A method, system and computer device for photovoltaic power forecasting

Abstract

The present application relates to the technical field of photovoltaic system, solve the technical problem that the prior art cannot accurately predict irradiance and temperature, the prediction process is easy to appear gradient disappearance or gradient explosion, especially relates to a kind of photovoltaic power prediction method, system and computer equipment, the method includes the following steps: S1, obtain the cloud atlas data set from the weather database in the place of photovoltaic equipment;S2, select several zenith cloud atlas from cloud atlas data set and obtain ground cloud atlas set consisting of several ground cloud atlas by preprocessing;The present application obtains the moving curve of cloud layer in the sky by image analysis to zenith cloud atlas, and establishes prediction model according to moving curve, can avoid the problem of gradient disappearance or gradient explosion of neural network model, also carries out real-time cloud layer detection to zenith cloud atlas, so that the prediction result is more accurate, improves the accuracy of irradiance, makes the prediction process more smooth, improves work efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of photovoltaic systems, and particularly relates to a photovoltaic power prediction method, system and computer device. BACKGROUND

[0002] The solar irradiance is the radiant energy per unit area per unit time of the sun's radiation after absorption, scattering and reflection of the atmosphere, and its unit is watt per square meter (W / m2). The photovoltaic power is related to the solar irradiance falling on its surface. The solar irradiance and temperature are very critical to the power generation of the photovoltaic system. The existing technology uses a neural network algorithm model to predict the irradiance and temperature, which is prone to inaccurate prediction. The main reason is that the neural network algorithm is prone to gradient disappearance or gradient explosion when predicting the nonlinear change of the irradiance. If many factors are considered for the prediction of the temperature, the processing speed is slow, which further leads to inaccurate prediction of the photovoltaic power and reduces the working efficiency of the photovoltaic power generation. SUMMARY

[0003] In view of the defects of the prior art, the present application provides a photovoltaic power prediction method, system and computer device, which solves the technical problems that the prior art cannot accurately predict the irradiance and temperature, and the gradient disappearance or gradient explosion is prone to occur in the prediction process. The present application achieves the purpose of accurately predicting the irradiance and temperature of the location of the photovoltaic system by using a new prediction model, and further improves the accuracy of the photovoltaic power prediction.

[0004] To solve the above technical problems, the present application provides the following technical scheme: a photovoltaic power prediction method, which comprises the following steps:

[0005] S1. Obtaining a cloud atlas data set composed of a plurality of zenith cloud maps from a meteorological database of a location of a photovoltaic device;

[0006] S2. Preprocessing the cloud atlas data set to obtain a ground-based cloud atlas set composed of a plurality of ground-based cloud maps;

[0007] S3. Selecting a random point on the ground-based cloud map as a cloud motion point coordinate (B a , C a ), generating a prediction trajectory S' according to the motion trajectory of the cloud motion point coordinate (B a , C a ), and optimizing the prediction trajectory S' to obtain an optimized trajectory S;

[0008] S4. Selecting a prediction time H, finding a prediction point E a (F e , K f ) on the optimized trajectory S matching the prediction time H, and obtaining a prediction result according to the prediction point E a(F e , K f ) to obtain a predicted irradiance value J;

[0009] S5, calculating a wind speed value V s in a region where the photovoltaic device is located in a current stage;

[0010] S6, calculating a ground heat value T according to the wind speed value V s ;

[0011] S7, predicting a power P y of the photovoltaic device according to the ground heat value T and the predicted irradiance value J.

[0012] Preferably, in step S2, the following steps are specifically implemented:

[0013] S21, obtaining a first zenith cloud image from a plurality of zenith cloud images, and filtering the first zenith cloud image to obtain a filtered image by a mean filtering method;

[0014] S22, defining a value of 365 days as a day number N, and defining a value of 24 hours as an hour value L;

[0015] S23, calculating a solar declination angle ε and a solar hour angle ω according to the day number N and the hour value L, and the calculation formula is as follows:

[0016]

[0017] ω = 15 × (L-12)

[0018] Wherein, ε represents the solar declination angle, and ω represents the solar hour angle;

[0019] S24, obtaining a latitude value of a location where the photovoltaic device is located from a meteorological database According to the solar declination angle ε and the solar hour angle ω, calculating a solar position angle γ s and a zenith angle θ a , and the calculation formula is as follows:

[0020]

[0021] Wherein, θ a represents the zenith angle, and γ s represents the solar position angle;

[0022] S25, according to the change of the solar position angle γ s and the zenith angle θ a , taking a time when the first zenith cloud image is located as a starting point, selecting a plurality of continuous adjacent time zenith cloud images and constructing a zenith cloud image set;

[0023] S26. Set the area with missing information in the first zenith cloud map as the missing area Ω, set the area with valid information in the zenith cloud map I as the valid area Φ, and set the boundary line between the missing area Ω and the valid area Φ as the boundary line ΦΩ.

[0024] S27. Convert the first zenith cloud image to an RGB image using an image processing library, and select any pixel on the boundary line ΦΩ as the point to be repaired. a Point p awaiting repair a Select the image block k to be repaired as the center;

[0025] S28. Randomly select a pixel in the effective region Φ as the matching point. And obtain the matching block k' that is the same size as the block k to be repaired, and calculate the point p to be repaired. a with matching points The color difference CCM between the two is calculated using the following formula:

[0026]

[0027] Where CCM represents the point p to be repaired. a with matching points The color difference between them, R(p) a ), G(p a ) and B(p a ) indicates the point p to be repaired. a The red, green, and blue color values, and Indicates matching point The red, green, and blue color values;

[0028] S29. Obtain the color difference threshold Y using an empirical method. s And based on the color difference threshold Y s Repair the image block k to be edited;

[0029] If CCM <Y s If so, then copy the matching block k' and replace the block k to be repaired;

[0030] If CCM≥Y s If the repair requirements are not met, the process returns to step S28.

[0031] S210. Repeat steps S28 to S29 until the zenith cloud atlas is repaired and the ground cloud atlas is obtained.

[0032] Preferably, in step S3, the specific implementation steps are as follows:

[0033] S31, select the ground cloud image as the first ground cloud image, and establish a two-dimensional coordinate system with the center point (e, f) of the first ground cloud image as the origin, and obtain the first cloud image on the first ground cloud image by edge detection method;

[0034] S32, randomly select a cloud motion point a in the center position of the cloud layer in the first cloud image S (B a , C a );

[0035] S33, select the second ground cloud image at time ∈-1 from the ground cloud image set, and obtain the second cloud image by step S31, and map the cloud motion point a of the second cloud image to the first cloud image as the mapping point a (B, C) with the center point of the second cloud image as the reference point; S S1 (B b , C b );

[0036] S34, repeat step 33 to obtain a plurality of mapping points a Sm = [a S1 , a S2 ……aS m] , according to the coordinates of a plurality of mapping points a Sm , a predicted trajectory S' is obtained by curve fitting method, and the expression of the predicted trajectory S' is:

[0037] S' = W g B x 2 +W k B x +b

[0038] Wherein, W g and W k represent the trajectory coefficient, and b represents the bias;

[0039] S35, find the predicted point (Y d , Z d ) matching the predicted time on the predicted trajectory S', find the ground cloud image at the predicted time in the ground cloud image set, and obtain the actual point (Y d’ , Z d’ ) in the ground cloud image;

[0040] S36, calculate the difference value V between the predicted point (Y d , Z d ) and the actual point (Y d’ , Z d’ ) by using the loss function;

[0041] S37, optimize the predicted trajectory S' according to the difference threshold M;​

[0042] If V < M, it is expected, end and get the optimized trajectory S;

[0043] If V ≥ M, it is not expected and return to step S33.

[0044] Preferably, the difference value V is calculated as follows:

[0045]

[0046] Wherein, V represents the difference value between the predicted point (Y d , Z d ) and the actual point (Y d’ , Z d’ ), and n represents the logarithm of the predicted point (Y d , Z d ) and the actual point (Y d’ , Z d’ ).

[0047] Preferably, in step S4, the specific implementation steps are as follows:

[0048] In step S4, the specific implementation steps are as follows:

[0049] S41, calculate the distance Δd 1j of the predicted point E a (F e , K f ) to the center point (e, f), and the calculation formula is as follows:

[0050]

[0051] Wherein, Δd 1j represents the distance of the jth predicted point E a (F e , K f ) to the center point (e, f);

[0052] S42, obtain the sun center point SU z (L h , Q k ) in the first cloud image, calculate the distance Δd 2j of the predicted point E a (F e , K f ) to the sun center point SU z (L h , Q k ), and the calculation formula is as follows:

[0053]

[0054] Wherein, Δd2j Represents the j-th prediction point E a (F a K f ) to the center of the sun SU z (L h Q k The distance;

[0055] S43, Based on the predicted point E a (F e K f The corresponding grayscale value Gr b Distance Δd 1j and distance Δd 2j Calculate cloud cover value LO f The calculation formula is as follows:

[0056]

[0057] Among them, Lo f The value represents cloud cover, and w represents the amount of data.

[0058] S44, Based on cloud cover value LO f The predicted irradiance value J is calculated using the following formula:

[0059] J = I po ×cosθ gd ×(1-LO f )

[0060] Where J represents the predicted irradiance value, I po θ represents clear-sky irradiance. gd This indicates the solar altitude angle.

[0061] Preferably, in step S5, the specific implementation steps are as follows:

[0062] S51. Select a first basic direction μ from several directions on the horizontal plane where the photovoltaic system is located, and take a direction in the plane that is perpendicular to the first basic direction μ as the second basic direction π.

[0063] S52. Measure the detection distance S between two detection points a and b in the first basic direction μ. μ The detection distance D between the two detection points c and e in the second fundamental direction π. π ;

[0064] S53. Record the downwind transmission time t of the ultrasonic wave between two detection points in the first basic direction μ. ab and headwind transmission time t ba ;

[0065] S54, Based on the detection distance S μ, the downwind transmission time t ab , the upwind transmission time t ba , the average wind speed of the first basic direction μ in the photovoltaic power generation time is calculated The calculation formula is as follows:

[0066]

[0067] Wherein, , the average wind speed in the photovoltaic power generation time, l represents the detection distance S μ , the downwind transmission time t ab , the upwind transmission time t ba , the number of groups, S μ , the μth detection distance;

[0068] S55, the downwind transmission time t ce , and the upwind transmission time t ec between two detection points of the ultrasonic wave in the second basic direction π is obtained;

[0069] S56, the average wind speed of the second basic direction π is calculated The calculation formula is as follows:

[0070]

[0071] Wherein, , the average wind speed of the second basic direction π, y represents the detection distance D π , the downwind transmission time t ce , the upwind transmission time t ec , the number of groups, D π , the πth detection distance;

[0072] S57, the wind speed value V s is calculated according to the average wind speed and the average wind speed , and the calculation formula is as follows:

[0073]

[0074] Wherein, V s represents the wind speed value.

[0075] Preferably, in step S6, the following steps are specifically implemented:

[0076] S61, the power generation time period of the photovoltaic power generation equipment is obtained, and the power generation time period is equally divided into a plurality of time units t f ;

[0077] S62, the ground surface reflected heat value R s is calculated according to the ground surface thermal energy data, and the calculation formula is as follows:

[0078] R s = α x J

[0079] wherein R s represents a ground surface reflected heat value, α represents a ground surface reflectivity, and J represents a predicted irradiance value;

[0080] S63, calculating a ground surface heat value T according to the ground surface reflected heat value R s , and a wind speed value V s , and a calculation formula is as follows:

[0081]

[0082] wherein t f1 and t f2 represent two time units, dT represents a differential of the ground surface heat value T, dt represents a differential of time, ∈ represents a ground surface emissivity, σ c represents a Stefan-Boltzmann constant, C p represents an air specific heat capacity, ρ represents an air density, T b represents an atmospheric temperature, Δz s represents an effective value of a ground surface thickness layer, and C m represents a ground surface heat capacity.

[0083] Preferably, in step S7, the following steps are specifically implemented:

[0084] S71, calculating a power current I p according to the predicted irradiance value J, and a calculation formula is as follows:

[0085]

[0086] wherein I p represents the power current, S N represents a standard irradiance, T represents the ground surface heat value, T1 represents a standard temperature, D T represents a current coefficient, I s represents a short-circuit current under the condition of the standard irradiance S N and the standard temperature T1, and J represents the predicted irradiance value;

[0087] S72, calculating a power voltage U p according to the power current I p , and a calculation formula is as follows:

[0088]

[0089] wherein U p represents the power voltage, B s represents a number of series of photovoltaic cells, and B pA represents a diode parameter, K represents a Boltzmann constant, T0 represents a power generation temperature, I0 represents an internal equivalent reverse diode saturation current of the photovoltaic cell, and E represents an elementary charge;

[0090] S73、According to the power voltage U p and the power current I p , the power P y is calculated, and the calculation formula is as follows:

[0091] P y =U p ×I p

[0092] wherein P y represents the power of the photovoltaic system.

[0093] The technical scheme also provides a system for the above-mentioned photovoltaic power prediction method, and the system comprises:

[0094] A data acquisition module acquires a cloud image data set from a meteorological database at a location of a photovoltaic device;

[0095] A preprocessing module selects a plurality of zenith cloud images from the cloud image data set and performs preprocessing to obtain a ground-based cloud image set composed of a plurality of ground-based cloud images;

[0096] A trajectory optimization module selects a point on the ground-based cloud image as a cloud movement point coordinate (B a , C a ), generates a prediction trajectory S' according to a movement trajectory of the cloud movement point coordinate (B a , C a ), and optimizes the prediction trajectory S' to obtain an optimized trajectory S;

[0097] An irradiance value prediction module selects a prediction time H, finds a prediction point E a (F e , K f ) on the optimized trajectory S that matches the prediction time H, and obtains a prediction irradiance value J according to the prediction point E a (F e , K f );

[0098] A wind speed analysis module calculates a wind speed value V s in a region where the photovoltaic device is located at a current stage;

[0099] A heat calculation module calculates a ground surface heat value T according to the wind speed value V s ;

[0100] A photovoltaic power prediction module predicts the power P y of the photovoltaic device according to the ground surface heat value T and the prediction irradiance value J.

[0101] The technical scheme also provides a computer device, which comprises a processor and a memory connected through a system bus.

[0102] By the above technical scheme, the application provides a photovoltaic power prediction method, system and computer device, which have at least the following beneficial effects:

[0103] 1. The application obtains the moving curve of the cloud layer in the sky through image analysis of the zenith cloud picture, establishes a prediction model according to the moving curve, predicts the cloud amount factor according to the relationship between the cloud layer and the position of the sun, and obtains the irradiance through the cloud amount factor and the solar elevation angle. This method can not only avoid the problems of gradient disappearance or gradient explosion of the neural network model, but also perform real-time cloud layer detection on the zenith cloud picture, so that the prediction result is more accurate, the accuracy of the irradiance is improved, the prediction process is smoother, and the work efficiency is improved.

[0104] 2. The application can more accurately obtain the wind speed of the photovoltaic system position by establishing the transmission time of the measuring ultrasonic wave in the upwind or downwind direction, and then obtain the temperature of the photovoltaic system through the accurate wind speed, so that the predicted photovoltaic power is more accurate, the accuracy of the photovoltaic power prediction is improved, and the operation efficiency of the photovoltaic power station is higher.

[0105] 3. In the measurement of the ground surface heat value, the measurement value is calculated in time units, and the ground surface radiation and reflected heat are combined, so that the ground surface heat value can be more accurate, the situation that the calculation deviation of the ground surface heat value is too large due to the change of the daytime temperature is avoided, and the prediction result of the photovoltaic power is affected. BRIEF DESCRIPTION OF DRAWINGS

[0106] The accompanying drawings, which are included to provide a further understanding of the application and constitute a part of this application, illustrate certain illustrative embodiments of the application and together with the description serve to explain the application. In the drawings:

[0107] Figure 1 It is a flow chart of the photovoltaic power prediction method of the application;

[0108] Figure 2 It is a structural block diagram of the photovoltaic power prediction system of the application;

[0109] Figure 3 It is a structural block diagram of the computer device of the application;

[0110] Figure 4 It is an example diagram of the zenith cloud picture of the application;

[0111] Figure 5 It is an example diagram of the repaired zenith cloud picture of the application. DETAILED DESCRIPTION

[0112] In order to make the above objectives, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments. By this means, the implementation process of how to apply technical means to solve technical problems and achieve technical effects can be fully understood and implemented.

[0113] Since the prior art cannot accurately predict irradiance and temperature, the prediction process is prone to the technical problems of gradient disappearance or gradient explosion. Please refer to Figure 1 Figure 5 The embodiment provides a photovoltaic power prediction method, system and computer device, which can accurately predict the irradiance and temperature of the location of a photovoltaic system through a new prediction model, and further improves the accuracy of photovoltaic power prediction. The method comprises the following steps:

[0114] S1, obtaining a cloud atlas data set composed of a plurality of zenith cloud maps from a meteorological database of a location of a photovoltaic device, the cloud atlas data set comprising a plurality of images of zenith cloud maps, as shown in Figure 4 The shooting of the zenith cloud map is through a camera at the top, which vertically shoots pictures towards the hemispherical mirror surface below. The reflection of the hemispherical mirror surface can shoot the cloud layer state of the entire sky. The superposition of a plurality of cloud layer pictures can better reflect the movement track of the cloud layer. Each zenith cloud map is marked with an accurately recorded time, which is convenient for later prediction.

[0115] S2, obtaining a ground cloud atlas set composed of a plurality of ground cloud maps by preprocessing the cloud atlas data set. The cloud atlas data is collected by a cloud photographing device. In actual application, a full-sky imager TSI-880 is used to photograph the cloud atlas of the sky. Since a sun position baffle is arranged in the image of the full-sky imager in order to block the sun, and the support of the camera can also be seen in the picture, as shown in Figure 4 Due to the blocking effect of the sun position baffle and the camera support, there must be a missing area in the image, which is prone to cause inconvenience to the fitting of the cloud layer movement track in the later stage. Therefore, the picture needs to be preprocessed. In step S2, the specific implementation steps are as follows:

[0116] S21, obtaining a first zenith cloud map from a plurality of zenith cloud maps, and filtering the first zenith cloud map by a mean filtering method to obtain a filtered image. The present application only needs to use the mean filtering method to achieve the expected effect, and the mean filtering method is a commonly used filtering method, which will not be described here.

[0117] S22, defining the value of 365 days as a day number N, and defining the value of 24 hours as an hour value L. In order to facilitate calculation, only the values of the day number and the hour number are used for calculation. ​

[0118] S23, calculate the solar declination angle ε and the solar hour angle ω before and after the time to be predicted according to the day number N and the hour value L, and the calculation formula is as follows:

[0119]

[0120] ω = 15 × (L-12)

[0121] Wherein, ε represents the solar declination angle, ω represents the solar hour angle, and in the later cloud track prediction, the time to be predicted is determined first, and the solar declination angle ε and the solar hour angle ω of three days to one week before and after the time to be predicted can be obtained in practice.

[0122] S24, obtain the latitude value of the location of the photovoltaic equipment from the meteorological database According to the solar declination angle ε and the solar hour angle ω, the solar position angle γ is calculated s And the zenith angle θ a , and the calculation formula is as follows:

[0123]

[0124] Wherein, θ a represents the zenith angle, γ s represents the solar position angle, and in the zenith cloud chart, the solar position can be better determined through the solar position angle γ s And the zenith angle θ a , which is also convenient for determining the position of the sun in the cloud image in the later stage.

[0125] S25, according to the change of the solar position angle γ s And the zenith angle θ a , take the time of the first zenith cloud chart as the starting point, select a plurality of continuous adjacent time zenith cloud charts and form a zenith cloud chart set, and the images in the cloud chart data set must be time adjacent and continuous ground cloud chart images, which can more accurately predict the cloud layer.

[0126] S26, set the region with missing information in the first zenith cloud chart as the missing region Ω, and set the region with effective information in the zenith cloud chart I as the effective region Φ, and set the boundary line of the missing region Ω and the effective region Φ as the boundary line ΦΩ. Due to the shielding of the equipment, the actual part of the image must be repaired.

[0127] S27, convert the first zenith cloud chart into an RGB image through an image processing library, select an arbitrary pixel point on the boundary line ΦΩ as a repair point p a , and select a repair block k centered on the repair point p a , and the point plane method can make the repair process more accurate and improve the prediction accuracy in the later stage.

[0128] S28. Randomly select a pixel in the effective region Φ as the matching point. And obtain the matching block k' that is the same size as the block k to be repaired, and calculate the point p to be repaired. a with matching points The color difference CCM between the two is calculated using the following formula:

[0129]

[0130] Where CCM represents the point p to be repaired. a with matching points The color difference between them, R(p) a ), G(p a ) and B(p a ) indicates the point p to be repaired. a The red, green, and blue color values, and Indicates matching point The red, green, and blue color values;

[0131] S29. Obtain the color difference threshold Y using an empirical method. s And based on the color difference threshold Y s Repair the image block k to be edited;

[0132] If CCM <Y s If so, then copy the matching block k' and replace the block k to be repaired;

[0133] If CCM≥Y s If the repair requirements are not met, the process returns to step S28.

[0134] S210. Repeat steps S28 to S29 until the zenith cloud atlas is repaired and a ground-based cloud atlas is obtained. It is necessary to repair all zenith cloud images here to facilitate later prediction of cloud movement trajectories. Based on the above steps, missing information in the zenith cloud images can be quickly repaired. Figure 4 The image shown is the repaired zenith cloud image, which makes the zenith cloud image more accurate in the later step of fitting the cloud movement trajectory and avoids the situation where the prediction step is terminated or fails due to incomplete images.

[0135] S3. Select any point on the ground-based cloud map as the coordinates of the cloud movement point (B). a C a According to the coordinates of the cloud moving point (B) a C a) to generate a prediction trajectory S', and the prediction trajectory S' is optimized to obtain an optimized trajectory S. The motion trajectory of the cloud layer is predicted by fitting a curve, and the cloud center points at different times are mapped on the same ground cloud image to complete the fitting point taking. Since the existing technology cannot accurately predict the irradiance and temperature, the technical problems of gradient disappearance or gradient explosion are prone to occur in the prediction process. In order to obtain more accurate prediction results, in step S3, the specific implementation steps are as follows:

[0136] S31, the selected ground cloud image is taken as a first ground cloud image, a two-dimensional coordinate system is established with the center point (e, f) of the first ground cloud image as the origin, and a first cloud image is obtained on the first ground cloud image by edge detection method. The present application can obtain the expected effect only by the edge detection method, and the edge detection method is a commonly used method for obtaining edges, which will not be described here. The cloud head image can be obtained by the edge detection method.

[0137] S32, the center position of the cloud layer in the first cloud image is arbitrarily selected as a cloud motion point a S (B a , C a ), in the point taking process, the geometric center of the cloud layer is generally obtained by a geometric method, and a plurality of motion points can also be obtained on the cloud layer, and a plurality of curves are synthesized to make the prediction of the cloud layer more accurate. The present application takes one motion point as an illustrative object.

[0138] S33, taking the time of the first cloud image as ∈, selecting a second ground cloud image at time ∈-1 from the ground cloud image set and obtaining a second cloud image by step S31, and mapping the cloud motion point a S of the second cloud image to the first cloud image as a mapping point a S1 (B d , C b ) with the center point of the second cloud image as the reference point;

[0139] S34, repeating step 33 to obtain a plurality of mapping points a Sm =[a S1 , a S2 ……a Sm ], and the prediction trajectory S' is obtained by curve fitting according to the coordinates of the plurality of mapping points a Sm . The expression of the prediction trajectory S' is:

[0140] S'=W g B x 2 +W k B x +b

[0141] Wherein, W g and Wk represents the trajectory coefficient, b represents the bias, the zenith cloud chart is a circular spherical chart, and the movement of the cloud layer is generally a parabola or a circular arc, and the parabola is used for illustration, and the trajectory of the circular arc can be modified as a circular curve expression.

[0142] S35, find a prediction point (Y d , Z d ) matching the prediction time on the predicted trajectory S'; d d , find a ground cloud chart of the prediction time in the ground cloud chart set and obtain an actual point (Y a’ , Z d’ ) in the ground cloud chart;

[0143] S36, calculate the difference V between the prediction point (Y d , Z d ) and the actual point (Y d’ , Z d’ ), and the calculation formula is as follows:

[0144]

[0145] Wherein, V represents the difference between the prediction point (Y d , Z d ) and the actual point (Y d’ , Z d’ ), and n represents the logarithm of the prediction point (Y d , Z d ) and the actual point (Y d’ , Z d’ ).

[0146] S37, optimize the predicted trajectory S' according to the difference threshold M;

[0147] If V < M, it meets the expectation, ends and obtains the optimized trajectory S;

[0148] If V ≥ M, it does not meet the expectation and returns to step S33, and steps S36-S37 can further optimize the fitting curve, so that the fitting curve is more accurate, the accuracy of the final prediction is increased, this step can obtain the preliminary trajectory curve of the cloud layer in the sky, and the preliminary prediction trajectory S' of the cloud center point trajectory fitting curve is established according to the trajectory curve, and the optimized trajectory S is finally obtained through optimization of the preliminary trajectory curve, so that the prediction result of the method for the cloud layer trajectory is more accurate.

[0149] S4, select a prediction time H, find a prediction point E a (F e , K f ) matching the prediction time H on the optimized trajectory S, and obtain a preliminary trajectory S' of the cloud center point trajectory fitting curve according to the prediction point E a (F e , Kf ) to obtain the predicted irradiance value J, since the prior art cannot accurately predict irradiance and temperature, the prediction process is prone to technical problems of gradient disappearance or gradient explosion, the embodiment takes time as the prediction variable, more accurately predicts the moving track of the cloud layer, in step S4, the specific implementation steps are as follows:

[0150] S41, calculate the prediction point E a (F e , K f ) to the center point (e, f) Distance Δd 1j , the calculation formula is as follows:

[0151]

[0152] Where, Δd 1j represents the distance from the jth prediction point E a (F e , K f ) to the center point (e, f) ;

[0153] S42, obtain the sun center point SU z (L h , Q k ) in the first cloud image, calculate the distance Δd a (F e , K f ) from the prediction point E z (L h , Q k ) to the sun center point SU 2j , the calculation formula is as follows:

[0154]

[0155] Where, Δd 2j represents the distance from the jth prediction point E a (F e , K f ) to the sun center point SU z (L h , Q k ) ;

[0156] S43, calculate the cloud amount value LO a (F e , K f ) according to the gray value Gr b , the distance Δd 1j and the distance Δd 2j corresponding to the prediction point E f , the calculation formula is as follows:

[0157]

[0158] wherein, LO f represents the cloud cover value, w represents the data quantity;

[0159] S44, according to the cloud cover value LO f , the predicted irradiance value J is calculated, and the calculation formula is as follows:

[0160] J=I po ×cosθ gd ×(1-LO f )

[0161] wherein, J represents the predicted irradiance value, I po represents the clear sky irradiance, θ gd represents the solar elevation angle, the clear sky irradiance I po and the solar elevation angle θ gd can be directly obtained through a meteorological database, the present application obtains the moving curve of the cloud layer in the sky through image analysis of the zenith nephogram, establishes a prediction model according to the moving curve, predicts the cloud cover factor according to the relationship between the cloud layer and the position of the sun, and obtains the irradiance through the cloud cover factor and the solar elevation angle. This method avoids the problems of gradient disappearance or gradient explosion of the neural network model, and performs real-time cloud layer detection on the zenith nephogram, so that the prediction result is more accurate, the accuracy of the irradiance is improved, the prediction process is more smooth, and the work efficiency is improved.

[0162] S5, the wind speed value V s in the region where the photovoltaic device is located in the current stage is calculated, in actual measurement, the wind speed measured by using the single-direction method lacks accuracy due to the change of the wind speed, in order to obtain more accurate measurement value, in step S5, the specific implementation steps are as follows:

[0163] S51, a first basic direction mu is selected from a plurality of directions of a horizontal plane where the photovoltaic system is located, a direction perpendicular to the first basic direction mu in the plane is taken as a second basic direction pi, the first basic direction mu can be randomly taken on the horizontal plane, and the two perpendicular directions can facilitate the final calculation of the wind speed.

[0164] S52, the detection distance S μ between two detection points a and b in the first basic direction mu and the detection distance D π between two detection points c and e in the second basic direction pi are measured, the distance between the detection points a and b or the detection points c and e can be measured by various methods, for example, the method of using sound wave ranging or the method of manual measurement can obtain the detection distance.

[0165] S53, the downwind transmission time t ab and the upwind transmission time t ba, the transmission time of the ultrasonic wave is accurately timed by a timing device, the timing device includes a timing chip of model TDC-GP2, other methods capable of timing can be used, the transmission time of the ultrasonic wave is different when the ultrasonic wave propagates in the wind and against the wind, the physical phenomenon can make the wind speed measurement more accurate, and the ultrasonic wave used in the application is the most commonly used medium frequency ultrasonic wave.

[0166] S54, according to the detection distance S μ , the transmission time t ab and the transmission time t ba of the wind, calculate the average wind speed of the first basic direction μ in the photovoltaic power generation time The calculation formula is as follows:

[0167]

[0168] Among them, indicates the average wind speed in the photovoltaic power generation time, and l indicates the detection distance S μ , the transmission time t ab and the transmission time t ba of the wind, S μ indicates the μth detection distance, since the measured distance does not consider the direction factor, the average wind speed is calculated.

[0169] S55, obtain the transmission time t ce and the transmission time t ec of the wind between two detection points of the second basic direction π of the ultrasonic wave;

[0170] S56, calculate the average wind speed of the second basic direction π The calculation formula is as follows:

[0171]

[0172] Among them, indicates the average wind speed of the second basic direction π, and y indicates the detection distance D π , the transmission time t ec and the transmission time t ec of the wind, D π indicates the πth detection distance,

[0173] S57, according to the average wind speed and the average wind speed , calculate the wind speed value V s , the calculation formula is as follows:

[0174]

[0175] Among them, V s indicates the wind speed value, since the average wind speed and the average wind speed The direction is vertical, so when calculating the wind speed value, the Pythagorean theorem is used to directly calculate the acquisition, the present application can more accurately obtain the wind speed of the photovoltaic system position by establishing the transmission time of the measuring ultrasonic wave upwind or downwind, and then the temperature of the photovoltaic system is obtained through the accurate wind speed, so that the predicted photovoltaic power is more accurate, the accuracy of the photovoltaic power prediction is improved, and the operation efficiency of the photovoltaic power station is higher.

[0176] S6, according to the wind speed value V s Calculate the ground heat value T, the ground heat value of the position of the photovoltaic system has a direct impact on the power, when calculating the ground heat value, the predicted irradiance value J and the wind speed value V s The prior art is too simple in calculating the heat value, and the reference physical quantity is too small, so the calculation of the heat value is not accurate, in order to improve the accuracy of the heat value, in step S6, the specific implementation steps are as follows:

[0177] S61, obtain the power generation time period of the photovoltaic power generation equipment, and divide the power generation time period into several time elements t f Dividing the time into several time elements can improve the calculation accuracy, so that the ground heat value T is more accurate, and the ground heat value T is prevented from fluctuating to cause a large error in the result.

[0178] S62, calculate the ground reflected heat value R according to the ground heat energy data s The calculation formula is as follows:

[0179] R s =α×J

[0180] Wherein, R s Indicates the ground reflected heat value, alpha indicates the ground reflectivity, J indicates the predicted irradiance value;

[0181] S63, calculate the ground heat value T according to the ground reflected heat value R s , the wind speed value V s The calculation formula is as follows:

[0182]

[0183] Wherein, t f1 And t f2 Indicate two time elements, dT indicates the differential of the ground heat value T, dt indicates the differential of time, epsilon indicates the ground emissivity, sigma c Indicates the Stefan-Boltzmann constant, C p Indicates the specific heat capacity of air, rho indicates the air density, T b Indicates the atmospheric temperature, Delta z S Indicates the effective value of the ground thickness layer, Cm The ground heat value T is calculated by using the integral method after the time is equally divided into several time units, the inaccurate result caused by simple calculation can be avoided, the measurement value is calculated by time unit in the measurement of the ground heat value, and the heat of the ground radiation and reflection is combined, so that the ground heat value can be more accurate, the situation that the ground heat value calculation deviation is too large caused by the temperature change in the day is avoided, and the prediction result of the photovoltaic power is affected.

[0184] S7, predict the power P of the photovoltaic device according to the ground heat value T and the predicted irradiance value J y The photovoltaic power is a relatively complex physical quantity, which needs to be calculated according to a large number of values, and since the prior art cannot accurately predict the irradiance and temperature, the technical problems of gradient disappearance or gradient explosion are prone to occur in the prediction process, in order to make the photovoltaic power prediction more accurate, in step S7, the following steps are specifically implemented:

[0185] S71, calculate the power current I according to the predicted irradiance value J p The calculation formula is as follows:

[0186]

[0187] Wherein, I p represents the power current, S N represents the standard irradiance, T represents the ground heat value, T1 represents the standard temperature, D T represents the current coefficient, I s represents the short-circuit current under the condition of standard irradiance S N and standard temperature T1, J represents the predicted irradiance value, and the physical quantity related to the current voltage can be directly obtained from the photovoltaic power generation database, without complex calculation.

[0188] S72, calculate the power voltage U p according to the power current I p The calculation formula is as follows:

[0189]

[0190] Wherein, U p represents the power voltage, B s represents the number of photovoltaic cell series, B p represents the number of photovoltaic cell parallel, A represents the diode parameter, K represents the Boltzmann constant, T0 represents the power generation temperature, I0 represents the internal equivalent reverse diode saturation current of the photovoltaic cell, E represents the elementary charge, and the size of the elementary charge is 1.6×10 -19 C, the Boltzmann constant is 1.380649×10 -23J / K, the diode parameters are obtained according to the nameplate of the diode in the photovoltaic system.

[0191] S73, according to the power voltage U p and the power current I p Calculate the power P y , the calculation formula is as follows:

[0192] P y = U p * I p

[0193] Wherein, P y represents the power of the photovoltaic system, the present application obtains the moving curve of the cloud layer in the sky through image analysis of the zenith cloud picture, and establishes a prediction model according to the moving curve, and then predicts the cloud amount factor according to the relationship between the cloud layer and the position of the sun, and obtains the irradiance through the cloud amount factor and the solar elevation angle. This method can not only avoid the problems of gradient disappearance or gradient explosion of neural network model, but also detect the cloud layer of the zenith cloud picture in real time, so that the prediction result is more accurate, the accuracy of the irradiance is improved, the prediction process is more smooth, and the work efficiency is improved.

[0194] Please refer to Figure 2 , which is a structural block diagram of the photovoltaic power prediction system provided by the embodiment, which comprises a data acquisition module, a pretreatment module, a trajectory optimization module, an irradiance value prediction module, a wind speed analysis module, a heat calculation module and a photovoltaic power prediction module.

[0195] The data acquisition module acquires the cloud picture data set from the meteorological database of the location of the photovoltaic equipment, the pretreatment module selects several zenith cloud pictures from the cloud picture data set and pretreats to obtain a ground cloud picture set composed of several ground cloud pictures, the trajectory optimization module selects a point on the ground cloud picture as a cloud moving point coordinate (B a , C a ), generates a prediction trajectory S' according to the motion trajectory of the cloud moving point coordinate (B a , C a ), and optimizes the prediction trajectory S' to obtain an optimized trajectory S, the irradiance value prediction module selects a prediction time H, finds a prediction point E a (F e , K f ) on the optimized trajectory S matching the prediction time H, and obtains a predicted irradiance value J according to the prediction point E a (F e , K f ), the wind speed analysis module calculates the wind speed value V s of the region where the photovoltaic equipment is located at the current stage, the heat calculation module calculates the wind speed value V sA ground heat value T is calculated by a photovoltaic power prediction module, which predicts a power P of a photovoltaic device according to the ground heat value T and a predicted irradiance value J y .

[0196] Please refer to Figure 3 The embodiment further provides a computer device, which comprises a processor, a memory and a storage medium connected through a system bus. The processor is used to provide computing and control capabilities. The memory comprises the storage medium and the internal memory. The storage medium can be a non-volatile storage medium or a volatile storage medium. The storage medium stores an operating system and computer readable instructions. When the computer readable instructions are executed by the processor, the processor can implement the photovoltaic power prediction method. The internal memory provides an environment for the operating system and the computer readable instructions in the storage medium. The internal memory can also store computer readable instructions, which can be executed by the processor to implement the photovoltaic power prediction method. The network interface of the computer device is used to communicate with an external server through a network connection. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0197] Those skilled in the art can understand that all or part of the steps of the above-mentioned embodiment methods can be completed by programs instructing relevant hardware, therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product in the form of being implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0198] The above embodiments are used to introduce the present application in detail, and the principles and implementation manners of the present application are described by applying specific examples. The above embodiment descriptions are only used to help understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will be changed, and the above description should not be understood as limiting the present application.

Claims

1. A method of forecasting photovoltaic power, characterized in that, The method comprises the following steps: S1, obtaining a cloud atlas dataset composed of several zenith cloud maps from a meteorological database of a location of a photovoltaic device; S2, preprocessing the cloud atlas dataset to obtain a ground-based cloud atlas set composed of several ground-based cloud maps; S3, select a point on the ground cloud image as the cloud motion point coordinate (B a , C a ), generate a predicted trajectory S' according to the motion trajectory of the cloud motion point coordinate (B a , C a ), and optimize the predicted trajectory S' to obtain an optimized trajectory S; S4, selecting a prediction time H, finding a prediction point E on the optimized trajectory S that matches the prediction time H a (F e , K f ) and obtaining a predicted irradiation value J from the prediction point E a (F e , K f ) S5, calculate the wind speed value V in the region where the current stage photovoltaic device is located s ; S6. According to the wind speed value V s calculating the surface heat value T; S7. predicting the power P of the photovoltaic device from the surface heat value T and the predicted irradiance value J y .

2. The method of prediction of photovoltaic power according to claim 1, characterized in that, In step S2, the following steps are implemented: S21, obtaining a first zenith cloud map from several zenith cloud maps, and filtering the first zenith cloud map by a mean filtering method to obtain a filtered image; S22, defining a value corresponding to 365 days as a day number N, and a value corresponding to 24 hours as an hour value L; S23, calculating a solar declination angle ε and a solar hour angle ω according to the day number N and the hour value L, and the calculation formula is as follows: ω = 15 × (L-12) Wherein, ε represents the solar declination angle, and ω represents the solar hour angle; S24. Obtain a latitude value of a location of the photovoltaic device from a meteorological database Calculate the solar position angle γ according to the solar declination ε and the solar hour angle ω s and the zenith angle θ a The calculation formula is as follows: where θ a denotes the zenith angle, γ s denotes the solar position angle; S25, according to the solar position angle γ s and the change of zenith angle θ a , taking the time when the first zenith cloud picture is located as the starting point, a number of continuous adjacent time zenith cloud pictures are selected and constitute a zenith cloud picture set; S26, setting a region with missing information in the first zenith cloud map as a missing region Ω, setting a region with effective information in the zenith cloud map I as an effective region Φ, and setting an intersection line of the missing region Ω and the effective region Φ as an intersection line ΦΩ; S27, convert the first day top cloud picture into an RGB image through an image processing library, select any pixel point on the boundary line ΦΩ as a to-be-repaired point p a , take the to-be-repaired point p a as the center to select a to-be-repaired block k; S28, select a pixel point in the effective area Φ as a matching point and obtain a matching block k' equal in size to the to-be-repaired block k, and calculate the to-be-repaired point p a and the color distance CCM between the matching point , and the calculation formula is as follows: where CCM denotes the color distance between the matching point p a and the point p to be inpainted, R(p a ), G(p a ) and B(p a ) denote the red, green and blue color values of the point p a to be inpainted, and and denote the red, green and blue color values of the matching point p . S29, obtain color difference threshold Y by empirical method s and according to color difference threshold Y s repair the to-be-repaired patch k; If CCM<Y s then the matching block k' is copied and replaces the patch k to be repaired; If CCM≥ Y s then the repair request is not met and return to step S28; S210, repeating steps S28 to S29 until the zenith cloud atlas is repaired, and obtaining the ground-based cloud atlas set.

3. The method of prediction of photovoltaic power according to claim 1, characterized in that, In step S3, the following steps are implemented: S31, selecting a ground-based cloud map as a first ground-based cloud map, establishing a two-dimensional coordinate system with a center point (e, f) of the first ground-based cloud map as an origin, and obtaining a first cloud layer image on the first ground-based cloud map by an edge detection method; S32, randomly selecting a cloud movement point a in the center position of the cloud layer in the first cloud layer image S (B a , C a ); S33, taking the time of the first cloud layer image as ∈, selecting the second ground cloud image of time ∈-1 from the ground cloud image set and obtaining the second cloud layer image through step S31, taking the midpoint of the second cloud layer image as a reference point, mapping the cloud movement point a S of the second cloud layer image to the first cloud layer image as a mapping point a S1 (B b , C b ); S34, repeat step 33 to obtain several mapping points a Sm = [a S1 , a S2 ... a Sm ], a predicted trajectory S' is obtained by curve fitting according to the coordinates of the several mapping points a Sm , and the expression of the predicted trajectory S' is: S' = W g B x 2 +W k B x +b where W g and W k denote trajectory coefficients, and b denotes a bias. S35, find a predicted point (Y d , Z d ) on the predicted trajectory S' that matches the predicted time, and find a ground cloud chart at the predicted time in the set of ground cloud charts and obtain an actual point (Y d’ , Z d’ ) in the ground cloud chart. S36, calculate the difference V between the predicted point (Y d , Z d ) and the actual point (Y d’ , Z d’ ); S37, optimizing the predicted trajectory S according to the difference threshold value M; If V < M, it meets the expectation, and the optimization is ended to obtain an optimized trajectory S; If V ≥ M, it does not meet the expectation and returns to step S33.

4. The method of prediction of photovoltaic power according to claim 3, characterized in that, The calculation formula of the difference value V is as follows: where V represents the difference between the predicted point (Y d , Z d ) and the actual point (Y d’ , Z d’ ), and n represents the logarithm of the predicted point (Y d , Z d ) and the actual point (Y d’ , Z d’ ).

5. The method of prediction of photovoltaic power according to claim 1, characterized in that, In step S4, the following steps are implemented: S41, calculate the prediction point E a (F e , K f ) to the center point (e, f) distance Δd 1j , the calculation formula is as follows: where Δd 1j represents the distance from the jth prediction point E a (F e , K f ) to the center point (e, f); S42, obtaining a sun center point SU in the first cloud layer image z (L h , Q k ), calculating a distance Δd a (F e , K f ) of the prediction point E z (L h , Q k ) to the sun center point SU 2j , and the calculation formula is as follows: where Δd 2j represents the distance from the jth prediction point E a e , K f to the sun center point SU z h , Q k .​​ S43、According to the prediction point E a (F e , K f ) corresponding gray value Gr b , distance Δd 1j and distance Δd 2j Calculate cloud value LO f , the calculation formula is as follows: where LO f represents the cloudiness value, w represents the data quantity; S44, according to the cloudiness value LO f The predicted irradiance value J is calculated according to the following formula: J = I po x cos θ gd x (1 - LO f ) where J represents the predicted irradiance value, I po represents the clear sky irradiance, θ gd represents the solar elevation angle.

6. The method of predicting photovoltaic power according to claim 1, characterized in that, In step S5, the following steps are implemented: S51, selecting a first basic direction μ from several directions of a horizontal plane where the photovoltaic system is located, and selecting a direction perpendicular to the first basic direction μ in the plane as a second basic direction π; S52, measure the detection distance S between the two detection points a, b of the first base direction μ μ , and the detection distance D between the two detection points c, e of the second base direction π π ; S53, recording the downwind transmission time t of the ultrasound waves between the two detection points in the first base direction μ ab and the upwind transmission time t ba ; S54、According to the detection distance S μ , the downwind transmission time t ab , and the upwind transmission time t ba Calculate the average wind speed of the first base direction μ in the photovoltaic power generation time The calculation formula is as follows: wherein represents the average wind speed during the photovoltaic power generation time, and l represents the detection distance S μ , the downwind transmission time t ab , and the upwind transmission time t ba , the number of groups, S μ represents the μth detection distance; S55, obtain the downwind transmission time t of the ultrasonic wave between the two detection points in the second base direction π ce and the upwind transmission time t ec ; S56, calculate the average wind speed of the second base direction π The calculation formula is as follows: wherein represents the average wind speed of the second base direction π, y represents the detection distance D π , the downwind transmission time t ce , and the upwind transmission time t ec , the number of groups, D π represents the πth detection distance; S57. Calculate the average wind speed and the average wind speed Calculate the wind speed value V s The calculation formula is as follows: where V s represents the wind speed value.

7. The method of predicting photovoltaic power according to claim 1, characterized in that, In step S6, the following steps are implemented: S61. Obtain the power generation time period of the photovoltaic power generation equipment, and divide the power generation time period into several time elements t. f ; S62, calculating the ground reflected heat value R according to the ground heat energy data s The calculation formula is as follows: R s = a x J wherein R s represents the ground surface reflection heat value, a represents the ground surface reflectivity, and J represents the predicted irradiance value. S63, calculating a surface heat value T based on the surface reflected heat value R s , a wind speed value V s The surface heat value T is calculated according to the following formula: where t f1 and t f2 represent two time elements, dT represents a differential of the surface heat value T, dt represents a differential of time, ∈ represents a surface emissivity, σ c represents a Stefan-Boltzmann constant, C p represents an air specific heat, ρ represents an air density, T b represents an atmospheric temperature, Δz s represents a surface thickness layer effective value, C m represents a surface heat capacity.

8. The method of predicting photovoltaic power according to claim 1, characterized in that, In step S7, the following steps are implemented: S71, calculating the power current I from the predicted irradiance value J p The calculation formula is as follows: where I p represents a power current, S N represents a standard irradiance, T represents a terrestrial heat value, T1 represents a standard temperature, D T represents a current coefficient, I s represents a short-circuit current under a standard irradiance S N and a standard temperature T1, and J represents a predicted irradiance value; S72, according to the power current I p The power voltage U is calculated p The calculation formula is as follows: wherein U p represents the power voltage, B s represents the number of photovoltaic cells in series, B p represents the number of photovoltaic cells in parallel, A represents a diode parameter, K represents the Boltzmann constant, T0represents the temperature of power generation, I0represents the internal equivalent reverse diode saturation current of the photovoltaic cell, and E represents an elementary charge; S73, according to the power voltage U p and the power current I p the power P is calculated y The calculation formula is as follows: P y = U p x I p where P y represents the power of the photovoltaic system.

9. A system comprising the method of predicting photovoltaic power according to any of the preceding claims 1 to 8, characterized in that, The system comprises: A data acquisition module, which acquires a cloud atlas dataset from a meteorological database of a location of a photovoltaic device; A preprocessing module, which selects several zenith cloud maps from the cloud atlas dataset and preprocesses the zenith cloud maps to obtain a ground-based cloud atlas set composed of several ground-based cloud maps; a trajectory optimization module, taking an optional point on the ground cloud image as a cloud motion point coordinate (B a , C a ), generating a prediction trajectory S' according to a motion trajectory of the cloud motion point coordinate (B a , C a ), and optimizing the prediction trajectory S' to obtain an optimized trajectory S; an irradiation value prediction module, selecting a prediction time H, finding a prediction point E on the optimized trajectory S matching the prediction time H a (F e , K f ) and obtaining a predicted irradiation value J from the prediction point E a (F e , K f ) a wind speed analysis module, which calculates a wind speed value V in the area in which the photovoltaic plant is located at the current stage s ; a heat calculation module for calculating a ground heat value T based on the wind speed value V s a heat calculation module for calculating a ground heat value T based on the wind speed value V a photovoltaic power prediction module to predict a power P of the photovoltaic device from a ground heat value T, a predicted irradiance value J y .

10. A computer device comprising a processor, a memory connected by a system bus, characterized in that, The processor implements the photovoltaic power prediction method according to claims 1-8 when executing the computer program.

Images