CN111596384A - A Slope Radiation Prediction Method Based on Effective Weather Type Identification - Google Patents

Abstract

The invention discloses an inclined plane radiation prediction method based on weather type effective identification. The method of the invention comprises the following steps: s1: acquiring historical data of weather and radiation; s2: calculating a direct incidence ratio and a corrected definition index, and calculating a weather type index SCF and a solar altitude; s3: judging whether the solar altitude is greater than 10 degrees, if so, entering the next step, and otherwise, returning to the previous step; s4: classifying weather, namely classifying the weather type index SCF by using a K-means clustering algorithm; s5: acquiring actually measured data, calculating a weather type index SCF and judging the weather type of the index SCF; s6: and selecting different direct and scattered separation models according to different weather types. According to the invention, the accuracy reduction phenomenon of a single model under a specific weather type is avoided through calculation, and the most suitable weather condition of each model is selected, so that the accuracy of the prediction model is obviously improved.

Description

A Slope Radiation Prediction Method Based on Effective Weather Type Identification

technical field

The invention relates to the field of photovoltaic systems, in particular to a slope radiation prediction method based on weather type identification effectively.

Background technique

The main source of greenhouse gases is traditional fossil fuels, and the development and application of renewable energy provides a feasible solution for reducing greenhouse gas emissions. Among them, solar energy resources have great potential for development, and thus receive more and more attention. At present, photovoltaic power generation technology is the most mature in solar energy resource development technology. According to the 2019-2023 Photovoltaic Review and Outlook Report released by the European Photovoltaic Industry Association, the new photovoltaic installed capacity in 2020 will achieve a growth of 26%, reaching 20-21GW; in 2021, the new photovoltaic installed capacity will reach 21.9GW, close to the annual photovoltaic installed capacity. The all-time high value of capacity. According to statistics from the National Energy Administration, as of the end of 2019, the total installed capacity of photovoltaic power generation in the country was about 190 million kWh, and the national photovoltaic power generation in the first three quarters of 2019 was 171.5 billion kWh, a year-on-year increase of 28%.

The total incident radiation directly determines the output power of the photovoltaic array. The total incident radiation is mainly composed of three parts, namely direct radiation, scattered radiation and reflected radiation. In order to maximize the received energy, PV arrays in the northern hemisphere are usually placed at an inclination to the south (optimal inclination). The radiation data measured in the weather station is generally on the horizontal plane, which requires the model to convert the radiation data from the horizontal plane to the inclined plane.

Literature: S Armstrong, W G Hurley. A new methodology to optimise solarenergy extraction under cloudy conditions [J]. Renewable Energy, 2010, 35: 780-787. In view of the fact that the scattered radiation accounts for a large proportion under cloudy conditions, it introduces two parameters, the percentage of sunshine (ie the ratio of sunshine hours to the hours of sunshine) and the total cloud amount. Through the calculation of historical data, different weather conditions can be obtained. The average proportion of the type to calculate the radiation dose, but the calculated value is lower than the actual value.

Literature: Demain C, Journée M, Bertrand C, et al.Evaluation of different models to estimate the global solar radiation on inclined surfaces[J].Renewable Energy,2013,50:710-721. It introduces the concept of the revised sharpness index, which reduces the influence of the sun's altitude on the sharpness index. By dividing the weather by subsections of the modified clarity index, the slope radiation model with the highest accuracy in the subsection is selected, and the combined prediction is carried out. The results show that the accuracy of the combined model is significantly improved compared with the single model.

Literature: Li Fen, Hu Chao, Ma Nianjun, et al. Study on the direct-scatter separation model considering PM2.5 under different weather types [J]. Chinese Journal of Solar Energy, 2017, 38(12): 3339-3347. It combines the three parameters of PM2.5, clarity index and sunshine percentage to establish a BP neural network model for predicting direct-scatter separation in Beijing, and tests the optimal parameter combination under different weather conditions.

Literature: Li F, Li C, Shi J, et al. An evaluation index system for photovoltaic systems statistical characteristics under hazy weatherconditions in central China [J]. IET Renewable Power Generation, 2017, 11(14): 1794-1803. Its comparison Five slope radiation models were studied, and the Liu&Jordan model was found to be the most suitable for the Wuhan area.

In my country's meteorological industry standards, ground meteorological observation specifications and operational meteorological forecasts, a single index, such as total cloud cover or sunshine hours, is usually used to divide the weather. However, the observation of total cloud cover mainly relies on manual observation, and there is human error, and using a single indicator, the accuracy of weather type classification is lacking.

The existing solar radiation models are mainly proposed by foreigners, and the data for establishing the models are mainly from the United States and Europe. These models may have large errors when used directly in my country.

Different slope radiation models are suitable for different weather types. Some models are more accurate under cloudy conditions, while others are the opposite. If a single model is used, it will inevitably bring about a reduction in accuracy.

SUMMARY OF THE INVENTION

The present invention aims to overcome the above-mentioned defects of the prior art. Based on the radiation data of the horizontal plane and the inclined plane and the meteorological environment data, combined with the meteorological environment factors such as the correction sharpness index, the direct sunlight ratio and the total cloud amount, a comprehensive weather type index is proposed. According to the index, the weather types are effectively divided into five categories, and the corresponding optimal models are analyzed for different weather types, so as to propose a high-precision slope radiation prediction method based on the effective identification of weather types.

To this end, the present invention adopts the following technical scheme: a method for predicting radiation on inclined planes based on the effective identification of weather types, which includes the steps:

S1: Obtain historical data of meteorology and radiation;

S2: Calculate the direct sunlight ratio and the corrected clarity index, calculate the weather type index SCF and the sun altitude;

S3: Determine whether the sun altitude angle is greater than 10 degrees, if the sun altitude angle is greater than 10 degrees, go to the next step, otherwise return to the previous step;

S4: classify the weather, that is, use the K-means clustering algorithm to classify the weather type index SCF;

S5: Obtain the measured data, calculate the weather type index SCF and determine the weather type to which it belongs;

S6: According to different weather types, choose different direct-scatter separation models.

Further, in step S1, the radiation data includes the total radiation on the horizontal plane, the scattered radiation, the direct radiation, the reflected radiation, and the total radiation data on the inclined plane facing due south and the inclination angle is the Beijing latitude; the meteorological data includes the total cloud amount. and visibility; total cloud cover is defined as the percentage of the sky field that is obscured by clouds, indicating the percentage of the area covered by cloud cover in the total sky area.

Further, in step S2, the data are screened and quality controlled, and the sharpness index and the corrected sharpness index are calculated according to the observed values of radiation on the horizontal plane and astronomical factors (such as latitude, solar altitude, etc.).

Further, in step S2, the clarity index is the ratio of the total solar radiation I on the horizontal plane to the solar radiation I _o on the horizontal plane outside the atmosphere, namely:

The sharpness index is not only related to the meteorological conditions, but also to the position of the sun in the sky. In order to reduce the influence of the sun altitude on the sharpness index, the correction is as follows:

where k' _T is the corrected clarity index and m is the air mass.

The Clarity Index can be used to characterize the attenuation of solar radiation by the atmosphere and is a preferred weather type classification indicator. The larger the clarity index, the higher the transparency of the atmosphere, the less the atmosphere attenuates the solar radiation, and the greater the solar radiation reaching the ground.

Further, in step S2, the three data of the revised clarity index, the direct ratio and the total cloud cover are used to normalize and weight respectively to obtain a comprehensive index factor, which is named SCF (sky condition factor). , the weather type index, the specific calculation formula is as follows:

SCF=w ₁ Bd+w ₂ k' _T +w ₃ (1-C),

Among them, the sum of the weights w ₁ , w ₂ and w ₃ is 1, and the specific value is calculated according to the local geographical location and radiation data; C represents the total cloud cover; Bd represents the direct radiation ratio, that is, the direct radiation radiation on the horizontal plane percentage of the radiation exposure.

Further, in step S4, the weather type index SCF is divided into 5 categories by K-means clustering algorithm: type I, 0.58≤SCF<1; type II, 0.44≤SCF<0.58; type III, SCF value 0.3≤SCF< 0.44; Type IV, 0.15≤SCF<0.3; Type V, 0<SCF<0.15, the richness of the direct radiation component of its solar radiation decreases sequentially.

Weather type I has the best conditions, which is sunny; weather type II is the second, which is sunny to cloudy; weather type III is sunny to cloudy; weather type IV mainly includes cloudy, cloudy to cloudy, cloudy to cloudy, etc.; weather type V is bad Weather, including light rain, showers, light snow, light fog, haze, moderate rain and above, or moderate snow and above, etc.

Further, the specific content of step S6 is:

S61) if the current weather type obtained in step S5 is Type I, select the Hay model to calculate;

S62) if the current weather type obtained in step S5 is type II, III or IV, select the Perez model to calculate;

S63) If it is obtained in step S5 that the current weather type is V type, Liu & Jordan model is selected for calculation.

Further, in step S1, the time period of the historical data is two years.

The beneficial effects of the present invention are as follows: 1) Weather types can be accurately and effectively classified. Compared with the existing method of using total cloud cover alone or using the combination of clarity index and cloud cover, it is more accurate.

2) The calculation formula of the weather type index is simple, easy to quantify, the calculation amount is small, and the discrimination is easy.

3) Compared with the single model, the comprehensive model avoids the reduction of the accuracy of the single model under specific weather types through calculation, and selects the most suitable weather conditions for each model, thereby significantly improving the accuracy of the prediction model.

The method of the present invention is verified by using the observation data in the Beijing area, which shows that the method is more suitable for the geographical and climatic characteristics of our country.

Description of drawings

Fig. 1 is the contrast diagram of each model under weather type I of the present invention;

Fig. 2 is the contrast diagram of each model under weather type II of the present invention;

Fig. 3 is the contrast diagram of each model under the weather type III of the present invention;

Fig. 4 is the contrast diagram of each model under weather type IV of the present invention;

FIG. 5 is a comparison diagram of each model under the weather type V of the present invention.

Detailed ways

In order to make those skilled in the art better understand the present invention, the present invention will be further described below with reference to specific embodiments, but the protection scope of the present invention is not limited to the following examples. Any modifications and changes made to the present invention within the spirit of the present invention and the protection scope of the claims fall into the protection scope of the present invention.

The present embodiment provides a slope radiation prediction method based on effective identification of weather types, which includes the steps:

S1: Obtain historical data of meteorology and radiation, and the time period of historical data is two years.

S2: Calculate the direct sunlight ratio and the corrected clarity index, calculate the weather type index SCF and the sun altitude;

S3: Determine whether the sun altitude angle is greater than 10 degrees, if the sun altitude angle is greater than 10 degrees, go to the next step, otherwise return to the previous step;

S4: classify the weather, that is, use the K-means clustering algorithm to classify the weather type index SCF;

S5: Obtain the measured data, calculate the weather type index SCF and determine the weather type to which it belongs;

S6: According to different weather types, choose different direct-scatter separation models.

In step S1, the radiation data includes horizontal plane total radiation, scattered radiation, direct radiation, and reflected radiation data; the meteorological data includes total cloud cover and visibility; total cloud cover is defined as the percentage of clouds covering the sky field of view, and represents The percentage of the sky covered by cloud cover as a percentage of the total sky extent.

In step S2, the data is screened and quality controlled, and the sharpness index and the corrected sharpness index are calculated according to the observed values of radiation on the horizontal plane and astronomical factors (such as latitude, solar altitude, etc.).

In step S2, the clarity index is the ratio of the total solar radiation I on the horizontal plane to the solar radiation I _o on the horizontal plane outside the atmosphere, namely:

The sharpness index is not only related to the meteorological conditions, but also to the position of the sun in the sky. In order to reduce the influence of the sun altitude on the sharpness index, the correction is as follows:

where k' _T is the corrected clarity index and m is the air mass.

In step S2, the three data of the corrected clarity index, the direct ratio and the total cloud cover are used to normalize and weight respectively to obtain a comprehensive index factor, which is named SCF, that is, the weather type index. The calculation formula is as follows:

SCF=w ₁ Bd+w ₂ k' _T +w ₃ (1-C),

Among them, the sum of the weights w ₁ , w ₂ and w ₃ is 1, and the specific value is calculated according to the local geographic location and radiation data; C represents the total cloud cover; Bd represents the direct sunlight ratio.

In step S4, the weather type index SCF is divided into 5 categories by K-means clustering algorithm: type I, 0.58≤SCF<1; type II, 0.44≤SCF<0.58; type III, SCF value 0.3≤SCF<0.44; IV Type V, 0.15≤SCF<0.3; Type V, 0<SCF<0.15, the richness of its direct solar radiation component decreases sequentially.

The specific content of step S6 is:

S61) if the current weather type obtained in step S5 is Type I, select the Hay model to calculate;

S62) if the current weather type obtained in step S5 is type II, III or IV, select the Perez model to calculate;

S63) If it is obtained in step S5 that the current weather type is V type, Liu & Jordan model is selected for calculation.

Comparison of several representative models:

Several representative models are selected in the present invention, namely the isotropic Liu&Jordan model, the anisotropic Temps&Clouson model, the Perez model, the Kulcher model, the Hay model, and the Reindl model. The calculation, analysis and verification are carried out through the measured data of the total radiation on the slope of the southern suburbs of Beijing facing due south and the inclination angle is equal to the latitude of Beijing. Under the weather type I, the calculated values of the isotropic and most anisotropic models are lower than the measured values, and the Hay model is the closest. This is mainly because the radiation component of the weather type I is mainly direct radiation; in the weather type II , III and IV, the error between the Perez model and the measured value is the smallest, and the calculated values of other models are low; in weather type V, the calculated values of the anisotropic model are slightly larger than the measured values, and only the isotropic Liu&Jordan model It is closest to the measured value, which is mainly due to the poor weather conditions in weather type V, and the solar radiation is mainly scattered radiation, which is evenly distributed in the sky and tends to be isotropic.

The comparison of each model under different weather conditions is shown in the figure below.

Among them, the standard deviation, also known as the mean square error, is the square root of the arithmetic mean of the square of the deviation between the sample and its mean; the coefficient of variation is the ratio of the standard deviation of the original data to the mean of the original data, which can be used to compare the degree of dispersion of the two groups of data; The square root of the square root of the deviation between the predicted value and the true value and the ratio of the number of observations n (or the number of samples) is often used to measure the deviation between the observed value and the true value; the mean absolute error is the arithmetic mean of all individual observations and the The average of the absolute value of the deviation can avoid the problem that the errors cancel each other, so it can accurately reflect the size of the actual prediction error.

Claims (8)

1. A method for predicting inclined plane radiation based on weather type effective identification is characterized by comprising the following steps:

s1: acquiring historical data of weather and radiation;

s2: calculating a direct incidence ratio and a corrected definition index, and calculating a weather type index SCF and a solar altitude;

s3: judging whether the solar altitude is greater than 10 degrees, if so, entering the next step, and otherwise, returning to the previous step;

s4: classifying weather, namely classifying the weather type index SCF by using a K-means clustering algorithm;

s5: acquiring actually measured data, calculating a weather type index SCF and judging the weather type of the index SCF;

s6: and selecting different direct and scattered separation models according to different weather types.

2. The method for predicting inclined plane radiation based on weather type effective recognition of claim 1, wherein in step S1, the radiation data comprises horizontal plane total radiation, scattered radiation, direct radiation and reflected radiation data; the meteorological data comprises total cloud cover and visibility; the total cloud cover is defined as the number of cloud-covered sky views, and represents the percentage of the range covered by the cloud cover in the sky to the total sky range.

3. The method for predicting the inclined plane radiation based on the effective weather type identification as claimed in claim 1, wherein in step S2, the data are screened and quality controlled, and the clarity index and the corrected clarity index are calculated according to the horizontal plane radiation observation value and the astronomical geographic factor.

4. The method for predicting solar radiation on an inclined plane based on weather type effective recognition of claim 3, wherein in step S2, the definition indexes are total solar radiation I on a horizontal plane and solar radiation I on a horizontal plane outside the atmosphere_oThe ratio of the components is as follows:

the clarity index is not only related to meteorological conditions but also to the position of the sun in the sky, and is modified as follows in order to reduce the influence of the solar altitude on the clarity index:

wherein, k'_TIs the sharpness index after correction, and m is the atmospheric mass.

5. The method for predicting the inclined plane radiation based on the effective weather type identification as claimed in claim 4, wherein in step S2, the corrected three data of the sharpness index, the direct incidence ratio and the total cloud amount are respectively normalized and weighted to obtain a comprehensive index factor, which is named as SCF, namely the weather type index, and the specific calculation formula is as follows:

SCF＝w₁Bd+w₂k'_T+w₃(1-C)，

wherein the weight w₁，w₂And w₃The sum of (1) is calculated according to the local geographic position and the radiation data; c represents total cloud cover; bd represents the direct ratio.

6. The method for predicting the inclined plane radiation based on the effective identification of the weather types as claimed in claim 1, wherein in the step S4, the weather type index SCF is classified into 5 categories by a K-means clustering algorithm: type I, SCF <1 > is more than or equal to 0.58; type II, SCF is more than or equal to 0.44 and less than 0.58; type III, SCF value is more than or equal to 0.3 and less than 0.44; IV type, SCF is more than or equal to 0.15 and less than 0.3; form v, 0< SCF <0.15, with successively less richness of the direct radiation component of solar radiation.

7. The method for predicting the radiation of the inclined plane based on the effective identification of the weather types as claimed in claim 6, wherein the specific content of the step S6 is as follows:

s61) if the current weather type obtained in the step S5 is I type, selecting a Hay model for calculation;

s62) if the current weather type obtained in the step S5 is II, III or IV, selecting a Perez model for calculation;

s63) if the current weather type obtained in the step S5 is V type, selecting a Liu & Jordan model for calculation.

8. The method for predicting the radiation of the inclined plane based on the effective identification of the weather type as claimed in claim 1, wherein the time of the historical data is two years in step S1.

Images