Photovoltaic power generation prediction method and system based on coupling model error correction

Abstract

The invention discloses a photovoltaic power generation prediction method and system based on coupling model error correction, and the method comprises the steps: carrying out the climate data observation and preprocessing of a to-be-researched target region, and obtaining a climate model data set; performing selection processing on the climate model data set according to a preset selection condition to obtain a screened climate model data set; performing error correction processing on the screened climate model data set through a Bayesian average model and a Delta method to obtain a corrected climate model data set; and performing photovoltaic power generation prediction on the corrected climate model data set in combination with the photovoltaic cell efficiency to obtain a photovoltaic power generation prediction result. According to the invention, the climate data simulation error and the simulation difference of climate mode data on different areas can be reduced, so that the accuracy of photovoltaic power generation prediction is improved. The photovoltaic power generation prediction method and system based on coupling model error correction can be widely applied to the technical field of photoelectric data prediction.

Description

Technical Field

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

[0002] With the increasing demand for renewable energy and the concern for environmental sustainability, renewable energy has developed rapidly. Therefore, making good use of photovoltaic power generation technology will be an effective way to reduce carbon emissions and reduce dependence on traditional fossil fuels. Climate change and photovoltaic cell materials have an important impact on solar power generation. However, in the current GCM model, due to the different regional divisions and modes of regional climate simulation in GCM models in different regions, there are large differences in simulation effects in different regions and large deviations in climate simulation data, resulting in inaccurate prediction results for photovoltaic power generation. Summary of the invention

[0003] In order to solve the above technical problems, the purpose of the present invention is to provide a photovoltaic power generation prediction method and system based on coupled model error correction, which can reduce the simulation error of climate data and the difference in simulation of climate model data for different regions, thereby improving the accuracy of photovoltaic power generation prediction.

[0004] The first technical solution adopted by the present invention is: a photovoltaic power generation prediction method based on coupling model error correction, comprising the following steps:

[0005] Observe and preprocess climate data for the target area to be studied, and obtain climate model data sets;

[0006] Selecting and processing the climate model data set according to preset selection conditions to obtain a screened climate model data set;

[0007] The error correction process of the selected climate model data set is carried out by using the Bayesian mean model and the delta method to obtain a corrected climate model data set;

[0008] The photovoltaic power generation prediction is carried out on the modified climate model data set in combination with the photovoltaic cell efficiency to obtain the photovoltaic power generation prediction result.

[0009] Furthermore, the step of observing and preprocessing the climate data of the target area to be studied to obtain the climate model data set specifically includes:

[0010] Identify the target area to be studied;

[0011] The target area to be studied is divided according to the longitude and latitude division conditions to obtain the divided target area;

[0012] Obtain climate simulation data for the divided target areas through a global climate model;

[0013] Obtain the climate values ​​of the divided target area and average them to obtain the observation data of the divided target area;

[0014] The climate simulation data of the divided target area is combined with the observation data of the divided target area and a bilinear difference supplementation process is performed to obtain a climate model data set.

[0015] Furthermore, the step of screening the climate model data set according to the preset selection conditions to obtain the screened climate model data set specifically includes:

[0016] The Pearson correlation coefficient, the centralized root mean square error and the standard deviation are calculated for the climate model data set, respectively, to obtain the Pearson correlation coefficient value R of the climate model data set, the centralized root mean square error value RMSE of the climate model data set and the standard deviation value SD of the climate model data set;

[0017] Constructing a preset selection condition, wherein the preset selection condition includes a first preset condition and a second preset condition, wherein the first preset condition is R>85%, RMSE<40, SD∈[50,60], and the second preset condition is R>98%, RMSE<4, SD∈[10,12];

[0018] According to the preset selection conditions, the climate model data set is evaluated for downward shortwave radiation and near-surface temperature to obtain evaluation results;

[0019] The evaluation results are visualized and screened through Taylor diagrams to obtain a filtered climate model data set.

[0020] Furthermore, the step of performing error correction processing on the screened climate model data set by using the Bayesian mean model and the delta method to obtain a corrected climate model data set specifically includes:

[0021] Performing average calculation processing on the screened climate model data group based on the Bayesian average model to obtain the average value of the climate model data group;

[0022] determining a posterior probability for a climate model data set based on the mean of the climate model data set;

[0023] According to the posterior probability of the climate model data group, the delta method is introduced to perform periodic division processing on the climate model data group to obtain the divided climate model data group;

[0024] The divided climate model data set is subjected to error correction processing to obtain a corrected climate model data set.

[0025] Furthermore, the expression for error correction processing of the filtered climate model data set by using the Bayesian mean model and the delta method is specifically as follows:

[0026]

[0027]

[0028] In the above formula, represents the data set after the error correction by the Delta method divided by month, represents a simulated data set, y emp represents the observed data, x emp represents the simulated dataset used for evaluation, represents the simulated data set after being processed by the Bayesian averaging method, represents the simulated data set used for evaluation after being processed by the Bayesian averaging method, Represents the average value of the data during the observation period, represents the posterior probability of the selected model, n1 represents the number of values ​​of observation data for each month, n2 represents the number of values ​​of observation data, n3 represents the number of selected climate models, x represents the data of a time node, m represents the time divided by month, and t represents the simulation time.

[0029] Furthermore, the expression for predicting photovoltaic power generation based on the modified climate model data set in combination with photovoltaic cell efficiency is specifically as follows:

[0030] P pv =R RSDS ×E P

[0031] E P =E Tref [1-C ref (T P -25)+γlog 10 R RSDS ]

[0032]

[0033] In the above formula, P pv represents the photovoltaic potential, R RSDS represents the surface solar radiation, E P Represents the photovoltaic cell power generation efficiency, T a Indicates the ambient surface temperature, T P Indicates the battery temperature, T NT Indicates the battery temperature in NOCT mode, E Tref represents the reference photovoltaic cell efficiency, C refrepresents the temperature coefficient and γ represents the radiation coefficient.

[0034] The second technical solution adopted by the present invention is: a photovoltaic power generation prediction system based on coupling model error correction, comprising:

[0035] The first module is used to observe and preprocess climate data in the target area to be studied and obtain the climate model data set;

[0036] The second module is used to select and process the climate model data set according to the preset selection conditions to obtain a screened climate model data set;

[0037] The third module is used to perform error correction processing on the screened climate model data set by using the Bayesian average model and the delta method to obtain a corrected climate model data set;

[0038] The fourth module is used to predict photovoltaic power generation based on the revised climate model data set in combination with photovoltaic cell efficiency to obtain photovoltaic power generation prediction results.

[0039] The beneficial effects of the method and system of the present invention are as follows: the present invention observes and preprocesses the climate data of the target area to be studied, further selects and processes the climate model data group according to preset selection conditions, obtains a screened climate model data group, selects high-quality input data, and further performs error correction processing on the screened climate model data group through the Bayesian average model and the delta method, thereby reducing the impact of climate data simulation errors and climate model data on simulation differences in different regions, and finally combines the efficiency of photovoltaic cells to predict photovoltaic power generation for the corrected climate model data group, thereby improving the accuracy of photovoltaic power generation prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flow chart of the steps of a photovoltaic power generation prediction method based on coupling model error correction of the present invention;

[0041] Figure 2 It is a structural block diagram of a photovoltaic power generation prediction system based on coupling model error correction of the present invention;

[0042] Figure 3 It is a schematic diagram of the framework of the steps of photovoltaic power generation prediction provided by a specific embodiment of the present invention;

[0043] Figure 4 is a schematic diagram of photovoltaic power generation prediction provided by a specific embodiment of the present invention;

[0044] Figure 5 It is a Taylor diagram of RSDS and TAS in a certain area provided by a specific embodiment of the present invention. DETAILED DESCRIPTION

[0045] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only provided for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0046] Reference Figure 1 The present invention provides a photovoltaic power generation prediction method based on coupling model error correction, the method comprising the following steps:

[0047] S100, observing and preprocessing climate data in the target area to be studied, and obtaining a climate model data set;

[0048] Specifically, the target area to be studied is determined; the target area to be studied is divided according to the longitude and latitude division conditions to obtain the divided target area; the climate simulation data of the divided target area is obtained through the global climate model; the climate values ​​of the divided target area are obtained and averaged to obtain the observed data of the divided target area; the climate simulation data of the divided target area is combined with the observed data of the divided target area and a bilinear difference supplementation process is performed to obtain a climate model data set.

[0049] In this embodiment, the research object (or region) is first selected for regional division, and the continuous GCM model data and observation data are screened according to the regional longitude and latitude. As shown in Table 1, the present invention selects CMIP6 (the sixth international coupled model comparison program) data of five target regions as the basis of GCM simulation data (about 3-5 climate models per state), and divides them by longitude and latitude, traverses the CMIP6 climate model values ​​in each provincial region according to the longitude and latitude range, and averages the regional values, and uses the average data to represent the historical gas phase data of the continuity of the simulation and observation of the research object (or region), as shown in Table 1. Figure 4 a. The meteorological data involved include the monthly average RSDS and TAS. After analysis, it was found that the resolution of the GCM climate model is greater than 1.25°×1.875°, and there will be no collected values ​​in the area with a longitude range of less than 2.5° and a latitude range of less than 1.66°. The data can be supplemented by the bilinear difference method.

[0050] Table 1 CMIP6 climate model information for target regions

[0051] Mode Name Country / Region R&D institution abbreviation Mode split ratio ACCESS-CM2 A CSIRO-ARCCSS 1.25°×1.875° ACCESS-ESM1.5 A CSIRO 1.25°×1.875° BCC-CSM2-MR B BCC 1.12°×1.12° CAMS-CSM1.0 B CAMS 1.12°×1.12° CanESM5 C CCC 2.77°×2.81° CAS-ESM2.0 B IAP 1.12°×1.12° CESM2 D NCAR 0.9°×1.25° CESM2-WACCM D NCAR 0.9°×1.25° FGOALS-f3-L B CAS 1°×1.25° GFDL-ESM4 D NOAA-GFDL 1°×1.25° INM-CM4.8 E INM 1.5°×2° INM-CM5.0 E INM 1.5°×2° IPSL-CM6A-LR F IPSL 1.27°×2.5° KACE-1.0-G G NIMS-KMA 1.25°×1.875° MIROC6 H MIROC 1.4°×1.4° MRI-ESM2.0 H MRI 1.12°×1.12°

[0052] S200, selecting and processing the climate model data set according to preset selection conditions to obtain a screened climate model data set;

[0053] Specifically, the Pearson correlation coefficient, the centered root mean square error and the standard deviation are calculated for the climate model data group, respectively, to obtain the Pearson correlation coefficient value R of the climate model data group, the centered root mean square error value RMSE of the climate model data group and the standard deviation value SD of the climate model data group; a preset selection condition is constructed, and the preset selection condition includes a first preset condition and a second preset condition, the first preset condition is R>85%, RMSE<40, SDF∈[50,60], and the second preset condition is R>98%, RMSE<4, SD∈[10,12]; according to the preset selection conditions, the climate model data group is evaluated for downward shortwave radiation from the surface and near-surface temperature to obtain an evaluation result; the evaluation result is visualized and screened through a Taylor diagram to obtain a screened climate model data group.

[0054] In this embodiment, the Pearson correlation coefficient (R), the central root mean square error (RMSE) and the standard deviation (SD) of the selected different GCM models and the observation data in each region are calculated to evaluate the model for the surface downward shortwave radiation (RSDS) and the near-surface air temperature (TAS), and the GCM model evaluation results of each region are displayed through the Taylor diagram, and the model array set with better model evaluation is selected, such as Figure 4 b. The selection criteria are as follows: the R of RSDS is greater than 85%, and the RMSE is less than 40W m -2 and SD between 50 and 60W m -2 6-10 model arrays with TAS R greater than 98%, RMSE less than 4℃ and SD between 10 and 12℃; based on the cosine relationship principle of the Taylor diagram, the first 35%-50% model array sets with the shortest distance in the Taylor diagram are selected.

[0055] S300, performing error correction processing on the screened climate model data set by using a Bayesian average model and a delta method to obtain a corrected climate model data set;

[0056] Specifically, based on the Bayesian average model, the average value of the filtered climate model data group is calculated to obtain the average value of the climate model data group; the posterior probability of the climate model data group is determined according to the average value of the climate model data group; according to the posterior probability of the climate model data group, the delta method is introduced to perform periodic division processing on the climate model data group to obtain the divided climate model data group; the divided climate model data group is error corrected to obtain the corrected climate model data group.

[0057] In this embodiment, the error correction of the BMA and Delta model coupling is performed based on the selected optimal model evaluation GCM array. The present invention assigns weights to each selected GCM climate model by the BMA model, and performs weighted averaging, and then divides it into 12-month periodicities by the Delta method to finally determine the predicted value, such as Figure 4 d. According to the RSDS and TAS climate data, the calculation is done independently according to the regional division, and the expression is:

[0058]

[0059] In the above formula, represents the data set after the error correction by the Delta method divided by month, represents a simulated data set, y emp represents the observed data, x emp represents the simulated dataset used for evaluation, represents the simulated data set after being processed by the Bayesian averaging method, represents the simulated data set used for evaluation after being processed by the Bayesian averaging method, Represents the average value of the data during the observation period, represents the posterior probability of the selected model, n1 represents the number of values ​​of observation data for each month, n2 represents the number of values ​​of observation data, n3 represents the number of selected climate models, x represents the data of a time node, m represents the time divided by month, and t represents the simulation time.

[0060] In summary, the number of models selected for RSDS and TAS is determined. First, the average data of RSDS and TAS of different GCM models in the observation data period is calculated. Then the posterior probability of the selected model is calculated. After calculating the posterior probability of the RSDS model and the TAS model, the climate simulation data sets in the observation data period and the simulation data period are further calculated respectively. Finally, it is divided into monthly cycles, and finally the historical gas phase data of the continuity of simulation and observation of the research object (or region) after error correction is calculated.

[0061] We further selected 16 GCM models for experiments in a certain region. After error correction, the average R of RSDS increased by 11.1% and the RMSE decreased by 21.44 W m -2 , SD decreased by 1.42W m -2 , the average R of TAS increased by 1.5%, RMSE decreased by 1.88℃, and SD decreased by 0.7℃. Figure 5 Shown

[0062] S400, performing photovoltaic power generation prediction on the modified climate model data set in combination with photovoltaic cell efficiency to obtain a photovoltaic power generation prediction result.

[0063] In this embodiment, based on the error-corrected TAS and RSDS, the PV power generation is calculated and predicted by a physical method based on TAS and RSDS. The main input data include RSDS, TAS and PV cell efficiency, such as Figure 4 c. The photovoltaic power generation prediction formula is as follows:

[0064] P pv =R RSDS ×E P

[0065] E P =E Tref [1-C ref (T P -25)+γlog 10 R RSDS ]

[0066]

[0067] In the above formula, P pv represents the photovoltaic potential, R RSDS represents the surface solar radiation, E P Represents the photovoltaic cell power generation efficiency, T a Indicates the ambient surface temperature, T P Indicates the battery temperature, T NT Indicates the battery temperature in NOCT mode, usually defined at 800Wm -2 Model temperature under solar radiation and wind speed of 1m / s, E Tref Represents the reference photovoltaic cell efficiency (at a temperature of 25°C and a solar irradiance of 1000Wm -2 Conditions), C ref represents the temperature coefficient, and γ represents the radiation coefficient, which is determined by the performance of the photovoltaic cell material.

[0068] In the present invention, the changes in the reference efficiency of photovoltaic cells are divided, including the selection of cell efficiency in 2024-2029 to determine that the average efficiency of the three major categories of photovoltaic cells is about 25%, the average efficiency of photovoltaic cells in 2030-2059 is 50%, and the average efficiency of photovoltaic cells in 2060-2099 reaches 75%.

[0069] Finally, 16 GCM models were selected for experimental verification of a certain country, which was divided into 32 provincial regions. It was concluded that if the installed photovoltaic power generation area of ​​a certain country reaches 15,000 square kilometers in the future, the photovoltaic cell conversion rate will reach 50%.

[0070] In summary, reference Figure 3The embodiment of the present invention first selects 3-5 global climate model (GCM) data as simulation data according to the target area division, selects the National Tibetan Plateau Science Data Center (TPDC) data corresponding to more than 20 years as observation data, performs RSDS and TAS climate data analysis, and divides the acquired data into regions according to administrative regions such as provinces, autonomous regions and municipalities. Secondly, the Pearson correlation coefficient (R), centralized root mean square error (RMSE) and standard deviation (SD) of different GCM models in each region are calculated to evaluate the model, and the better climate model is selected according to the following standards, where the RSDS R is greater than 85%, and the RMSE is less than 40W m -2 and SD between 50 and 60W m -2 6-10 model arrays with TAS R greater than 98%, RMSE less than 4℃ and SD between 10 and 12℃ were selected, and the first 35%-50% model array sets with the shortest distance in the Taylor diagram were selected through the cosine relationship principle of the Taylor diagram. Then, the Delta and BMA (Bayesian average model) coupling models divided by monthly cycles were used to correct the errors of GCM simulation data, so as to reduce the simulation errors of climate data and reduce the impact of GCM on simulation differences in different regions. Finally, a new method for predicting photovoltaic power generation potential was established by combining the physical photovoltaic power generation calculation based on RSDS and TAS.

[0071] Reference Figure 2 , a photovoltaic power generation prediction system based on coupling model error correction, comprising:

[0072] The first module 201 is used to observe and preprocess the climate data of the target area to be studied, and obtain a climate model data set;

[0073] The second module 202 is used to select and process the climate model data set according to the preset selection conditions to obtain a filtered climate model data set;

[0074] The third module 203 is used to perform error correction processing on the filtered climate model data set by using the Bayesian mean model and the delta method to obtain a corrected climate model data set;

[0075] The fourth module 204 is used to perform photovoltaic power generation prediction on the modified climate model data set in combination with the photovoltaic cell efficiency to obtain a photovoltaic power generation prediction result.

[0076] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0077] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A photovoltaic power generation prediction method based on coupling model error correction, characterized in that: The following steps are involved: Observe and preprocess climate data for the target area to be studied, and obtain climate model data sets; Selecting and processing the climate model data set according to preset selection conditions to obtain a screened climate model data set; The error correction process of the screened climate model data set is carried out by using the Bayesian mean model and the delta method to obtain a corrected climate model data set; The photovoltaic power generation prediction is carried out on the modified climate model data set in combination with the photovoltaic cell efficiency to obtain the photovoltaic power generation prediction result.

2. A photovoltaic power generation prediction method based on coupling model error correction according to claim 1, characterized in that: The step of observing and preprocessing the climate data of the target area to be studied to obtain the climate model data set specifically includes: Identify the target area to be studied; The target area to be studied is divided according to the longitude and latitude division conditions to obtain the divided target area; Obtain climate simulation data for the divided target areas through a global climate model; Obtain the climate values ​​of the divided target area and average them to obtain the observation data of the divided target area; The climate simulation data of the divided target area is combined with the observation data of the divided target area and a bilinear difference supplementation process is performed to obtain a climate model data set.

3. A photovoltaic power generation prediction method based on coupling model error correction according to claim 2, characterized in that: The step of screening the climate model data set according to the preset selection conditions to obtain the screened climate model data set specifically includes: The Pearson correlation coefficient, the centralized root mean square error and the standard deviation are calculated for the climate model data set, respectively, to obtain the Pearson correlation coefficient value R of the climate model data set, the centralized root mean square error value RMSE of the climate model data set and the standard deviation value SD of the climate model data set; Constructing a preset selection condition, wherein the preset selection condition includes a first preset condition and a second preset condition, wherein the first preset condition is R>85%, RMSE<40, SD∈[50,60], and the second preset condition is R>98%, RMSE<4, SD∈[10,12]; According to the preset selection conditions, the climate model data set is evaluated for downward shortwave radiation and near-surface temperature to obtain evaluation results; The evaluation results are visualized and screened through Taylor diagrams to obtain a filtered climate model data set.

4. The photovoltaic power generation prediction method based on coupling model error correction according to claim 3 is characterized in that: The step of performing error correction processing on the screened climate model data set by using the Bayesian mean model and the delta method to obtain a corrected climate model data set specifically includes: Performing average calculation processing on the filtered climate model data group based on the Bayesian average model to obtain the average value of the climate model data group; determining a posterior probability for a climate model data set based on the mean of the climate model data set; According to the posterior probability of the climate model data group, the delta method is introduced to perform periodic division processing on the climate model data group to obtain the divided climate model data group; The divided climate model data set is subjected to error correction processing to obtain a corrected climate model data set.

5. A photovoltaic power generation prediction method based on coupling model error correction according to claim 4, characterized in that: The expression for error correction processing of the filtered climate model data set by using the Bayesian mean model and the delta method is specifically as follows: In the above formula, represents the data set after the error correction by the Delta method divided by month, represents a simulated data set, y emp represents the observed data, x emp represents the simulated dataset used for evaluation, represents the simulated data set after being processed by the Bayesian averaging method, represents the simulated data set used for evaluation after being processed by the Bayesian averaging method, Represents the average value of the data during the observation period, represents the posterior probability of the selected model, n1 represents the number of values ​​of observation data for each month, n2 represents the number of values ​​of observation data, n3 represents the number of selected climate models, x represents the data of a time node, m represents the time divided by month, and t represents the simulation time.

6. A photovoltaic power generation prediction method based on coupling model error correction according to claim 5, characterized in that: The expression for predicting photovoltaic power generation based on the modified climate model data set in combination with photovoltaic cell efficiency is specifically as follows: P pv =R RSDS ×E P E P =E Tref [1-C ref (T P -25)+γlog 10 R RSDS ] In the above formula, P pv represents the photovoltaic potential, R RSDS represents the surface solar radiation, E P Represents the photovoltaic cell power generation efficiency, T a Indicates the ambient surface temperature, T P Indicates the battery temperature, T NT Indicates the battery temperature in NOCT mode, E Tref represents the reference photovoltaic cell efficiency, C ref represents the temperature coefficient and γ represents the radiation coefficient.

7. A photovoltaic power generation prediction system based on coupling model error correction, characterized in that: Includes the following modules: The first module is used to observe and preprocess climate data in the target area to be studied and obtain the climate model data set; The second module is used to select and process the climate model data set according to the preset selection conditions to obtain a screened climate model data set; The third module is used to perform error correction processing on the screened climate model data set by using the Bayesian average model and the delta method to obtain a corrected climate model data set; The fourth module is used to predict photovoltaic power generation based on the revised climate model data set in combination with photovoltaic cell efficiency to obtain photovoltaic power generation prediction results.

Images