Distributed photovoltaic power prediction system and method based on cloud layer data

Abstract

The invention discloses a distributed photovoltaic power prediction system and method based on cloud layer data, relates to the technical field of photovoltaic power prediction, and solves the problem that the power is difficult to dynamically predict according to the short-time change of the cloud layer thickness at the next time point during the photovoltaic power prediction. And different dynamic monitoring intervals are difficult to set for cloud layers and wind speeds of different time and space. According to the method, the analysis time point of the photovoltaic panel is set, and the wind condition data of the next analysis time point is predicted through the wind condition data of the current analysis time point; predicting cloud layer data of the target area at the next analysis time point based on the cloud layer data at the current analysis time point and the wind regime data at the next analysis time point; analyzing the cloud layer data at the next analysis time point to obtain the light transmittance at the next analysis time point, and obtaining a photovoltaic power prediction value based on the light transmittance at the next analysis time point; dynamic prediction data can be provided in the impact analysis of photovoltaic power generation on the power grid, and the operation safety of the power grid is improved.

Description

Technical Field

[0001] The present invention belongs to the field of new energy power, and relates to photovoltaic power prediction technology, specifically a distributed photovoltaic power prediction system and method based on cloud data. Background Art

[0002] With the global energy transformation and the development of new power systems, distributed photovoltaic power stations have been widely used due to their flexibility and environmental friendliness. However, the widespread access of distributed photovoltaic panels has brought challenges to the planning and operation of the power grid. Especially in areas with large fluctuations in photovoltaic power, frequent power fluctuations can lead to voltage fluctuations and harmonic pollution in the power grid, affecting the power quality. For the power fluctuations of photovoltaic power generation, cloud changes are an important influencing factor. The occlusion of clouds will cause a decrease in solar radiation intensity, thereby reducing the amount of sunlight received by the photovoltaic panels and further reducing the power generation. This impact may be instantaneous or continuous, bringing a greater impact to the power grid.

[0003] Currently, most distributed photovoltaic power prediction systems and methods based on cloud data, when analyzing the impact of cloud changes on the power generation of photovoltaic panels, only analyze according to the cloud thickness, moving speed, and coverage area, ignoring the dynamic change of the thickness of the same cloud in a fixed area. It is difficult to dynamically predict the power according to the dynamic change of the cloud thickness, which may bring the risk of power grid fluctuations caused by short-term cloud mutations. At the same time, most distributed photovoltaic power prediction systems and methods based on cloud data are difficult to set different dynamic monitoring intervals for clouds and wind speeds in different time and space, and only predict the power according to a fixed monitoring interval, resulting in the analysis of the system not being able to dynamically adapt to the time and space conditions of the location of the photovoltaic panels, reducing the accuracy of the prediction of photovoltaic power generation. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. Therefore, the present invention provides a distributed photovoltaic power prediction system and method based on cloud data, which are used to solve the technical problems that it is difficult to dynamically predict the power according to the short-term change of the cloud thickness at the next time point and it is difficult to set different dynamic monitoring intervals for clouds and wind speeds in different time and space during photovoltaic power prediction. The present invention solves the above problems by setting an analysis time point of the current photovoltaic panel, predicting the wind condition data at the next analysis time point through the wind condition data at the current analysis time point; predicting the cloud data in the target area at the next analysis time point based on the cloud data at the current analysis time point and the wind condition data at the next analysis time point; analyzing the cloud data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtaining the photovoltaic power prediction value based on the light transmittance at the next analysis time point, and setting the warning level based on the photovoltaic power prediction value.

[0005] To achieve the above object, a first aspect of the present invention provides a distributed photovoltaic power prediction system based on cloud layer data, including: a cloud layer analysis module, and a data collection module, a power prediction module and a database connected thereto;

[0006] The data collection module: is used to set the target area corresponding to the photovoltaic panel and obtain the basic data of the target area; wherein, the basic data includes cloud layer data and wind condition data; the cloud layer data includes cloud layer thickness and cloud layer type, and the wind condition data includes wind direction and wind speed;

[0007] The cloud layer analysis module: is used to set the analysis time point of the current photovoltaic panel, predict the wind condition data at the next analysis time point through the wind condition data at the current analysis time point; predict the cloud layer data of the target area at the next analysis time point based on the cloud layer data at the current analysis time point and the wind condition data at the next analysis time point;

[0008] The power prediction module: is used to analyze the cloud layer data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtain the photovoltaic power prediction value based on the light transmittance at the next analysis time point, and set the warning level based on the photovoltaic power prediction value.

[0009] Preferably, the setting of the target area corresponding to the photovoltaic panel includes:

[0010] Extract the geographical latitude φ of the photovoltaic panel, as well as the solar declination δ and hour angle ω corresponding to the current time, from the database, and obtain the solar altitude angle α based on formula (1), and formula (1) is specifically:

[0011] sinα = sinφ·sinδ + cosφ·cosδ·cosω (1);

[0012] Obtain the solar azimuth angle β based on formula (2), and formula (2) is specifically:

[0013]

[0014] Extract the current cloud layer average moving speed V and cloud layer average height H of the location where the photovoltaic panel is located from the historical data, and obtain the maximum viewing angle radius R based on formula (3), and formula (3) is specifically:

[0015] R = arctan(V·Δt / H) (3);

[0016] Wherein, Δt is the manually set time window;

[0017] Taking the location where the photovoltaic panel is located as the observation point, mark the sky area with the real-time solar position (α, β) as the center and the maximum viewing angle radius R as the angular range as the target area.

[0018] Preferably, the obtaining of the basic data of the target area includes:

[0019] Obtaining the cloud thickness and cloud type within the target area in real time through a space remote sensing system, and obtaining the wind direction and wind speed within the target area in real time through the ground meteorological station where the photovoltaic panel is located.

[0020] Preferably, the setting of the analysis time point of the current photovoltaic panel includes:

[0021] A1: Obtain the current time, and determine whether there is an analysis time point within the next n seconds; if yes, mark the analysis time point existing within n seconds as the current analysis time point, and jump to A3; if no, jump to A2;

[0022] A2: Determine whether the time obtained by adding the standard time interval BG to the previous analysis time point is before the current time; if yes, set the current time as the current analysis time point, and jump to A3; if no, set the time where the previous analysis time point plus the standard time interval BG is located as the current analysis time point, and jump to A3; where n is obtained through manual setting;

[0023] A3: Determine whether there are clouds in the target area at the current time and whether the wind speed is not 0; if yes, jump to A4; if no, perform time addition calculation on the current analysis time point and the standard time interval BG to obtain the next analysis time point of the current photovoltaic panel;

[0024] A4: Determine whether the cloud type of the current photovoltaic panel in the current target area is an easily interfering type; if yes, jump to A5; if no, perform time addition calculation on the current analysis time point and the standard time interval BG to obtain the next analysis time point of the current photovoltaic panel; where the standard time interval BG is obtained through manual setting; the easily interfering types include: cirrus cloud, stratocumulus cloud, altocumulus cloud, fractostratus cloud, and fractocumulus cloud;

[0025] A5: Mark the cloud type of the current photovoltaic panel in the current target area as the target type, extract several historical cloud thicknesses and historical wind speeds corresponding to the target type in the historical data of the current target area from the database, integrate the historical cloud thicknesses into the B1 data group, and integrate the historical wind speeds into the B2 data group; sequentially obtain the variances of each data group, and determine whether the variance of the data group is less than the corresponding judgment threshold; if yes, perform average value calculation on the data group to obtain the characteristic value; if no, remove the data in the data group that differs the most from the mode, and re-judge the variance until the variance of the data group is less than the judgment threshold, and then perform average value calculation on the remaining data in the data to obtain the characteristic value; where the judgment threshold is obtained through empirical setting;

[0026] A6: Mark the eigenvalue of the B1 data group as the characteristic thickness TH, and mark the eigenvalue of the B2 data group as the characteristic wind speed TF; extract the cloud layer thickness YH and wind speed FS within the current target area, and obtain the dynamic time interval DG based on the formula DG = BG × β × exp(-(γ1 × TH / YH + γ2 × FS / TF)); where, BH is the standard thickness set manually, BF is the standard wind speed set manually; β is the amplitude adjustment coefficient of the manually set exp() function, and the value range of β is (0, 2]; γ1 and γ2 are both proportional adjustment coefficients greater than 0, and γ1 + γ2 = 1;

[0027] A7: Perform time addition calculation on the current analysis time point and the dynamic time interval DG to obtain the next analysis time point of the current photovoltaic panel.

[0028] Preferably, the prediction of the wind condition data at the next analysis time point through the wind condition data at the current analysis time point includes:

[0029] Extract the wind direction, wind speed, air pressure, temperature and humidity in the historical reference data of the target area from the database, as well as the wind direction and wind speed at the corresponding next analysis time point;

[0030] Integrate the wind direction, wind speed, air pressure, temperature and humidity, as well as the wind direction and wind speed at the next analysis time point into several training data and test data; use the training data to train the artificial intelligence model, use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a wind condition prediction model with the wind direction, wind speed, air pressure, temperature and humidity as the input and the wind direction and wind speed at the corresponding next analysis time point as the output; where, the artificial intelligence model includes an RNN neural network model and a Transformer neural network model;

[0031] Input the current wind direction, wind speed, air pressure, temperature and humidity of the target area into the wind condition prediction model to obtain the wind direction and wind speed at the next analysis time point.

[0032] Preferably, the prediction of the cloud layer data in the target area at the next analysis time point includes:

[0033] Extract the cloud layer thickness, cloud layer type, wind direction and wind speed in the historical data of the target area from the database, as well as the cloud layer type and cloud layer thickness at the corresponding next analysis time point;

[0034] Integrate the cloud thickness, cloud type, wind direction and wind speed, as well as the cloud type and cloud thickness at the next analysis time point into a number of training data and test data; use the training data to train the artificial intelligence model, use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a thickness prediction model with the cloud thickness, cloud type, wind direction and wind speed as the input and the cloud type and cloud thickness at the corresponding next analysis time point as the output; wherein, the artificial intelligence model includes an RNN neural network model and a Transformer neural network model;

[0035] Input the current cloud thickness, cloud type, wind direction and wind speed of the target area into the thickness prediction model to obtain the cloud type and cloud thickness at the next analysis time point.

[0036] Preferably, the analysis of the cloud data at the next analysis time point to obtain the transmittance at the next analysis time point includes:

[0037] C1: Extract the cloud thickness at the next analysis time point, and judge whether the cloud thickness exceeds the defined maximum value; if yes, update the cloud thickness with the defined maximum value; if no, perform a zero-update on the cloud thickness and jump to C2; wherein, the defined maximum value is obtained through experience;

[0038] C2: Extract the extinction coefficient θ of the cloud type at the next analysis time point and the updated cloud thickness L from the database, and obtain the transmittance LT at the next analysis time point based on the formula LT = exp(-(θ×L)).

[0039] Preferably, the obtaining of the photovoltaic power prediction value based on the transmittance at the next analysis time point includes:

[0040] Extract the transmittance LT at the next analysis time point and the ground solar irradiance G0 when there is no cloud in the target area, and obtain the irradiance G at the next analysis time point based on the formula G = G0×LT;

[0041] Obtain a number of battery temperatures W1 of historical irradiances identical to the irradiance G from the database, and the corresponding battery temperature W2 at the next analysis time point;

[0042] Based on the formula CW = W2 - W1 to obtain a number of temperature differences CW, and based on the formula to obtain a number of difference change rates CB; extract the percentile value DZ1, average value JZ1 and mode ZS1 of the number of temperature differences CW, and based on the formula to obtain the standard temperature difference BW; extract the percentile value DZ2, average value JZ2 and mode ZS2 of the number of difference change rates CB, and based on the formula Obtain the standard change rate BL; extract the current cell temperature DW of the photovoltaic panel, and obtain the cell temperature WD at the next analysis time point based on the formula WD = DW + (ρ3×BW + ρ4×DW×BL); where ρ1, ρ2, ρ3, and ρ4 are all proportional adjustment coefficients greater than 0, and ρ1 + ρ2 = 1, ρ3 + ρ4 = 1; the percentile of the percentile value is obtained by manual setting.

[0043] Obtain the area M of the current photovoltaic panel, based on the formula Obtain the predicted value P of the photovoltaic power at the next analysis time point; where σ is the nominal efficiency of the photovoltaic module, is the temperature coefficient, F1 is the system loss factor, and BD is the standard test temperature set manually.

[0044] Preferably, setting the warning level based on the predicted value of photovoltaic power includes:

[0045] Extract the usage duration of the current photovoltaic panel in use, and extract the duration correction coefficient from the database based on the usage duration; perform a multiplication operation on the duration correction coefficient and the standard power threshold to obtain the dynamic power threshold; where the standard power threshold is obtained by empirical setting.

[0046] Extract the predicted value P of the photovoltaic power at the next analysis time point, and determine whether the predicted value P of the photovoltaic power exceeds the dynamic power threshold; if so, mark the warning level as level one and issue a warning of exceeding the photovoltaic power standard at the next analysis time point; if not, mark the warning level as level two and do not issue a warning.

[0047] The second aspect of the present invention provides a distributed photovoltaic power prediction method based on cloud data, including the following steps:

[0048] S1: Set the target area corresponding to the photovoltaic panel, and obtain the basic data of the target area.

[0049] S2: Set the analysis time point of the current photovoltaic panel, and predict the wind condition data at the next analysis time point through the wind condition data at the current analysis time point; predict the cloud data in the target area at the next analysis time point based on the cloud data at the current analysis time point and the wind condition data at the next analysis time point.

[0050] S3: Analyze the cloud data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtain the predicted value of photovoltaic power based on the light transmittance at the next analysis time point, and set the warning level based on the predicted value of photovoltaic power.

[0051] Compared with the prior art, the beneficial effects of the present invention are:

[0052] 1. The present invention predicts the wind condition data at the next analysis time point through the wind condition data at the current analysis time point by setting the analysis time point of the current photovoltaic panel; predicts the cloud layer data in the target area at the next analysis time point based on the cloud layer data at the current analysis time point and the wind condition data at the next analysis time point; analyzes the cloud layer data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtains the predicted photovoltaic power value based on the light transmittance at the next analysis time point, and sets the warning level based on the predicted photovoltaic power value, solving the technical problems that it is difficult to dynamically predict the power according to the short-term change of the cloud layer thickness at the next time point and it is difficult to set different dynamic monitoring intervals for the cloud layer and wind speed in different time and space during photovoltaic power prediction; the present invention can provide dynamic prediction data in the impact analysis of photovoltaic power generation on the power grid.

[0053] 2. When monitoring the predicted photovoltaic power value, the present invention dynamically adjusts the monitoring interval according to the cloud layer and wind speed in different time and space, enabling the analysis and power prediction of the photovoltaic panel to dynamically adapt to the time and space conditions of the location of the photovoltaic panel; by dynamically adjusting the data monitoring frequency, it is possible to reduce the data transmission volume during invalid periods and extend the battery life cycle or the endurance cycle of the solar power supply system; at the same time, it can increase the monitoring frequency during critical periods, improve the data acquisition volume, and increase the accuracy of system analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0055] Figure 1 It is a schematic diagram of the operation steps of the present invention;

[0056] Figure 2 It is a schematic diagram of the system module of the present invention;

[0057] Figure 3 It is a schematic diagram of the operation steps for obtaining the cloud layer data at the next analysis time point of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] The following will clearly and completely describe the technical solutions of the present invention in combination with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0059] Please refer to Figure 1 - Figure 2, an embodiment of the first aspect of the present invention provides a distributed photovoltaic power prediction system based on cloud data, including: a cloud analysis module, and a data collection module, a power prediction module, and a database connected thereto;

[0060] Data collection module: used to set the target area corresponding to the photovoltaic panel and obtain the basic data of the target area; wherein, the basic data includes cloud data and wind condition data; the cloud data includes cloud thickness and cloud type, and the wind condition data includes wind direction and wind speed;

[0061] Cloud analysis module: used to set the analysis time point of the current photovoltaic panel, predict the wind condition data at the next analysis time point through the wind condition data at the current analysis time point; predict the cloud data in the target area at the next analysis time point based on the cloud data at the current analysis time point and the wind condition data at the next analysis time point;

[0062] Power prediction module: used to analyze the cloud data at the next analysis time point to obtain the transmittance at the next analysis time point, obtain the photovoltaic power prediction value based on the transmittance at the next analysis time point, and set the warning level based on the photovoltaic power prediction value.

[0063] In this application, setting the target area corresponding to the photovoltaic panel includes:

[0064] Extract the geographical latitude φ of the photovoltaic panel from the database, as well as the solar declination δ and hour angle ω corresponding to the current time, and obtain the solar altitude angle α based on formula (1). The specific formula (1) is:

[0065] sinα = sinφ·sinδ + cosφ·cosδ·cosω (1);

[0066] Obtain the solar azimuth angle β based on formula (2). The specific formula (2) is:

[0067]

[0068] Extract the current average cloud movement speed V and average cloud height H at the current time of the location where the photovoltaic panel is located from the historical data, and obtain the maximum viewing radius R based on formula (3). The specific formula (3) is:

[0069] R = arctan(V·Δt / H) (3);

[0070] Wherein, Δt is a manually set time window;

[0071] Taking the location where the photovoltaic panel is located as the observation point, mark the sky area with the real-time solar position (α, β) as the center and the maximum viewing radius R as the angular range as the target area.

[0072] It should be noted that the solar declination δ corresponding to the current time is the solar declination δ of the date where the current time is located, and the specific calculation formula is: where N is the day of the year.

[0073] It should be noted that the calculation formula for the hour angle ω is: ω = 15°×(LST - 12); where LST is the local time.

[0074] The basic data of the target area obtained in this application includes:

[0075] The cloud thickness and cloud type within the target area are obtained in real time through a space remote sensing system, and the wind direction and wind speed within the target area are obtained in real time through the ground meteorological station at the location of the photovoltaic panel.

[0076] The analysis time points for the current photovoltaic panel set in this application include:

[0077] A1: Obtain the current time, and determine whether there is an analysis time point within the next n seconds; if yes, mark the analysis time point existing within n seconds as the current analysis time point, and jump to A3; if no, jump to A2;

[0078] A2: Determine whether the time obtained by adding the standard time interval BG to the previous analysis time point is before the current time; if yes, set the current time as the current analysis time point, and jump to A3; if no, set the time where the previous analysis time point plus the standard time interval BG is located as the current analysis time point, and jump to A3; where n is obtained through manual setting;

[0079] A3: Determine whether there are clouds in the target area at the current time and whether the wind speed is not 0; if yes, jump to A4; if no, perform a time addition calculation on the current analysis time point and the standard time interval BG to obtain the next analysis time point of the current photovoltaic panel;

[0080] A4: Determine whether the cloud type of the current photovoltaic panel in the current target area is an easily interfering type; if yes, jump to A5; if no, perform a time addition calculation on the current analysis time point and the standard time interval BG to obtain the next analysis time point of the current photovoltaic panel; where the standard time interval BG is obtained through manual setting; the easily interfering types include: cirrus, stratocumulus, altocumulus, fractostratus, and fractocumulus;

[0081] A5: Mark the cloud type of the current photovoltaic panel in the current target area as the target type, extract several historical cloud thicknesses and historical wind speeds corresponding to the target type in the historical data of the current target area from the database, integrate the historical cloud thicknesses into the B1 data group, and integrate the historical wind speeds into the B2 data group; sequentially obtain the variances of each data group, and determine whether the variance of the data group is less than the corresponding judgment threshold; if yes, calculate the average value of the data group to obtain the characteristic value; if not, remove the data in the data group that has the largest difference from the mode, and re-judge the variance until the variance of the data group is less than the judgment threshold, and then calculate the average value of the remaining data in the data to obtain the characteristic value; among them, the judgment threshold is set through experience.

[0082] A6: Mark the characteristic value of the B1 data group as the characteristic thickness TH, and mark the characteristic value of the B2 data group as the characteristic wind speed TF; extract the cloud thickness YH and wind speed FS in the current target area, and obtain the dynamic time interval DG based on the formula DG = BG × β × exp(-(γ1 × TH / YH + γ2 × FS / TF)); where, BH is the standard thickness set manually, BF is the standard wind speed set manually; β is the amplitude adjustment coefficient of the exp() function set manually, and the value range of β is (0, 2]; both γ1 and γ2 are proportional adjustment coefficients greater than 0, and γ1 + γ2 = 1.

[0083] A7: Perform time addition calculation on the current analysis time point and the dynamic time interval DG to obtain the next analysis time point of the current photovoltaic panel.

[0084] It should be noted that when the present invention monitors the photovoltaic power prediction value, the monitoring interval is dynamically adjusted according to the clouds and wind speeds in different time and space, so that the analysis and power prediction of the photovoltaic panel can dynamically adapt to the time and space conditions of the location of the photovoltaic panel; by dynamically adjusting the data monitoring frequency, the data transmission volume can be reduced during invalid periods, and the battery or solar power supply system's endurance cycle can be extended; at the same time, the monitoring frequency can be increased during critical periods, the data acquisition volume can be increased, and the accuracy of system analysis can be improved.

[0085] It should be noted that in obtaining the dynamic time interval DG based on the formula DG = BG × β × exp(-(γ1 × TH / YH + γ2 × FS / TF)), the dependent variable of the function exp(-x) decreases as the independent variable increases, which is a downward trend curve; and because the thinner the cloud or the greater the wind speed, the easier the cloud thickness change is affected by the wind, and more frequent observations are required at this time; therefore, the independent variable cloud thickness YH is used as the denominator and the wind speed FS is used as the numerator, so that the dependent variable dynamic time interval DG decreases as the cloud thickness YH thins and decreases as the wind speed FS increases.

[0086] It should be noted that cloud types include cirrus clouds, stratocumulus clouds, altocumulus clouds, fractostratus clouds, fractocumulus clouds, cumulonimbus clouds, etc.; the types of clouds that are prone to interference include: clouds such as cirrus clouds, stratocumulus clouds, altocumulus clouds, fractostratus clouds, and fractocumulus clouds that are easily dispersed by the wind.

[0087] It should be noted that β is the amplitude adjustment coefficient of the artificially set exp() function. β is used to adjust the influence degrees of cloud thickness YH and wind speed F on the dynamic time interval DG; when other conditions remain unchanged, the larger β is, the larger the value of the dynamic time interval DG is, and the smaller β is, the smaller the value of the dynamic time interval DG is.

[0088] In this application, the wind condition data at the next analysis time point is predicted based on the wind condition data at the current analysis time point, including:

[0089] Extract the wind direction, wind speed, air pressure, temperature, and humidity of the target area in the historical reference data from the database, as well as the wind direction and wind speed at the corresponding next analysis time point;

[0090] Integrate the wind direction, wind speed, air pressure, temperature, and humidity, as well as the wind direction and wind speed at the next analysis time point into a number of training data and test data; use the training data to train the artificial intelligence model, use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally, obtain a wind condition prediction model with the wind direction, wind speed, air pressure, temperature, and humidity as the input and the wind direction and wind speed at the corresponding next analysis time point as the output; among them, the artificial intelligence model includes an RNN neural network model and a Transformer neural network model;

[0091] Input the current wind direction, wind speed, air pressure, temperature, and humidity of the target area into the wind condition prediction model to obtain the wind direction and wind speed at the next analysis time point.

[0092] Specifically, the specific steps of using the test data to test the trained artificial intelligence model and adjusting the artificial intelligence model according to the test results are as follows:

[0093] Input the wind direction, wind speed, air pressure, temperature, and humidity in the inspection data into the trained artificial intelligence model to obtain the wind direction and wind speed at the corresponding next analysis time point. Compare the wind direction and wind speed at the corresponding next analysis time point with the wind direction and wind speed at the corresponding next analysis time point in the inspection data. When the angular deviation of the wind direction and the difference in wind speed are less than their respective set thresholds, no parameter adjustment is required, and the inspection of the next set of inspection data is carried out; if either the angular deviation or the difference in wind speed is not less than the corresponding threshold, the corresponding parameters are adjusted until both the angular deviation and the difference in wind speed are within the corresponding thresholds, and then the inspection of the next set of inspection data is carried out. When the number of inspection data with angular deviation and wind speed difference less than the threshold obtained from all inspection data accounts for 90% or more of the total amount of inspection data, a wind condition prediction model with input of wind direction, wind speed, air pressure, temperature, and humidity and output of wind direction and wind speed at the next analysis time point is obtained.

[0094] It should be noted that the historical reference data includes wind direction, wind speed, air pressure, temperature, and humidity, as well as the wind direction and wind speed at the next analysis time point set manually according to the wind direction, wind speed, air pressure, temperature, and humidity.

[0095] Please refer to Figure 3 , in this application, the cloud data of the target area at the next analysis time point includes:

[0096] Extract the cloud thickness, cloud type, wind direction, and wind speed in the historical data of the target area from the database, as well as the cloud type and cloud thickness at the corresponding next analysis time point;

[0097] Integrate the cloud thickness, cloud type, wind direction, and wind speed, as well as the cloud type and cloud thickness at the next analysis time point into several training data and inspection data; use the training data to train the artificial intelligence model, use the inspection data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally, obtain a thickness prediction model with input of cloud thickness, cloud type, wind direction, and wind speed and output of cloud type and cloud thickness at the corresponding next analysis time point; among them, the artificial intelligence model includes an RNN neural network model and a Transformer neural network model;

[0098] Input the current cloud thickness, cloud type, wind direction, and wind speed of the target area into the thickness prediction model to obtain the cloud type and cloud thickness at the next analysis time point.

[0099] Specifically, the specific steps of using the inspection data to test the trained artificial intelligence model and adjusting the artificial intelligence model according to the test results are as follows:

[0100] Input the cloud thickness, cloud type, wind direction, and wind speed in the inspection data into the trained artificial intelligence model to obtain the cloud type and cloud thickness at the corresponding next analysis time point, and compare whether the cloud type at the corresponding next analysis time point is consistent with the cloud type at the corresponding next analysis time point in the inspection data;

[0101] Yes, compare the cloud thickness at the corresponding next analysis time point with the cloud thickness at the corresponding next analysis time point in the inspection data. When the difference between the two is less than the set threshold, no parameter adjustment is required, and the inspection of the next set of inspection data is carried out; if the difference between the two is not less than the set threshold, the corresponding parameters are adjusted until the difference between the two is less than the set threshold, and then the inspection of the next set of inspection data is carried out;

[0102] No, adjust the corresponding parameters until the cloud type at the corresponding next analysis time point is consistent with the cloud type at the corresponding next analysis time point in the inspection data. Then compare the cloud thickness at the corresponding next analysis time point with the cloud thickness at the corresponding next analysis time point in the inspection data. When the difference between the two is less than the set threshold, no parameter adjustment is required, and the inspection of the next set of inspection data is carried out; if the difference between the two is not less than the set threshold, the corresponding parameters are adjusted until the difference between the two is less than the set threshold, and then the inspection of the next set of inspection data is carried out;

[0103] When the number of inspection data with the difference in cloud thickness less than the threshold accounts for 90% or more of the total amount of inspection data, a thickness prediction model with the input of cloud thickness, cloud type, wind direction, and wind speed and the output of the cloud type and cloud thickness at the next analysis time point is obtained.

[0104] It should be noted that the historical data includes cloud thickness, cloud type, wind direction, and wind speed, as well as the cloud type and cloud thickness at the next analysis time point set manually according to the cloud thickness, cloud type, wind direction, and wind speed.

[0105] In this application, the transmittance at the next analysis time point is obtained by analyzing the cloud data at the next analysis time point, including:

[0106] C1: Extract the cloud thickness at the next analysis time point and determine whether the cloud thickness exceeds the defined maximum value; if yes, update the cloud thickness using the defined maximum value; if no, update the cloud thickness by adding zero and jump to C2; where the defined maximum value is obtained through experience;

[0107] C2: Extract the extinction coefficient θ of the cloud type at the next analysis time point and the updated cloud thickness L from the database, and obtain the transmittance LT at the next analysis time point based on the formula LT = exp(-(θ × L)).

[0108] Exemplarily, in this embodiment, if the extinction coefficient θ of the cloud type at the next analysis time point is 0.005 m-1 and the updated cloud thickness L is 100 m, then based on the formula LT = exp(-(θ × L)) = exp(-(0.005 × 100)) = 0.606, the transmittance LT at the next analysis time point is obtained as 0.606.

[0109] In this application, the photovoltaic power prediction value is obtained based on the transmittance at the next analysis time point, including:

[0110] Extract the transmittance LT at the next analysis time point and the ground solar irradiance G0 when there is no cloud in the target area, and obtain the irradiance G at the next analysis time point based on the formula G = G0 × LT;

[0111] Obtain several battery temperatures W1 of historical irradiances identical to the irradiance G from the database, and the corresponding battery temperature W2 at the next analysis time point;

[0112] Based on the formula CW = W2 - W1, obtain several temperature differences CW, and based on the formula obtain several difference change rates CB; extract the percentile value DZ1, average value JZ1, and mode ZS1 of several temperature differences CW, and based on the formula obtain the standard temperature difference BW; extract the percentile value DZ2, average value JZ2, and mode ZS2 of several difference change rates CB, and based on the formula obtain the standard change rate BL; extract the current battery temperature DW of the photovoltaic panel, and based on the formula WD = DW + (ρ3 × BW + ρ4 × DW × BL), obtain the battery temperature WD at the next analysis time point; where ρ1, ρ2, ρ3, and ρ4 are all proportional adjustment coefficients greater than 0, and ρ1 + ρ2 = 1, ρ3 + ρ4 = 1; the percentile of the percentile value is obtained by manual setting;

[0113] Obtain the area M of the current photovoltaic panel, and based on the formula obtain the photovoltaic power prediction value P at the next analysis time point; where σ is the nominal efficiency of the photovoltaic module, is the temperature coefficient, F1 is the system loss factor, and BD is the standard test temperature set manually.

[0114] It should be noted that the present invention uses the percentile value when calculating the standard temperature difference and the standard change rate. The reason for adding this value is that: the system can set different percentiles according to the requirements of the photovoltaic panel, so that the obtained characteristic values can adapt to different photovoltaic panels, increasing the flexibility of the system.

[0115] It should be noted that there may be outliers or extreme points in the data set, which may have a greater impact on statistics such as the average. Using percentile values can reduce the interference of these outliers and improve the accuracy of system analysis.

[0116] It should be noted that in the present invention, the percentile value is the value at a specific percentile; for example, if the data is sorted from small to large, and the artificially set percentile is 60%, then the data at the 60% position in the sorted order is the percentile value. If there is no data at the 60% position, the data closest to the 60% position is used as the percentile value.

[0117] Exemplarily, in this embodiment, the area M of the photovoltaic panel is 50 square meters, the nominal efficiency σ of the photovoltaic module is 0.18, and the temperature coefficient The irradiance G is 1000 W / square meter, the system loss factor F1 is 0.9, the battery temperature WD at the next analysis time point is 40 °C, and the standard test temperature BD is 25 °C. Based on the formula

[0118] In this application, the warning level is set based on the predicted value of photovoltaic power, including:

[0119] Extract the usage duration of the current photovoltaic panel in use, and extract the duration correction coefficient from the database based on the usage duration; perform a multiplication operation on the duration correction coefficient and the standard power threshold to obtain the dynamic power threshold; wherein, the standard power threshold is set through experience;

[0120] Extract the predicted value P of photovoltaic power at the next analysis time point, and determine whether the predicted value P of photovoltaic power exceeds the dynamic power threshold; if yes, mark the warning level as level one and issue a warning that the photovoltaic power at the next analysis time point exceeds the standard; if no, mark the warning level as level two and do not issue a warning.

[0121] An embodiment of the second aspect of the present invention provides a distributed photovoltaic power prediction method based on cloud data, including the following steps:

[0122] S1: Set the target area corresponding to the photovoltaic panel and obtain the basic data of the target area;

[0123] S2: Set the analysis time point of the current photovoltaic panel, and predict the wind condition data at the next analysis time point through the wind condition data at the current analysis time point; predict the cloud data in the target area at the next analysis time point based on the cloud data at the current analysis time point and the wind condition data at the next analysis time point;

[0124] S3: Analyze the cloud data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtain the predicted value of the photovoltaic power based on the light transmittance at the next analysis time point, and set the warning level based on the predicted value of the photovoltaic power.

[0125] Some of the data in the above formula are calculated by removing the dimension and taking their numerical values. The formula is obtained by software simulation of a large amount of collected data to get a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0126] The working principle of the present invention:

[0127] The present invention sets the target area corresponding to the photovoltaic panel and obtains the basic data of the target area; sets the analysis time point of the current photovoltaic panel, which dynamically adjusts the monitoring interval according to the clouds and wind speed in different time and space, so that the analysis and power prediction of the photovoltaic panel can dynamically adapt to the time and space conditions of the location of the photovoltaic panel; predicts the wind condition data at the next analysis time point based on the wind condition data at the current analysis time point; predicts the cloud data in the target area at the next analysis time point based on the cloud data at the current analysis time point and the wind condition data at the next analysis time point; analyzes the cloud data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtains the predicted value of the photovoltaic power based on the light transmittance at the next analysis time point, which can reduce the risk of power grid fluctuations caused by sudden changes in short-term clouds; sets the warning level based on the predicted value of the photovoltaic power.

[0128] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A distributed photovoltaic power prediction system based on cloud layer data, characterized in that Including: A cloud layer analysis module, as well as a data collection module, a power prediction module, and a database connected thereto; The data collection module: used to set the target area corresponding to the photovoltaic panel and obtain the basic data of the target area; wherein, the basic data includes cloud layer data and wind condition data; the cloud layer data includes cloud layer thickness and cloud layer type, and the wind condition data includes wind direction and wind speed; The cloud layer analysis module: used to set the analysis time point of the current photovoltaic panel, predict the wind condition data at the next analysis time point through the wind condition data at the current analysis time point; predict the cloud layer data of the target area at the next analysis time point based on the cloud layer data at the current analysis time point and the wind condition data at the next analysis time point; The power prediction module: used to analyze the cloud layer data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtain the photovoltaic power prediction value based on the light transmittance at the next analysis time point, and set the warning level based on the photovoltaic power prediction value.

2. The distributed photovoltaic power prediction system based on cloud layer data according to claim 1, wherein The setting of the target area corresponding to the photovoltaic panel includes: Extracting the geographical latitude φ of the photovoltaic panel, as well as the solar declination δ and hour angle ω corresponding to the current time from the database, and obtaining the solar altitude angle α based on formula (1), and formula (1) is specifically: sinα = sinφ·sinδ + cosφ·cosδ·cosω (1); Obtaining the solar azimuth angle β based on formula (2), and formula (2) is specifically: Extracting the average cloud layer moving speed V and average cloud layer height H at the current time of the location of the photovoltaic panel from historical data, and obtaining the maximum viewing angle radius R based on formula (3), and formula (3) is specifically: R = arctan(V·Δt / H) (3); Wherein, Δt is the time window; Taking the location of the photovoltaic panel as the observation point, marking the sky area with the real-time solar position (α,β) as the center and the maximum viewing angle radius R as the angular range as the target area.

3. The distributed photovoltaic power prediction system based on cloud layer data according to claim 1, characterized in that The obtaining of the basic data of the target area includes: Real-time obtaining of the cloud layer thickness and cloud layer type in the target area through a space remote sensing system, and real-time obtaining of the wind direction and wind speed in the target area through a ground meteorological station at the location of the photovoltaic panel.

4. The distributed photovoltaic power prediction system based on cloud layer data according to claim 1, wherein, The setting of the analysis time point of the current photovoltaic panel includes: A1: Obtain the current time, and judge whether there is an analysis time point within the next n seconds; if yes, mark the analysis time point existing within n seconds as the current analysis time point, and jump to A3; if no, jump to A2; A2: Judge whether the time obtained by adding the standard time interval BG to the previous analysis time point is before the current time; if yes, set the current time as the current analysis time point, and jump to A3; if no, set the time where the previous analysis time point plus the standard time interval BG is located as the current analysis time point, and jump to A3; A3: Judge whether there are clouds in the target area at the current time and whether the wind speed is not 0; if yes, jump to A4; if no, perform time addition calculation on the current analysis time point and the standard time interval BG to obtain the next analysis time point of the current photovoltaic panel; A4: Determine whether the cloud type of the current photovoltaic panel in the current target area is an interference-prone type; if yes, jump to A5; if no, perform time addition calculation on the current analysis time point and the standard time interval BG to obtain the next analysis time point of the current photovoltaic panel; where the interference-prone types include: cirrus clouds, stratocumulus clouds, altocumulus clouds, fractostratus clouds, and fractocumulus clouds; A5: Mark the cloud type of the current photovoltaic panel in the current target area as the target type, extract several historical cloud thicknesses and historical wind speeds corresponding to the target type in the historical data of the current target area from the database, integrate the historical cloud thicknesses into the B1 data group, and integrate the historical wind speeds into the B2 data group; sequentially obtain the variances of each data group, and determine whether the variance of the data group is less than the corresponding judgment threshold; if yes, perform average value calculation on the data group to obtain the characteristic value; if no, remove the data in the data group that has the largest difference from the mode, and re-perform variance judgment until the variance of the data group is less than the judgment threshold, and then perform average value calculation on the remaining data in the data to obtain the characteristic value; A6: Mark the characteristic value of the B1 data group as the characteristic thickness TH, and mark the characteristic value of the B2 data group as the characteristic wind speed TF; extract the cloud thickness YH and wind speed FS in the current target area, and obtain the dynamic time interval DG based on the formula DG = BG × β × exp(-(γ1 × TH / YH + γ2 × FS / TF)); where BH is the standard thickness, BF is the standard wind speed; β is the amplitude adjustment coefficient of the exp() function, and the value range of β is (0, 2]; both γ1 and γ2 are proportional adjustment coefficients greater than 0, and γ1 + γ2 = 1; A7: Perform time addition calculation on the current analysis time point and the dynamic time interval DG to obtain the next analysis time point of the current photovoltaic panel.

5. The distributed photovoltaic power prediction system based on cloud layer data according to claim 1, characterized in that The prediction of the wind condition data at the next analysis time point through the wind condition data at the current analysis time point includes: Extract the wind direction, wind speed, air pressure, temperature, and humidity in the historical reference data of the target area from the database, as well as the wind direction and wind speed at the corresponding next analysis time point; Integrate the wind direction, wind speed, air pressure, temperature, and humidity, as well as the wind direction and wind speed at the next analysis time point into several training data and test data; use the training data to train the artificial intelligence model, use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a wind condition prediction model with the input of wind direction, wind speed, air pressure, temperature, and humidity and the output of the wind direction and wind speed at the corresponding next analysis time point; where the artificial intelligence model includes an RNN neural network model and a Transformer neural network model; Input the current wind direction, wind speed, air pressure, temperature, and humidity of the target area into the wind condition prediction model to obtain the wind direction and wind speed at the next analysis time point.

6. The distributed photovoltaic power prediction system based on cloud layer data according to claim 1, wherein The prediction of the cloud data in the target area at the next analysis time point includes: Extract the cloud thickness, cloud type, wind direction, and wind speed in the historical data of the target area from the database, as well as the cloud type and cloud thickness at the corresponding next analysis time point; Integrate cloud thickness, cloud type, wind direction and wind speed, as well as cloud type and cloud thickness at the next analysis time point into a number of training data and test data; use the training data to train an artificial intelligence model, use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a thickness prediction model with cloud thickness, cloud type, wind direction and wind speed as inputs and cloud type and cloud thickness at the corresponding next analysis time point as outputs; wherein, the artificial intelligence model includes an RNN neural network model and a Transformer neural network model; Input the current cloud thickness, cloud type, wind direction and wind speed of the target area into the thickness prediction model to obtain the cloud type and cloud thickness at the next analysis time point.

7. The distributed photovoltaic power prediction system based on cloud layer data according to claim 6, characterized in that, The analysis of the cloud data at the next analysis time point to obtain the transmittance at the next analysis time point includes: C1: Extract the cloud thickness at the next analysis time point, and judge whether the cloud thickness exceeds the defined maximum value; if yes, update the cloud thickness with the defined maximum value; if no, perform a zero addition update on the cloud thickness and jump to C2; C2: Extract the extinction coefficient θ of the cloud type at the next analysis time point and the updated cloud thickness L from the database, and obtain the transmittance LT at the next analysis time point based on the formula LT = exp(-(θ × L)).

8. The distributed photovoltaic power prediction system based on cloud layer data according to claim 1, characterized in that The obtaining of the photovoltaic power prediction value based on the transmittance at the next analysis time point includes: Extract the transmittance LT at the next analysis time point and the ground solar irradiance G0 when the target area is cloudless, and obtain the irradiance G at the next analysis time point based on the formula G = G0 × LT; Obtain several battery temperatures W1 of historical irradiances identical to the irradiance G and the corresponding battery temperature W2 at the next analysis time point from the database; Based on the formula CW = W2 - W1, a number of temperature differences CW are obtained. Based on the formula a number of difference change rates CB are obtained; the percentile value DZ1, average value JZ1, and mode value ZS1 of a number of temperature differences CW are extracted. Based on the formula the standard temperature difference BW is obtained; the percentile value DZ2, average value JZ2, and mode value ZS2 of a number of difference change rates CB are extracted. Based on the formula the standard change rate BL is obtained; the current cell temperature DW of the photovoltaic panel is extracted. Based on the formula WD = DW + (ρ3×BW + ρ4×DW×BL), the cell temperature WD at the next analysis time point is obtained; where ρ1, ρ2, ρ3, and ρ4 are all proportional adjustment coefficients greater than 0, and ρ1 + ρ2 = 1, ρ3 + ρ4 = 1; Obtain the area M of the current photovoltaic panel, and based on the formula obtain the predicted value P of the photovoltaic power at the next analysis time point; where, σ is the nominal efficiency of the photovoltaic module, is the temperature coefficient, F1 is the system loss factor, and BD is the standard test temperature.

9. The distributed photovoltaic power prediction system based on cloud layer data according to claim 8, characterized in that, The setting of the warning level based on the photovoltaic power prediction value includes: Extract the service life of the current photovoltaic panel in use, and extract the duration correction coefficient from the database based on the service life; perform a multiplication operation on the duration correction coefficient and the standard power threshold to obtain the dynamic power threshold; Extract the photovoltaic power prediction value P at the next analysis time point, and judge whether the photovoltaic power prediction value P exceeds the dynamic power threshold; if yes, mark the warning level as level one and issue a warning that the photovoltaic power at the next analysis time point exceeds the standard; if no, mark the warning level as level two and do not issue a warning.

10. A distributed photovoltaic power prediction method based on cloud data, which operates based on the distributed photovoltaic power prediction system according to any one of claims 1 to 9, and is characterized in that: S1: Set the target area corresponding to the photovoltaic panel and obtain the basic data of the target area; S2: Set the analysis time point of the current photovoltaic panel, and predict the wind condition data at the next analysis time point through the wind condition data at the current analysis time point; predict the cloud data of the target area at the next analysis time point based on the cloud data at the current analysis time point and the wind condition data at the next analysis time point; S3: Analyze the cloud layer data at the next analysis time point to obtain the light transmittance at the next analysis time point, obtain the predicted value of photovoltaic power based on the light transmittance at the next analysis time point, and set the warning level based on the predicted value of photovoltaic power.

Images