Photovoltaic power generation power prediction method, system and device based on improved SVR, and medium

Abstract

The invention provides a photovoltaic power generation power prediction method, system and device based on improved SVR and a medium, and the method comprises the following steps: collecting historical data of photovoltaic power generation power, and carrying out the preprocessing; performing feature screening on the preprocessed data to obtain a historical data key feature set; performing prediction based on the historical data key feature set in the SVR model to obtain an initial prediction value, and performing time granularity mean residual feedback on the initial prediction value to obtain correction feedback; on the basis of an improved sparrow search algorithm and correction feedback, the SVR model is optimized, and a TGRF-ISSA-SVR prediction model is obtained; performing photovoltaic power generation power prediction based on the TGRF-ISSA-SVR prediction model to obtain an actual photovoltaic power generation power prediction value; according to the method, the operation speed and precision of the algorithm are improved, the use of clean energy by the building is facilitated, and the technical problems of complex photovoltaic power generation efficiency prediction process and low efficiency in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of photovoltaic power generation system power generation prediction, and in particular to a photovoltaic power generation power prediction method, system, equipment and medium based on improved SVR. Background Art

[0002] The rapid development of science and technology has increased people's energy demand, but the problem of insufficient resource supply is becoming more and more serious. Clean energy has become a rapidly developing field and plays an increasingly important role in the energy industry. The best solution to the energy problem depends on the possible combination of renewable energy generation and load, which can be predicted using historical statistical data or empirical inference. Model predictive control has become a well-known control technology for dealing with energy problems. Accurate prediction of photovoltaic power generation can guide the optimal energy management strategy, so the prediction of building power generation is considered to be an important area of ​​future research.

[0003] Prediction methods are mainly divided into three categories: physical models or white box methods, including EnergyPlus, eQuest, Ecoect, etc., whose heavy reliance on physical mechanisms makes this method time-consuming, labor-intensive and complex in practical applications; statistical models or gray box methods, such as temperature frequency method, gradient method, etc., but gray box methods have no advantages in accuracy and universality; data-driven models or black box methods, including neural networks, support vector machines (SVM), decision trees (DT), regression models, etc. Data-driven methods have been widely used in energy prediction in recent years due to the continuous improvement of data operation and maintenance systems and the versatility, flexibility and high accuracy of their algorithms. Some deep learning methods in machine learning can simultaneously mine the feature information and time series information of data, but the training time is too long, more data is required, and the structure is complex and time-consuming. Summary of the invention

[0004] In order to solve the problems existing in the prior art, the present invention provides a photovoltaic power prediction method, system, device and medium based on improved SVR, which are used to solve the technical problems existing in the prior art that the photovoltaic power generation efficiency prediction process is complicated and inefficient.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A photovoltaic power generation power prediction method based on improved SVR includes the following steps:

[0007] Collect historical data of photovoltaic power generation and perform preprocessing;

[0008] Performing feature screening on the preprocessed data to obtain a key feature set of historical data;

[0009] In the SVR model, prediction is performed based on the key feature set of historical data to obtain the initial prediction value, and the time-granular mean residual feedback is performed on the initial prediction value to obtain the correction feedback;

[0010] Based on the improved sparrow search algorithm and the correction feedback, the SVR model is optimized to obtain the TGRF-ISSA-SVR prediction model;

[0011] Photovoltaic power generation is predicted based on the TGRF-ISSA-SVR prediction model to obtain an actual photovoltaic power generation prediction value.

[0012] Furthermore, the preprocessing process is:

[0013] Based on the sunshine time data of the photovoltaic location, the data of effective photovoltaic power generation within the preset sunshine time is screened from the collected historical data of photovoltaic power generation and the photovoltaic power generation data to be predicted.

[0014] Furthermore, the process of feature screening the preprocessed data and the key feature set of historical data is as follows:

[0015] The random number forest method is used to screen the features of the preprocessed data and the key feature set of historical data;

[0016] The feature screening is factors that affect photovoltaic power generation, and the factors that affect photovoltaic power generation include cloud coverage, temperature and wind speed.

[0017] Furthermore, the process of performing prediction based on the key feature set of historical data in the SVR model to obtain an initial prediction value, performing time-granular mean residual feedback on the initial prediction value, and obtaining a correction feedback is as follows:

[0018] The key feature set of historical data is divided into a training set TD and a prediction set PD.

[0019] Calculate the time granularity average residual of the training set TD, and obtain the residual feature filling value of the prediction set PD based on the time granularity average residual of the training set TD;

[0020] The correlation coefficient between the residual feature filling value and the actual photovoltaic power prediction value based on the prediction set PD;

[0021] Based on the correlation coefficient, the time granularity mean residuals significantly related to the output are retained, and the time granularity mean residuals irrelevant to the output are deleted to obtain the correction feedback.

[0022] Furthermore, the TGRF-ISSA-SVR prediction model uses Latin hypercube to initialize the population, and the process of Latin hypercube initializing the population is:

[0023] Sampling n points in a two-dimensional n×n square matrix, wherein the n points do not overlap and are at a maximum distance;

[0024] The n points are defined as points (c i ,g i )∈{1,2,…,n} 2 , (i=1,2,…,n);

[0025] where c i ≠c j ,g i ≠g j ,i≠j;

[0026] In an n×n square, the points are determined column by column, and only one point can be specified in each column or row.

[0027] Furthermore, in the communication mechanism of the TGRF-ISSA-SVR prediction model, the follower position in the sparrow search algorithm is updated as follows:

[0028]

[0029] in, is the worst global position, It is the best place to find sparrows. is a random discoverer, A is a 1×d matrix in which each element is randomly 1 or -1, s is the number of sparrows, and k is the ratio of random follower sparrows.

[0030] Furthermore, the sparrow ratio dynamic control of the TGRF-ISSA-SVR prediction model is:

[0031]

[0032] Among them, FP max is the maximum proportion of discoverers, FP min is the minimum proportion of discoverers, t is the current number of iterations, It m is the maximum number of iterations.

[0033] A photovoltaic power generation prediction system based on improved SVR, including

[0034] The preprocessing unit is configured as follows:

[0035] Used to collect historical data of photovoltaic power generation and perform preprocessing;

[0036] The screening unit is configured to:

[0037] Used to perform feature screening on the preprocessed data to obtain a key feature set of historical data;

[0038] The correction unit is configured as follows:

[0039] It is used to make predictions based on the key feature set of historical data in the SVR model to obtain initial prediction values, and to perform time-granular mean residual feedback on the initial prediction values ​​to obtain correction feedback;

[0040] The optimization unit is configured as:

[0041] Used to optimize the SVR model based on the improved sparrow search algorithm and the correction feedback to obtain the TGRF-ISSA-SVR prediction model;

[0042] Output unit, configured as:

[0043] It is used to predict photovoltaic power generation based on the TGRF-ISSA-SVR prediction model to obtain the actual photovoltaic power generation prediction value.

[0044] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the photovoltaic power prediction method based on improved SVR are implemented.

[0045] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the photovoltaic power prediction method based on improved SVR are implemented.

[0046] Compared with the prior art, the present invention has the following beneficial technical effects:

[0047] The invention provides a photovoltaic power prediction method, system, device and medium based on improved SVR, comprising the following steps: collecting historical data of photovoltaic power and performing preprocessing; performing feature screening on the preprocessed data to obtain a key feature set of historical data; performing prediction based on the key feature set of historical data in an SVR model to obtain an initial prediction value, performing time granularity mean residual feedback on the initial prediction value to obtain a correction feedback; optimizing the SVR model based on an improved sparrow search algorithm and the correction feedback to obtain a TGRF-ISSA-SVR prediction model; performing photovoltaic power prediction based on the TGRF-ISSA-SVR prediction model to obtain an actual photovoltaic power prediction value; the method improves the operation speed and accuracy of the algorithm, and is experimentally verified to be able to more accurately reflect the photovoltaic power size of buildings in solar-rich areas in actual applications, thereby facilitating the use of clean energy by buildings, and solving the technical problems of a complex and low-efficiency photovoltaic power generation efficiency prediction process in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flow chart of a photovoltaic power generation prediction method based on improved SVR according to an embodiment of the present disclosure is shown;

[0049] Figure 2 A schematic diagram showing an improved communication mechanism in an improved sparrow search algorithm according to an embodiment of the present disclosure is shown;

[0050] Figure 3 A feature screening diagram of the random number forest method according to an embodiment of the present disclosure is shown;

[0051] FIG. 4 shows a comparison diagram of photovoltaic power generation prediction results according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0052] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0053] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0054] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0055] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0056] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0057] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0058] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0059] Figure 1 A flow chart of a photovoltaic power generation prediction method based on improved SVR according to an embodiment of the present disclosure is shown. Figure 1 As shown, the following steps are included:

[0060] Step S1: Collect historical data of photovoltaic power generation and perform preprocessing;

[0061] Preferably, in the embodiment of the present disclosure, the preprocessing process is:

[0062] Based on the sunshine time data of the photovoltaic location, the data of effective photovoltaic power generation within the preset sunshine time is screened from the collected historical data of photovoltaic power generation and the photovoltaic power generation data to be predicted.

[0063] Specifically, in some embodiments, the preset time refers to the sunrise time and sunset time of the area where the photovoltaic module is located.

[0064] Step S2: performing feature screening on the preprocessed data to obtain a key feature set of historical data;

[0065] Preferably, in the disclosed embodiment, the process of performing feature screening on the preprocessed data, the key feature set of historical data and the key feature set of the data to be predicted is:

[0066] The random number forest method is used to screen the features of the preprocessed data, the key feature set of historical data and the key feature set of the data to be predicted;

[0067] The feature screening is factors that affect photovoltaic power generation, and the factors that affect photovoltaic power generation include cloud coverage, temperature and wind speed.

[0068] Specifically, the cloud cover as an influencing factor specifically refers to: clouds will block part of the sunlight, reducing the light intensity reaching the photovoltaic panels, thereby affecting the power generation. Although photovoltaic panels will generate heat under light, excessively high temperatures will reduce the conversion efficiency of photovoltaic cells, because the voltage of photovoltaic cells decreases with increasing temperature. Although the current may increase slightly, the overall output power will decrease. Although the direct impact of wind speed is small, strong winds may bring pollutants such as sand and dust, affecting the cleanliness of photovoltaic panels and indirectly affecting power generation.

[0069] To be more specific, since photovoltaic power generation is significantly affected by weather, different weather types may lead to different importance of various features to photovoltaic power generation. The features correctly constructed by the feature extraction method can reduce the calculation time of the data-driven model. Therefore, it is necessary to screen features separately for different weather types.

[0070] Photovoltaic data has strong temporal characteristics, so the relevant factors of each influencing factor one moment, two moments, and three moments before it are generated as input features, and the importance is calculated by the random forest algorithm, such as Figure 3 As shown, the redundant feature variables are eliminated through recursive elimination, and the results are shown in Table 1.

[0071] Table 1 Feature screening table

[0072]

[0073]

[0074] According to the results of the analysis in Table 1, all the feature names and their corresponding symbols in the table, as well as the feature variables that need to be retained for the four weather types, are shown in the table with 1 for retention and 0 for deletion.

[0075] Step S3: In the SVR model, prediction is performed based on the key feature set of historical data to obtain an initial prediction value, and time-granular mean residual feedback is performed on the initial prediction value to obtain correction feedback;

[0076] Preferably, in the disclosed embodiment, the process of performing prediction based on the key feature set of historical data in the SVR model to obtain an initial prediction value, performing time granularity mean residual feedback on the initial prediction value, and obtaining a correction feedback is as follows:

[0077] The key feature set of historical data is divided into a training set TD and a prediction set PD.

[0078] Calculate the time granularity average residual of the training set TD, and obtain the residual feature filling value of the prediction set PD based on the time granularity average residual of the training set TD;

[0079] The correlation coefficient between the residual feature filling value and the actual photovoltaic power prediction value based on the prediction set PD;

[0080] Based on the correlation coefficient, the time granularity mean residuals significantly related to the output are retained, and the time granularity mean residuals irrelevant to the output are deleted to obtain the correction feedback.

[0081] Specifically, the process of calculating the time-granular average residual of the training set TD is:

[0082] e in =y′-y

[0083] Among them, e in is the residual, y=(y 1 …y n ) is the actual value of the sample in the training set TD, y′=(y 1 ′…y n ′) is the sample prediction value of the training set TD.

[0084]

[0085]

[0086] Among them, N h , t hour.i = the number of samples and the mean residual of h; N d , t day.i = the number of samples and the mean residual error when d; N m , t month.i = m. t (e hour 、e day 、emonth ) T is the time granularity mean residual.

[0087] The process of obtaining the residual feature filling value of the prediction set PD based on the time granularity average residual of the training set TD is:

[0088]

[0089] in, is the residual feature of the test set, is the daily residual feature of the test set, is the monthly residual feature of the training set. The mean residuals padded for the test set.

[0090] The process of calculating the correlation coefficient between the residual feature filling value based on the prediction set PD and the actual photovoltaic power generation power prediction value is as follows:

[0091]

[0092] Among them, r is the correlation coefficient, E i and Y i is the time granularity mean residual and the actual load value, and is the average of the time granularity mean residual and the actual load value, and n is the number of samples.

[0093] The process of retaining the time granularity mean residual significantly related to the output based on the correlation coefficient and deleting the time granularity mean residual irrelevant to the output to obtain the corrected feedback is:

[0094]

[0095] Among them, K q (k 1 ,k 2 ,k 3 ) is e hour 、e day 、e month Correlation with the actual value of power prediction, when Pearson correlation occurs, k=r=1, where k and r are both correlation coefficients, and k=0 when Pearson is irrelevant. t.q , is the time granularity mean residual of TD and PD fed back for the qth time.

[0096] Output J and E t.q ,

[0097]

[0098] Among them, when If at least one of them is significantly correlated with the output, J = 1, and the residual feedback is continued to be called. Otherwise, the residual feedback is no longer called.

[0099] E t.q Incorporate X' E As the training set input, Incorporate X' E As the test set input, the training set and test set are used as correction feedback.

[0100] Step S4: Based on the improved sparrow search algorithm and the correction feedback, the SVR model is optimized to obtain the TGRF-ISSA-SVR prediction model;

[0101] Preferably, in the disclosed embodiment, the TGRF-ISSA-SVR prediction model uses Latin hypercube to initialize the population, and the process of Latin hypercube initializing the population is:

[0102] Sampling n points in a two-dimensional n×n square matrix, wherein the n points do not overlap and are at a maximum distance;

[0103] The n points are defined as points (c i ,g i )∈{1,2,…,n} 2 , (i=1,2,…,n);

[0104] where c i ≠c j ,g i ≠g j ,i≠j;

[0105] In an n×n square, the points are determined column by column, and only one point can be specified in each column or row.

[0106] Furthermore, in the communication mechanism of the TGRF-ISSA-SVR prediction model, the follower position in the sparrow search algorithm is updated as follows:

[0107]

[0108] in, is the worst global position, It is the best place to find sparrows. is a random discoverer, A is a 1×d matrix in which each element is randomly 1 or -1, s is the number of sparrows, and k is the ratio of random follower sparrows.

[0109] It should be noted that the sparrow search algorithm includes a finder, a follower, and a warning value. The embodiment of the present disclosure only improves the follower, and the position update of the finder and the warning value is consistent with the original update method, for example:

[0110] The finder location is updated to:

[0111]

[0112] Among them, t is the current iteration number, I iter.max is the maximum number of iterations, is the position of the i-th sparrow before the update. Q is a random number that follows a normal distribution. L is an all-one matrix. R 2 is the warning value, ST is the safety threshold, and its range is [0.5, 1].

[0113] The Sentinel position is updated to:

[0114]

[0115] in, is the optimal global position of the tth iteration, β is the step size control parameter, which is a random number that follows a standard normal distribution, K∈[-1,1], f g 、f i and f w They represent the global optimal, fitness and worst fitness of the current sparrow respectively. ∈ is a very small constant used to avoid the denominator appearing 0 in the formula, which is generally set to 10e -8 .

[0116] Further, if Figure 2 As shown, the dynamic control of the sparrow ratio of the TGRF-ISSA-SVR prediction model is:

[0117]

[0118] Among them, FP max is the maximum proportion of discoverers, FP min is the minimum value of the proportion of discoverers, t is the current number of iterations, It m is the maximum number of iterations.

[0119] Step S5: Predicting photovoltaic power generation based on the TGRF-ISSA-SVR prediction model to obtain a predicted value of actual photovoltaic power generation.

[0120] According to the screening results of the random forest algorithm, several important variables under different weather types, such as the radiation in the previous three moments and the power value at the previous moment, are obtained as the input of the TGRF-ISSA-SVR model, and the corresponding photovoltaic power generation value at the next moment is output.

[0121] Finally, the present invention was experimentally verified by collecting real data of photovoltaic power generation of a building in Tibet's solar energy-enriched area from January 1, 2023 to December 31, 2023, indicating that the present application can accurately predict the changes in the photovoltaic power generation power of the building in practical applications, which is helpful to achieve efficient utilization of clean energy, thereby providing effective data support and theoretical reference for the energy-saving operation of the building.

[0122] According to the weather characteristics of the area, the number of training data samples used is the historical data set of the building every 15 minutes from 6 am to 10 pm every day in 2023, totaling 34,849 samples. The data mainly includes component temperature (℃), temperature (℃), air pressure (hPa), humidity (%), radiation (W / m 2 The final evaluation index is the determination coefficient (R 2 ), root mean square error (RMSE) and mean absolute error (MAE), the formulas are as follows:

[0123]

[0124] Where N′ is the total number of sample data sets, y′ i is the actual value, is y′ i The average value of is the corresponding predicted value.

[0125] Compared with other models and the SVR model improved by other algorithms, its related evaluation indicators are shown in Table 2.

[0126] Table 2 Comparison of model prediction accuracy

[0127]

[0128]

[0129] The comparison of the photovoltaic power point prediction results using TGRF-ISSA-SVR with other models such as LSTM and GRNN is shown in Figure 4, wherein Figure 4(a) shows a comparison chart of the photovoltaic power prediction values ​​using the present application and existing related technical means under sunny weather conditions, Figure 4(b) shows a comparison chart of the photovoltaic power prediction values ​​using the present application and existing related technical means under cloudy weather conditions, Figure 4(c) shows a comparison chart of the photovoltaic power prediction values ​​using the present application and existing related technical means under sudden weather conditions, and Figure 4(d) shows a comparison chart of the photovoltaic power prediction values ​​using the present application and existing related technical means under rainy and snowy weather conditions.

[0130] The horizontal axis is the sampling point every 15 minutes, and the vertical axis is the building photovoltaic power generation value. The black one is the actual building photovoltaic power generation value, and the red one is the building photovoltaic power generation value predicted by the model.

[0131] The overall upward and downward trend of the prediction value curve of this application can accurately reflect the real load curve. The combined model TGRF-ISSA-SVR in the prediction model has an average RMSE reduction of 53.235% compared to the TGRF-SVR model and an average reduction of 34.89% compared to the TGRF-SSA-SVR model, indicating that the improvement measures are effective, and TGRF-ISSA-SVR has higher accuracy and lower error than other models, indicating that the model has more advantages in practical applications.

[0132] The embodiment of the present disclosure also provides a photovoltaic power generation prediction system based on improved SVR, including a preprocessing unit, which is configured to:

[0133] Used to collect historical data of photovoltaic power generation and photovoltaic power generation data to be predicted, and perform preprocessing;

[0134] The screening unit is configured to:

[0135] Used to perform feature screening on the preprocessed data to obtain a key feature set of historical data and a key feature set of data to be predicted;

[0136] The correction unit is configured as follows:

[0137] It is used to make predictions based on the key feature set of historical data in the SVR model to obtain initial prediction values, and to perform time-granular mean residual feedback on the initial prediction values ​​to obtain correction feedback;

[0138] The optimization unit is configured as:

[0139] Used to optimize the SVR model based on the improved sparrow search algorithm and the correction feedback to obtain the TGRF-ISSA-SVR prediction model;

[0140] Output unit, configured as:

[0141] It is used to predict photovoltaic power generation based on the TGRF-ISSA-SVR prediction model and the key feature set of the data to be predicted to obtain the final prediction value.

[0142] In another embodiment of the present invention, a computer device is provided, the computer device comprising a processor and a memory, the memory being used to store a computer program, the computer program comprising program instructions, and the processor being used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions in a computer storage medium to implement a corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used for the operation of the photovoltaic power generation power prediction method based on the improved SVR.

[0143] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the photovoltaic power prediction method based on the improved SVR described in the above embodiment.

[0144] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0146] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0147] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0148] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

[0149] In addition, it should be understood that although this specification is described in accordance with the implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation modes that can be understood by those skilled in the art. The above content is only to illustrate the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A photovoltaic power prediction method based on improved SVR, It is characterized in that The following steps are involved: Collect historical data of photovoltaic power generation and perform preprocessing; Performing feature screening on the preprocessed data to obtain a key feature set of historical data; In the SVR model, prediction is performed based on the key feature set of historical data to obtain the initial prediction value, and the time-granular mean residual feedback is performed on the initial prediction value to obtain the correction feedback; Based on the improved sparrow search algorithm and the correction feedback, the SVR model is optimized to obtain the TGRF-ISSA-SVR prediction model; Photovoltaic power generation is predicted based on the TGRF-ISSA-SVR prediction model to obtain an actual photovoltaic power generation prediction value.

2. The photovoltaic power generation prediction method based on improved SVR according to claim 1, It is characterized in that The pre-processing process is: Based on the sunshine time data of the photovoltaic location, the data of effective photovoltaic power generation within the preset sunshine time is screened from the collected historical data of photovoltaic power generation and the photovoltaic power generation data to be predicted.

3. The photovoltaic power generation prediction method based on improved SVR according to claim 1, It is characterized in that The process of feature screening the preprocessed data and the key feature set of historical data is as follows: The random number forest method is used to screen the features of the preprocessed data and the key feature set of historical data; The feature screening is factors that affect photovoltaic power generation, and the factors that affect photovoltaic power generation include cloud coverage, temperature and wind speed.

4. The photovoltaic power generation prediction method based on improved SVR according to claim 1, It is characterized in that The process of making predictions based on the key feature set of historical data in the SVR model to obtain initial prediction values, performing time-granular mean residual feedback on the initial prediction values, and obtaining correction feedback is as follows: The key feature set of historical data is divided into a training set TD and a prediction set PD. Calculate the time granularity average residual of the training set TD, and obtain the residual feature filling value of the prediction set PD based on the time granularity average residual of the training set TD; The correlation coefficient between the residual feature filling value and the actual photovoltaic power prediction value based on the prediction set PD; Based on the correlation coefficient, the time granularity mean residuals significantly related to the output are retained, and the time granularity mean residuals irrelevant to the output are deleted to obtain the correction feedback.

5. The photovoltaic power generation prediction method based on improved SVR according to claim 1, It is characterized in that The TGRF-ISSA-SVR prediction model uses Latin hypercube to initialize the population. The process of Latin hypercube initialization of the population is: Sampling n points in a two-dimensional n×n square matrix, wherein the n points do not overlap and are at a maximum distance; The n points are defined as points (c i ,g i )∈{1,2,…,n} 2 , (i=1,2,…,n); where c i ≠c j ,g i ≠g j ,i≠j; In an n×n square, the points are determined column by column, and only one point can be specified in each column or row.

6. The photovoltaic power generation prediction method based on improved SVR according to claim 1, It is characterized in that In the communication mechanism of the TGRF-ISSA-SVR prediction model, the follower position in the sparrow search algorithm is updated as follows: in, is the worst global position, It is the best place to find sparrows. is a random discoverer, A is a 1×d matrix in which each element is randomly 1 or -1, s is the number of sparrows, and k is the ratio of random follower sparrows.

7. The photovoltaic power generation prediction method based on improved SVR according to claim 1, It is characterized in that The dynamic control of sparrow proportion of the TGRF-ISSA-SVR prediction model is: Among them, FP max is the maximum proportion of discoverers, FP min is the minimum value of the proportion of discoverers, t is the current number of iterations, It m is the maximum number of iterations.

8. A photovoltaic power generation prediction system based on improved SVR, It is characterized in that The photovoltaic power generation power prediction method based on the improved SVR according to any one of claims 1 to 7 comprises: The preprocessing unit is configured as follows: Used to collect historical data of photovoltaic power generation and perform preprocessing; The screening unit is configured to: Used to perform feature screening on the preprocessed data to obtain a key feature set of historical data; The correction unit is configured as follows: It is used to make predictions based on the key feature set of historical data in the SVR model to obtain initial prediction values, and to perform time-granular mean residual feedback on the initial prediction values ​​to obtain correction feedback; The optimization unit is configured as: Used to optimize the SVR model based on the improved sparrow search algorithm and the correction feedback to obtain the TGRF-ISSA-SVR prediction model; Output unit, configured as: It is used to predict photovoltaic power generation based on the TGRF-ISSA-SVR prediction model to obtain the actual photovoltaic power generation prediction value.

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

10. A computer-readable storage medium storing a computer program. It is characterized in that When the computer program is executed by a processor, the steps of the photovoltaic power prediction method based on improved SVR as described in any one of claims 1 to 7 are implemented.

Images