Method and device for discriminating output mode of photovoltaic power station based on clear sky judgment of cloud map

Abstract

The present invention belongs to the technical field of photovoltaic power station power prediction, and discloses a method and device for discriminating the output mode of a photovoltaic power station based on clear sky judgment of cloud images. It includes: selecting satellite cloud images; detecting whether there are clouds in the clear sky at the location of the photovoltaic power station; if there are clouds, the output of the photovoltaic power station will fluctuate in the next 4 hours; if there are no clouds, search for cloud mass regions around the photovoltaic power station, merge the cloud mass regions representing the same cloud mass movement trend found, and calculate the cloud masses that may move over the photovoltaic power station; perform clear sky prediction on the cloud masses that may move over the photovoltaic power station, and judge the impact on the photovoltaic power station according to the clear sky prediction result. It effectively improves the accuracy of photovoltaic power prediction at sunny times, improves the output capacity of the photovoltaic power station, and has no reference significance for power system dispatching.

Description

Technical Field

[0001] The present invention belongs to the technical field of photovoltaic power station power prediction, and particularly relates to a method and device for discriminating the output mode of a photovoltaic power station based on clear sky judgment of cloud images. Background Art

[0002] By making a judgment on the future weather at the location of the photovoltaic power station, the future output mode of the photovoltaic power station can be better predicted. If the moments when there is no cloud mass blocking the power station in the future can be judged, then the output mode of the photovoltaic system can be accurately predicted according to the result combined with the clear sky model. Currently, the information sources for weather judgment in the methods applied to photovoltaic power prediction mostly use numerical weather prediction or ground-based sky cloud images. For ultra-short-term photovoltaic power prediction, it is required to predict the output of the photovoltaic power station in the next four hours every 15 minutes. Due to the problems of update time and frequency, numerical weather prediction cannot well adapt to the ultra-short-term prediction time scale of four hours. For ground-based sky cloud images, only a small part of the sky image can be captured, as an estimate of the future situation within the next half hour to one hour, and it also cannot well adapt to the ultra-short-term prediction of photovoltaic power.

[0003] The most advanced atmospheric observation satellite in China at present is equipped with multi-channel scanning imaging radiometers, interferometric atmospheric vertical sounders, lightning imagers, space environment monitoring instruments, etc., and can obtain 14-channel cloud images in total. It has made color satellite cloud images for the first time, generates regional observation images as fast as once a minute, and the highest spatial resolution can reach 500m. According to the different band frequencies of the satellite scanner, satellite cloud images can be divided into two categories: visible light satellite cloud images, water vapor channels, and infrared satellite cloud images. The wavelength range of visible light cloud images is 0.55 - 0.75um, and the wavelength range of infrared cloud images is 10.5 - 12.5um. The one between them is generally called the water vapor channel, and the ones commonly used corresponding to ground irradiance are mostly visible light cloud images and infrared cloud images.

[0004] Visible light cloud images are obtained by using the scanning radiometer of a meteorological satellite to capture the sunlight reflected by the cloud top and the ground. Therefore, they can only be imaged during the day when there is light irradiating the cloud top or the ground. The thickness information of the cloud layer can be obtained from visible light cloud images. Thick cloud layers have strong reflection ability and appear white in visible light cloud images with large gray values; on the contrary, when the cloud layer is thinner, the cloud image appears dark gray with smaller image gray values.

[0005] An infrared cloud image can be obtained by converting the radiation in the infrared band into an image with an infrared measuring instrument, which can represent the temperature at the corresponding position. When the cloud layer is higher, the temperature of the cloud layer will be relatively low, and it appears bright white on the infrared cloud image; conversely, when the cloud mass is close to the ground, the temperature of the cloud layer will be relatively high, and the corresponding position on the infrared cloud image appears dark gray. Therefore, the infrared cloud image can judge the temperature of the cloud mass through the gray scale and then judge the height of the cloud top. Since infrared remote sensing does not require the reflection of visible light, the infrared cloud image is not limited by the fact that the visible light cloud image can only be imaged during the day. It can perform remote sensing day and night and provide more information than the visible light cloud image. However, due to current technical limitations, the resolution of the infrared cloud image is often lower than that of the visible light cloud image. Taking the Fengyun-4 satellite as an example, only the visible light cloud image has a resolution of up to 500m, while the resolution of the infrared cloud image near 12um is at most 4km.

[0006] In summary, the visible light cloud image can provide effective information only when there is sufficient visible light irradiation during the day. The infrared cloud image can provide the position information of the cloud mass at all times. However, due to current technical limitations, the resolution of the infrared cloud image is often lower than that of the visible light cloud image. Therefore, neither of the above two methods can well solve the problem of low prediction accuracy of photovoltaic power under different output modes, and cannot provide a suitable application scenario for the sunny-day power prediction algorithm of photovoltaic power plants, cannot effectively improve the prediction accuracy of photovoltaic power at sunny moments, and the accuracy of the clear sky prediction model is relatively low, which has no reference significance for power system scheduling. Summary of the Invention

[0007] The purpose of the present invention is to provide a method and device for discriminating the output mode of a photovoltaic power plant based on cloud image clear sky judgment to solve the problems in the prior art;

[0008] To achieve the above purpose, the present invention provides the following technical solutions:

[0009] A method for discriminating the output mode of a photovoltaic power plant based on cloud image clear sky judgment, including:

[0010] Select the corresponding satellite cloud image area according to the longitude and latitude of the power plant to be predicted and the prediction time;

[0011] Detect whether there are clouds over the location of the photovoltaic power station within the selected satellite cloud image area; if there are clouds, the output of the photovoltaic power station will fluctuate in the next 4 hours; if there are no clouds, search for the cloud cluster areas around the photovoltaic power station, and merge the cloud cluster areas representing the same cloud cluster movement trend, and judge whether the merged cloud cluster will move over the photovoltaic power station and cause fluctuations in the output of the photovoltaic power station in the next 4 hours. If the merged cloud cluster moves over the photovoltaic power station, the output of the photovoltaic power station will fluctuate in the next 4 hours; if the merged cloud cluster cannot move over the photovoltaic power station, calculate the cloud clusters that may move over the photovoltaic power station; perform clear sky prediction on the cloud clusters that may move over the photovoltaic power station, and judge the output mode of the photovoltaic power station according to the clear sky prediction result.

[0012] Further, the method for selecting the satellite cloud image is as follows:

[0013] According to the longitude and latitude of the location of the photovoltaic power station and the date of the moment to be predicted, judge the sunrise time when the photovoltaic power station starts to generate power with the sunrise and the sunset time when the power generation starts to decrease to 0 with the sunset on the same day; calculate the cloud image at the prediction time t between the preset time after sunrise and the preset time before sunset using the infrared cloud image, and calculate the cloud image for the rest of the day using the visible light cloud image, and select two cloud images imgt and img(t - 1) with sampling times within one hour of the prediction time t as the cloud images for analysis and detection.

[0014] Further, the preset time is 1.5 hours.

[0015] Further, the method for detecting whether there are clouds in the clear sky at the location of the photovoltaic power station is as follows:

[0016] Taking the photovoltaic power station as the center, detect cloud pixels in the cloud image within a range of 300×300 km 2 around the photovoltaic power station;

[0017] If cloud pixels are detected, it is a non-clear sky mode, there are clouds in the clear sky at the location of the photovoltaic power station, and the output of the photovoltaic power station will fluctuate within the next 4 hours; if no cloud pixels are detected, there are no clouds in the clear sky at the location of the photovoltaic power station, and search for the cloud cluster areas around the photovoltaic power station.

[0018] Further, the method for searching for the cloud cluster areas around the photovoltaic power station is as follows:

[0019] Taking the photovoltaic power station as the center, arbitrarily select one of the two cloud images imgt and img(t - 1), and start a search every 15 degrees from the exact right of the selected cloud image, and search for a cloud image area of 60×60 km 2 with a cloud pixel ratio greater than 20% to determine the cloud cluster edge around the power station, and determine the cloud cluster area according to the determined cloud cluster edge.

[0020] Furthermore, the method for calculating the cloud masses that may move over the PV power station is as follows:

[0021] Based on the results of cloud detection in the cloud images imgt and img(t - 1), the displacement vectors of each detected cloud mass region are calculated using the block matching method; the cloud mass regions with the included angle between the displacement vector and the vector pointing to the power station less than 45 degrees are selected to form the set of cloud masses that may move over the power station.

[0022] Furthermore, the method for merging the cloud mass regions representing the movement trend of the same large cloud mass is as follows:

[0023] In the set of cloud masses that may move over the PV power station, any arbitrarily specified cloud mass region is taken as the first cluster; by calculating the included angle between the cloud displacement vector of the remaining cloud mass regions and the vector pointing to the power station, and comparing the calculated included angle with 45 degrees, if it is less than 45 degrees, the cloud mass region belongs to this cluster; if it is greater than or equal to 45 degrees, the cloud mass region has no cluster belonging, and then it is further determined whether there are cloud mass regions without cluster belonging among the remaining cloud mass regions in the cloud mass set.

[0024] If the judgment result is negative, no cloud mass region merging is required;

[0025] If the judgment result is positive, the cloud mass regions without cluster belonging are removed from the set.

[0026] Furthermore, the method for predicting and judging the future clear sky of the PV power station is as follows:

[0027] The speed is recalculated for the obtained independent cloud mass edge regions to obtain the displacement vectors of the cloud mass edges around the PV power station, and it is judged whether the future cloud mass edges will block the PV power station by calculating the cloud displacement vectors;

[0028] If it is blocked, it is not a clear day;

[0029] If it is not blocked, it is a clear day.

[0030] A discriminator for the output mode of a PV power station based on clear sky judgment of cloud images, comprising:

[0031] A selection module, configured to select the corresponding satellite cloud image region according to the longitude and latitude of the power station to be predicted and the prediction time;

[0032] A judgment module is used to detect whether there are clouds over the location of the photovoltaic power station within the selected operational satellite cloud image area; if there are clouds, the output of the photovoltaic power station will fluctuate in the next 4 hours; if there are no clouds, search the cloud cluster areas around the photovoltaic power station, and merge the cloud cluster areas representing the same cloud cluster movement trend, and judge whether the merged cloud cluster will move over the photovoltaic power station and cause fluctuations in the output of the photovoltaic power station in the next 4 hours. If the merged cloud cluster moves over the photovoltaic power station, the output of the photovoltaic power station will fluctuate in the next 4 hours; if the merged cloud cluster cannot move over the photovoltaic power station, calculate the cloud clusters that may move over the photovoltaic power station; conduct clear sky prediction on the cloud clusters that may move over the photovoltaic power station, and judge the output mode of the photovoltaic power station according to the clear sky prediction result.

[0033] Compared with the prior art, the advantages of the present invention are as follows:

[0034] A method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment of the present invention detects clouds over the location of the photovoltaic power station, searches and merges the areas with clouds, and predicts the output of the photovoltaic power station according to whether the movement trend of the merged clouds affects the clear sky at the location of the photovoltaic power station, effectively improving the accuracy of photovoltaic power prediction at sunny times, enhancing the output prediction ability of the photovoltaic power station, and having reference significance for power system dispatching. Description of the Drawings

[0035] The specification drawings forming a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0036] Figure 1 is a schematic flow chart of a method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment provided by an embodiment of the present invention;

[0037] Figure 2 is a schematic flow chart of cloud cluster area merging of a method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment of the present invention;

[0038] Figure 3 is a schematic diagram of the cloud detection result of the maximum between-class variance method of visible light cloud map of a method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment of the present invention;

[0039] Figure 4 is a schematic diagram of the cloud detection result of the maximum between-class variance method of infrared cloud map of a method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment of the present invention;

[0040] Figure 5It is a schematic diagram of the cloud cluster area search result of a method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment in the present invention;

[0041] Figure 6 It is a schematic diagram of the cloud cluster area merging result of a method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment in the present invention. Detailed implementation manners

[0042] The present invention will be described in detail below with reference to the drawings and in conjunction with embodiments. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0043] The following detailed descriptions are all exemplary descriptions, aiming to provide further detailed descriptions of the present invention. Unless otherwise specified, all technical terms adopted in the present invention have the same meaning as commonly understood by those of ordinary skill in the art to which the present invention belongs. The terms used in the present invention are only for describing specific implementation manners, and are not intended to limit the exemplary embodiments according to the present invention.

[0044] Embodiment 1

[0045] A method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment in the present invention detects clouds over the location of the photovoltaic power station, searches for and merges the areas with clouds, and predicts the output of the photovoltaic power station based on whether the movement trend of the merged clouds affects the clear sky at the location of the photovoltaic power station, effectively improving the accuracy of photovoltaic power prediction at sunny times, enhancing the output prediction ability of the photovoltaic power station, and having reference significance for power system dispatching.

[0046] A method for discriminating the output mode of a photovoltaic power station based on cloud map clear sky judgment in the present invention, as Figure 1 shown, includes:

[0047] Select the corresponding satellite cloud map area according to the longitude and latitude of the power station to be predicted and the prediction time; detect whether there are clouds over the location of the photovoltaic power station in the selected satellite cloud map area; if there are clouds, the output of the photovoltaic power station will fluctuate in the next 4 hours; if there are no clouds, search for the cloud cluster areas around the photovoltaic power station, merge the cloud cluster areas representing the same cloud cluster movement trend, and judge whether the merged cloud cluster will move over the photovoltaic power station to cause fluctuations in the output of the photovoltaic power station in the next 4 hours. If the merged cloud cluster moves over the photovoltaic power station, the output of the photovoltaic power station will fluctuate in the next 4 hours; if the merged cloud cluster cannot move over the photovoltaic power station, calculate the cloud clusters that may move over the photovoltaic power station; conduct clear sky prediction on the cloud clusters that may move over the photovoltaic power station, and judge the output mode of the photovoltaic power station according to the clear sky prediction result.

[0048] More specifically, the method for selecting satellite cloud images is as follows: Based on the longitude and latitude of the location of the photovoltaic power station and the date of the moment to be predicted, determine the sunrise time when the photovoltaic power station starts to generate power with the sunrise and the sunset time when the power generation starts to decrease to 0 with the sunset on the same day; calculate the cloud images at the prediction moment t between the preset time after sunrise and the preset time before sunset using infrared cloud images, calculate the cloud images at the remaining time of the day using visible light cloud images, and select two cloud images imgt and img(t - 1) with sampling times within one hour of the prediction moment t as the cloud images for analysis and detection.

[0049] Preferably, the preset time is 1.5 hours.

[0050] Further, the method for detecting whether there are clouds in the clear sky at the location of the photovoltaic power station is as follows: Taking the photovoltaic power station as the center, detect cloud pixels in the cloud image within a range of 300×300 km of the photovoltaic power station. 2 If cloud pixels are detected, it is a non-clear sky mode, there are clouds in the clear sky at the location of the photovoltaic power station, and the power generation of the photovoltaic power station fluctuates within the next 4 hours; if no cloud pixels are detected, there are no clouds in the clear sky at the location of the photovoltaic power station, and search for the cloud mass area around the photovoltaic power station.

[0051] Further, the method for searching for the cloud mass area around the photovoltaic power station is as follows: Taking the photovoltaic power station as the center, arbitrarily select one of the two selected cloud images, and start a search every 15 degrees from the exact right of the selected cloud image, search for a cloud image area of 60×60 km where the proportion of cloud pixels is greater than 20% to determine the edge of the cloud mass around the power station, and determine the cloud mass area according to the determined edge of the cloud mass. 2 Based on the determined edge of the cloud mass, determine the cloud mass area.

[0052] Further, the method for calculating the cloud mass that may move over the photovoltaic power station is as follows:

[0053] According to the results of cloud detection of the cloud images imgt and img(t - 1), calculate the displacement vectors of each searched cloud mass area using the block matching method; select the cloud mass areas where the included angle between the displacement vector and the vector pointing to the power station is less than 45 degrees to form the cloud mass that may move over the power station.

[0054] Further, the method for merging cloud mass areas representing the movement trend of the same large cloud mass is as follows:

[0055] In the set of cloud mass areas that may move over the photovoltaic power station, arbitrarily specify a cloud mass area as the first cluster; by calculating the magnitude of the included angle between the cloud displacement vector of the remaining cloud mass areas and the vector pointing to the power station and comparing it with 45 degrees, if it is less than 45 degrees, the cloud mass area belongs to this cluster, otherwise the cloud mass area has no cluster belonging; determine whether there are areas without cluster belonging among the remaining cloud mass areas in the cloud mass set.

[0056] If the judgment result is negative, cloud cluster region merging is not required;

[0057] If the judgment result is positive, cloud cluster regions without cluster affiliation are removed from the set.

[0058] Furthermore, the method for predicting clear skies in a future photovoltaic power station is as follows:

[0059] Recalculate the speed for the obtained independent cloud cluster edge regions to obtain the displacement vector of the cloud cluster edge around the photovoltaic power station, and judge whether the future cloud cluster edge will block the photovoltaic power station by calculating the cloud displacement vector;

[0060] If it is blocked, it is not a clear day;

[0061] If it is not blocked, it is a clear day.

[0062] More specifically, a method for discriminating the power output mode of a photovoltaic power station based on cloud map clear sky judgment according to the present invention includes:

[0063] Step 1, select a satellite cloud map;

[0064] Calculate the solar altitude angle H according to the longitude and latitude of the photovoltaic power station and the date of the moment to be predicted.

[0065] sinH s = sinφ × sinδ + cosφ × cosδ × cost (1)

[0066] Where φ is the latitude, t is the solar hour angle, and δ is the solar declination angle. The calculation formulas are as follows:

[0067] δ = 0.006918 - 0.399912cos(b) + 0.070257sin(b) - 0.006758cos(2b) + 0.000907sin(2b) - 0.002697cos(3b) + 0.00148sin(3b) (2)

[0068]

[0069]

[0070] t = 15×(real_sun_time - 12) (5)

[0071] According to the solar altitude angle H being equal to 0 degrees and 180 degrees, select the moments when the solar altitude angle H is equal to 0 degrees and 180 degrees, and select the corresponding satellite cloud image area according to the longitude and latitude of the power station to be predicted and the prediction moment; determine the sunrise moment when the photovoltaic power station starts to generate power with the sunrise and the sunset moment when the power generation starts to decrease to 0 with the sunset on the same day; when the prediction moment t is before 1.5 hours after sunrise and after 1.5 hours before sunset, use the infrared cloud image for subsequent calculation and analysis, and use the visible light cloud image for calculation at other times; select the two cloud images imgt and img(t - 1) closest to the prediction moment t for subsequent analysis.

[0072] Specifically, the power data used in this embodiment are the actual operation data of a certain photovoltaic power station from 2017 to 2018. The time resolution of the power data is 15 minutes. The satellite cloud images use certain cloud image data released by the National Meteorological Satellite Center, and the time resolution varies from four minutes to one hour. The spatial resolution of the infrared cloud image is 4 km, and the spatial resolution of the visible light cloud image is 500 m.

[0073] Step 2, cloud detection at the location of the photovoltaic power station;

[0074] Use the Otsu method to perform cloud detection on the cloud image selected in Step 1, and determine whether cloud pixels are detected in the cloud image within a range of 300×300 km centered on the power station: If cloud pixels are detected, it is determined that there will be large fluctuations in the power generation of the photovoltaic power station within the next 4 hours, and the power generation mode of the photovoltaic power station is non-clear sky; if no cloud pixels are detected, proceed to Step 3; 2 The Otsu method has a good segmentation effect on unimodal images and can meet the division between clouds (foreground) and the ground (background) required for cloud detection. The Otsu method divides the image into two parts, foreground and background, according to the gray characteristics of the image, and determines the division threshold by maximizing the variance between the two classes of foreground and background of the picture, so as to ensure that misclassification does not occur as much as possible. Specifically, in this embodiment, let T be the segmentation threshold between the foreground and the background, the proportion of foreground pixel number be ω0, and the average gray level be u0; the proportion of background points in the image be ω1, and the average gray level be u1, the total average gray level of the image be u, and the variance between the foreground and background images be g, then there are:

[0075] u = ω0u0 + ω1u1

[0076] u = ω0u0 + ω1u1

[0077] g = ω0(u0 - u) 2 + ω1(u1 - u) 2

[0078] Combining the above two equations, we can get:

[0079] g = ω0ω1(u0 - u1) 2

[0080] When the variance g reaches the maximum, the difference between the foreground and background of the image also reaches the maximum, and the gray level at this time is the segmentation threshold to be found.

[0081] Figure 3 、 Figure 4 In [reference], a, b, c, and d are the maximum between-class variance cloud detection results of the visible light cloud image and the infrared cloud image in this embodiment, respectively, verifying that the maximum between-class variance method has good cloud detection effects on both types of satellite cloud images.

[0082] Step 3, Search for the cloud cluster area around the photovoltaic power station;

[0083] According to the cloud detection result of the cloud image imgt in Step 2, starting from the due right with the photovoltaic power station as the center, a search is carried out every 15 degrees to search for a 60×60 km 2 cloud image area where the proportion of the cloud cluster is greater than 20% to determine the edge of the cloud cluster around the power station.

[0084] Step 4, Calculate the cloud cluster area that may move over the power station;

[0085] For the cloud detection results of the cloud images imgt and img(t - 1) obtained in Step 2, use the block matching method to calculate the displacement vectors of each cloud cluster area searched in Step 3 in imgt and img(t - 1); select the cloud cluster areas where the included angle between the displacement vector and the vector pointing to the power station is less than 45 degrees to form a set of cloud clusters that may move over the power station.

[0086] In this embodiment, the block matching algorithm is an algorithm for estimating the motion vectors between partial regions inside two adjacent images in time. The basic assumption of the block matching method is that the patterns of the corresponding objects and the background in one frame of the image move within the frame, forming the corresponding objects in the next frame of the image.

[0087] Based on the motion vector calculation of the block matching method, it is carried out between two adjacent frames of the current frame and the reference frame. During the block matching process, the image of the current frame is divided into a non-overlapping multiple sub-regions. These block regions are compared with the nearby regions of the corresponding blocks in the adjacent frame after being translated pixel by pixel to obtain the most similar block region, and then the motion vector is obtained.

[0088] Specifically, during the motion vector calculation of the block matching method, it is necessary to quantify the similarity degree between regions in two adjacent images. There are four most commonly used matching criteria for measuring similarity in the block matching method, namely the mean square error (MSE), the mean absolute difference (MAD), the peak signal-to-noise ratio (PSNR), and the sum of absolute differences (SAD):

[0089]

[0090]

[0091]

[0092]

[0093] where (N×N) is the size of the block region, C i,j and R i,j are the pixel values in the current image and the reference image respectively. Among all evaluation metrics, MSE is the most commonly used.

[0094] Specifically, Figure 5 This is the result display of the cloud clusters that may move over the power station in this embodiment: The red five-star is the location of the power station, and the yellow dots with arrows around it are the centers of the cloud cluster regions around the power station searched according to the algorithm. The direction of the arrow points to the direction of the motion vector calculated based on two adjacent cloud images, and the length of the arrow represents the magnitude of the motion vector. The algorithm accurately finds the boundary between the cloud clusters around the power station and the sky.

[0095] Step 5: The process of merging the cloud cluster regions representing the motion trend of the same large cloud cluster in this embodiment is as Figure 2 shown, and the main steps include:

[0096] Step A1, in the set of cloud clusters that may move over the power station obtained in Step 4, arbitrarily specify a cloud cluster region as the first cluster;

[0097] Step A2, determine whether there is a region in the remaining cloud cluster regions that has no cluster affiliation: If the judgment result is no, then no cloud cluster region merging is required; if the judgment result is yes, then go to Step A3;

[0098] Step A3, determine whether the condition that the angle between the displacement vector of the region without cluster affiliation and all regions within the cluster is less than 45 degrees and the angle between the vectors pointing to the power station is less than 90 degrees is satisfied: If the judgment result is yes, then this region is incorporated into the cluster that meets the condition; if the judgment result is no, then this region forms a new sub-cluster by itself.

[0099] Specifically, Figure 6 This is the result display of the cloud cluster region merging in this embodiment: The edges of some cloud clusters whose motion vectors point to the power station in Figure 5 are merged into a large cloud cluster edge within the red square. On the one hand, the cloud cluster edge features within this region are more unique, so the motion vector results calculated according to the block matching method will be more accurate; on the other hand, after the region merging, the motion direction of this part of the cloud cluster edge changes from multiple previous ones to a more accurate one, which is convenient for backward inference of the cloud conditions around the power station at future times along this direction.

[0100] Step 6, future clear sky prediction judgment;

[0101] For the edge region of the independent cloud clusters obtained in step 5, recalculate the velocity to obtain the displacement vector of the cloud cluster edge around the power station.

[0102] Based on the direction of the displacement vector, extrapolate the cloud cluster edge that will affect the power station in the future.

[0103] Based on this edge and the magnitude of the displacement vector, it can be inferred whether there will be large cloud pixels within 70 pixels around the power station within the next 4 hours: If there are cloud pixels, it is determined that there will be large fluctuations in the photovoltaic power station within the next 4 hours, and the output mode is non-clear sky; if there are no cloud pixels, there will be no large fluctuations in the photovoltaic power station within the next 4 hours, and the output mode is clear sky.

[0104] Specifically, in this embodiment, considering the time of processing the cloud map, 30 days throughout 2018 with clear, cloudy, and clear turning to cloudy weather conditions for a certain power station are selected as test data. Starting 4 hours before sunrise, weather type judgment and prediction are performed every 15 minutes. The weather judgment results are shown in Table 1:

[0105] Table 1

[0106]

[0107] Among them, actually clear sky means that the error between the corresponding sample and the clear sky model is less than 5%, and actually non-clear sky means that the error between the corresponding sample and the clear sky model is greater than 5%. The proportion of samples predicted as clear sky but actually non-clear sky is relatively small, avoiding the error caused by mispredicting as clear sky during non-clear sky and then using the clear sky model for prediction.

[0108] Embodiment 2

[0109] A photovoltaic power station output mode discrimination device based on cloud map clear sky judgment, comprising:

[0110] A selection module for selecting the corresponding satellite cloud map area according to the longitude and latitude of the power station to be predicted and the prediction time;

[0111] A judgment module for detecting whether there are clouds over the location of the photovoltaic power station in the selected satellite cloud map area; if there are clouds, the output of the photovoltaic power station in the next 4 hours will have fluctuating changes; if there are no clouds, search for the cloud cluster area around the photovoltaic power station, and merge the cloud cluster areas representing the same cloud cluster movement trend, and judge whether the merged cloud cluster will move over the photovoltaic power station to cause fluctuations in the output of the photovoltaic power station in the next 4 hours. If the merged cloud cluster moves over the photovoltaic power station, the output of the photovoltaic power station in the next 4 hours will have fluctuating changes; if the merged cloud cluster cannot move over the photovoltaic power station, calculate the cloud cluster that may move over the photovoltaic power station; perform clear sky prediction on the cloud cluster that may move over the photovoltaic power station, and judge the output mode of the photovoltaic power station according to the clear sky prediction result.

[0112] As is known by common technical knowledge, the present invention can be implemented by other embodiments without departing from its spirit or essential features. Therefore, the above-disclosed embodiments are illustrative in all aspects and not exclusive. All changes within the scope of the present invention or within the scope equivalent to the present invention are encompassed by the present invention.

[0113] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0114] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0115] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that realizes the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0116] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent substitutions can still be made to the specific implementation manners of the present invention, and any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for discriminating the output mode of a photovoltaic power station based on clear sky judgment of cloud images, characterized in that, Including: Selecting a corresponding satellite cloud image area according to the longitude and latitude of the power station to be predicted and the prediction time; Detecting whether there are clouds above the location of the photovoltaic power station within the selected satellite cloud image area; if there are clouds, the output of the photovoltaic power station will fluctuate in the next 4 hours; if there are no clouds, searching the cloud cluster areas around the photovoltaic power station, merging the cloud cluster areas representing the same cloud cluster movement trend, and judging whether the merged cloud cluster will move above the photovoltaic power station to cause fluctuations in the output of the photovoltaic power station in the next 4 hours. If the merged cloud cluster moves above the photovoltaic power station, the output of the photovoltaic power station will fluctuate in the next 4 hours; if the merged cloud cluster cannot move above the photovoltaic power station, calculating the cloud clusters that may move above the photovoltaic power station; making a clear sky prediction for the cloud clusters that may move above the photovoltaic power station, and judging the output mode of the photovoltaic power station according to the clear sky prediction result; The method for merging cloud cluster areas representing the same large cloud cluster movement trend is: In the set of cloud clusters that may move above the photovoltaic power station, arbitrarily specify a cloud cluster area as the first cluster; by calculating the magnitude of the angle between the cloud displacement vector of the remaining cloud cluster areas and the vector pointing to the power station, and comparing the calculated vector angle magnitude with 45 degrees. If it is less than 45 degrees, the cloud cluster area belongs to this cluster; if it is greater than or equal to 45 degrees, the cloud cluster area has no cluster belonging, and further judge whether there are areas without cluster belonging among the remaining cloud cluster areas in the cloud cluster set; If the judgment result is no, no cloud cluster area merging is required; If the judgment result is yes, the cloud cluster areas without cluster belonging are removed from the set.

2. The discriminant method for the output power mode of a photovoltaic power station based on clear sky judgment of a nephogram according to claim 1, wherein, The method for selecting the satellite cloud image is: According to the longitude and latitude of the location of the photovoltaic power station and the date of the prediction time, judging the sunrise time when the photovoltaic power station starts to have output with the sunrise and the sunset time when the output starts to decrease to 0 with the sunset on the same day; calculating the cloud image in the infrared cloud image for the prediction time t between the preset time after sunrise and the preset time before sunset, and calculating the cloud image in the visible light cloud image for the rest of the day. Select two cloud images imgt and img(t - 1) with sampling times within one hour of the prediction time t as the cloud images for analysis and detection.

3. A method for discriminating the output power mode of a photovoltaic power station based on clear sky judgment of a cloud map, characterized in that, The preset time is 1.5 hours.

4. A discriminant method for the output power mode of a photovoltaic power station based on clear sky judgment of a nephogram, characterized in that, The method for detecting whether there are clouds in the clear sky at the location of the photovoltaic power station is: Centered on a photovoltaic power station, detect cloud pixels in the cloud map within 300×300 km of the photovoltaic power station 2 within the range; If cloud pixels are detected, it is a non-clear sky mode, there are clouds in the clear sky at the location of the photovoltaic power station, and the output of the photovoltaic power station will fluctuate within the next 4 hours; if no cloud pixels are detected, there are no clouds in the clear sky at the location of the photovoltaic power station, and search the cloud cluster areas around the photovoltaic power station.

5. The discriminant method for the output power mode of a photovoltaic power station based on clear sky judgment of a nephogram according to claim 2, wherein The method for searching the cloud cluster areas around the photovoltaic power station is: Centered on a photovoltaic power station, arbitrarily select one of the two cloud images of imgt and img(t - 1) that are selected, and start a search every 15 degrees starting from the exact right side of the selected cloud image to search for a cloud image area of 60×60 km where the proportion of cloud pixels is greater than 20% 2 to determine the edge of the cloud cluster around the power station, and determine the cloud cluster area based on the determined edge of the cloud cluster.

6. The discriminant method for the output power mode of a photovoltaic power station based on clear sky judgment of a nephogram according to claim 2, characterized in that, The method for calculating the cloud clusters that may move above the photovoltaic power station is: According to the cloud detection results of the cloud images imgt and img(t - 1), calculating the displacement vectors of the searched cloud cluster areas by the block matching method; selecting the cloud cluster areas with the angle between the displacement vector and the vector pointing to the power station less than 45 degrees to form a set of cloud clusters that may move above the power station.

7. A discriminant method for the output power mode of a photovoltaic power station based on clear sky judgment of a cloud map, characterized in that The method for judging the future clear sky prediction of the photovoltaic power station is: Recalculate the velocity of the obtained independent cloud edge region to obtain the displacement vector of the cloud edge around the photovoltaic power station, and judge whether the future cloud edge will block the photovoltaic power station by calculating the cloud displacement vector; If it is blocked, it is not a sunny day; If it is not blocked, it is a sunny day.

8. A photovoltaic power station output mode discrimination device based on clear sky judgment of cloud images, characterized in that, It includes: A selection module for selecting a corresponding satellite cloud image area according to the longitude and latitude of the power station to be predicted and the prediction time; A judgment module for detecting whether there are clouds over the location of the photovoltaic power station in the selected satellite cloud image area; if there are clouds, the output of the photovoltaic power station will fluctuate in the next 4 hours; if there are no clouds, search for the cloud area around the photovoltaic power station, and merge the cloud areas representing the same cloud movement trend, and judge whether the merged cloud will move to the sky above the photovoltaic power station and cause fluctuations in the output of the photovoltaic power station in the next 4 hours. If the merged cloud moves to the sky above the photovoltaic power station, the output of the photovoltaic power station will fluctuate in the next 4 hours; if the merged cloud cannot move to the sky above the photovoltaic power station, calculate the cloud that may move to the sky above the photovoltaic power station; perform clear sky prediction on the cloud that may move to the sky above the photovoltaic power station, and judge the output mode of the photovoltaic power station according to the clear sky prediction result; The method for merging cloud areas representing the same large cloud movement trend is: In the set of cloud areas that may move to the sky above the photovoltaic power station, arbitrarily specify a cloud area as the first cluster; by calculating the magnitude of the included angle between the cloud displacement vector of the remaining cloud areas and the vector pointing to the power station, and comparing the calculated vector included angle magnitude with 45 degrees, if it is less than 45 degrees, the cloud area belongs to this cluster; if it is greater than or equal to 45 degrees, the cloud area has no cluster belonging, and further judge whether there are cloud areas without cluster belonging in the remaining cloud areas in the cloud set; If the judgment result is no, no cloud area merging is required; If the judgment result is yes, remove the cloud area without cluster belonging from the set.

Images