CN116542383A - Distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud image - Google Patents
Distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud image Download PDFInfo
- Publication number
- CN116542383A CN116542383A CN202310532250.9A CN202310532250A CN116542383A CN 116542383 A CN116542383 A CN 116542383A CN 202310532250 A CN202310532250 A CN 202310532250A CN 116542383 A CN116542383 A CN 116542383A
- Authority
- CN
- China
- Prior art keywords
- data
- photovoltaic system
- power
- cloud
- weather
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims abstract description 15
- 238000003062 neural network model Methods 0.000 claims abstract description 3
- 230000033001 locomotion Effects 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 239000013598 vector Substances 0.000 claims description 10
- 230000005855 radiation Effects 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000005259 measurement Methods 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000002310 reflectometry Methods 0.000 claims description 2
- 238000006467 substitution reaction Methods 0.000 claims description 2
- 238000012795 verification Methods 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims 1
- 238000006243 chemical reaction Methods 0.000 description 7
- 230000003287 optical effect Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000012937 correction Methods 0.000 description 5
- 238000013507 mapping Methods 0.000 description 3
- 230000032683 aging Effects 0.000 description 2
- 239000000443 aerosol Substances 0.000 description 1
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000007620 mathematical function Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000000087 stabilizing effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Educational Administration (AREA)
- Biomedical Technology (AREA)
- Operations Research (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Game Theory and Decision Science (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a distributed photovoltaic system output prediction method based on a small fluctuation weather satellite cloud picture, which comprises the following steps: step 1: dividing the regional distributed photovoltaic system according to grids; step 2: recording data; step 3: history data processing; step 4: correcting power data of a power clear sky model; step 5: calculating cloud index; step 6: normalizing the data of different time delay scales of the output power of each grid, the corrected clear sky model power, the solar space-time model data, the air quality data, the weather forecast information and the satellite cloud index operator data, and further respectively establishing an Elman neural network model by using a random training method; step 7: and inputting the latest calculated input variable parameters into the model, and superposing the predicted data of each grid to serve as output predicted data of the regional photovoltaic system.
Description
Technical Field
The invention relates to a distributed photovoltaic system output prediction method based on a small fluctuation weather satellite cloud picture, which mainly comprises a photovoltaic system power ultra-short-term prediction method under a small fluctuation scene. The method also comprises a method for modeling the regional grid segmentation of the distributed photovoltaic system, a data error processing method, a strategy for fully utilizing satellite images to improve prediction accuracy and an Elman neural network algorithm utilizing a random training method.
Background
With the development of new energy power generation industry of the photovoltaic system, the grid-connected scale of the distributed photovoltaic system is larger and larger. Because the output force of the photovoltaic system has randomness, intermittence and fluctuation, the establishment of the photovoltaic system generated power prediction system with reliable precision is particularly important. Most of the existing output prediction methods only consider a single photovoltaic system, and cannot adapt to the situation that a plurality of distributed photovoltaic systems exist in a power grid in actual situations. And secondly, the existing model cannot reflect the change information immediately under the condition of large change of output power, so that energy scheduling of a power grid is not timely, and voltage fluctuation is generated. On the other hand, part of prediction systems adopt expensive cloud image measuring instruments, and cannot meet the economic requirements.
Classical clear sky irradiance models, such as a heliosa 2-4 model, a REST2 model, an ASHRAE model and the like, give ideal solar irradiance reaching the ground surface without cloud group shielding, and can be used as a reference value for removing deterministic components of a photovoltaic system sequence. However, there is not an accurate proportional relationship between clear sky irradiance and clear sky power, and clear sky irradiance results in clear sky indexes that are not accurate enough for the accuracy required for ultra-short term predictions.
Disclosure of Invention
Considering that the weather with small fluctuation in one year has quite a lot of occupation ratio, the prediction method has higher value for the output prediction of the photovoltaic system in the weather with small fluctuation, and has greater significance for improving the overall prediction precision of the photovoltaic system. And the satellite cloud image is combined to develop, so that more accurate judgment is made on future weather conditions, and the scene suitable for the model application is prejudged in advance.
The invention provides an improved clear sky power model based on parameter online updating, and ultra-short-term prediction is carried out on the power of a photovoltaic system in small fluctuation weather based on the model. The main contributions of the invention are as follows: (1) a calculation model capable of describing the clear sky ideal output of the photovoltaic system is provided. The model does not need a large number of power station parameters, is obtained by calculation of historical data, has good adaptability to power stations, can be used as a reference value for stabilizing deterministic components in power removal of a photovoltaic system, and can also be used for other photovoltaic system power prediction methods instead of similar daily results; (2) the ultra-short-term power prediction method for the on-line update of the small fluctuation weather is provided based on the clear sky model, and the prediction accuracy of the small fluctuation weather under the scale of 3-4 hours is improved better.
The regional distributed photovoltaic system output prediction method based on the satellite cloud picture utilizes the satellite cloud picture, the air quality information and the weather forecast to predict the power output of the ultra-short-term photovoltaic system under a small fluctuation scene by utilizing the Internet weather service resources, the ground photovoltaic system power data information and the cloud altitude data.
The embodiment of the invention provides a regional short time scale power prediction method, which comprises the following steps: the grid division area distributed photovoltaic system output prediction method; cloud layer index algorithm of cloud movement; introducing a clear sky model power correction calculation formula; elman neural network algorithm using a random training method.
According to the regional distributed photovoltaic system generated power prediction method provided by the invention, solar space-time model data, corrected power clear sky model data, satellite cloud image data, cloud height data and different time lag scale historical power data and weather forecast data obtained by processed real-time power of a designated region are used as input variables. Compared with the traditional power prediction, the method fully considers the utilization of open network data and theoretical data, improves the accuracy of the output prediction of the regional distributed photovoltaic system, and reduces the cost of constructing the sensor. The predictive data may provide data support for smooth operation of the grid system.
The specific embodiment of the invention provides the following technical scheme:
1. the distributed photovoltaic system output prediction method based on the small fluctuation weather satellite cloud image comprises the following steps,
s1, according to a grid division area distributed photovoltaic system, grid division is carried out on the distributed photovoltaic system in a designated area, so that the distance of a minimum unit of the grid division accords with the distance indicated by a single pixel of a satellite cloud image, and when an actual photovoltaic system is overlapped with grid lines, the grid with a large dividing part occupation ratio is used as a grid where the grid is positioned;
s2, recording measured power data of all photovoltaic systems in the area, satellite cloud pictures of Internet weather service websites, air quality, weather forecast information and cloud altimeter measurement data;
s3, carrying out data error processing on measured power data of each distributed photovoltaic system, satellite cloud pictures of internet weather service websites, air quality, weather forecast information and data measured by a cloud altimeter, wherein the measured power data are subjected to data error processing to obtain data with different time lag scales of output power;
s4, calculating solar space-time model data by using the space geographic information and the time information of the distributed photovoltaic system, and calculating solar time angle omega, solar zenith angle theta and solar radiance theoretical data G CLR Corrected clear sky model power data;
s5, intercepting a local cloud image with a proper area size from the original satellite cloud image; using PIV to measure the speed, distributing a single motion vector to all pixel points of each image, further calculating the average value of the motion vectors and calculating to obtain a satellite cloud image cloud index operator;
s6, normalizing the output power different time lag scale data, the corrected clear sky model power data, the solar space-time model data, the air quality data after error processing, the weather forecast data after error processing and the satellite cloud image cloud index operator data according to the obtained output power different time lag scale data after the processing;
s7, regional output prediction of the distributed photovoltaic system.
The invention has the following beneficial effects:
1. the prediction method fully utilizes network open resources, improves the prediction precision of the regional distributed photovoltaic system generated power in a short time scale, provides data support for intelligently controlling regional power, reducing the rotating reserve capacity of the power system and reducing the running cost.
2. The pixel area is corrected and a new cloud index operator is introduced by using an improved method based on satellite cloud image information, so that the method is more efficient than the traditional cloud index calculation.
3. And an Elman neural network model is established by utilizing a random training method, so that local overfitting of the neural network is avoided, and modeling efficiency is improved.
4. The clear sky power model provided by the invention is very similar to actual measurement power data in small fluctuation weather, and can better describe the power characteristics of the photovoltaic system to obtain the deterministic component of the power output of the photovoltaic system.
5. The prediction algorithm is used for carrying out iterative correction on the model according to the data newly appearing on the same day in a prediction day on the basis of the proposed clear sky power model, and is used for predicting the output of the power station for 4 hours in the future at the target moment, the prediction result is improved compared with other prediction models, wherein the improvement ratio of the prediction precision under a longer time scale is larger than that under a short time scale, and the prediction algorithm has important significance for the operation scheduling of an actual system.
6. Compared with the similar day method, the result of the prediction model is reduced in the mean value of the prediction error, and the condition of larger error in the similar day method is improved. In the data set with large annual peak power fluctuation, the prediction accuracy of the method provided by the invention is greatly improved compared with that of a similar day method. Considering that the weather with small fluctuation in one year has quite a lot of occupation ratio, the method provided by the invention has higher value for the output prediction of the photovoltaic system in the weather with small fluctuation and has greater significance for improving the integral prediction precision of the photovoltaic system. The next step of work is to develop in combination with the satellite cloud image, and more accurate judgment is made on future weather conditions, so that the scene suitable for the model application is prejudged in advance.
Drawings
FIG. 1 is a computational flow chart of the satellite cloud application of the present invention.
FIG. 2 is a flow chart of the regional distributed photovoltaic system prediction of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The principles of the invention will be further described with reference to the drawings and specific examples.
Referring to fig. 1-2, a regional distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud patterns comprises the following steps:
step 1, according to a grid division area distributed photovoltaic system, grid division is carried out on the distributed photovoltaic system in a designated area, so that the distance of a minimum unit of the grid division accords with the distance pointed by a single pixel of a satellite cloud image, and when an actual photovoltaic system is overlapped with grid lines, the grid with a large dividing part occupation ratio is used as a grid where the grid is positioned;
step 2, recording measured power data of all photovoltaic systems in the area, satellite cloud pictures of internet weather service websites, air quality, weather forecast information and cloud altimeter measurement data;
step 3, carrying out data error processing on the measured power data of each distributed photovoltaic system, the satellite cloud picture of an Internet weather service website, the air quality, weather forecast information and cloud altimeter measurement data photovoltaic systems, and adopting the following method when missing measurement and abnormal data appear in recorded data:
firstly, according to the sample size selected in advance, the number of data under normal condition of each data type is calculated. It is checked whether the number of types of data coincides with the calculated number. If the number of the data is not equal, searching the breakpoint of the data and marking the breakpoint. And checks whether the start time and the end time of each data are consistent.
Filling the missing or abnormal power data by a linear interpolation method or a front value substitution method; replacing power data less than zero with zero; missing or anomalous weather data may be corrected by other weather elements according to correlation principles.
Step 4, calculating solar space-time model related theoretical data by using the space geographic information and the time information of the distributed photovoltaic system, and calculating solar time angle omega, solar zenith angle theta and solar radiance theoretical data G CLR The process is as follows:
ω＝15×(ST-12)
where ST is time, in 24 hours.
wherein ,to calculate the latitude of the point, delta declination angle.
Where k is an empirically determined normal number, a i The empirical coefficients obtained for fitting; further searching historical power data of each distributed photovoltaic system nearest to sunny weatherRepresenting the power at time n of the ith distributed photovoltaic system, cl represents clear sky weather.
Further correcting the power data by utilizing the theoretical radiance data ratio of the day to be detected to the historical data date, and taking the corrected power data as the power data under the clear sky model;
in the formula ,Pi cl For clear sky power of the time to be predicted, G' CLR Calculating radiance, G 'for model of time to be predicted' n Calculating the radiance for the model closest to the corresponding time of the sunny weather;
the clear sky model is:
(1) and establishing a mapping model for calculating the clear sky irradiance and the clear sky output of the actual photovoltaic system by using the existing ASHRAE model, so as to obtain the clear sky power of the photovoltaic system.
The model requires an atmospheric parameter of only the optical thickness τ of the direct radiation b And optical thickness τ of scattered radiation d . τ of a certain place b and τd Is constantly changing throughout the year, representing atmospheric parameters such as altitude, moisture content and aerosols, associated with local clear sky solar radiation.
Under clear sky conditions, the solar radiation finally reaching the ground can be divided into direct radiation I b And scattered radiation I d The calculation formula is as follows:
I sc taking 1366.1W/m2 as a solar constant; z is the optical quality of the atmosphere; z and s are respectivelyAtmospheric optical correction index of direct and scattered radiation, e denotes the exponential function.
The atmospheric optical quality Z can be approximately calculated through the solar altitude H, and the atmospheric optical quality correction indexes Z and s can be calculated according to tau b and τd The following expressions were obtained:
z＝1.219－0.043τ b －0.15τ d －0.204τ b τ d
s＝0.202+0.852τ b －0.007τ d －0.357τ b τ d
based on the calculation, clear sky surface irradiance I of ASHRAE model can be obtained t The method comprises the following steps:
I t ＝I b sin H+I d
(2) modeling of clear sky irradiance to clear sky power mapping equation
Because of the influence of factors such as temperature, irradiance and the like on conversion efficiency, the relation between the irradiance of the earth surface and the output power of the photovoltaic system is not a simple linear relation, and the distinction between clear sky irradiance and clear sky power mainly shows 2 aspects, namely amplitude and shape. In addition, the conversion relationship between irradiance and power varies periodically with changes in weather conditions due to changes in the solar's day period and year period. This section will analyze and fit these factors one by one. Maximum empirical output power P for a photovoltaic system at a fixed location PV The method comprises the following steps:
P PV ＝ηA PV I t
η is the photoelectric conversion efficiency of the photovoltaic system; a is that PV Is the area of the photovoltaic system array.
For practical photovoltaic systems, the photoelectric conversion efficiency is mainly affected by 2 aspects. On the one hand, the grid-connected inverters of the photovoltaic system are controlled by tracking the maximum power point, and when the grid-connected inverters run at the maximum power point, the efficiency of the photovoltaic system is subjected to irradiance I t And operating temperature T C Modeling this, the resulting parametric equation model is as follows:
η shape (I t ，T C )＝η MPP (I t )η T (T C )＝(a 1 +a 2 I t +a 3 lnI t )(1－0.005(T C －25))
in the formula ：ηT (T C ) To determine the temperature-affected part of the irradiance-to-power conversion relationship; η (eta) MPP (I t ) To obtain the operation mode related part of the irradiance-to-power conversion relationship; a, a 1 、a 2 、a 3 The undetermined parameters are parameter equations; η (eta) shape (I t ，T C ) Differences in clear sky irradiance and clear sky power curve shape caused by photovoltaic system panel operation and temperature are described.
On the other hand, the photoelectric conversion efficiency is also influenced by factors such as the operating environment of a power station, the ageing of elements and the like, and the difference of the actual installed capacity can influence the parameter A in PV . This part of the influencing factors are difficult to obtain, besides the easy availability of the installed capacity. Therefore, the relation between the ratio of the measured power peak value of the small fluctuation weather to the clear sky irradiance peak value and the product day in one year is fitted according to the historical data by analyzing the measured power data in the small fluctuation weather, and finally, a parameter equation eta in the form of a trigonometric function with the minimum fitting error is selected altitude (n) to describe this relationship, namely:
wherein: m, alpha and beta are constants and are undetermined parameters of the equation; n is the nth day of the year. η -type lens altitude (n) contains the formula P PV Medium parameter A PV And the efficiency of the operating and aging factors.
Combined P PV 、η shape (I t ，T C )、η altitude (n) heddleConsider eta shape (I t ，T C) and ηaltitude (n) the effect of these 2 aspects, the invention builds a parameter mapping model of clear sky irradiance to clear sky power as follows.
Step 5, firstly, intercepting a local cloud image with a proper area size from an original satellite cloud image;
using PIV to measure the speed, all pixel points of each image are distributed with a single motion vector, and then the average value is calculated;
wherein N is the number of pixel points, u i and Vi Representing the speeds in the horizontal and vertical directions, e x and ey Refers to two directions;meaning the average value of the motion vectors of all pixels.
Further, calculating the number M of pixel points passing through by cloud motion in a specified time scale by using a cloud motion average vector, wherein the number M of pixel points is a rounding value;
further, a cloud index is obtained by analyzing each pixel of the relevant area, the pixel intensity being denoted by E,
E(t)＝I 0 ρcos(θ(t))^((1+α))
where t is time, E (t) is pixel intensity at a particular time, and θ (t) is the angle at a particular time;
ρ＝E(t)/(I 0 cos(θ(t))^((1+α)))
wherein ρ is the pixel reflectivity, α is the empirical coefficient, I 0 Is the solar radiation constant outside the earth;
n= (ρ - ρ_max)/(ρ_max- ρ_min), n is the cloud index,
wherein ρ_max and ρ_min are the maximum and minimum values calculated from the history picture, respectively;
further correcting a pixel area corresponding to the ground distributed photovoltaic system by using the solar space-time model, respectively converting deltay and deltax into pixel points according to proportion, and correcting the offset e in the transverse direction by using deltax x Correction of the offset e in the longitudinal direction by Δy y ；
Δx＝-htanθsinω
Δy＝-htanθcosω
Wherein h is the measured cloud height;
further utilizing the cloud index and the number M of the pixel points, selecting the pixel points according to the opposite direction of the cloud movement direction and solving the average value n of the cloud indexes;
further constructing a new operator of the average cloud index by using the basic mathematical function: n, n 2 、1/n、ln(n)、e n 。
Step 6, normalizing the data of different time lag scales of the output power, the corrected clear sky model power data, the solar space-time model data, the air quality data, the weather forecast data and the satellite cloud image cloud index operator data obtained by processing in the steps 2, 3, 4 and 5;
the processing formula is as follows:
X′(i)＝(X(i)-X min (i))/(X max (i)-X min (i))
wherein X' (i) is normalized value, X (i), X max (i)、X min (i) Respectively, a sample actual value, a sample maximum value and a sample minimum value.
The Elman neural network is further trained, and the training method of the Elman neural network is a random training verification method, and the process is as follows:
6.1 70% of the data packets to be learned are used for training, and the rest of the data are used for testing;
6.2 The data of the training part are further divided into 20 subsets according to the sunny days, cloudy days and cloudy days of weather conditions, and the data in the subsets are randomly ordered;
6.3 Random selection of a plurality of homogeneous subsets for training Elman neural networks;
6.4 A trained Elman neural network, testing through prepared test data;
6.5 And stopping training after the test result meets the actual requirement.
Step 7, predicting regional output of the distributed photovoltaic system;
and inputting the latest calculated input variable parameters into the model to obtain a prediction result, and further accumulating the prediction data of each grid to be used as regional photovoltaic system output prediction data.
While the foregoing has been described in relation to illustrative embodiments thereof, so as to facilitate the understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, but is to be construed as limited to the spirit and scope of the invention as defined and defined by the appended claims, as long as various changes are apparent to those skilled in the art, all within the scope of which the invention is defined by the appended claims.
Claims (6)
1. A distributed photovoltaic system output prediction method based on a small fluctuation weather satellite cloud picture is characterized by comprising the following steps of: the method specifically comprises the following steps of,
s1, according to a grid division area distributed photovoltaic system, grid division is carried out on the distributed photovoltaic system in a designated area, so that the distance of a minimum unit of the grid division accords with the distance indicated by a single pixel of a satellite cloud image, and when an actual photovoltaic system is overlapped with grid lines, the grid with a large dividing part occupation ratio is used as a grid where the grid is positioned;
s2, recording measured power data of all photovoltaic systems in the area, satellite cloud pictures of Internet weather service websites, air quality, weather forecast information and cloud altimeter measurement data;
s3, carrying out data error processing on measured power data of each distributed photovoltaic system, satellite cloud pictures of internet weather service websites, air quality, weather forecast information and data measured by a cloud altimeter, wherein the measured power data are subjected to data error processing to obtain data with different time lag scales of output power;
s4, calculating solar space-time model data by using the space geographic information and the time information of the distributed photovoltaic system, and calculating solar time angle omega, solar zenith angle theta and solar radiance theoretical data G CLR Corrected clear sky model power data;
s5, intercepting a local cloud image with a proper area size from the original satellite cloud image; using PIV to measure the speed, distributing a single motion vector to all pixel points of each image, further calculating the average value of the motion vectors and calculating to obtain a satellite cloud image cloud index operator;
s6, normalizing the output power different time lag scale data, the corrected clear sky model power data, the solar space-time model data, the air quality data after error processing, the weather forecast data after error processing and the satellite cloud image cloud index operator data according to the obtained output power different time lag scale data after the processing;
and S7, obtaining regional output prediction data of the distributed photovoltaic system.
2. The method for predicting output of a distributed photovoltaic system based on a small fluctuation weather satellite cloud image according to claim 1, wherein in step S3, when missing and abnormal data occur in the recorded data, the following method is adopted: firstly, calculating the number of data under the normal condition of each data type according to the size of a sample selected in advance, checking whether the number of each type of data accords with the calculated number, if the number of the data is not equal, searching data break points and marking, checking whether the starting time and the ending time of each data are consistent, and filling missing or abnormal power data by a linear interpolation method or a front value substitution method; zero replaces power data less than zero.
3. The method for predicting output of a distributed photovoltaic system based on a small-fluctuation weather satellite cloud cover as set forth in claim 1, wherein in step S4,
the calculation process is as follows: ω=15× (ST-12), where ST is time, in 24 hours;
wherein ,for calculating the latitude of the point, delta is the declination angle;
where k is an empirically determined normal number, a i The empirical coefficients obtained for fitting;
searching historical power data of each distributed photovoltaic system power station nearest to sunny weather Representing the power of the ith distributed photovoltaic system at the moment n, wherein cl represents sunny weather;
correcting power data by utilizing the theoretical radiance data ratio of the day to be detected to the historical data date, and taking the corrected power data as power data under a clear sky model;
in the formula ,Pi cl For clear sky power of the time to be predicted, G' CLR Calculating radiance, G 'for model of time to be predicted' n And calculating the radiance for the model closest to the corresponding time of the sunny weather.
4. The method for predicting output of a distributed photovoltaic system based on a small fluctuation weather satellite cloud image according to claim 1, wherein in step S5, a motion vector average value is calculated by adopting the following formula:
wherein N is the number of pixel points, u i and Vi Representing the speeds in the horizontal and vertical directions, e x and ey Refers to two directions;meaning the average value of motion vectors of all pixels;
calculating the number M of pixel points passing through by cloud motion in a specified time scale by using a cloud motion average vector, wherein the number M of pixel points is a rounding value;
analyzing each pixel of the relevant area to obtain a cloud index;
the pixel intensity is denoted by E and,
E(t)＝I 0 ρcos(θ(t))^((1+α))
where t is time, E (t) is pixel intensity at a particular time, and θ (t) is the angle at a particular time;
ρ＝E(t)/(I 0 cos(θ(t))^((1+α)))
wherein ρ is the pixel reflectivity, α is the empirical coefficient, I 0 Is the solar radiation constant outside the earth;
n＝(ρ-ρ_max)/(ρ_max-ρ_min)
wherein ρ_max and ρ_min are the maximum and minimum values calculated from the history picture, respectively;
correcting a pixel area corresponding to the ground distributed photovoltaic system by using a solar space-time model, respectively converting deltay and deltax into pixel points according to proportion, and correcting e by using deltax x Offset in the transverse direction, e is corrected by Δy y Offset in the longitudinal direction;
Δx＝-h tanθsinω
Δy＝-h tanθcosω
wherein h is the measured cloud height;
selecting pixel points according to the opposite direction of the cloud motion direction by utilizing the cloud index of each pixel point and the number M of the pixel points, and calculating an average value n of the cloud indexes;
constructing a cloud index operator: n, n 2 、1/n、ln(n)、e n 。
5. The method for predicting output of a distributed photovoltaic system based on a small fluctuation weather satellite cloud image according to claim 1, wherein in step S6, a normalization processing formula is as follows:
X′(i)＝(X(i)-X min (i))/(X max (i)-X min (i))
wherein X' (i) is normalized value, X (i), X max (i)、X min (i) Respectively a sample actual value, a sample maximum value and a sample minimum value;
the Elman neural network is trained, and the training method of the Elman neural network is a random training verification method, and the process is as follows:
6.1 70% of the data packets to be learned are used for training, and the rest of the data are used for testing;
6.2 The data of the training part are further divided into 20 subsets according to the sunny days, cloudy days and cloudy days of weather conditions, and the data in the subsets are randomly ordered;
6.3 Random selection of a plurality of homogeneous subsets for training Elman neural networks;
6.4 A trained Elman neural network, testing through prepared test data;
6.5 And stopping training after the test result meets the actual requirement.
6. The method for predicting output of a distributed photovoltaic system based on a small fluctuation weather satellite cloud image according to claim 1, wherein in step S7, the calculated input variable parameters are input to an Elman neural network model to obtain a prediction result, and the prediction data of each grid is accumulated to be used as regional photovoltaic system output prediction data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310532250.9A CN116542383A (en) | 2023-05-12 | 2023-05-12 | Distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310532250.9A CN116542383A (en) | 2023-05-12 | 2023-05-12 | Distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116542383A true CN116542383A (en) | 2023-08-04 |
Family
ID=87450232
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310532250.9A Pending CN116542383A (en) | 2023-05-12 | 2023-05-12 | Distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116542383A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116780533A (en) * | 2023-08-16 | 2023-09-19 | 北京东润环能科技股份有限公司 | Photovoltaic ultra-short-term forecasting method and device, electronic equipment and storage medium |
| CN116865263A (en) * | 2023-08-30 | 2023-10-10 | 天津市普迅电力信息技术有限公司 | Distributed photovoltaic power prediction method based on space grid |
-
2023
- 2023-05-12 CN CN202310532250.9A patent/CN116542383A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116780533A (en) * | 2023-08-16 | 2023-09-19 | 北京东润环能科技股份有限公司 | Photovoltaic ultra-short-term forecasting method and device, electronic equipment and storage medium |
| CN116780533B (en) * | 2023-08-16 | 2023-11-03 | 北京东润环能科技股份有限公司 | Photovoltaic ultra-short-term forecasting method and device, electronic equipment and storage medium |
| CN116865263A (en) * | 2023-08-30 | 2023-10-10 | 天津市普迅电力信息技术有限公司 | Distributed photovoltaic power prediction method based on space grid |
| CN116865263B (en) * | 2023-08-30 | 2024-01-09 | 天津市普迅电力信息技术有限公司 | Distributed photovoltaic power prediction method based on space grid |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106779154B (en) | Regional distributed photovoltaic output prediction method based on satellite cloud picture | |
| CN116542383A (en) | Distributed photovoltaic system output prediction method based on small fluctuation weather satellite cloud image | |
| CN109543721B (en) | Solar irradiance ultra-short-term forecasting method under sunny cloudy weather condition | |
| Ulbricht et al. | First steps towards a systematical optimized strategy for solar energy supply forecasting | |
| Carneiro et al. | Review on photovoltaic power and solar resource forecasting: current status and trends | |
| CN111033318A (en) | Local weather prediction | |
| CN106372749B (en) | Ultra-short term photovoltaic power prediction technique based on the analysis of cloud variation | |
| AU2014389497B2 (en) | Combing multiple trending models for photovoltaic plant output forecasting | |
| CN107491833B (en) | Photovoltaic power generation power prediction method and device | |
| CN109884896B (en) | Photovoltaic tracking system optimized tracking method based on similar day irradiation prediction | |
| CN108196317B (en) | Meteorological prediction method for micro-grid system | |
| Lorenz et al. | Overview of irradiance and photovoltaic power prediction | |
| CN114707688A (en) | Photovoltaic power ultra-short-term prediction method based on satellite cloud chart and space-time neural network | |
| CN114510513A (en) | Short-term meteorological forecast data processing method for ultra-short-term photovoltaic power prediction | |
| Kayri et al. | Predicting power production from a photovoltaic panel through artificial neural networks using atmospheric indicators | |
| CN110781458A (en) | Method for predicting surface solar irradiance based on mixed regression model | |
| Panagopoulos et al. | Predicting the power output of distributed renewable energy resources within a broad geographical region | |
| CN115546657A (en) | Regional earth surface radiation intensity prediction method based on cloud cluster motion trail | |
| CN115049163A (en) | Photovoltaic region prediction method and device based on tabu search and deep stable learning | |
| Constantinescu et al. | Unit commitment with wind power generation: integrating wind forecast uncertainty and stochastic programming. | |
| CN114418378A (en) | Photovoltaic power generation internet data checking method based on LOF outlier factor detection algorithm | |
| CN106815655B (en) | Photovoltaic output 2D interval prediction method based on fuzzy rule | |
| CN117526274A (en) | New energy power prediction method, electronic equipment and storage medium in extreme climate | |
| CN116484998A (en) | Distributed photovoltaic power station power prediction method and system based on meteorological similar day | |
| Ulbricht et al. | Systematical evaluation of solar energy supply forecasts |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |