CN104729430A - Plane shape detection method of heliostat for tower type solar power generation - Google Patents

Abstract

A plane shape detection method of a heliostat for tower type solar power generation includes the steps that an industrial personal computer (4) controls the heliostat (5) to be detected, a CCD camera (2) can collect a deformed image, reflected by the heliostat (5), of a complex stripe image screen (3), a computer (1) controls the CCD camera (2) to collect a deformed complex stripe image reflected by the heliostat (5), and the slope distribution situation of the heliostat plane is obtained in combination with constructed virtual reference plane phase distribution and a heliostat plane slope and phase mapping model. The plane shape of the heliostat plane of the heliostat (5) to be detected can be obtained on the basis of the generalized Hermite interpolation method.

Description

A kind of tower type solar energy thermal power generation heliostat surface testing method

Technical field

The present invention relates to a kind of heliostat surface testing method.

Background technology

Tower type solar energy thermal power generation has that light concentrating times is high, working temperature is high, heat trnasfer distance is short, thermal losses is few, system synthesis efficiency advantages of higher, its output characteristics and conventional thermal power generation the most close, be one of photo-thermal power generation development course of being expected most of prospect.Tower type solar energy thermal power generation forms with thermal energy transfer systems, electricity generation system 3 part primarily of heliostat condensing system, heat absorption.Sunshine convergence is reflexed to the heat dump be positioned on tower by the heliostat of One's name is legion by heliostat condensing system, converts luminous energy to heat energy, then generate electricity through thermodynamic cycle by heat dump.Heliostat is the core component in tower type solar energy thermal power generation station, accounts for the 40-50% of power station cost of investment, is the carrier realizing system high efficiency photothermal deformation.The principal feature of tower type solar energy thermal power generation station heliostat is: 1) aperture area is comparatively large, from tens square meters to square meter up to a hundred; 2) heliostat enormous amount (thousands of, even hundreds of thousands face) in mirror field, and from heat dump distance, heliostat focal length comparatively large (far away to upper km from tens meters, very high to optically focused accuracy requirement); 3) heliostat by the sub-mirror of multi-surface cell be spliced to form complex-curved by sunshine converge reflex to heat dump, integral face shape require high.Therefore, high precision mirror shape ensures that the sunshine that heliostat converges accurately reflects the prerequisite arriving heat dump.But, on the one hand, between the actual integral face shape that heliostat unit mirror in installation process is spliced to form and Theoretical Design face shape, there is certain deviation; On the other hand, heliostat is at outdoor run duration, be subject to the factor impacts such as wind load, self gravitation, temperature and make surface deformation, cause heliostat minute surface cannot form desirable curved surface, the mirror shape little deviation caused thus makes the solar radiation of reflecting depart from heat dump daylight opening, reduces mirror field light gathering efficiency.Make the hot spot converging reflection offset, even deviate from heat dump by above-mentioned error component, affect the safety of tower, reduce light gathering efficiency.Therefore, quick heliostat minute surface surface testing ensures spotlight effect, improves mirror field installation effectiveness, reduces costs, improves the key of mirror field light gathering efficiency and power station operational efficiency.At present, this type of mirror shape detects and usually adopts three coordinate machine, laser radar, photogrammetric and streak reflex method.Plant measuring method relative to first three, the advantages such as streak reflex method has high sensitivity, detects fast, system is simple and easy, are applicable to detect all kinds of minute surface.Document " S.Ulmer, et.al, Solar Energy, 2011,85 (4) 681-687 " adopts streak reflex method to detect 39.6 square meter settled date mirror surface-shapeds in Spain PSA power station, obtains good measurement result.But, the method employing four-stepped switching policy utilizes the projector arranged in mirror field to the white stripe pattern shielding projected horizontal, vertical both direction respectively on tower, at least project 16 spoke print images, the real-time detected is lower, and whole measuring process can only be carried out at night and be subject to the impact of external light source simultaneously.Therefore a kind of efficient, quick, accurate, easy detection method is needed to detect settled date mirror surface-shaped.

Summary of the invention

The object of the invention is the deficiency existed to overcome prior art, a kind of efficient, quick, accurate, easy tower type solar energy thermal power generation heliostat surface testing method is provided.

Tower type solar energy thermal power generation heliostat surface testing method of the present invention comprises the following steps:

1, compound bar print image screen is arranged on the below of heat dump mouth on heat absorption tower, adjusts the position of the CCD camera on solar eyepiece to be determined and heat absorption tower, make CCD camera can observe the deformation bonding stripe pattern of solar eyepiece to be determined reflection; Compound bar print image screen is diffuse reflection screen, and the compound bar print image fringe frequency in the horizontal direction and the vertical direction presented is identical;

2, utilize CCD camera to gather the deformation bonding stripe pattern of a solar eyepiece reflection to be determined, this picture signal is transferred to computing machine;

3, phase developing method is guided to ask for the deformation bonding stripe pattern PHASE DISTRIBUTION in the horizontal and vertical directions of solar eyepiece to be determined reflection based on window Fourier filtering and quality.Utilize Zernike polynomial expression to obtain to distribute with the virtual reference plane phase of solar eyepiece homalographic to be determined, according to the PHASE DISTRIBUTION of virtual reference plane and solar eyepiece to be determined, obtain the phase offset caused by day to be determined mirror surface-shaped;

4, based on the relative space position of CCD camera, heliostat, compound bar print image screen, set up the numerical relationship model of phase offset and the mirror slope caused by heliostat minute surface, the phase offset caused by heliostat minute surface in integrating step 3 obtains heliostat minute surface slope distribution in both the horizontal and vertical directions;

5, broad sense Hermite interpolation algorithm is utilized to carry out the 3 d shape reconstruct of heliostat.

Detection system of the present invention comprises CCD camera, compound bar print image screen, solar eyepiece to be determined, computing machine etc.Compound bar print image screen is placed in below heat absorption mouth, solar eyepiece to be determined in mirror field is by industrial computer controlling run, the CCD camera being arranged on tower-type solar thermal power generating system heat absorption tower top end is controlled by computing machine, gathers the deformation bonding stripe pattern of solar eyepiece to be determined reflection.Compared with existing detection method, the present invention has the following advantages:

1, the present invention adopts compound bar print image to shield, phase developing method is guided to obtain the PHASE DISTRIBUTION of heliostat in level and vertical both direction based on window Fourier filtering and quality, only need gather the phase pushing figure that piece image can obtain heliostat minute surface, compared with the existing detection method based on four-step phase-shifting, the present invention has that real-time is higher, strong interference immunity, the feature such as lower affected by environment;

2, the present invention utilizes Zernike polynomial expression to obtain the distribution of virtual reference plane phase.Because solar eyepiece minute surface area to be determined is larger, obtain and be difficult to realize with the PHASE DISTRIBUTION of the standard flat mirror of solar eyepiece homalographic size to be determined, therefore first obtain the PHASE DISTRIBUTION of compound bar print image screen, obtain the PHASE DISTRIBUTION with the virtual reference plane of heliostat homalographic in conjunction with Zernike polynomial expression; Compared with existing streak reflex detection method, dirigibility of the present invention is higher.

3, the present invention adopts the deformation bonding stripe pattern of window Fourier filtering to mirror-reflection to carry out phase extraction, compared with conventional Fourier transform method, has that anti-interference is stronger and what obtain is PHASE DISTRIBUTION in level and vertical both direction.

4, the present invention is based on broad sense Hermite interpolation algorithm and three-dimensionalreconstruction is carried out to settled date mirror surface-shaped.By gradient or normal vector to the restructuring procedure of 3 d shape mainly based on integral and calculating.In the ideal case, integration and path have nothing to do, but owing to being subject to the impact such as detection noise, system deviation in actual measurement process, measured gradient is non-conservative field, and path of integration and choosing of integral algorithm will directly cause different reconstructed results.Therefore, for this type of big data quantity, band noise, lack sampling gradient discrete data, utilize broad sense Hermite interpolation algorithm to carry out 3 d shape reconstruction, local and overall reconstruction precision can be guaranteed simultaneously.

Accompanying drawing explanation

The detection system structural representation of Fig. 1 application solar energy thermal-power-generating heliostat of the present invention surface testing method;

The compound bar print image of Fig. 2 solar energy thermal-power-generating heliostat of the present invention surface testing method.

Embodiment

The present invention is further illustrated below in conjunction with the drawings and specific embodiments.

As shown in Figure 1, solar energy thermal-power-generating heliostat surface testing method of the present invention based on detection system comprise CCD camera 2, compound bar print image screen 3, industrial computer 4 and computing machine 1.Described computing machine 1 controls CCD camera 2 and gathers image, and processes obtained image; Described CCD camera 2 is placed in heat absorption tower top end; Described compound bar print image screen 3 is arranged on heat absorption tower; Described industrial computer 4 controls the operation of solar eyepiece 5 to be determined, makes CCD camera 2 can gather the deformation pattern of the compound bar print image screen 3 that heliostat 5 reflects.

Solar energy thermal-power-generating heliostat surface testing method of the present invention comprises the following steps:

1, compound bar print image screen 3 is arranged on the below of heat dump mouth on heat absorption tower, adjusts the position of solar eyepiece 5 to be determined and CCD camera 2, make CCD camera 2 can observe the deformation bonding stripe pattern of solar eyepiece 5 to be determined reflection;

2, phase developing method is guided to ask for the deformation bonding stripe pattern PHASE DISTRIBUTION in the horizontal and vertical directions of solar eyepiece to be determined reflection based on window Fourier filtering and quality.Utilize Zernike polynomial expression to obtain to distribute with the virtual reference plane phase of solar eyepiece homalographic to be determined, according to the PHASE DISTRIBUTION of virtual reference plane and solar eyepiece to be determined, obtain the phase offset caused by day to be determined mirror surface-shaped.

The compound bar print image expression formula that compound bar print image screen 3 presents is:

$I_0(x,y)=A(x,y)+B_1(x,y)\cos \left(\frac{2\mathrm{\πx}}{p_x}\right)+B_2(x,y)\cos \left(\frac{2\mathrm{\πy}}{p_y}\right)---\left(1\right)$

Wherein (x, y) is the coordinate of compound bar print image screen 3, p _xand p _ybe respectively the fringe period in horizontal direction and vertical direction, I ₀(x, y) is fringe intensity, and A (x, y) is bias light intensity, B ₁(x, y) and B ₂(x, y) be respectively striped in the horizontal direction with the amplitude of vertical direction.

The wrapped phase of the compound bar print image that this compound bar print image screen 3 presents is obtained based on window Fourier filtering method:

First carry out window Fourier transform to expression formula (1), its conversion is as follows:

$\mathrm{Sf}(u,v,\mathrm{\ξ},\mathrm{\η})={\∫}_{-\∞}^{\∞}{\∫}_{-\∞}^{\∞}{I}_0(x,y)g(x-u,y-v)e^{(-\mathrm{j\ξx}-\mathrm{j\ξy})}\mathrm{dxdy}---\left(2\right)$

Wherein sf (u, v, ξ, η) is I ₀the Fourier transform of (x, y), (u, v) is volume coordinate, and (ξ, η) is frequency domain coordinates, and window function g (x, y) is Gaussian function:

$g(x,y)=\frac{1}{\sqrt{\mathrm{\π}{\mathrm{\σ}}_x{\mathrm{\σ}}_y}}\exp (-\frac{x^2}{2{\mathrm{\σ}}_x^2}-\frac{y^2}{2{\mathrm{\σ}}_y^2})$

σ _xand σ _ybe respectively Gaussian function in x and y standard deviation upwards.

Secondly, formula (2) is carried out filtering:

$\stackrel{\&OverBar;}{\mathrm{Sf}}(u,v,\mathrm{\&xi;},\mathrm{\&eta;})=\left\{\begin{array}{cc}\mathrm{Sf}(u,v,\mathrm{\&xi;},\mathrm{\&eta;})& \left|\mathrm{Sf}(u,v,\mathrm{\&xi;},\mathrm{\&eta;})\right|\&GreaterEqual;\mathrm{\&theta;}\\ 0,& \left|\mathrm{Sf}(u,v,\mathrm{\&xi;},\mathrm{\&eta;})\right|<\mathrm{\&theta;}\end{array}\right.---\left(3\right)$

Wherein θ is threshold value.

Choose the frequency spectrum fundamental component in the horizontal and vertical directions after filtering respectively simultaneously.Therefore, described compound bar print image wrapped phase in the horizontal and vertical directions be respectively:

$\left\{\begin{array}{c}\stackrel{\&OverBar;}{f_x}(x,y)=\frac{1}{4{\mathrm{\&pi;}}^2}{\&Integral;}_{-\&infin;}^{\&infin;}{\&Integral;}_{-\&infin;}^{\&infin;}{\&Integral;}_{-\&infin;}^{\&infin;}{\&Integral;}_{-\&infin;}^{\&infin;}{\stackrel{\&OverBar;}{\mathrm{Sf}}}_x(u,v,\mathrm{\&xi;},\mathrm{\&eta;})g(x-u,y-v)e^{(\mathrm{j\&xi;x}+\mathrm{j\&xi;y})}\mathrm{d\&xi;d\&eta;dudv}\\ \stackrel{\&OverBar;}{f_y}(x,y)=\frac{1}{4{\mathrm{\&pi;}}^2}{\&Integral;}_{-\&infin;}^{\&infin;}{\&Integral;}_{-\&infin;}^{\&infin;}{\&Integral;}_{-\&infin;}^{\&infin;}{\&Integral;}_{-\&infin;}^{\&infin;}\stackrel{\&OverBar;}{{\mathrm{Sf}}_y}(u,v,\mathrm{\&xi;},\mathrm{\&eta;})g(x-u,y-v)e^{(\mathrm{j\&xi;x}+\mathrm{j\&xi;y})}\mathrm{d\&xi;d\&eta;dudv}\end{array}\right.---\left(5\right)$

Wherein, with be respectively x and y wrapped phase upwards, (x, y) is volume coordinate, with be respectively at x and y fundamental component upwards, with be respectively with inverse Fourier transform, g (x, y) is the Gaussian function in formula (2).

Finally, phase-unwrapping algorithm is guided to obtain wrapped phase based on quality with pHASE DISTRIBUTION with

By above-mentioned continuous phase distribution phi _0x(x, y) and φ _0y(x, y) is expressed as the polynomial linear combination expression formula of Zernike:

$\left\{\begin{array}{c}{\mathrm{\&phi;}}_{0x}(x,y)=\underset{n=0}{\overset{\&infin;}{\mathrm{\&Sigma;}}}{k}_n^1{z}_n(x,y)=k_0^1+k_1^1{z}_1(x,y)+...+k_n^1{z}_n(x,y)\\ {\mathrm{\&phi;}}_{0y}(x,y)=\underset{n=0}{\overset{\&infin;}{\mathrm{\&Sigma;}}}{k}_n^2{z}_n(x,y)=k_0^2+k_1^2{z}_1(x,y)+...+k_n^2{z}_n(x,y)\end{array}\right.---\left(6\right)$

Wherein z _nbe n-th Zernike polynomial expression, be n-th Zernike multinomial coefficient.

Coefficient is solved based on least-squares algorithm utilize the PHASE DISTRIBUTION φ ' of this coefficients to construct virtual plane _0x(x, y) and φ ' _0y(x, y).

3, based on the relative space position of CCD camera, heliostat, compound bar print image screen, set up the numerical relationship model of phase offset and the mirror slope caused by heliostat minute surface, the phase offset caused by heliostat minute surface in integrating step 3 obtains heliostat minute surface slope distribution in both the horizontal and vertical directions.

Industrial computer 4 controls solar eyepiece 5 to be determined, makes CCD camera, and the 2 compound bar print images that can gather heliostat 5 reflection shield the deformation pattern of 3.Computing machine 1 controls the deformation bonding stripe pattern that CCD camera gathers heliostat 5 reflection, and the expression formula of deformation bonding stripe pattern is as follows:

I ₁(x,y)＝a(x,y)+b ₁(x,y)cos[φ _x(x,y)]+b ₂(x,y)cos[φ _y(x,y)] (7)

Wherein I ₁the light distribution that (x, y) is stripe pattern, a (x, y) is background light intensity, b ₁(x, y) and b ₂(x, y) represents the degree of modulation in horizontal direction and vertical direction respectively, φ _x(x, y) and φ _y(x, y) represents the phase place by solar eyepiece 5 mirror shape modulation to be determined in horizontal direction and vertical direction respectively.

Formula (2) formula (7) is utilized to guide phase-unwrapping algorithm to obtain heliostat minute surface phase in the horizontal direction and the vertical direction respectively to the computation process of formula (5) and quality _1x(x, y) and φ _1y(x, y).

Detection system is as shown in Figure 1 known, and the pass of heliostat mirror slope and phase place is:

$\left\{\begin{array}{c}{\mathrm{Slope}}_x=\tan \left(\frac{2p_x({\mathrm{\&phi;}}_{1x}-{\mathrm{\&phi;}}_{0x})}{8\mathrm{\&pi;L}+p_y({\mathrm{\&phi;}}_{1y}-{\mathrm{\&phi;}}_{0y})\sin \left(2\mathrm{\&alpha;}\right)}\right)\\ {\mathrm{Slope}}_y=\tan \left(\frac{p_y({\mathrm{\&phi;}}_{1y}-{\mathrm{\&phi;}}_{0y}){\cos }^2\left(\mathrm{\&alpha;}\right)}{4\mathrm{\&pi;L}+p_y({\mathrm{\&phi;}}_{1y}-{\mathrm{\&phi;}}_{0y})\sin \left(2\mathrm{\&alpha;}\right)}\right)\end{array}\right.---\left(8\right)$

Wherein α is the acquisition angles of CCD camera, Slope _xfor minute surface slope distribution in the horizontal direction, Slope _yfor minute surface slope distribution in vertical direction.

Based on the minute surface phase obtained _1x(x, y) and φ _1ythe phase of (x, y) and virtual plane ' _0x(x, y) and φ ' _0y(x, y), convolution (8) can obtain minute surface slope distribution Slope in the horizontal direction and the vertical direction _xand Slope _y.

4, broad sense Hermite interpolation algorithm is utilized to reconstruct the 3 d shape of heliostat.

The broad sense Hermite interpolation algorithm based on radial basis function is adopted to carry out three-dimensionalreconstruction to the solar eyepiece mirror slope to be determined obtained.Definition interpolation method is as follows:

$s\left(X\right)=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_i{\mathrm{\&psi;}}_x(X-X_i)+\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{\&beta;}}_i{\mathrm{\&psi;}}_y(X-X_i)---\left(9\right)$

Wherein X=(x, y) ^t, α _iand β _i(1≤i≤N) is unknowm coefficient, ψ: R ²→ R is radial basis function, ψ _xand ψ _ybe respectively radial basis function ψ at x and y partial derivative upwards, s (X) is day to be determined minute surface shape function.Set up this interpolation method resolve derivative with survey slope data Slope _xand Slope _ymatching relationship:

$\left\{\begin{array}{c}{s}_x\left(X_j\right)=\mathrm{Slop}{e}_x\left(X_j\right)\\ s_y\left(X_j\right)={\mathrm{Slope}}_y\left(X_j\right)\end{array}\right.,1\&le;j\&le;N---\left(10\right)$

Wherein j is a jth sampled point.Namely following linear equation is solved:

Based on the factor alpha that formula (11) is asked for _iand β _i, apply it in formula (9) interpolation and reconstruct face shape, obtain the 3 d shape s (X) of solar eyepiece minute surface to be determined.

Claims (3)

1. a tower type solar energy thermal power generation heliostat surface testing method, is characterized in that: described detection method comprises the following steps:

(1) CCD camera is utilized to gather the deformation bonding stripe pattern of solar eyepiece to be determined reflection, and store in a computer, then combined window Fourier filtering method and quality guide phase developing method to obtain the deformation bonding stripe pattern PHASE DISTRIBUTION in the horizontal and vertical directions of solar eyepiece to be determined reflection;

(2) utilize Zernike polynomial expression to obtain the PHASE DISTRIBUTION with the virtual reference plane of solar eyepiece homalographic to be determined, according to the PHASE DISTRIBUTION of virtual reference plane and solar eyepiece to be determined, obtain the phase offset caused by day to be determined mirror surface-shaped;

(3) according to the relative space position of the CCD camera in detection system, solar eyepiece to be determined and compound bar print image screen, set up phase offset and mirror slope numerical relationship model that solar eyepiece minute surface to be determined causes, PHASE DISTRIBUTION in integrating step (1) and step (2), obtains heliostat minute surface slope distribution in the horizontal direction and the vertical direction;

(4) based on heliostat slope distribution in broad sense Hermite interpolation method and step (3), solar eyepiece mirror shape to be determined reconstruct is obtained.

2. detection method as claimed in claim 1, is characterized in that: in described step (2), utilizes Zernike polynomial expression to obtain with the method for the PHASE DISTRIBUTION of the virtual reference plane of solar eyepiece homalographic to be determined as follows:

The wrapped phase distribution phase distribution phi of the compound bar print image that described compound bar print image screen (3) presents is obtained based on window Fourier filtering method _0x(x, y) and φ _0y(x, y), is expressed as the polynomial linear combination expression formula of Zernike by this PHASE DISTRIBUTION:

$\left\{\begin{array}{c}{\mathrm{\&phi;}}_{0x}(x,y)=\underset{n=0}{\overset{\&infin;}{\mathrm{\&Sigma;}}}{k}_n^1{z}_n(x,y)=k_0^1+k_1^1{z}_1(x,y)+...+k_n^1{z}_n(x,y)\\ {\mathrm{\&phi;}}_{0y}(x,y)=\underset{n=0}{\overset{\&infin;}{\mathrm{\&Sigma;}}}{k}_n^2{z}_n(x,y)=k_0^2+k_1^2{z}_1(x,y)+...+k_n^2{z}_n(x,y)\end{array}\right.$

Wherein z _nbe n-th Zernike polynomial expression, be n-th Zernike multinomial coefficient;

Coefficient is solved based on least-squares algorithm utilize the PHASE DISTRIBUTION φ ' of this coefficients to construct virtual plane _0x(x, y) and φ ' _0y(x, y).

3. detection method as claimed in claim 1, is characterized in that: the method that described step (4) reconstructs solar eyepiece minute surface 3 d shape to be determined based on broad sense Hermite interpolation and heliostat slope distribution is as follows:

Obtain solar eyepiece minute surface to be determined slope in the horizontal direction and the vertical direction by described step (3) and be respectively Slope _xand Slope _y, definition Hermite interpolating function is:

$s\left(X\right)=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{\&alpha;}}_i{\mathrm{\&psi;}}_x(X-X_i)+\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{\mathrm{\&beta;}}_i{\mathrm{\&psi;}}_y(X-X_i)$

Wherein X=(x, y) ^t, α _iand β _i(1≤i≤N) is unknowm coefficient, ψ: R ²→ R is radial basis function, ψ _xand ψ _ybe respectively radial basis function ψ at x and y partial derivative upwards,

S (X) is day to be determined minute surface shape function;

Set up this interpolation method resolve derivative with survey slope data Slope _xand Slope _ymatching relationship:

$\left\{\begin{array}{c}{s}_x\left(X_j\right)={\mathrm{Slope}}_x\left(X_j\right)\\ s_y\left(X_j\right)={\mathrm{Slope}}_y\left(X_j\right)\end{array},1\&le;j\&le;N\right.$

Wherein j is a jth sampled point;

Namely following linear equation is solved:

Based on the factor alpha asked for _iand β _i, namely obtain the 3 d shape s (X) of solar eyepiece minute surface to be determined.