Solar Irradiance Prediction Method Considering Cloud Physical Properties

Abstract

This invention relates to a solar irradiance prediction method that considers cloud physical parameters, belonging to the field of solar photovoltaic power generation output prediction. It addresses the problems of long prediction times and slow response speeds in existing methods that rely on cloud information. The invention focuses on measuring cloud physical parameters and predicting the solar irradiance incident on the photovoltaic panel using an atmospheric radiative transfer model. During the prediction process, the cloud physical parameters input into the atmospheric radiative transfer model are corrected in real time based on measured cloud and irradiance data, improving prediction accuracy. The method is time-efficient and has a fast response speed. This invention is primarily used for predicting solar irradiance.

Description

Technical Field

[0001] This invention belongs to the field of solar photovoltaic power generation output prediction. Background Technology

[0002] my country's installed capacity of green energy sources such as solar power is rapidly increasing, and the solar energy industry is developing rapidly. However, the intermittency, volatility, and randomness of solar energy are the main factors restricting its large-scale grid connection. Accurately predicting changes in photovoltaic power is one of the effective means to reduce the adverse effects of its instability on the integrated energy system and to achieve precise scheduling and intelligent management of the integrated energy system. Many studies have shown that the output power of photovoltaic power generation is linearly related to the solar irradiance received by photovoltaic modules. Therefore, predicting the surface solar irradiance is an important prerequisite for ensuring accurate prediction of photovoltaic power generation.

[0003] Currently, solar irradiance prediction methods mainly include data-driven methods and numerical weather prediction methods. Data-driven surface solar irradiance prediction models establish a mapping relationship based on statistical meteorological and historical surface irradiance data to predict surface irradiance. However, this method lacks a description of the underlying mechanisms, typically relies on a large number of training samples, and is generally only applicable to specific environments. Numerical weather prediction models need to comprehensively consider various physical processes such as atmospheric dynamics, thermodynamics, radiative transfer, cloud microphysics, and land surface processes, as well as the influence of factors such as Earth's rotation and atmospheric turbulence. This method can provide relatively accurate weather prediction results, but solving the equations involves a large amount of computation, a long solution process, and high computational costs. While numerical weather prediction methods can predict cloud formation, movement, and dissipation, their long computation time and slow response speed make them difficult to meet the needs of precise scheduling and intelligent management of integrated energy systems.

[0004] Atmospheric radiation transfer processes are mainly referred to in [reference]. Figure 1 Solar radiation is primarily affected by scattering and absorption from absorbing gases, aerosols, and clouds during its transmission through the atmosphere. Studies have shown that cloud scattering and absorption are the main causes of solar radiation attenuation. Among these, clouds cause the greatest attenuation of solar irradiance, followed by absorbing gases and aerosols. The timescale refers to the duration of the influence of absorbing gases, aerosols, and clouds on solar radiation. The timescale for absorbing gases is measured in years, for aerosols in days, and for clouds in hours. Of the three factors, clouds have the most significant impact on irradiance, and their optical properties change most rapidly.

[0005] Therefore, it is urgent to solve the problems of long time consumption and slow response speed in existing solar irradiance prediction methods that take cloud information into account. Summary of the Invention

[0006] The purpose of this invention is to solve the problems of long time consumption and slow response speed in existing solar irradiance prediction methods that take cloud information into account. This invention provides a solar irradiance prediction method that takes cloud physical parameters into account.

[0007] A method for predicting solar irradiance considering cloud physical properties, comprising the following steps:

[0008] Step 1: Obtain the physical property parameters of aerosols and absorbing gases in the irradiated area where the photovoltaic panel is located based on literature or meteorological databases. The physical property parameters include scattering coefficient, absorption coefficient and scattering phase function.

[0009] Step 2: Divide the irradiation geometry of the photovoltaic panel within the atmosphere into multiple layers from top to bottom; calculate the physical properties of the clouds that block the photovoltaic panel in each layer at the current time and the next time.

[0010] Step 3: The atmospheric radiative transfer model obtains the transmittance based on the physical properties of aerosols, absorbing gases, and clouds at the current moment. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted solar irradiance incident on the photovoltaic panel at the current moment. The atmospheric radiative transfer model is implemented using the Monte Carlo ray tracing method.

[0011] Among them, the aerosol physical property parameters of each layer at each time point are the same, and are all aerosol physical property parameters obtained in step 1. The physical property parameters of the absorbent gas of each layer at each time point are the same, and are all absorbent gas physical property parameters obtained in step 1.

[0012] Step 4: Divide the actual solar irradiance value at the current moment by the predicted solar irradiance value to obtain the coefficient k; determine whether the coefficient k meets the threshold. If the result is no, use the coefficient k to correct the cloud property parameters at the next moment to obtain the corrected cloud property parameters at the next moment, and proceed to step 5; if the result is yes, proceed to step 6.

[0013] Step 5: The atmospheric radiative transfer model obtains the transmittance based on the aerosol, absorbing gas, and corrected cloud properties of each layer at the next moment. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted solar irradiance incident on the photovoltaic panel at the next moment, thus completing the real-time prediction of solar irradiance.

[0014] Step 6: The atmospheric radiation transfer model obtains the transmittance based on the physical properties of aerosols and absorbing gases in each layer at the next moment, as well as the physical properties of clouds in each layer at the next moment obtained in Step 2. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted value of solar irradiance incident on the photovoltaic panel at the next moment, thus completing the real-time prediction of solar irradiance.

[0015] Preferably, the absorbent gases include H2O, CO2, CH4, CO, O2, N2O, and O3.

[0016] Preferably, the threshold range of coefficient k is 0.95 < k ≤ 1.05.

[0017] Preferably, the method for determining the current moment when the cloud causes shading of the photovoltaic panel is as follows:

[0018] A1. Establish a local rectangular coordinate system with the location of the all-sky imager in the photovoltaic power station as the reference point, and assume that the cloud shape and height remain unchanged during the prediction period;

[0019] A2. The position of each cloud in the local rectangular coordinate system at the current moment is collected by the TWS-CC all-sky imager, and the position of each point in the cloud shadow in the local rectangular coordinate system at the current moment is determined based on the position of each cloud in the local rectangular coordinate system at the current moment.

[0020] A3. Treat the position of the photovoltaic panel in the local rectangular coordinate system as a point, and determine whether the current position coordinate of the photovoltaic panel is within the range of the cloud shadow. If the result is yes, it is determined that the cloud is blocking the photovoltaic panel; if the result is no, it is determined that the cloud is not blocking the photovoltaic panel.

[0021] Preferably, the method for determining the cloud's shading of the photovoltaic panel in the next moment is as follows:

[0022] B1. Establish a local rectangular coordinate system with the location of the all-sky imager in the photovoltaic power station as the reference point, and assume that the cloud shape and height remain unchanged during the prediction period;

[0023] B2. The position and speed of each cloud in the local rectangular coordinate system at the current moment are collected by the TWS-CC all-sky imager. Based on the position of each cloud in the local rectangular coordinate system, the speed of the cloud and the time interval between the current moment and the next moment, the position of the cloud in the local rectangular coordinate system at the next moment is obtained.

[0024] And based on the position of each cloud in the local rectangular coordinate system at the next moment, determine the position of each point within the cloud shadow in the local rectangular coordinate system at the next moment;

[0025] B3. Treat the position of the photovoltaic panel in the local rectangular coordinate system as a point, and determine whether the position coordinates of the photovoltaic panel at the next moment are within the range of the cloud shadow. If the result is yes, it is determined that the cloud is blocking the photovoltaic panel; if the result is no, it is determined that the cloud is not blocking the photovoltaic panel.

[0026] Preferably, the method for determining the position of each point within the cloud's shadow in the local rectangular coordinate system based on the position of each cloud in the local rectangular coordinate system is as follows:

[0027] x p =h c1 ·tanz1·sina s1 +x

[0028] y p =h c1 ·tanz1·cosa s1 +y;

[0029] Where, x p and y p Let x and y be the x and y coordinates of any point within the cloud shadow, and h be the x and y coordinates of any point within the cloud. c1 Z1 is the solar altitude angle, z2 is the solar celestial azimuth angle, and a is the solar altitude angle. s1 This is the solar azimuth angle.

[0030] Preferably, the physical property parameters of the clouds that shade the photovoltaic panels at any given time are realized as follows:

[0031] Obtain cloud thickness, cloud top height, cloud droplet size, and cloud water content; where cloud water content includes liquid water content and ice water content;

[0032] The cloud's physical properties were obtained by analyzing cloud thickness, cloud top height, cloud droplet size, and cloud water content using Mie theory.

[0033] Preferably, it is assumed that the cloud thickness, cloud top height, cloud droplet size, and cloud water content are the same at two adjacent time points;

[0034] The cloud physical properties parameters at the next time step are corrected using the coefficient k, which is equivalent to correcting the cloud thickness at the next time step.

[0035] The correction of the cloud thickness at the next moment is specifically achieved by multiplying the coefficient k by the cloud thickness collected at the current moment.

[0036] Preferably, cloud thickness and cloud top height are collected using a ground-based laser cloud measuring instrument;

[0037] The cloud droplet size and cloud water content were obtained by inverting the echo signal from the ground-based lidar.

[0038] Preferably, the actual solar irradiance value is collected using an irradiance meter.

[0039] The beneficial effects of this invention are as follows:

[0040] This invention focuses on measuring cloud physical properties and predicting surface irradiance (i.e., solar irradiance incident on photovoltaic panels) using an atmospheric radiative transfer model. Cloud movement is highly random, resulting in almost no clear causal relationship between fluctuations in solar irradiance and historical data. Therefore, traditional data-driven methods struggle to accurately account for the impact of clouds on surface solar irradiance.

[0041] This invention utilizes an atmospheric radiative transfer model to replace the traditional numerical weather prediction model, proposing a solar irradiance prediction method that combines atmospheric radiative transfer model and cloud property parameter prediction, taking into account cloud property parameters. This method targets clouds, which are highly random and have the greatest impact, and can be monitored through a real-time meteorological monitoring system. Based on measured cloud and irradiance data, the cloud property parameters input into the atmospheric radiative transfer model are corrected in real time, improving prediction accuracy.

[0042] The method of the present invention also has the following advantages:

[0043] 1) Fast calculation speed, capable of providing predictions at the hourly or minute level;

[0044] 2) Low computational cost, no need for large computing servers, ordinary servers or computers can meet the needs of prediction simulation calculations;

[0045] 3) The calculation accuracy is high, and the prediction accuracy can reach more than 90%. This method can overcome the shortcomings of traditional data-driven methods in calculating cloud impact and numerical weather prediction methods in terms of time consumption and high cost. It plays an important role in promoting the precise scheduling and intelligent management of integrated energy systems. Attached Figure Description

[0046] Figure 1 This is a schematic diagram illustrating the principle of atmospheric radiation transfer.

[0047] Figure 2 This is a flowchart of the solar irradiance prediction method that considers cloud physical properties as described in this invention;

[0048] Figure 3 This is a schematic diagram of cloud shadow calculation;

[0049] Figure 4 This is a schematic diagram of the TWS-CC all-sky imager. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0052] Solar radiation, during its atmospheric transmission, is absorbed, reflected, and scattered multiple times by atmospheric components (absorbing gases, aerosols, clouds, etc.) before reaching the Earth's surface and ultimately being received by the photovoltaic panels of a photovoltaic power station. To study the surface irradiance of photovoltaic power stations under real-world conditions, this invention provides a method for predicting solar irradiance that considers cloud physical properties.

[0053] The concept of this invention is as follows: Calculate the physical properties of clouds that obstruct photovoltaic panels in each layer at the current and next time points. Divide the actual solar irradiance at the current time by the predicted solar irradiance to obtain a coefficient k. Determine the accuracy of the current atmospheric radiative transfer model prediction. Then, determine whether coefficient k meets a threshold.

[0054] The result was negative, proving that the atmospheric radiative transfer model had low prediction accuracy. The main factor causing the low prediction accuracy was the influence of clouds. Therefore, the cloud physical property parameters calculated at the next time step were corrected using the coefficient k. The atmospheric radiative transfer model obtained the transmittance based on the aerosols, absorbing gases and corrected cloud physical property parameters of each layer at the next time step, and thus obtained the predicted value of solar irradiance incident on the photovoltaic panel at the next time step.

[0055] The result proves that the accuracy of the atmospheric radiative transfer model prediction meets the requirements. The atmospheric radiative transfer model obtains the transmittance based on the physical properties of aerosols and absorbing gases in each layer at the next time step, as well as the physical properties of clouds in each layer at the next time step, and thus obtains the predicted value of solar irradiance incident on the photovoltaic panel at the next time step.

[0056] Specific Implementation Method 1: The following is combined with... Figure 2 This embodiment describes a solar irradiance prediction method that considers cloud physical properties. The method includes the following steps:

[0057] Step 1: Obtain the physical property parameters of aerosols and absorbing gases in the irradiated area where the photovoltaic panel is located based on literature or meteorological databases. The physical property parameters include scattering coefficient, absorption coefficient and scattering phase function.

[0058] Aerosols are composed of various solid or liquid particles suspended in the atmosphere, with particle sizes ranging from 0.001 μm to 100 μm and time scales measured in days. The physical properties of aerosols in the area to be measured can be obtained using historical meteorological data or relevant literature.

[0059] Studies have shown that the attenuation of solar irradiance by absorbing gases is mainly related to the radiophysical properties of the absorbing gases. The time scale of the absorbing gas parameters is in years, and these properties are relatively stable, simplifying them to time-independent parameters in atmospheric radiative transfer models. As an example, the absorbing gases considered in this invention mainly include H2O, CO2, CH4, CO, O2, N2O, and O3. In this case, the physical properties of the absorbing gases in step 1 include the scattering coefficient, absorption coefficient, and scattering phase function of H2O; the scattering coefficient, absorption coefficient, and scattering phase function of CO2; the scattering coefficient, absorption coefficient, and scattering phase function of CH4; the scattering coefficient, absorption coefficient, and scattering phase function of CO; the scattering coefficient, absorption coefficient, and scattering phase function of O2; the scattering coefficient, absorption coefficient, and scattering phase function of N2O; and the scattering coefficient, absorption coefficient, and scattering phase function of O3.

[0060] In terms of scattering, absorbing gases are mainly subjected to Rayleigh scattering. Therefore, in specific applications, the Rayleigh scattering coefficient and Rayleigh scattering phase function of the absorbing gas are used.

[0061] Clouds, a major factor contributing to the attenuation of solar radiation, are polymers composed of water droplets, supercooled water droplets, and ice crystals, typically covering about 60% of the Earth's atmosphere. Based on their temperature and particle phase, they can be classified into ice clouds composed of ice crystal particles and liquid water clouds. Ice clouds consist of layered, banded, or fibrous non-spherical ice crystal particles, with temperatures generally below -20°C, and are typically located in the upper atmosphere. Water clouds are formed by the condensation of water vapor in the air and are usually located below 7000 meters in the atmosphere; the distribution and concentration of water droplets within them need to be measured. Due to the relatively small timescale of clouds, their radiometric properties fluctuate significantly. Their movement, deformation, and dissipation all cause drastic fluctuations in surface solar irradiance. Low-level cloud cover plays a decisive role in obstructing the surface solar energy resources of the target area.

[0062] Studies have shown that cloud scattering and absorption are the main causes of solar radiation attenuation, and the scattering and absorption of solar radiation by clouds are related to cloud thickness, cloud top height, cloud droplet size, and cloud water content. Different cloud amounts (cloud thicknesses) result in different cross-sectional areas for reflecting and absorbing solar radiation, leading to varying degrees of solar irradiance attenuation. Furthermore, the solar radiation transmission path is affected by cloud height, leading to different radiation budgets. Ice clouds reflect and absorb both solar and infrared radiation. The radiometric properties of clouds can be calculated using Mie theory.

[0063] Step 2: Divide the irradiation geometry of the photovoltaic panel within the atmosphere into multiple layers from top to bottom; calculate the physical properties of the clouds that block the photovoltaic panel in each layer at the current time and the next time.

[0064] Among them, the physical properties of clouds include the cloud scattering coefficient, the cloud absorption coefficient, and the cloud scattering phase function;

[0065] Step 3: The atmospheric radiative transfer model obtains the transmittance based on the physical properties of aerosols, absorbing gases, and clouds at the current moment. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted solar irradiance incident on the photovoltaic panel at the current moment. The atmospheric radiative transfer model is implemented using the Monte Carlo ray tracing method.

[0066] Among them, the aerosol physical property parameters of each layer at each time point are the same, and are all aerosol physical property parameters obtained in step 1. The physical property parameters of the absorbent gas of each layer at each time point are the same, and are all absorbent gas physical property parameters obtained in step 1.

[0067] Step 4: Divide the actual solar irradiance value at the current moment by the predicted solar irradiance value to obtain the coefficient k; determine whether the coefficient k meets the threshold. If the result is no, use the coefficient k to correct the cloud property parameters at the next moment to obtain the corrected cloud property parameters at the next moment, and proceed to step 5; if the result is yes, proceed to step 6.

[0068] Step 5: The atmospheric radiative transfer model obtains the transmittance based on the aerosol, absorbing gas, and corrected cloud properties of each layer at the next moment. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted solar irradiance incident on the photovoltaic panel at the next moment, thus completing the real-time prediction of solar irradiance.

[0069] Step 6: The atmospheric radiation transfer model obtains the transmittance based on the physical properties of aerosols and absorbing gases in each layer at the next moment, as well as the physical properties of clouds in each layer at the next moment obtained in Step 2. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted value of solar irradiance incident on the photovoltaic panel at the next moment, thus completing the real-time prediction of solar irradiance.

[0070] In this embodiment, by introducing the physical property parameters of clouds that shade the photovoltaic panels in each layer at the current and next time moments, the actual solar irradiance value at the current time is compared with the predicted solar irradiance value. A correction coefficient k is introduced to determine whether to use the physical property parameters of clouds that shade the photovoltaic panels in each layer at the next time moment calculated in step 2 for solar irradiance prediction, or to use the physical property parameters of clouds that shade the photovoltaic panels in each layer at the next time moment calculated in step 2 to correct the solar irradiance prediction before making a solar irradiance prediction. The entire process relies on historical data to judge the accuracy of the model prediction. The entire prediction method is simple, short in time, and has a fast response and calculation speed, and can achieve prediction speeds at the hour or minute level.

[0071] In practical applications, the threshold range for coefficient k is 0.95 < k ≤ 1.05.

[0072] Furthermore, in practical applications, the solar irradiance incident on the photovoltaic panel is related to the cloud position and the solar zenith angle. That is, it is necessary to determine whether the cloud is shading the photovoltaic panel of the photovoltaic power station, infer the position of the cloud shadow based on the cloud position and movement trajectory, and calculate the physical property parameters of the cloud that is shading the photovoltaic panel in each layer at the current time and the next time after determining that the cloud is shading the photovoltaic panel at the current time and the next time.

[0073] (I) See Figure 3 and Figure 4 The mechanism by which clouds in each layer shade the photovoltaic panels at the current moment is determined as follows:

[0074] A1. Establish a local rectangular coordinate system with the location of the all-sky imager in the photovoltaic power station as the reference point, and assume that the cloud shape and height remain unchanged during the prediction period;

[0075] A2, see also Figure 4 The positions of each cloud in the local rectangular coordinate system at the current moment are collected by the TWS-CC all-sky imager, and the positions of each point within the cloud shadow in the local rectangular coordinate system at the current moment are determined based on the positions of each cloud in the local rectangular coordinate system at the current moment.

[0076] A3. Treat the position of the photovoltaic panel in the local rectangular coordinate system as a point, and determine whether the current position coordinates of the photovoltaic panel are within the range of the cloud shadow. If the result is yes, it is determined that the cloud is shading the photovoltaic panel; if the result is no, it is determined that the cloud is not shading the photovoltaic panel. See [link to relevant documentation]. Figure 3 .

[0077] (II) See Figure 3 and Figure 4 The mechanism by which clouds in each layer shade the photovoltaic panels at the next moment is determined as follows:

[0078] B1. Establish a local rectangular coordinate system with the location of the all-sky imager in the photovoltaic power station as the reference point, and assume that the cloud shape and height remain unchanged during the prediction period;

[0079] B2, see also Figure 4 The system uses a TWS-CC all-sky imager to collect the positions and speeds of each cloud in the local Cartesian coordinate system at the current moment. Based on the positions of each cloud in the local Cartesian coordinate system, its speed, and the time interval between the current moment and the next moment, the system obtains the position of the cloud in the local Cartesian coordinate system at the next moment. Based on the positions of each cloud in the local Cartesian coordinate system at the next moment, the system determines the positions of each point within the cloud shadow in the local Cartesian coordinate system at the next moment.

[0080] B3. Treat the position of the photovoltaic panel in the local rectangular coordinate system as a point, and determine whether the position coordinates of the photovoltaic panel at the next moment are within the range of the cloud shadow. If the result is yes, it is determined that the cloud is shading the photovoltaic panel; if the result is no, it is determined that the cloud is not shading the photovoltaic panel. See [link to relevant documentation]. Figure 3 .

[0081] The TWS-CC all-sky imager is primarily used to acquire all-sky image data of a target area in a bottom-up manner. Image processing can reveal cloud position, shape, and velocity. Additionally, the device can obtain parameters such as illumination intensity, high-frequency components, transmittance, and zenith distance. Cloud position, shape, and velocity can be indirectly determined from continuous all-sky images, while parameters such as illumination intensity, high-frequency components, transmittance, and zenith distance are mainly used to provide data for cloud shadows and data-driven analysis.

[0082] The cloud positions and zenith distances were obtained using a TWS-CC all-sky imager. A local rectangular coordinate system was established with the TWS-CC all-sky imager at the photovoltaic power station as a reference, assuming that the shape and height of the clouds remained unchanged over a short period of time.

[0083] The coordinates of clouds in the local coordinate system are obtained using an all-sky imager. Then, using continuous time-series cloud coordinates, the average velocity of the previous cloud movement is calculated to predict the cloud's next position. Assuming the cloud shadow and cloud shape are similar, and the cloud's coordinates on the ground-based cloud image are (x, y), the coordinates of the cloud shadow (x...y) are... p ,y p );

[0084] See Figure 3 The method for determining the position of each point within the cloud's shadow in the local Cartesian coordinate system, based on the cloud's position in the local Cartesian coordinate system, is as follows:

[0085]

[0086] Where, x p and y pLet x and y be the x and y coordinates of any point within the cloud shadow, and h be the x and y coordinates of any point within the cloud. c1 Z1 is the solar altitude angle, z2 is the solar zenith angle, and a is the solar altitude angle. s1 This is the solar azimuth angle.

[0087] Furthermore, the physical property parameters of the clouds that shade the photovoltaic panels at any given time are realized as follows:

[0088] Obtain cloud thickness, cloud top height, cloud droplet size, and cloud water content; where cloud water content includes liquid water content and ice water content;

[0089] The cloud's physical properties were obtained by analyzing cloud thickness, cloud top height, cloud droplet size, and cloud water content using Mie theory.

[0090] In this preferred embodiment, cloud thickness, cloud top height, cloud droplet size, and cloud water content are analyzed using Mie theory, which is achieved through existing technology.

[0091] To measure cloud physical properties, this invention establishes a real-time cloud monitoring system. This system primarily obtains information on cloud thickness, cloud top height, cloud droplet size, and cloud water content. The methods employed include high-speed imaging technology (such as the TWS-CC all-sky imager), ground-based laser cloud measuring instruments, and ground-based lidar. This invention directly utilizes the TWS-CC all-sky imager to obtain sky cloud image data, focusing on obtaining a full-sky image. Other similar ground-based imaging devices can also be used.

[0092] In practical applications, ground-based imaging equipment acquires image sequences of clouds over the target area from a bottom-up perspective. Irrelevant pixels and noise are removed through image preprocessing. Additionally, data such as light intensity, high-frequency components, transmittance, zenith distance, and cloud characteristics are acquired, processed, and stored. A fixed threshold discrimination method based on the red-to-blue ratio of image pixels is used to identify clouds in the images, and finally, thin clouds are filtered using brightness parameters. Based on the assumption that cloud shape and movement speed remain unchanged at minute-level scales, cloud features are identified and matched in the image sequence, and cloud movement speed is calculated. Extrapolation based on sky images is used to predict the location of clouds over photovoltaic power stations. Then, based on the measurement results of the TWS-CC all-sky imager, the geometric relationship between the sun, clouds, and the ground surface is determined, and finally, the ground shadow position of the clouds is calculated.

[0093] In a specific application, as an example: to obtain cloud thickness and cloud top height, a ground-based laser cloud meter is used to collect cloud thickness and cloud top height.

[0094] To obtain the cloud droplet size and cloud water content, the cloud droplet size and cloud water content are specifically calculated by inverting the echo signal of a ground-based lidar.

[0095] To obtain the actual value of solar irradiance, an irradiance meter is used to collect the actual value of solar irradiance.

[0096] Furthermore, when using the coefficient k to correct the cloud physical properties at the next time step, it is first assumed that the cloud thickness, cloud top height, cloud droplet size and cloud water content are the same at two adjacent time steps.

[0097] The cloud physical properties parameters at the next time step are corrected using the coefficient k, which means: the cloud thickness at the next time step is corrected.

[0098] The correction of the cloud thickness at the next moment is specifically achieved by multiplying the coefficient k by the cloud thickness collected at the current moment.

[0099] While the invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A method for predicting solar irradiance considering cloud physical properties, characterized in that, The method includes the following steps: Step 1: Obtain the physical property parameters of aerosols and absorbing gases in the irradiated area where the photovoltaic panel is located based on literature or meteorological databases. The physical property parameters include scattering coefficient, absorption coefficient and scattering phase function. Step 2: Divide the irradiation geometry of the photovoltaic panel within the atmosphere into multiple layers from top to bottom; calculate the physical properties of the clouds that block the photovoltaic panel in each layer at the current time and the next time. The method by which the cloud currently shades the photovoltaic panels is determined is as follows: A1. Establish a local rectangular coordinate system with the location of the all-sky imager in the photovoltaic power station as the reference point, and assume that the cloud shape and height remain unchanged during the prediction period; A2. The position of each cloud in the local rectangular coordinate system at the current moment is collected by the TWS-CC all-sky imager, and the position of each point in the cloud shadow in the local rectangular coordinate system at the current moment is determined based on the position of each cloud in the local rectangular coordinate system at the current moment. A3. Treat the position of the photovoltaic panel in the local rectangular coordinate system as a point, and determine whether the current position coordinate of the photovoltaic panel is within the range of the cloud shadow. If the result is yes, it is determined that the cloud is blocking the photovoltaic panel; if the result is no, it is determined that the cloud is not blocking the photovoltaic panel. The method by which the cloud will shade the photovoltaic panels in the next moment is determined to be: B1. Establish a local rectangular coordinate system with the location of the all-sky imager in the photovoltaic power station as the reference point, and assume that the cloud shape and height remain unchanged during the prediction period; B2. The position and speed of each cloud in the local rectangular coordinate system at the current moment are collected by the TWS-CC all-sky imager. Based on the position of each cloud in the local rectangular coordinate system, the speed of the cloud and the time interval between the current moment and the next moment, the position of the cloud in the local rectangular coordinate system at the next moment is obtained. And based on the position of each cloud in the local rectangular coordinate system at the next moment, determine the position of each point within the cloud shadow in the local rectangular coordinate system at the next moment; B3. Treat the position of the photovoltaic panel in the local rectangular coordinate system as a point, and determine whether the position coordinates of the photovoltaic panel at the next moment are within the range of the cloud shadow. If the result is yes, it is determined that the cloud is blocking the photovoltaic panel; if the result is no, it is determined that the cloud is not blocking the photovoltaic panel. The method for realizing the physical properties of clouds that shade the photovoltaic panels at any given time in each layer is as follows: Obtain cloud thickness, cloud top height, cloud droplet size, and cloud water content; where cloud water content includes liquid water content and ice water content. The cloud's physical properties were obtained by analyzing cloud thickness, cloud top height, cloud droplet size, and cloud water content using Mie theory. Step 3: The atmospheric radiative transfer model obtains the transmittance based on the physical properties of aerosols, absorbing gases, and clouds at the current moment. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted solar irradiance incident on the photovoltaic panel at the current moment. The atmospheric radiative transfer model is implemented using the Monte Carlo ray tracing method. Among them, the aerosol physical property parameters of each layer at each time point are the same, and are all aerosol physical property parameters obtained in step 1. The physical property parameters of the absorbent gas of each layer at each time point are the same, and are all absorbent gas physical property parameters obtained in step 1. Step 4: Divide the actual solar irradiance value at the current moment by the predicted solar irradiance value to obtain the coefficient k; determine whether the coefficient k meets the threshold. If the result is no, use the coefficient k to correct the cloud property parameters at the next moment to obtain the corrected cloud property parameters at the next moment, and proceed to step 5; if the result is yes, proceed to step 6. Step 5: The atmospheric radiative transfer model obtains the transmittance based on the aerosol, absorbing gas, and corrected cloud properties of each layer at the next moment. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted solar irradiance incident on the photovoltaic panel at the next moment, thus completing the real-time prediction of solar irradiance. Step 6: The atmospheric radiation transfer model obtains the transmittance based on the physical properties of aerosols and absorbing gases in each layer at the next moment, as well as the physical properties of clouds in each layer at the next moment obtained in Step 2. This transmittance is multiplied by the actual solar irradiance incident on the atmosphere to obtain the predicted value of solar irradiance incident on the photovoltaic panel at the next moment, thus completing the real-time prediction of solar irradiance.

2. The solar irradiance prediction method considering cloud physical parameters according to claim 1, characterized in that, Absorbent gases include H2O, CO2, CH4, CO, O2, N2O, and O3.

3. The solar irradiance prediction method considering cloud physical parameters according to claim 1, characterized in that, The threshold range for coefficient k is 0.95 < k ≤ 1.

05.

4. The solar irradiance prediction method considering cloud physical parameters according to claim 1, characterized in that, The method for determining the position of each point within the cloud's shadow in the local Cartesian coordinate system, based on the position of each cloud in the local Cartesian coordinate system, is as follows: ； in, and Let x and y be the coordinates of any point within the cloud shadow. and Let x and y be the coordinates of any point within the cloud. The solar altitude angle, Determine the angle of the sun's celestial sphere. This is the solar azimuth angle.

5. The solar irradiance prediction method considering cloud physical parameters according to claim 1, characterized in that, Assume that the cloud thickness, cloud top height, cloud droplet size, and cloud water content are the same at two adjacent time points; The cloud physical properties parameters at the next time step are corrected using the coefficient k, which is equivalent to correcting the cloud thickness at the next time step. The correction of the cloud thickness at the next moment is specifically achieved by multiplying the coefficient k by the cloud thickness collected at the current moment.

6. The solar irradiance prediction method considering cloud physical parameters according to claim 1, characterized in that, Cloud thickness and cloud top height were collected using a ground-based laser cloud measuring instrument; The cloud droplet size and cloud water content were obtained by inverting the echo signal from the ground-based lidar.

7. The solar irradiance prediction method considering cloud physical parameters according to claim 1, characterized in that, The actual values ​​of solar irradiance were collected using an irradiance meter.

Images