CN104217259A - Regional ground surface irradiance distribution predicting method - Google Patents

Abstract

The invention relates to a regional ground surface irradiance distribution predicting method which is used for determining minute-level ground surface irradiance regional distribution prediction within future 1 hour. The regional ground surface irradiance distribution predicting method comprises the following steps of: determining a local cloud base height; performing image process on a foundation cloud picture; predicting local ground surface irradiance distribution. The regional ground surface irradiance distribution predicting method provides calculation of the cloud base height by utilizing the ground surface regular meteorological observation data, realizes ground surface irradiance and photovoltaic power prediction within the local range through combination of foundation cloud picture data, is free from the dependence from a traditional method to the observation of the cloud base height, and improves calculation accuracy of cloud position and size at the absence of a ceilometer and the water level total irradiance prediction accuracy of a photovoltaic power station.

Description

Earth's surface, a kind of region irradiance distribution Forecasting Methodology

Technical field:

The present invention relates to earth's surface, a kind of region irradiance distribution Forecasting Methodology, more specifically relate to earth's surface, the region irradiance distribution Forecasting Methodology of a kind of combination height of cloud base calculating and ground cloud Picture.

Background technology:

Along with increasing sharply of large-scale photovoltaic electricity generation grid-connecting capacity, the randomness of photovoltaic generation and undulatory property cause day by day serious impact to the safe operation of electric system, scheduling and controlling.Harm electrical network being brought for reducing photovoltaic uncertainty, the utilization ratio of raising sun power, first needs photovoltaic generation power to predict exactly.

A precondition of photovoltaic power prediction is earth's surface irradiance prediction accurately, and the earth's surface irradiance prediction of main flow at present mainly adopts two kinds of methods, be the Forecasting Methodology based on statistics and machine learning algorithm, another kind is the Forecasting Methodology based on weather forecast.These two kinds of methods can be compared with accurate forecast earth's surface irradiance under fine day condition, but under cloudy, rainy weather, therefore owing to cannot obtaining the distributed intelligence of local sky cloud cluster and movement tendency, can not Accurate Prediction go out earth's surface irradiance and photovoltaic power sudden change that the occlusion effect due to cloud causes.

In order to overcome the defect of above Forecasting Methodology, academia has proposed surface radiation and the photovoltaic power Forecasting Methodology based on ground cloud atlas in recent years.The method is carried out cloud cluster identification and movement tendency prediction according to the ground cloud atlas image of taking in real time, can the decay of Accurate Prediction cloud cluster to surface radiation.But existing algorithm depends on the observation of ceilometer to obtaining of height of cloud base information.In the situation that there is no ceilometer observation data, existing algorithm cannot obtain real-time height of cloud base information, can only utilize assembly average or empirical value to substitute, cause cloud cluster size to calculate inaccurate, cannot accurately obtain cloud cluster shade in information such as the size on ground and positions.Therefore propose earth's surface, a kind of region irradiance distribution Forecasting Methodology, overcome the problems referred to above.

Summary of the invention:

The object of this invention is to provide earth's surface, a kind of region irradiance distribution Forecasting Methodology, the method has been broken away from the dependence of classic method to height of cloud base observation, improved cloud cluster position and big or small calculating accuracy under without ceilometer condition, and photovoltaic plant photovoltaic plant surface level solar global irradiance prediction accuracy.

For achieving the above object, the present invention by the following technical solutions: earth's surface, a kind of region irradiance distribution Forecasting Methodology, the method is for determining minute level earth's surface irradiance areal distribution prediction in following 1 hour; Said method comprising the steps of:

(1) the local height of cloud base determines;

(2) ground cloud atlas image is processed;

(3) earth's surface, region irradiance distribution prediction.

Earth's surface, a kind of region provided by the invention irradiance distribution Forecasting Methodology, is characterized in that: definite the comprising the following steps of described step (1):

(1-1) obtain the real-time meteorological measuring of photovoltaic plant;

(1-2) determine the mean value of surface air temperature, relative humidity and surface pressure in following 1 hour;

(1-3) determine air mass megadyne temperature and in subaerial specific humidity;

(1-4) determine temperature, saturation vapour pressure and the saturation specific humidity of isentropic condensation clevel;

(1-5) air pressure of isentropic condensation clevel;

(1-6) determine isentropic condensation clevel value.

Earth's surface, a kind of region provided by the invention irradiance distribution Forecasting Methodology, the data in described step (1-1) comprise surface air temperature T, ground relative humidity RH, surface pressure P; Reject scarce survey data, unreasonable data recording wherein; Described data zone of reasonableness is defined as: surface air temperature scope-55 ℃～50 ℃, relative humidity scope 0%～100%, surface pressure scope 600hPa～1050hPa.

Another earth's surface, preferred a kind of region irradiance distribution Forecasting Methodology provided by the invention, the mean value of surface air temperature, relative humidity and surface pressure in described step (1-2) is determined by following formula (1)-(3):

$T_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{T}_i---\left(1\right)$

${\mathrm{RH}}_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{\mathrm{RH}}_i---\left(2\right)$

$P_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{P}_i---\left(3\right)$

Wherein, T _s, RH _s, P _srepresent respectively the mean value of surface air temperature, relative humidity and surface pressure in following 1 hour, T _i, RH _i, P _ifor i effective surface air temperature, relative humidity, surface pressure value observed in 1 hour future, N is effective meteorological observation number of samples in following 1 hour;

Air mass megadyne temperature in described step (1-3) is determined by following formula (6):

$\mathrm{\θ}=T_s{\left(\frac{1000}{P_s}\right)}^{0.286}---\left(4\right)$

Described definite by following formula (5) and (6) in subaerial specific humidity:

$\begin{array}{c}{e}_s={\mathrm{RH}}_s\·E_s\\ ={\mathrm{RH}}_s\·6.1078\exp \left(17.269\frac{T_s-273.16}{T_s-35.86}\right)\end{array}---\left(5\right)$

$q_s=622\frac{e_s}{P_s-0.387e_s}\≈622\frac{e_s}{P_s}---\left(6\right)$

Wherein, e _sfor steam pressure near the ground, q _sfor in subaerial specific humidity.

Earth's surface, a preferred a kind of region irradiance distribution Forecasting Methodology more provided by the invention, the saturation specific humidity in described step (1-4) is determined by following formula (7):

$T_{\mathrm{LCL}}=\mathrm{\θ}{\left(\frac{P_{\mathrm{LCL}}}{1000}\right)}^{0.286}---\left(7\right)$

Described saturation vapour pressure is determined by following formula (8):

$E_{\mathrm{LCL}}=6.1078\·\exp \left(17.269\frac{T_{\mathrm{LCL}}-273.16}{T_{\mathrm{LCL}}-35.86}\right)---\left(8\right)$

The temperature of described isentropic condensation clevel is determined by following formula (9):

$q_{\mathrm{LCL}}=q_s=622\frac{E_{\mathrm{LCL}}}{P_{\mathrm{LCL}}}---\left(9\right)$

Wherein, P _lCLfor the air pressure of isentropic condensation clevel in described step (1-5) and according to table for reference, by ground observation to megadyne temperature and specific humidity obtain;

Described table for reference is by specific humidity q _sabout megadyne temperature and P _lCLfunction expression (10) determine;

q _s＝F(θ,P _LCL) (10)

To θ according to 0.1 ° of interval and P _lCLaccording to 0.1hPa interval, set up table for reference, calculate corresponding q _s, the q that recycling calculates _sand θ, interpolation obtains about P _lCLtable for reference, as shown in the formula (11):

P _LCL＝F _inv(θ,q _s) (11)。

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, the definite isentropic condensation clevel value in described step (1-6) is determined by following formula (11):

$Z_{\mathrm{LCL}}=\frac{R}{1000g}3.4965({P_s}^{0.286}-{P_{\mathrm{LCL}}}^{0.286})---\left(11\right)$

Wherein, R is air gas constant.

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, the processing procedure of described step (2) comprises the following steps:

(2-1) determine the coordinate of ground cloud atlas zenith point;

(2-2) sky peripheral images is filtered;

(2-3) block information restores.

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, the coordinate in described step (2-1) is determined by following formula (12):

$(X_c,Y_c)=(\frac{x_1+x_2}{2},\frac{y_1+y_2}{2})---\left(12\right)$

Wherein, on ground cloud atlas image, make horizontal, the profile tangent on border, sky disc area upper and lower, left and right, obtain the boundary coordinate of disc area in X, Y-direction and be respectively x ₁, x ₂, y ₁, y ₂;

Filter process in described step (2-2) is: the cloud atlas zenith point of take is justified as the center of circle, and the maximum radius of the object that the needs of usining are rejected, as radius, is designated as R _max, will be greater than R apart from zenith point _maxpixel RGB color assignment be [0,0,0], filter.

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, described step (2-3) block information restores and adopts the mode of linear interpolation to carry out;

The mode of described linear interpolation is: first the linear pattern object in image is carried out to edge identification, and the center line on the long limit of definite object; Respectively 5 of the object both sides of the edge pixel value that blocks perpendicular to center line is averaging, is designated as A ₁and A ₂; If A _xfor A ₁and A ₂certain middle some pixel value, has:

$A_x=\frac{A_1{l}_2+A_2{l}_1}{l_1+l_2}---\left(13\right)$

Wherein, l ₁and l ₂be respectively A _xto A ₁and A ₂distance.

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, the forecasting process in described step (1-3) comprises the following steps:

(3-1) identification of ground cloud atlas classifying type cloud cluster;

(3-2) cloud cluster ground area shading coordinate determines;

(3-3) calibration of cloud radiation attenuation coefficient;

(3-4) prediction of cloud cluster shade motion;

(3-5) region surface water plane irradiance distribution prediction.

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, the identifying in described step (3-1) is:

To each pixel in cloud atlas image, calculate its normalization red-blue ratio value R:

$R=\frac{b-r}{b+r}---\left(14\right)$

Wherein, b is pixel blue channel brightness value; R is pixel red channel brightness value;

Translucidus and light tight cloud are set respectively to red-blue ratio threshold value η _skyand η _thin, the type of pixel of sky is identified:

In described step (3-2), the deterministic process of coordinate is:

If the zenith angle of any point P is α in cloud cluster, position angle is γ, and this is at the coordinate (x of take in the local rectangular coordinate system that imager is initial point ₀, y ₀) be:

$\left\{\begin{array}{c}{x}_0=H\·\tan \mathrm{\α}\·\sin \mathrm{\γ}\\ y_0=H\·\tan \mathrm{\α}\·\cos \mathrm{\γ}\end{array}\right.---\left(16\right)$

Obtain after vertical projection coordinate, according to this solar zenith angle θ and position angle φ constantly, calculate this at the shade coordinate (x on ground _s, y _s) be:

$\left\{\begin{array}{c}{x}_s=x_0-H\·\tan \mathrm{\θ}\·\cos \mathrm{\φ}\\ y_s=y_0-H\·\tan \mathrm{\θ}\·\cos \mathrm{\φ}\end{array}\right.---\left(17\right)$

Wherein, H is for utilizing ground cloud atlas image and the local height of cloud base after cloud cluster Classification and Identification.

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, the calibration process of the coefficient in described step (3-3) is:

According to cloud cluster occlusion state and the cloud cluster type in surface total radiation meter historical record, corresponding period, corresponding place, statistics obtains the attenuation coefficient of dissimilar cloud cluster to surface level solar global irradiance;

By following formula (18), determine corresponding built-up radiation observation of this period clear sky solar global irradiance I constantly _clear:

I _clear＝c ₁I ² _TOA+c ₂I _TOA+c ₃ (18)

Wherein, I _tOAfor section, atmospheric envelope top solar irradiance, c ₁, c ₂, c ₃for statistical fit parameter, by earth's surface solar global irradiance under fine day condition and corresponding I _tOAcarrying out fitting of a polynomial obtains;

According to occlusion state, ground solar global irradiance observational record and clear sky solar global irradiance are screened, by following formula (19), determine each translucidus and light tight cloud irradiance attenuation coefficient constantly:

Above each moment result is averaging, obtains average irradiance attenuation coefficient:

$\left\{\begin{array}{c}{d}_{\mathrm{thin}}=\frac{1}{N_1}{\mathrm{\Σ}}_{i=1}^{N_1}{d^i}_{\mathrm{thin}}\\ d_{\mathrm{opaque}}=\frac{1}{N_2}{\mathrm{\Σ}}_{i=1}^{N_2}{d^i}_{\mathrm{opaque}}\end{array}\right.---\left(20\right)$

Wherein, I _realfor the prediction of region surface water plane irradiance distribution, i is i valid data, N ₁for the effective sample number that translucidus blocks, N ₂for light tight effective sample number of blocking.

Earth's surface, another preferred a kind of region provided by the invention irradiance distribution Forecasting Methodology, the forecasting process of the cloud cluster shade motion in described step (3-4) is;

According to the cloud cluster shadow positions of adjacent two ground cloud atlas, change, calculate the t movement velocity of cloud cluster shade constantly:

$v_t=\frac{{(x,y)}_t-{(x,y)}_{t-1}}{\mathrm{\Δt}}---\left(21\right)$

Utilize t cloud cluster shade speed linear extrapolation constantly, the cloud cluster shadow positions that obtains the t+1 moment is:

(x,y) _t+1＝(x,y) _t+v _t·Δt (22)

Wherein, Δ t is that adjacent two Zhang Yuns scheme to take the mistiming constantly;

Region surface water plane irradiance distribution forecasting process in described step (3-5) is:

According to irradiance I in ground under fine day condition _clear, translucidus built-up radiation attenuation coefficient d _thinwith light tight cloud built-up radiation attenuation rate d _opaque, in conjunction with the prediction of cloud cluster shadow positions, obtain region surface water plane irradiance distribution prediction:

With immediate prior art ratio, the invention provides technical scheme and there is following excellent effect:

1, the present invention utilizes ground routine meteorological measuring to calculate the height of cloud base, in conjunction with ground cloud atlas data, realizes surface radiation and photovoltaic power prediction in local scope;

2, in the present invention, break away from the dependence of classic method to height of cloud base observation, improved cloud cluster position and big or small calculating accuracy under without ceilometer condition;

3, photovoltaic plant photovoltaic plant surface level solar global irradiance prediction accuracy in the present invention;

4, the applicability of the inventive method is wider than existing methodical applicability;

5, the present invention reduces the harm that photovoltaic uncertainty is brought electrical network, improves the utilization ratio of sun power;

6, the present invention contributes to the safe operation of electric system, scheduling and controlling.

Accompanying drawing explanation

Fig. 1 is the inventive method process flow diagram;

Fig. 2 is that cloud cluster shade coordinate of the present invention calculates schematic diagram;

Fig. 3 is cloud radiation attenuation coefficient calibration result schematic diagram of the present invention;

Fig. 4 is the surface level irradiance of the present invention schematic diagram that predicts the outcome.

Embodiment

Below in conjunction with embodiment, the invention will be described in further detail.

Embodiment 1:

As Figure 1-4, earth's surface, a kind of region of the invention of this example irradiance distribution Forecasting Methodology, the method is for determining minute level earth's surface irradiance areal distribution prediction in following 1 hour; Said method comprising the steps of:

(1) the local height of cloud base determines;

(2) ground cloud atlas image is processed;

(3) earth's surface, region irradiance distribution prediction.

1) calculating of the local height of cloud base

Step1: obtain the real-time meteorological measuring of photovoltaic plant, comprise surface air temperature T, ground relative humidity RH, surface pressure P.Reject to lack and survey data, unreasonable data recording.Wherein data zone of reasonableness is defined as: surface air temperature scope-55 ℃～50 ℃, relative humidity scope 0%～100%, surface pressure scope 600hPa～1050hPa.

Step2: calculate surface air temperature, relative humidity, surface pressure mean value in 1h, computing formula is as follows:

$T_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{T}_i---\left(1\right)$

${\mathrm{RH}}_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{\mathrm{RH}}_i---\left(2\right)$

$P_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{P}_i---\left(3\right)$

Wherein, T _s, RH _s, P _srepresent respectively hourly average surface air temperature, relative humidity, surface pressure value, T _i, RH _i, P _ifor i in 1h effective observation surface air temperature, relative humidity, surface pressure value, N is effective meteorological observation number of samples in 1h.

Step3: utilize surface air temperature, air pressure, relative humidity, calculate air mass megadyne temperature with in subaerial specific humidity:

$\mathrm{\θ}=T_s{\left(\frac{1000}{P_s}\right)}^{0.286}---\left(4\right)$

$\begin{array}{c}{e}_s={\mathrm{RH}}_s\·E_s\\ ={\mathrm{RH}}_s\·6.1078\exp \left(17.269\frac{T_s-273.16}{T_s-35.86}\right)\end{array}---\left(5\right)$

$q_s=622\frac{e_s}{P_s-0.387e_s}\≈622\frac{e_s}{P_s}---\left(6\right)$

Step4: according to megadyne temperature conservation and steam conservation in atmosphere dry adiabatic lifting process, when known air mass rises to isentropic condensation clevel, megadyne temperature and specific humidity remain unchanged, and can obtain thus:

$T_{\mathrm{LCL}}=\mathrm{\θ}{\left(\frac{P_{\mathrm{LCL}}}{1000}\right)}^{0.286}---\left(7\right)$

$E_{\mathrm{LCL}}=6.1078\·\exp \left(17.269\frac{T_{\mathrm{LCL}}-273.16}{T_{\mathrm{LCL}}-35.86}\right)---\left(8\right)$

$q_{\mathrm{LCL}}=q_s=622\frac{E_{\mathrm{LCL}}}{P_{\mathrm{LCL}}}---\left(9\right)$

Wherein, P _lCL, E _lCL, q _lCL, T _lCLbe respectively air pressure, temperature, saturation vapour pressure, the saturation specific humidity of isentropic condensation clevel.

Step5: by (7), (8) two formula substitution formula (9), can obtain specific humidity q _sabout megadyne temperature and P _lCLfunction expression, be designated as:

q _s＝F(θ,P _LCL) (10)

To θ according to 0.1 ° of interval, to P _lCLaccording to 0.1hPa interval, set up table for reference, calculate corresponding q _s, the q that recycling calculates _sand θ, interpolation obtains about P _lCLtable for reference, be designated as:

P _LCL＝F _inv(θ,q _s) (11)

According to above table for reference, can by ground observation to megadyne temperature and steam specific humidity, obtain the air pressure of isentropic condensation clevel.

Step6: according to gas pressure-height formula, can calculate isentropic condensation clevel value:

$Z_{\mathrm{LCL}}=\frac{R}{1000g}3.4965({P_s}^{0.286}-{P_{\mathrm{LCL}}}^{0.286})---\left(11\right)$

Wherein, R is air gas constant.

2) ground cloud atlas image is processed

Step1: make horizontal, the profile tangent on border, sky disc area upper and lower, left and right on ground cloud atlas image, obtain the boundary coordinate of disc area in X, Y-direction and be respectively x ₁, x ₂, y ₁, y ₂, determine that thus the coordinate of ground cloud atlas zenith point is:

$(X_c,Y_c)=(\frac{x_1+x_2}{2},\frac{y_1+y_2}{2})---\left(12\right)$

Step2: sky peripheral images is filtered

Because ground cloud atlas outer peripheral areas is generally the objects such as ground, building, vegetation, need to reject.The cloud atlas zenith point of take is justified as the center of circle, usings the maximum radius of objects such as not comprising ground, building, vegetation as radius, is designated as R _max, will be greater than R apart from zenith point _maxpixel RGB color assignment be [0,0,0], filter.

Step3: shield portions information recovery

Except cloud atlas outer peripheral areas, in cloud atlas, also have the small-sized shelters such as photo-shield strip, shaft tower, this class is blocked, adopt the method for linear interpolation to carry out block information recovery.First the linear pattern object in image is carried out to edge identification, and the center line on the long limit of definite object.Respectively 5 of the object both sides of the edge pixel value that blocks perpendicular to center line is averaging, is designated as A ₁and A ₂.If A _xfor A ₁and A ₂certain middle some pixel value, has:

$A_x=\frac{A_1{l}_2+A_2{l}_1}{l_1+l_2}---\left(13\right)$

Wherein, l ₁and l ₂be respectively A _xto A ₁and A ₂distance.

3) earth's surface, region irradiance distribution prediction

Step1: ground cloud atlas classifying type cloud cluster identification

To each pixel in cloud atlas image, calculate its normalization red-blue ratio value R:

$R=\frac{b-r}{b+r}---\left(14\right)$

Wherein b is pixel blue channel brightness value; R is pixel red channel brightness value.Translucidus and light tight cloud are set respectively to red-blue ratio threshold value η _skyand η _thin, can identify the type of pixel of sky:

Step2: cloud cluster ground area shading coordinate calculates

Ground cloud atlas image and the local height of cloud base H of utilization after cloud cluster Classification and Identification, the cloud cluster ground vertical projection coordinate that can to obtain take ground imager be initial point.If the zenith angle of any point P is α in cloud cluster, position angle is γ, and this is at the coordinate (x of take in the local rectangular coordinate system that imager is initial point ₀, y ₀) be:

$\left\{\begin{array}{c}{x}_0=H\·\tan \mathrm{\α}\·\sin \mathrm{\γ}\\ y_0=H\·\tan \mathrm{\α}\·\cos \mathrm{\γ}\end{array}\right.---\left(16\right)$

Obtain after vertical projection coordinate, according to this solar zenith angle θ and position angle φ constantly, calculate this at the shade coordinate (x on ground _s, y _s) be:

$\left\{\begin{array}{c}{x}_s=x_0-H\·\tan \mathrm{\θ}\·\cos \mathrm{\φ}\\ y_s=y_0-H\·\tan \mathrm{\θ}\·\cos \mathrm{\φ}\end{array}\right.---\left(17\right)$

Cloud cluster shade coordinate calculates schematic diagram and sees accompanying drawing 2.

Step3: cloud radiation attenuation coefficient calibration

Cloud cluster occlusion state and cloud cluster type according to surface total radiation meter historical record with corresponding period, corresponding place, can add up and obtain the attenuation coefficient of dissimilar cloud cluster to surface level solar global irradiance.

First calculate corresponding built-up radiation observation of this period clear sky solar global irradiance I constantly _clear:

I _clear＝c ₁I ² _TOA+c ₂I _TOA+c ₃ (18)

I wherein _tOAfor section, atmospheric envelope top solar irradiance, c ₁, c ₂, c ₃for statistical fit parameter, by earth's surface solar global irradiance under fine day condition and corresponding I _tOAcarrying out fitting of a polynomial obtains.

Then utilize cloud cluster shade coordinate result of calculation, obtain the corresponding cloud cluster shade occlusion state in this place constantly, can be divided into that clear sky, translucidus block, state in light tight cloud block 3.According to occlusion state, ground solar global irradiance observational record and clear sky solar global irradiance are screened, calculate each translucidus and light tight cloud irradiance attenuation coefficient constantly:

Finally, above each moment result is averaging, obtains average irradiance attenuation coefficient:

$\left\{\begin{array}{c}{d}_{\mathrm{thin}}=\frac{1}{N_1}{\mathrm{\Σ}}_{i=1}^{N_1}{d^i}_{\mathrm{thin}}\\ d_{\mathrm{opaque}}=\frac{1}{N_2}{\mathrm{\Σ}}_{i=1}^{N_2}{d^i}_{\mathrm{opaque}}\end{array}\right.---\left(20\right)$

Step4: cloud cluster shade motion prediction

According to the cloud cluster shadow positions of adjacent two ground cloud atlas, change, calculate the t movement velocity of cloud cluster shade constantly:

$v_t=\frac{{(x,y)}_t-{(x,y)}_{t-1}}{\mathrm{\Δt}}---\left(21\right)$

Utilize t cloud cluster shade speed linear extrapolation constantly, the cloud cluster shadow positions that obtains the t+1 moment is:

(x,y) _t+1＝(x,y) _t+v _t·Δt (22)

Wherein, Δ t is that adjacent two Zhang Yuns scheme to take the mistiming constantly;

Step5: region surface water plane irradiance distribution prediction

According to irradiance I in ground under fine day condition _clear, translucidus built-up radiation attenuation coefficient d _thinwith light tight cloud built-up radiation attenuation rate d _opaque, in conjunction with the prediction of cloud cluster shadow positions, obtain region surface water plane irradiance distribution prediction:

In order to verify correctness and the applicability of this Forecasting Methodology, take certain photovoltaic plant as case, utilize surface air temperature, air pressure, relative humidity observation data, ground cloud atlas observation data, carries out the prediction of surface level irradiance.

Accompanying drawing 2 is the algorithm schematic diagram that cloud cluster ground area shading coordinate calculates, wherein ground imager position is true origin, cloud cluster position is a P, according to cloud cluster zenith angle, position angle, solar zenith angle, position angle, and height of cloud base H, utilizing formula (16) can calculate cloud cluster ground area shading is X to the distance of initial point _s.

Accompanying drawing 3 has shown the built-up radiation attenuation coefficient statistical distribution under dissimilar cloud block state.According to statistics, the radiation attenuation coefficient of translucidus is 0.2, and light tight cloud radiation attenuation coefficient is 0.7.

Accompanying drawing 4 has shown that the surface level irradiance being obtained by this Forecasting Methodology predicts the outcome, and contrasts with measured value.Result demonstration, the method can Accurate Prediction surface level solar global irradiance, the effectively process of predicting radiation saltus step.

Finally should be noted that: above embodiment is only in order to illustrate that technical scheme of the present invention is not intended to limit; although those of ordinary skill in the field are to be understood that with reference to above-described embodiment: still can modify or be equal to replacement the specific embodiment of the present invention; these do not depart from any modification of spirit and scope of the invention or are equal to replacement, within the claim protection domain of the present invention all awaiting the reply in application.

Claims (13)

1. earth's surface, a region irradiance distribution Forecasting Methodology, the method is for determining minute level earth's surface irradiance areal distribution prediction in following 1 hour; It is characterized in that: said method comprising the steps of:

(1) the local height of cloud base determines;

(2) ground cloud atlas image is processed;

(3) earth's surface, region irradiance distribution prediction.

2. earth's surface, a kind of region as claimed in claim 1 irradiance distribution Forecasting Methodology, is characterized in that: definite the comprising the following steps of described step (1):

(1-1) obtain the real-time meteorological measuring of photovoltaic plant;

(1-2) determine the mean value of surface air temperature, relative humidity and surface pressure in following 1 hour;

(1-3) determine air mass megadyne temperature and in subaerial specific humidity;

(1-4) determine temperature, saturation vapour pressure and the saturation specific humidity of isentropic condensation clevel;

(1-5) air pressure of isentropic condensation clevel;

(1-6) determine isentropic condensation clevel value.

3. earth's surface, a kind of region as claimed in claim 2 irradiance distribution Forecasting Methodology, is characterized in that: the data in described step (1-1) comprise surface air temperature T, ground relative humidity RH, surface pressure P; Reject scarce survey data, unreasonable data recording wherein; Described data zone of reasonableness is defined as: surface air temperature scope-55 ℃～50 ℃, relative humidity scope 0%～100%, surface pressure scope 600hPa～1050hPa.

4. earth's surface, a kind of region as claimed in claim 2 irradiance distribution Forecasting Methodology, is characterized in that: the mean value of surface air temperature, relative humidity and surface pressure in described step (1-2) is determined by following formula (1)-(3):

$T_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{T}_i---\left(1\right)$

${\mathrm{RH}}_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{\mathrm{RH}}_i---\left(2\right)$

$P_s=\frac{1}{N}{\mathrm{\Σ}}_{i=1}^N{P}_i---\left(3\right)$

Wherein, T _s, RH _s, P _srepresent respectively the mean value of surface air temperature, relative humidity and surface pressure in following 1 hour, T _i, RH _i, P _ifor i effective surface air temperature, relative humidity, surface pressure value observed in 1 hour future, N is effective meteorological observation number of samples in following 1 hour;

Air mass megadyne temperature in described step (1-3) is determined by following formula (6):

$\mathrm{\θ}=T_s{\left(\frac{1000}{P_s}\right)}^{0.286}---\left(4\right)$

Described definite by following formula (5) and (6) in subaerial specific humidity:

$\begin{array}{c}{e}_s={\mathrm{RH}}_s\·E_s\\ ={\mathrm{RH}}_s\·6.1078\exp \left(17.269\frac{T_s-273.16}{T_s-35.86}\right)\end{array}---\left(5\right)$

$q_s=622\frac{e_s}{P_s-0.387e_s}\≈622\frac{e_s}{P_s}---\left(6\right)$

Wherein, e _sfor steam pressure near the ground, q _sfor in subaerial specific humidity.

5. earth's surface, a kind of region as claimed in claim 4 irradiance distribution Forecasting Methodology, is characterized in that: the saturation specific humidity in described step (1-4) is determined by following formula (7):

$T_{\mathrm{LCL}}=\mathrm{\θ}{\left(\frac{P_{\mathrm{LCL}}}{1000}\right)}^{0.286}---\left(7\right)$

Described saturation vapour pressure is determined by following formula (8):

$E_{\mathrm{LCL}}=6.1078\·\exp \left(17.269\frac{T_{\mathrm{LCL}}-273.16}{T_{\mathrm{LCL}}-35.86}\right)---\left(8\right)$

The temperature of described isentropic condensation clevel is determined by following formula (9):

$q_{\mathrm{LCL}}=q_s=622\frac{E_{\mathrm{LCL}}}{P_{\mathrm{LCL}}}---\left(9\right)$

Wherein, P _lCLfor the air pressure of isentropic condensation clevel in described step (1-5) and according to table for reference, by ground observation to megadyne temperature and specific humidity obtain;

Described table for reference is by specific humidity q _sabout megadyne temperature and P _lCLfunction expression (10) determine;

q _s＝F(θ,P _LCL) (10)

To θ according to 0.1 ° of interval and P _lCLaccording to 0.1hPa interval, set up table for reference, calculate corresponding q _s, the q that recycling calculates _sand θ, interpolation obtains about P _lCLtable for reference, as shown in the formula (11):

P _LCL＝F _inv(θ,q _s) (11)。

6. earth's surface, a kind of region as claimed in claim 5 irradiance distribution Forecasting Methodology, is characterized in that: the definite isentropic condensation clevel value in described step (1-6) is determined by following formula (11):

$Z_{\mathrm{LCL}}=\frac{R}{1000g}3.4965({P_s}^{0.286}-{P_{\mathrm{LCL}}}^{0.286})---\left(11\right)$

Wherein, R is air gas constant.

7. earth's surface, a kind of region as claimed in claim 1 irradiance distribution Forecasting Methodology, is characterized in that: the processing procedure of described step (2) comprises the following steps:

(2-1) determine the coordinate of ground cloud atlas zenith point;

(2-2) sky peripheral images is filtered;

(2-3) block information restores.

8. earth's surface, a kind of region as claimed in claim 1 irradiance distribution Forecasting Methodology, is characterized in that: the coordinate in described step (2-1) is determined by following formula (12):

$(X_c,Y_c)=(\frac{x_1+x_2}{2},\frac{y_1+y_2}{2})---\left(12\right)$

Wherein, on ground cloud atlas image, make horizontal, the profile tangent on border, sky disc area upper and lower, left and right, obtain the boundary coordinate of disc area in X, Y-direction and be respectively x ₁, x ₂, y ₁, y ₂;

Filter process in described step (2-2) is: the cloud atlas zenith point of take is justified as the center of circle, and the maximum radius of the object that the needs of usining are rejected, as radius, is designated as R _max, will be greater than R apart from zenith point _maxpixel RGB color assignment be [0,0,0], filter.

9. earth's surface, a kind of region as claimed in claim 1 irradiance distribution Forecasting Methodology, is characterized in that: described step (2-3) block information restores and adopts the mode of linear interpolation to carry out;

The mode of described linear interpolation is: first the linear pattern object in image is carried out to edge identification, and the center line on the long limit of definite object; Respectively 5 of the object both sides of the edge pixel value that blocks perpendicular to center line is averaging, is designated as A ₁and A ₂; If A _xfor A ₁and A ₂certain middle some pixel value, has:

$A_x=\frac{A_1{l}_2+A_2{l}_1}{l_1+l_2}---\left(13\right)$

Wherein, l ₁and l ₂be respectively A _xto A ₁and A ₂distance.

10. earth's surface, a kind of region as claimed in claim 1 irradiance distribution Forecasting Methodology, is characterized in that: the forecasting process in described step (1-3) comprises the following steps:

(3-1) identification of ground cloud atlas classifying type cloud cluster;

(3-2) cloud cluster ground area shading coordinate determines;

(3-3) calibration of cloud radiation attenuation coefficient;

(3-4) prediction of cloud cluster shade motion;

(3-5) region surface water plane irradiance distribution prediction.

11. earth's surface, a kind of region as claimed in claim 10 irradiance distribution Forecasting Methodologies, is characterized in that: the identifying in described step (3-1) is:

To each pixel in cloud atlas image, calculate its normalization red-blue ratio value R:

$R=\frac{b-r}{b+r}---\left(14\right)$

Wherein, b is pixel blue channel brightness value; R is pixel red channel brightness value;

Translucidus and light tight cloud are set respectively to red-blue ratio threshold value η _skyand η _thin, the type of pixel of sky is identified:

In described step (3-2), the deterministic process of coordinate is:

If the zenith angle of any point P is α in cloud cluster, position angle is γ, and this is at the coordinate (x of take in the local rectangular coordinate system that imager is initial point ₀, y ₀) be:

$\left\{\begin{array}{c}{x}_0=H\·\tan \mathrm{\α}\·\sin \mathrm{\γ}\\ y_0=H\·\tan \mathrm{\α}\·\cos \mathrm{\γ}\end{array}\right.---\left(16\right)$

Obtain after vertical projection coordinate, according to this solar zenith angle θ and position angle φ constantly, calculate this at the shade coordinate (x on ground _s, y _s) be:

$\left\{\begin{array}{c}{x}_s=x_0-H\·\tan \mathrm{\θ}\·\cos \mathrm{\φ}\\ y_s=y_0-H\·\tan \mathrm{\θ}\·\cos \mathrm{\φ}\end{array}\right.---\left(17\right)$

Wherein, H is for utilizing ground cloud atlas image and the local height of cloud base after cloud cluster Classification and Identification.

12. earth's surface, a kind of region as claimed in claim 11 irradiance distribution Forecasting Methodologies, is characterized in that: the calibration process of the coefficient in described step (3-3) is:

According to cloud cluster occlusion state and the cloud cluster type in surface total radiation meter historical record, corresponding period, corresponding place, statistics obtains the attenuation coefficient of dissimilar cloud cluster to surface level solar global irradiance;

By following formula (18), determine corresponding built-up radiation observation of this period clear sky solar global irradiance I constantly _clear:

I _clear＝c ₁I ² _TOA+c ₂I _TOA+c ₃ (18)

Wherein, I _tOAfor section, atmospheric envelope top solar irradiance, c ₁, c ₂, c ₃for statistical fit parameter, by earth's surface solar global irradiance under fine day condition and corresponding I _tOAcarrying out fitting of a polynomial obtains;

According to occlusion state, ground solar global irradiance observational record and clear sky solar global irradiance are screened, by following formula (19), determine each translucidus and light tight cloud irradiance attenuation coefficient constantly:

Above each moment result is averaging, obtains average irradiance attenuation coefficient:

$\left\{\begin{array}{c}{d}_{\mathrm{thin}}=\frac{1}{N_1}{\mathrm{\Σ}}_{i=1}^{N_1}{d^i}_{\mathrm{thin}}\\ d_{\mathrm{opaque}}=\frac{1}{N_2}{\mathrm{\Σ}}_{i=1}^{N_2}{d^i}_{\mathrm{opaque}}\end{array}\right.---\left(20\right)$

Wherein, I _realfor the prediction of region surface water plane irradiance distribution, i is i valid data, N ₁for the effective sample number that translucidus blocks, N ₂for light tight effective sample number of blocking.

13. earth's surface, a kind of region as claimed in claim 12 irradiance distribution Forecasting Methodologies, is characterized in that: the forecasting process of the cloud cluster shade motion in described step (3-4) is;

According to the cloud cluster shadow positions of adjacent two ground cloud atlas, change, calculate the t movement velocity of cloud cluster shade constantly:

$v_t=\frac{{(x,y)}_t-{(x,y)}_{t-1}}{\mathrm{\Δt}}---\left(21\right)$

Utilize t cloud cluster shade speed linear extrapolation constantly, the cloud cluster shadow positions that obtains the t+1 moment is:

(x,y) _t+1＝(x,y) _t+v _t·Δt (22)

Wherein, Δ t is that adjacent two Zhang Yuns scheme to take the mistiming constantly;

Region surface water plane irradiance distribution forecasting process in described step (3-5) is:

According to irradiance I in ground under fine day condition _clear, translucidus built-up radiation attenuation coefficient d _thinwith light tight cloud built-up radiation attenuation rate d _opaque, in conjunction with the prediction of cloud cluster shadow positions, obtain region surface water plane irradiance distribution prediction: