Photovoltaic power prediction system and method

Abstract

The invention discloses a photovoltaic power prediction system and method, and belongs to the technical field of photovoltaic power prediction, and the system comprises a prediction terminal platform, a database, a data collection module, a data processing module, a data management module, a power prediction module, and an evaluation module. An interpolation method is adopted to fuse collected data, information from a plurality of different data sources is comprehensively processed to obtain more accurate, more comprehensive and more reliable data, and missing data values can be speculated according to known data points, so that a complete data set is supplemented, the continuity of the data is ensured, the photovoltaic power prediction effect is improved, and the photovoltaic power prediction efficiency is improved. The photovoltaic power prediction is evaluated, the influence of different factors on the photovoltaic power prediction is judged, the comprehensiveness and accuracy of the photovoltaic power prediction are improved, the appropriate model is selected for visual display, and the prediction result of the photovoltaic power prediction system is displayed more visually.

Description

Technical Field

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

[0002] Photovoltaic power is highly volatile, and its output power is affected by a variety of meteorological factors such as light intensity, temperature, and cloud cover. Accurate photovoltaic power forecasting can help the grid dispatching department prepare in advance and reasonably arrange the power generation plan of other power sources to ensure the balance of supply and demand of the grid and avoid fluctuations in grid frequency and voltage due to sudden changes in photovoltaic power. For operators of photovoltaic power stations, power forecasting helps optimize the operation and maintenance plan of the power station. The photovoltaic power forecasting system aims to accurately predict the power generation of photovoltaic power stations in the future based on meteorological data, historical photovoltaic power data and other information, which is of great significance for grid dispatching and optimizing energy management.

[0003] The existing photovoltaic power prediction system cannot integrate the collected data. The collected data may be missing, and the missing data value cannot be inferred based on the known data points. The data continuity is insufficient, which will affect the photovoltaic power prediction effect. The photovoltaic power prediction model is relatively simple, which will affect the comprehensiveness and accuracy of the photovoltaic power prediction. The prediction results of the photovoltaic power prediction system cannot be displayed intuitively.

[0004] In view of the above, this case proposes a photovoltaic power prediction system and method, which solves the above technical problems through a data processing module, a power prediction module and an evaluation module. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a photovoltaic power prediction system and method, which solves the above technical problems through a data processing module, a power prediction module and an evaluation module.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A photovoltaic power prediction system includes a prediction terminal platform, a database, a data acquisition module, a data processing module, a data management module, a power prediction module, and an evaluation module;

[0008] The prediction terminal platform is used to control the operation of the photovoltaic power prediction system;

[0009] The database is used to store photovoltaic power prediction system prediction related data;

[0010] The data acquisition module is used to collect and store real-time data required by the system;

[0011] The data processing module is used for the management and processing of raw data;

[0012] The data management module is used to query and manage the prediction data after data processing;

[0013] The power prediction module selects a sample set and optimized parameters for training and prediction based on the selected kernel function model and using the preprocessed data;

[0014] The evaluation module is used to evaluate and compare photovoltaic power prediction models.

[0015] Furthermore, the processing process of the data processing module is to clean the collected raw data, remove outliers and erroneous data, standardize the cleaned data, and fuse the data through interpolation.

[0016] Furthermore, the specific steps of the interpolation method are as follows: given a function y=f(x) at n+1 different points x0, x1, ..., x n , the function value at y0, y1, …, y n , use interpolation to find the polynomial P n (x), so that P n (x i )=y i , i = 0, 1, ..., n,, construct the basis function of the interpolation polynomial, for each i = 0, 1, ..., n, define the basis function:

[0017]

[0018] Interpolation polynomial P n (x) is:

[0019]

[0020] Get the function value and calculate the function value at the interpolation point.

[0021] Furthermore, the kernel function model selects a radial basis function, which is defined as a monotonic function of the Euclidean distance from any point x in space to the center point xc, denoted as:

[0022]

[0023] where x c Represents the center of the kernel function, σ represents the width parameter, and controls the radial range of the function.

[0024] Furthermore, the selection of the model parameters of the power prediction module adopts a grid search method, and cross-validation is required after the model parameters are selected.

[0025] Furthermore, the power prediction module performs photovoltaic power prediction by selecting a support vector machine regression method to establish a model, and the specific steps are as follows:

[0026] Step 1: predict a certain time t; the solar radiation intensity H at the predicted time G ; The power output value P at the same time in the previous five days with the same weather type as the forecast day h (P h1 , P h2 , P h3 , P h4 , P h5 ), the maximum temperature of the predicted day T max , minimum temperature T min and the average temperature T ave , wind speed S at the prediction time w ;

[0027] Step 2: predict the photovoltaic system power generation power P at the hour.

[0028] Furthermore, the solar radiation intensity H G Factors affecting the weather include the angle of incidence of sunlight, atmospheric transparency, atmospheric quality, latitude, altitude, slope and cloud cover.

[0029] Furthermore, the evaluation indicators of the evaluation module are root mean square error, mean absolute error and mean absolute percentage error. The fitting and prediction degree of the model are evaluated according to the evaluation indicators. The calculation formulas of each indicator are as follows:

[0030]

[0031]

[0032]

[0033] where Y t is a time series variable.

[0034] Furthermore, the evaluation result of the evaluation module is displayed through a visualization unit.

[0035] A prediction method for a photovoltaic power prediction system, the prediction method comprising the following steps:

[0036] S1. Install meteorological sensors in the photovoltaic power station to obtain real-time meteorological data. The data collection frequency is set according to the prediction accuracy requirements, and the real-time data required by the system is collected and stored;

[0037] S2. Clean the collected raw data, remove outliers and erroneous data, standardize the cleaned data, and fuse the data through interpolation;

[0038] S3, using the grid search method to select the model parameters of the power prediction module, and perform cross-validation, and using the support vector machine regression method to establish a model for photovoltaic power prediction;

[0039] S4. Evaluate the fitting and prediction degree of the model based on the comprehensive evaluation indicators of root mean square error, mean absolute error and mean absolute percentage error.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] 1. In the present invention, the interpolation method is used to fuse the collected data, and the information from multiple different data sources is comprehensively processed to obtain more accurate, comprehensive and reliable data. The missing data values ​​can be inferred based on the known data points, thereby supplementing the complete data set, ensuring the continuity of the data, and improving the photovoltaic power prediction effect;

[0042] 2. In the present invention, by evaluating the photovoltaic power prediction and judging the influence of different factors on the photovoltaic power prediction, the comprehensiveness and accuracy of the photovoltaic power prediction are improved, and the appropriate model is selected for visual display to more intuitively display the prediction results of the photovoltaic power prediction system. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 The present invention is a block diagram of a photovoltaic power prediction system and method. DETAILED DESCRIPTION

[0044] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] like Figure 1 As shown, a photovoltaic power prediction system includes a prediction terminal platform, a database, a data acquisition module, a data processing module, a data management module, a power prediction module, and an evaluation module;

[0046] The prediction terminal platform is used to control the operation of the photovoltaic power prediction system;

[0047] The database is used to store the prediction-related data of the photovoltaic power prediction system;

[0048] The data acquisition module is used to collect and store the real-time data required by the system;

[0049] The data processing module is used for the management and processing of raw data;

[0050] The data management module is used to query and manage the forecast data after data processing;

[0051] The power prediction module uses the preprocessed data according to the selected kernel function model, selects the sample set and the optimal parameters for training and prediction;

[0052] The evaluation module is used to evaluate and compare photovoltaic power prediction models.

[0053] It should be noted that the data information includes real-time monitoring data, data required for prediction and weather forecast data. The prediction results of the power prediction module are saved in the database. The data management module provides a manual correction function to adjust the erroneous data.

[0054] The processing process of the data processing module is to clean the collected raw data, remove outliers and erroneous data, standardize the cleaned data, and fuse the data through interpolation.

[0055] It should be noted that the cleaned data is standardized, for example, sensor data of different ranges are mapped to a unified numerical range; time series data are aligned and completed to ensure the integrity of the data in the time dimension.

[0056] The specific steps of the interpolation method are: given a function y = f(x) at n+1 different points x 0， x1,…,x n , the function value at y0,y1,…,y n , use interpolation to find the polynomial P n (x), so that P n (x i )=y i , i = 0, 1, ..., n,, construct the basis function of the interpolation polynomial, for each i = 0, 1, ..., n, define the basis function:

[0057]

[0058] Interpolation polynomial P n (x) is:

[0059]

[0060] Get the function value and calculate the function value at the interpolation point.

[0061] It should be noted that when n = 2, there are three points x0, x1, x2, then When x0=1, y0=2, x1=3, y1=4, x2=5, y2=6, first find the basis function but Simplify to get a polynomial, used to estimate the function value in the interval [1, 5], substitute the point x to be interpolated into P n (x), calculate the corresponding function value P n (x), for example, if we need to estimate the function value when x = 2, we can substitute x = 2 into P2(x) and obtain

[0062] The kernel function model selects the radial basis function, which is defined as a monotonic function of the Euclidean distance from any point x in space to the center point xc, denoted as:

[0063]

[0064] where x c Represents the center of the kernel function, σ represents the width parameter, and controls the radial range of the function.

[0065] It should be noted that k(x, x c ) in x c There is a unique maximum value at , as || xx c || increases k(x, x c ) decreases rapidly to 0 when x is far away from the center x c When k(x,x c ) becomes very small, only when k(x, x c ) has a corresponding adjustment effect when it is greater than a certain value.

[0066] The grid search method is used to select the model parameters of the power prediction module, and cross-validation is required after the model parameters are selected.

[0067] It should be noted that the advantage of using the grid search method is that a comprehensive parameter search can ensure that the vicinity of the global optimal parameters is found. When there are fewer parameters, the complexity of the grid search will not be much higher. Since each set of parameters is independent of each other, the grid search can be performed in parallel. Cross-validation is used to estimate the accuracy of the prediction model in solving practical problems.

[0068] The power prediction module uses the support vector machine regression method to establish a model for photovoltaic power prediction. The specific steps are as follows:

[0069] Step 1: predict a certain time t; the solar radiation intensity H at the predicted time G ; The power output value P at the same time in the previous five days with the same weather type as the forecast day h (P h1 , P h2 , P h3 , P h4 , P h5), the maximum temperature of the predicted day T max , minimum temperature T min and the average temperature T ave , wind speed S at the prediction time w ;

[0070] Step 2: predict the photovoltaic system power generation power P at the hour.

[0071] It should be noted that the solar radiation intensity H G The expression is H G =f(d, t, s), where d is the altitude of the sun, t is the day, and s is the hour. The expression for the power generation P of the photovoltaic system is P=I×A×μ, where I is the intensity of solar radiation on the slope of the photovoltaic panel, A is the area of ​​the photovoltaic cell, and u is the conversion efficiency of the photovoltaic cell.

[0072] Solar radiation intensity H G Factors affecting the weather include the angle of incidence of sunlight, atmospheric transparency, atmospheric quality, latitude, altitude, slope and cloud cover.

[0073] It should be noted that the greater the angle of incidence, the greater the total solar radiation. Low atmospheric transparency will reduce the solar radiation reaching the ground. The greater the mass, the less total solar radiation reaching the ground. The higher the latitude, the smaller the angle of incidence of sunlight, so the lower the total radiation. The higher the altitude, the stronger the direct solar radiation received by the ground. The south slope faces the sun and the north slope faces the shade, and it is more obvious in winter than in summer. The thicker the clouds and the larger the cloud cover, the weaker the direct solar radiation and the increased scattered radiation.

[0074] The evaluation indicators of the evaluation module are root mean square error, mean absolute error and mean absolute percentage error. The fitting and prediction degree of the model are evaluated according to the evaluation indicators. The calculation formulas of each indicator are as follows:

[0075]

[0076]

[0077]

[0078] where Y t is a time series variable.

[0079] It should be noted that through model screening, five exogenous variables, including average temperature, wind speed, precipitation, light intensity, and humidity, were selected to enter the model to obtain a new model. The new model can more accurately obtain the overall influence of each variable on the output power. Light intensity has the most significant impact, followed by temperature.

[0080] The evaluation results of the evaluation module are displayed through the visualization unit.

[0081] It should be noted that in order not to affect the user's comparison experience, the mobile zoom function is added so that users can choose the observation target at will to have an in-depth understanding of the performance of each model in different data intervals, select the appropriate model for visualization, and more intuitively display the prediction results of the photovoltaic power prediction system.

[0082] A prediction method for a photovoltaic power prediction system, the prediction method comprising the following steps:

[0083] S1. Install meteorological sensors in the photovoltaic power station to obtain real-time meteorological data. The data collection frequency is set according to the prediction accuracy requirements, and the real-time data required by the system is collected and stored;

[0084] S2. Clean the collected raw data, remove outliers and erroneous data, standardize the cleaned data, and fuse the data through interpolation;

[0085] S3, using the grid search method to select the model parameters of the power prediction module, and perform cross-validation, and using the support vector machine regression method to establish a model for photovoltaic power prediction;

[0086] S4. Evaluate the fitting and prediction degree of the model based on the comprehensive evaluation indicators of root mean square error, mean absolute error and mean absolute percentage error.

[0087] In the embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation; the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the method of this embodiment.

[0088] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A photovoltaic power prediction system, characterized in that: It includes prediction terminal platform, database, data acquisition module, data processing module, data management module, power prediction module and evaluation module; The prediction terminal platform is used to control the operation of the photovoltaic power prediction system; The database is used to store photovoltaic power prediction system prediction related data; The data acquisition module is used to collect and store real-time data required by the system; The data processing module is used for the management and processing of raw data; The data management module is used to query and manage the prediction data after data processing; The power prediction module selects a sample set and optimized parameters for training and prediction based on the selected kernel function model and using the preprocessed data; The evaluation module is used to evaluate and compare photovoltaic power prediction models.

2. A photovoltaic power prediction system according to claim 1, characterized in that: The processing process of the data processing module is to clean the collected raw data, remove outliers and erroneous data, standardize the cleaned data, and fuse the data through interpolation.

3. A photovoltaic power prediction system according to claim 2, characterized in that: The specific steps of the interpolation method are: given a function y=f(x) at n+1 different points x0, x1, ..., x n , the function value at y0,y1,…,y n , use interpolation to find the polynomial P n (x), so that P n (x i )=y i , i=0,1,…,n,, construct the basis function of the interpolation polynomial, and for each i=0,1,…,n, define the basis function: Interpolation polynomial P n (x) is: Get the function value and calculate the function value at the interpolation point.

4. A photovoltaic power prediction system according to claim 1, characterized in that: The kernel function model selects the radial basis function, which is defined as a monotonic function of the Euclidean distance from any point x in space to the center point xc, denoted as: where x c Represents the center of the kernel function, σ represents the width parameter, and controls the radial range of the function.

5. A photovoltaic power prediction system according to claim 1, characterized in that: The power prediction module model parameters are selected by a grid search method, and cross-validation is required after the model parameters are selected.

6. A photovoltaic power prediction system according to claim 4, characterized in that: The power prediction module performs photovoltaic power prediction by selecting a support vector machine regression method to establish a model, and the specific steps are: Step 1: predict a certain time t; the solar radiation intensity H at the predicted time G ; The power output value P at the same time in the previous five days with the same weather type as the forecast day h (P h1 ,P h2 ,P h3 ,P h4 ,P h5 ), the maximum temperature of the predicted day T max , minimum temperature T min and the average temperature T ave , wind speed S at the prediction time w ; Step 2: predict the photovoltaic system power generation power P at the hour.

7. A photovoltaic power prediction system according to claim 6, characterized in that: The solar radiation intensity H G Factors affecting the weather include the angle of incidence of sunlight, atmospheric transparency, atmospheric quality, latitude, altitude, slope and cloud cover.

8. A photovoltaic power prediction system according to claim 1, characterized in that: The evaluation indicators of the evaluation module are root mean square error, mean absolute error and mean absolute percentage error. The fitting and prediction degree of the model are evaluated according to the evaluation indicators. The calculation formulas of each indicator are as follows: where Y t is a time series variable.

9. A photovoltaic power prediction system according to claim 1, characterized in that: The evaluation result of the evaluation module is displayed through a visualization unit.

10. A prediction method of a photovoltaic power prediction system according to any one of claims 1 to 9, characterized in that: The prediction method includes the following steps: S1. Install meteorological sensors in the photovoltaic power station to obtain real-time meteorological data. The data collection frequency is set according to the prediction accuracy requirements, and the real-time data required by the system is collected and stored; S2. Clean the collected raw data, remove outliers and erroneous data, standardize the cleaned data, and fuse the data through interpolation; S3, using the grid search method to select the model parameters of the power prediction module, and perform cross-validation, and using the support vector machine regression method to establish a model for photovoltaic power prediction; S4. Evaluate the fitting and prediction degree of the model based on the comprehensive evaluation indicators of root mean square error, mean absolute error and mean absolute percentage error.

Images