode decomposition algorithm, the meteorological type power generation conditions are classified and processed, and the power generation subsequences of each condition are decomposed and processed to extract the power generation fluctuation characteristics, including the following steps: The classification and processing results of power generation conditions of different meteorological types are sorted and organized in chronological order; Performing time series decomposition on the meteorological type power generation condition classification processing results that have been sorted in time series, and obtaining the time series decomposition processing results of the power generation subsequences of each condition; Through the time series decomposition processing results of the power generation subsequences under each operating condition, the central frequency of the IMF of the power generation subsequence segment under each operating condition after decomposition is determined, and the power generation fluctuation characteristics are extracted.

[0014] Furthermore, constructing a photovoltaic power generation prediction model includes the following steps: Obtain historical meteorological data information and photovoltaic power generation data information of photovoltaic power stations, and classify and identify meteorological type power generation conditions; The classification and identification results of meteorological type power generation conditions are decomposed into power generation subsequences, and the power generation fluctuation characteristics are extracted; Based on the extracted power generation fluctuation characteristics, the initial photovoltaic power generation prediction model is trained; If the model training results are the same as the training sample results, the model training is completed; if the model training results are different from the training sample results, the model training is repeated.

[0015] Furthermore, based on the photovoltaic power generation prediction model, the photovoltaic power generation prediction processing is performed on the power generation fluctuation characteristics of the power generation subsequences under each meteorological type working condition, including the following steps: Encoding and marking the power generation fluctuation characteristics of the power generation subsequences under each meteorological type of working condition; Performing feature screening processing on the coding mark processing results; Photovoltaic power generation prediction is performed based on the feature screening processing results.

[0016] Furthermore, the feature screening process for the coding mark processing result includes the following steps: Perform feature activation value calculation on the encoding and marking processing results; The data entry state of the encoding tag processing result is determined based on the feature activation value calculation processing result, and the data entry state includes an input gate state, a forget gate state, a unit state, a candidate unit state and an output gate state.

[0017] The second aspect of the present application provides a method for predicting different meteorological power generation systems based on a photovoltaic power generation prediction model, including: A data acquisition unit, used to acquire meteorological data information and photovoltaic power generation data information of the photovoltaic power station; The first processing unit is used to perform meteorological factor photovoltaic power generation correlation analysis on the photovoltaic power station meteorological data information and photovoltaic power generation data information based on the Spearman algorithm; The second processing unit is used to classify the meteorological type power generation conditions based on the meteorological factor photovoltaic power generation correlation analysis results based on the FCM fuzzy clustering algorithm; The third processing unit is used to classify the meteorological type power generation conditions based on the VDM variational mode decomposition algorithm, perform power generation subsequence decomposition processing for each condition and extract power generation fluctuation characteristics; A fourth processing unit is used to perform photovoltaic power generation prediction processing on the power generation fluctuation characteristics of the power generation subsequences of each meteorological type working condition based on the photovoltaic power generation prediction model; The model building unit is used to build a photovoltaic power generation prediction model.

[0018] The beneficial effects of the present application are as follows: the FCM fuzzy clustering algorithm is used to classify photovoltaic power generation conditions under different meteorological conditions, thereby improving the adaptability of the model to various weather changes; the VMD variational mode algorithm is used to extract power generation characteristics, effectively process non-stationary signals and reduce modal aliasing, and provide stable feature input for the model; the LSTM deep learning algorithm is combined to establish a prediction model to capture long-term dependencies in time series data, and the LSTM model hyperparameters are optimized based on the GA algorithm to improve prediction efficiency and accuracy; while considering meteorological parameters, the photovoltaic backplane temperature is further incorporated into the feature parameters.

[0019] The adaptability and accuracy of the enhanced model were achieved: the power generation conditions of different meteorological types were classified based on the FCM fuzzy clustering algorithm, so that the model can adapt to the characteristics of photovoltaic power generation under various meteorological conditions. This method can more accurately describe the uncertainty and ambiguity of the data than the traditional hard clustering algorithm, thereby improving the adaptability of the model to changes in different meteorological conditions. Optimized feature extraction: The VMD variational mode algorithm is used to process non-stationary signals and extract deep features of photovoltaic power generation. This step helps to extract more accurate and representative features from complex power generation data, and can more effectively capture subtle changes and trends in the data than traditional methods. Enhanced model practicality: The meteorological data and photovoltaic power generation data are collected in real time through the constructed test platform, ensuring the timeliness of the data and the real-time update capability of the prediction model. The GA genetic algorithm is used to optimize the hyperparameters of the LSTM deep learning model and improve the generalization ability of the model.

[0020] The FCM fuzzy clustering algorithm is used to classify power generation conditions of different meteorological types, which helps to better understand and simulate the behavior of photovoltaic power generation systems under different meteorological conditions, thereby improving the adaptability and accuracy of the prediction model. The VMD variational mode algorithm is used to extract power generation features. This method can effectively process non-stationary signals and reduce modal aliasing, thereby providing more accurate and stable feature inputs for the prediction model. A photovoltaic power generation prediction model is established based on the LSTM deep learning algorithm that is good at processing time series data, and the GA genetic algorithm is used to optimize the hyperparameters of the LSTM model, which can improve the accuracy and generalization ability of photovoltaic power generation prediction. A test platform is built to collect measured meteorological data and photovoltaic power generation data, and further collect backplane temperature data that directly reflects the working status of photovoltaic panels. This provides reliable training and verification data for the model, which helps to improve the practicality of the prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in 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 only 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.

[0022] Figure 1 It is a schematic diagram of the steps of a method for predicting different meteorological power generation based on a photovoltaic power generation prediction model of the present invention; Figure 2 It is a schematic diagram of the IMF of each sub-sequence after the power generation is decomposed under the cloudy weather type power generation condition of the present invention. DETAILED DESCRIPTION

[0023] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0025] Example: A method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model comprises the following steps: S101, performing meteorological factor photovoltaic power generation correlation analysis on the photovoltaic power station meteorological data information and photovoltaic power generation data information based on the Spearman algorithm; Based on the time series, the meteorological data information and photovoltaic power generation data information of the photovoltaic power station are obtained, and the data preprocessing is performed on the meteorological data information and photovoltaic power generation data information of the photovoltaic power station, and the abnormal values ​​and missing values ​​in the data can be standardized. The meteorological data information of the photovoltaic power station includes total atmospheric radiation, temperature, humidity, air pressure, wind speed, wind direction and photovoltaic backplane temperature. Based on the Spearman algorithm, the meteorological factor photovoltaic power generation correlation analysis and processing are performed on the meteorological data information and photovoltaic power generation data information of the photovoltaic power station, and the meteorological factor photovoltaic power generation correlation analysis and processing results are obtained.

[0026] The meteorological factor photovoltaic power generation correlation analysis and processing based on the Spearman algorithm for the photovoltaic power station meteorological data information and photovoltaic power generation data information includes the following steps: Preprocess the meteorological data information and photovoltaic power generation data information of the photovoltaic power station; Based on the Spearman algorithm, the correlation coefficient is calculated for the pre-processed photovoltaic power station meteorological data information and photovoltaic power generation data information. The expression is as follows:

[0027] in, X i is the magnitude of the influencing factors of the i-th sample point, including total radiation, temperature, humidity, atmospheric pressure, wind speed, wind direction, and backplane temperature. Y i is the power generation size of the i-th sample point, X and Y are the influencing factors and the power generation N The average value among the sample points.

[0028] Based on the results of correlation coefficient calculation, the correlation analysis of meteorological factors and photovoltaic power generation is carried out to select parameters with greater correlation with power generation as characteristic parameters of the prediction model.

[0029] The correlation between each influencing factor and the power generation is shown in Table 1. Based on the correlation coefficient calculation results, the correlation analysis of meteorological factors and photovoltaic power generation is carried out. The closer the absolute value of the correlation coefficient calculation result is to 1, the greater the correlation with the power generation, and vice versa. Therefore, the parameters with a greater correlation with the power generation can be selected as the characteristic parameters of the prediction model, which are total radiation, temperature, humidity, wind speed, and backplane temperature.

[0030] Table 1 S102, based on the FCM fuzzy clustering algorithm, the results of the meteorological factor photovoltaic power generation correlation analysis are processed to classify the meteorological type power generation conditions; By analyzing the correlation between meteorological factors and photovoltaic power generation based on the Spearman algorithm, the meteorological factors and photovoltaic power generation correlation analysis results are obtained, and the meteorological data information of photovoltaic power stations with high correlation with photovoltaic power generation is screened out, that is, the total radiation, temperature, humidity, wind speed, backplane temperature and photovoltaic power generation are selected. The meteorological type power generation conditions are classified based on the FCM fuzzy clustering algorithm to obtain the classification results of meteorological type power generation conditions. Meteorological type power generation conditions include sunny meteorological type power generation conditions, cloudy meteorological type power generation conditions and rainy meteorological type power generation conditions. By classifying meteorological type power generation conditions, the photovoltaic power generation power under different meteorological conditions can be analyzed and predicted.

[0031] Based on the FCM fuzzy clustering algorithm, the results of the meteorological factor photovoltaic power generation correlation analysis are processed to classify the meteorological type power generation conditions, including the following steps: Taking irradiance as the clustering index, the results of the meteorological factor photovoltaic power generation correlation analysis are preliminarily divided into different meteorological types of power generation conditions; for example, the results of the meteorological factor photovoltaic power generation correlation analysis are preliminarily divided into sunny, cloudy and rainy meteorological types of power generation conditions, and the clustering centers of the three meteorological types of power generation conditions are obtained; The membership degree and distance calculation are performed on the preliminary classification results of meteorological type power generation conditions and the cluster centers of different meteorological type power generation conditions; Based on the membership degree and distance calculation processing results, the meteorological type power generation conditions are reclassified to iteratively optimize the preliminary classification processing results of the meteorological type power generation conditions. The iterative process is as follows: 1) Optimize the objective function through iteration J ( U , V ) to minimize the weighted sum of square distances between data points and cluster centers. The objective function is: ; 2) In order to ensure that the sum of the membership of each data point is 1, the constraint function is: ; 3) Randomly initialize the membership matrix U and cluster center V, and update the membership matrix u according to the following expression j and cluster center u ij : ; 4) Repeat step 3) until the membership matrix U converges or the maximum number of iterations is reached; In the formula, u ijis the sample point x i With cluster center v j The membership degree of is in the range of 0-1; m is the fuzzy index, and m>1, which determines the fuzziness of the clustering, and its value range is 1.5-2.5; d ij is the sample point x i With cluster center v j The distance between n is the sample size; k is the number of cluster centers.

[0032] It should be noted that the power generation of photovoltaic power generation systems varies significantly under different weather conditions. Therefore, by dividing the results of the meteorological factor photovoltaic power generation correlation analysis into different meteorological types of power generation conditions based on the FCM fuzzy clustering algorithm, the prediction accuracy of power generation under different conditions can be greatly improved.

[0033] S103, based on the VDM variational mode decomposition algorithm, the meteorological type power generation condition classification processing results are processed, the power generation subsequence of each condition is decomposed and the power generation fluctuation characteristics are extracted; By analyzing the correlation between meteorological factors and photovoltaic power generation based on the FCM fuzzy clustering algorithm, the meteorological type power generation conditions are classified and processed. The results of the classification of meteorological type power generation conditions are obtained by decomposing the power subsequences of each condition and extracting the power fluctuation characteristics, and obtaining the power fluctuation characteristics of each meteorological type power generation condition.

[0034] Based on the VDM variational mode decomposition algorithm, the classification processing results of meteorological type power generation conditions are processed, and the power generation subsequences of each condition are decomposed and the power generation fluctuation characteristics are extracted, which includes the following steps: The classification and processing results of power generation conditions of different meteorological types are sorted and organized in chronological order; The classification processing results of meteorological type power generation conditions that have been sorted in time sequence are decomposed in time sequence to obtain the time sequence decomposition processing results of the power generation subsequences of each condition, and the expression is as follows:

[0035] in,{ u k} is the set of variational modes, { w k} is the corresponding meteorological type power generation condition subsequence center frequency set, represents the time derivative, * represents convolution, j is an imaginary unit, d ( t ) is the Dirac delta function, , let the sum of all modal components equal the original signal f ( t ), , l ( t ) is a function that varies over the entire time domain and is used to ensure that the sum of the original signal and the modal is the same at all time points. l k} is a set of discrete Lagrange multipliers used to ensure that each mode satisfies certain constraints.

[0036] Through the time series decomposition processing results of the power generation subsequences under each operating condition, the central frequency of the IMF of the power generation subsequence segment under each operating condition after decomposition is determined, and the power generation fluctuation characteristics are extracted.

[0037] like Figure 2 As shown in the figure, taking the cloudy weather type power generation condition as an example, the classification processing results of the cloudy weather type power generation condition are sorted in chronological order based on the FCM fuzzy clustering algorithm, and the time series decomposition is performed to obtain the time series decomposition processing results of the power generation subsequences of each condition. The center frequency of the IMF of the power generation subsequence segments of each condition of the cloudy weather type after decomposition is 0.00023-0.00029, and the overall reconstruction error is about 4.53%, indicating that the decomposition effect is good.

[0038] S104, performing photovoltaic power generation prediction processing on the power generation fluctuation characteristics of the power generation subsequences under each meteorological type of working condition based on the photovoltaic power generation prediction model; Based on the VDM variational mode decomposition algorithm, the classification results of meteorological type power generation conditions are processed, and the power generation subsequences of each condition are decomposed and the power generation fluctuation characteristics are extracted, which effectively processes non-stationary signals and reduces modal aliasing, providing stable feature input data for the photovoltaic power generation prediction model. Based on the long short-term memory network LSTM and the genetic algorithm GA, the initial photovoltaic power generation prediction model is constructed, and the photovoltaic power generation is predicted based on the power generation fluctuation characteristics of the power generation subsequences of each meteorological type condition based on the photovoltaic power generation prediction model.

[0039] Building a photovoltaic power generation prediction model includes the following steps: Obtain historical meteorological data information and photovoltaic power generation data information of photovoltaic power stations, and classify and identify meteorological type power generation conditions; The classification and identification results of meteorological type power generation conditions are decomposed into power generation subsequences, and the power generation fluctuation characteristics are extracted; Based on the extracted power generation fluctuation characteristics, the initial photovoltaic power generation prediction model is trained; If the model training results are the same as the training sample results, the model training is completed; if the model training results are different from the training sample results, the model training is repeated.

[0040] Based on the photovoltaic power generation prediction model, the photovoltaic power generation prediction processing includes the following steps: Encoding and marking the power generation fluctuation characteristics of the power generation subsequences under each meteorological type of working condition; Performing feature screening processing on the coding mark processing results; Photovoltaic power generation prediction is performed based on the feature screening processing results.

[0041] The feature screening process for the coding mark processing result includes the following steps: Perform feature activation value calculation on the encoding and marking processing results; The data entry state of the encoding tag processing result is determined based on the feature activation value calculation processing result, and the data entry state includes an input gate state, a forget gate state, a unit state, a candidate unit state and an output gate state.

[0042] The feature activation value is calculated for the encoding and marking processing results, and the calculation expression is as follows:

[0043] in, i t is the activation value of the input gate, s is the sigmoid activation function, W i is the weight, h i-1 is the hidden state at the previous time step, x t is the input of the current time step, b i is the bias term, f t is the activation value of the forget gate, W f is the weight; C t is the cell state at the current time step, is the candidate cell state, ∗ denotes element-wise multiplication; O t is the activation value of the output gate, h t is the hidden state at the current time step, h t-1is the hidden state of the previous time step, w 0 is the weight matrix of the output gate, b i is the bias term of the input gate, b f is the bias term of the forget gate, C t-1 is the state of the memory cell at time step t−1, b 0 is the bias term of the output gate, h ( C t ) is the hidden state at time step t.

[0044] It should be noted that the activation value of the input gate determines which power generation fluctuation characteristic data information will be allowed to enter the unit state at each time step. The activation value of the forget gate determines which power generation fluctuation characteristic data information should be forgotten or retained from the unit state at each time step. The unit state activation value is used to update according to the activation values ​​of the input gate and the forget gate at each time step. The candidate unit activation value is a temporary value for the unit state update, which is calculated by the activation function of the current input and the hidden state of the previous time step. The output gate activation value determines the value of the next hidden state, which is based on the current unit state and the hidden state of the previous time step.

[0045] The above is a method for predicting different meteorological power generation methods based on a photovoltaic power generation prediction model provided in an embodiment of the present application. The following is a method for predicting different meteorological power generation systems based on a photovoltaic power generation prediction model provided in an embodiment of the present application.

[0046] A method for predicting different meteorological power generation systems based on a photovoltaic power generation prediction model includes: A data acquisition unit, used to acquire meteorological data information and photovoltaic power generation data information of the photovoltaic power station; The first processing unit is used to perform meteorological factor photovoltaic power generation correlation analysis on the photovoltaic power station meteorological data information and photovoltaic power generation data information based on the Spearman algorithm; The second processing unit is used to classify the meteorological type power generation conditions based on the meteorological factor photovoltaic power generation correlation analysis results based on the FCM fuzzy clustering algorithm; The third processing unit is used to classify the meteorological type power generation conditions based on the VDM variational mode decomposition algorithm, perform power generation subsequence decomposition processing for each condition and extract power generation fluctuation characteristics; A fourth processing unit is used to perform photovoltaic power generation prediction processing on the power generation fluctuation characteristics of the power generation subsequences of each meteorological type working condition based on the photovoltaic power generation prediction model; The model building unit is used to build a photovoltaic power generation prediction model.

[0047] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0048] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application 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 application.

Claims

1. A method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model, characterized in that: The following steps are involved: S101, performing meteorological factor photovoltaic power generation correlation analysis on the photovoltaic power station meteorological data information and photovoltaic power generation data information based on the Spearman algorithm; S102, classifying power generation conditions by meteorological type based on the results of meteorological factor photovoltaic power generation correlation analysis and processing based on a fuzzy clustering algorithm; S103, based on the variational mode decomposition algorithm, the meteorological type power generation condition classification processing results are processed, and the power generation subsequences of each condition are decomposed and the power generation fluctuation characteristics are extracted; S104, based on the photovoltaic power generation prediction model, the photovoltaic power generation prediction processing is performed on the power generation power fluctuation characteristics of the power generation power subsequences under each meteorological type working condition.

2. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 1, characterized in that: The method of performing meteorological factor photovoltaic power generation correlation analysis processing on the photovoltaic power station meteorological data information and photovoltaic power generation data information based on the Spearman algorithm includes the following steps: Preprocess the meteorological data information and photovoltaic power generation data information of the photovoltaic power station; Based on the Spearman algorithm, the correlation coefficient of the pre-processed photovoltaic power station meteorological data and photovoltaic power generation data is calculated and processed; Based on the results of correlation coefficient calculation, the correlation analysis of meteorological factors and photovoltaic power generation is carried out to select parameters with a greater correlation with power generation as characteristic parameters of the prediction model; the parameters with a greater correlation with power generation are total radiation, temperature, humidity, wind speed and backplane temperature.

3. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 2, characterized in that: The Spearman algorithm is used to calculate the correlation coefficient of the pre-processed photovoltaic power station meteorological data information and photovoltaic power generation data information, and the expression is as follows: in, X i is the magnitude of the influencing factors of the i-th sample point, including total radiation, temperature, humidity, atmospheric pressure, wind speed, wind direction, and backplane temperature. Y i is the power generation size of the i-th sample point, X and Y are the influencing factors and the power generation N The average value among the sample points.

4. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 1, characterized in that: The fuzzy clustering algorithm is used to analyze the correlation between meteorological factors and photovoltaic power generation, and the meteorological type power generation condition classification process is carried out, which includes the following steps: Taking irradiance as the clustering index, the results of the meteorological factor photovoltaic power generation correlation analysis are preliminarily divided into different meteorological types of power generation conditions; The membership degree and distance calculation are performed on the preliminary classification results of meteorological type power generation conditions and the cluster centers of different meteorological type power generation conditions; Based on the membership degree and distance calculation processing results, the meteorological type power generation conditions are reclassified to iteratively optimize the preliminary classification processing results of the meteorological type power generation conditions.

5. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 4, characterized in that: The meteorological type power generation conditions are reclassified based on the membership degree and distance calculation processing results to iteratively optimize the preliminary classification processing results of the meteorological type power generation conditions. The iterative process is as follows: 1) Optimize the objective function through iteration J ( U , V ) to minimize the weighted sum of square distances between data points and cluster centers. The objective function is: ； 2) In order to ensure that the sum of the membership of each data point is 1, the constraint function is: ； 3) Randomly initialize the membership matrix U and cluster center V, and update the membership matrix u according to the following expression j and cluster center u ij : ； 4) Repeat step 3) until the membership matrix U converges or the maximum number of iterations is reached; In the formula, u ij is the sample point x i With cluster center v j The membership degree of is in the range of 0-1; m is the fuzzy index, and m>1, which determines the fuzziness of the clustering, and its value range is 1.5-2.5; d ij is the sample point x i With cluster center v j The distance between n is the sample size; k is the number of cluster centers.

6. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 1, characterized in that: The method of classifying the meteorological type power generation conditions based on the variational mode decomposition algorithm, performing power generation subsequence decomposition processing for each condition and extracting power generation fluctuation characteristics comprises the following steps: The classification and processing results of power generation conditions of different meteorological types are sorted and organized in chronological order; Performing time series decomposition on the meteorological type power generation condition classification processing results that have been sorted in time series, and obtaining the time series decomposition processing results of the power generation subsequences of each condition; Through the time series decomposition processing results of the power generation subsequences under each operating condition, the central frequency of the IMF of the power generation subsequence segment under each operating condition after decomposition is determined, and the power generation fluctuation characteristics are extracted.

7. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 1, characterized in that: Constructing the photovoltaic power generation prediction model includes the following steps: Obtain historical meteorological data information and photovoltaic power generation data information of photovoltaic power stations, and classify and identify meteorological type power generation conditions; The classification and identification results of meteorological type power generation conditions are decomposed into power generation subsequences, and the power generation fluctuation characteristics are extracted; Based on the extracted power generation fluctuation characteristics, the initial photovoltaic power generation prediction model is trained; If the model training results are the same as the training sample results, the model training is completed; if the model training results are different from the training sample results, the model training is repeated.

8. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 1, characterized in that: The photovoltaic power generation prediction processing based on the photovoltaic power generation prediction model for the power generation power fluctuation characteristics of the power generation subsequences under each meteorological type working condition includes the following steps: Encoding and marking the power generation fluctuation characteristics of the power generation subsequences under each meteorological type of working condition; Performing feature screening processing on the coding mark processing results; Photovoltaic power generation prediction is performed based on the feature screening processing results.

9. The method for predicting power generation in different meteorological conditions based on a photovoltaic power generation prediction model according to claim 8, characterized in that: The feature screening process for the coding mark processing result comprises the following steps: Perform feature activation value calculation on the encoding and marking processing results; The data entry state of the encoding tag processing result is determined based on the feature activation value calculation processing result, and the data entry state includes an input gate state, a forget gate state, a unit state, a candidate unit state and an output gate state.

Images