CN104350676B - The concentrator of polychromatic light - Google Patents

Abstract

One example of a solar current concentrator has a primary Fresnel lens comprising multiple panels, each of which forms a Kohler integrator with a corresponding panel in the lens secondary. The resulting multiple integrators concentrate sunlight onto a common multi-junction photovoltaic cell. Integrators provide matched illumination in the different bands required by different junctions. It is also possible to use illuminators with similar geometries.

Description

polychromatic light concentrator

Cross References to Related Applications

This application requires Benitez and Filed on April 16, 2012 entitled "Domed Fresnel Concentrator", which is hereby incorporated by reference in its entirety.

refer to Filed on 18 November 2008 on behalf of et al. entitled " Concentrator" U.S. Provisional Patent Application No. 61/115,892 and U.S. Provisional Patent Application No. 61/268,129 of the same title filed on June 8, 2009 and filed on October 6, 2009 in the name of Benítez et al. Submission entitled " concentrator azimuthally combining radial- sub-concentrators", commonly assigned U.S. Provisional Patent Application No. 61/278,476, and the subsequent issued August 16, 2011, entitled " Concentrator", U.S. Patent No. 8,000,018, which is hereby incorporated by reference in its entirety.

referenced commonly assigned International Patent Application No. PCT/US 2006/029464 (WO 2007/016363) assigned to Benítez et al. It is hereby incorporated by reference in its entirety.

Embodiments of the apparatus shown and described in this application may be within the scope of one or more of the following U.S. patents and patent applications and/or equivalents in other countries: U.S. Patent 6,639,733 issued October 28, 2003 to Benítez et al., and U.S. Patent 7,460,985 issued December 2, 2008 to Benítez et al.; WO 2007/016363 mentioned above, and also in US 2008/0316761 of the same title published on December 25, 2008 in the name of Benítez et al. and entitled "Multi-Junction Solar Cells with a Homogenizer System and Coupled Non- - Imaging Light Concentrator" WO 2007/103994; US 2008/0223443 published September 18, 2008 entitled "Optical Concentrator Especially for Solar Photovoltaic" in the name of Benítez et al; and 3 in 2009 in the name of Benítez et al US2009/0071467 entitled "Multi-Junction Solar Cells with a Homogenizer System and Coupled Non-Imaging Light Concentrator" published on March 19.

Glossary

Primary Optical Element (POE)—An optical element (which may be a surface of a refractive element) that receives light from the sun or other source and concentrates it toward a secondary optical element.

Secondary Optical Element (SOE) - An optical element (which may be a surface of a refractive element) that receives light from the primary optical element and concentrates it toward a solar cell or other target.

Kohler integrator ( Integrator)—strictly speaking, an optical device in which a primary optic images a source onto a secondary optic and a secondary optic images the primary optic onto a target. In this specification, as explained below, the focal point of the primary optical element intentionally does not exactly coincide with the secondary optical element.

Parallel acceptance angle (α _p )—for a flat rectangular solar cell, the angle of incident light relative to the direction of perfect alignment in a plane containing the direction of perfect alignment and parallel to the two sides of the solar cell, compared to The equal incident irradiance in the quasi-direction drops by 10% compared to the cell photocurrent at this angle.

Diagonal Acceptance Angle (α _d )—the angle of incident light in a plane parallel to the direction of perfect alignment and containing one of the diagonals of the solar cell relative to the direction of perfect alignment at which the cell photocurrent drops by 10% .

Geometric concentration (Cg)—the ratio of the projected area of the POE perpendicular to the center of the sun to the battery area.

Concentration-Acceptance Product (CAP)—a parameter associated with any solar concentration architecture, and is defined as the square root of the geometric concentration multiplied by the acceptance (as the minimum of parallel and diagonal acceptance angles) The product of the sines of the angles. Some optical architectures have a higher CAP than others, enabling higher concentrations and/or higher acceptance angles. For a particular structure, CAP is nearly constant when the geometric concentration is changed, so increasing the value of one parameter decreases the value of the other.

Fresnel Facet - An element of a discontinuous-slanted lens that deflects light by refraction.

Cartesian Oval - A curve (strictly speaking, family of curves) used in imaging and non-imaging optics to transform a given set of light into another predetermined set. See reference [10], p. 185, reference [14].

Perfectly aligned position - the central direction of incident collimated light or incident sunlight (source diameter 0.5° centered in the nominal direction), away from which performance degrades in all directions. In all the embodiments described below, the perfectly aligned position is the intersection of the planes of symmetry of the overall concentrator, but this is not necessarily the case for individual segments. However, the rate of performance degradation may be different in different planes, eg see α _p and α _d above.

Uniformity - the ratio of minimum to maximum irradiance on the cell with the sun centered on a perfectly aligned location.

Background technique

Triple-junction photovoltaic solar cells are expensive, so it is desirable to operate with as much concentration as practical. However, the efficiency of currently available multi-junction photovoltaic solar cells deteriorates when the local concentration of incident radiation exceeds 2,000-3,000 suns (sun). Some concentrator designs of the prior art have substantial non-uniformity with respect to the flux distribution across the cell such that as much as 20x average concentration occurs (9,000-11,000x concentration with 500x average concentration) The "hot spot" of , thus greatly restricting the maximum average concentration that is commercially feasible.

By using a light-pipe homogenizer, a method well known in classical optics, it is possible to obtain good irradiance uniformity across solar cells. See reference [1]. When using a light pipe homogenizer, the solar cell is attached to one end of the light pipe and the light reaches the cell after some bounce off the light pipe wall. As the light pipe length is increased, the light distribution on the cell becomes more uniform. However, the use of light pipes for concentrating photovoltaics (CPV) has some drawbacks. The first drawback is that in the case of high illumination angles, the reflective surface of the light pipe needs to be metallized, which reduces the optical efficiency relative to the near-perfect reflectivity of total internal reflection of the polished surface in dielectric based light pipes . A second drawback is that for good homogenization a relatively long light pipe is necessary, but increasing the length of the light pipe both increases its absorption and reduces the mechanical stability of the device. A third drawback is that light pipes are not suitable for relatively thick (small) cells due to side light spillage from the edges of the bonding material (usually made of silicone rubber) used to hold the cell to the end of the light pipe. Finally, the amount of bonding material used in the adhesive layer is critical. Too little material and there will be an air gap over part of the cell, causing losses due to Fresnel reflections. Too much material will cause the overflow problem mentioned above. The smaller the area of the cell, the greater proportion of solar radiation is lost through overflow. Even so, light pipes have been proposed several times in CPV systems, see references [2], [3], [4], [5], [6], and [7], their usage is much longer than the cell size The length of the light pipe is usually 4-5 times.

Another strategy for achieving good uniformity across the cell is the Kohler illuminator. This technique can solve or at least alleviate the uniformity problem without compromising acceptance angles or increasing assembly difficulty.

Sandia Labs proposed the first photovoltaic concentrators using Koehler integrals in the late eighties (see Ref. [8]), and was subsequently commercialized by Alpha Solarco. The design uses a standard radial concentric Fresnel lens as its primary optical element (POE), and an imaging single-surface lens (called SILO (SIngLeOptical), for single optical surface) encapsulating the photovoltaic cell as its secondary optical element ( SOE). This approach uses two imaging optics (Fresnel lens and SILO), where the SILO is arranged at the focal plane of the Fresnel lens and the SILO images the Fresnel lens (which is non-uniformly illuminated) to the photovoltaic cell. Therefore, if the cell is square, the primary element can be square cut without loss of optical efficiency.

Despite the simplicity and high uniformity of illumination of the cell, the practical application of the Sandia Labs system is still limited to low concentrations due to its low concentration acceptance area of approximately 0.3° (±1° at 300× ). This is due to the inability of the imaging secondary to capture high illumination angles on the cell, thus resulting in low acceptance angles even at an intensity ratio of only 300×.

Another previously proposed Kohler approach uses 4 optical surfaces to obtain a photovoltaic concentrator with high acceptance angle and relatively uniform radiation distribution over the solar cell (see Ref. [9]). The POE of the concentrator is a double aspheric imaging lens that images the sun onto the aperture of the SOE. Suitable secondary optics are the RX concentrators designed by SMS described in refs [10], [11] and [12]. This is an imaging element operating near the thermodynamic limit of the concentrator. This concentrator is of academic interest only because both the dual aspheric element and the RX concentrator are not economically viable for practical applications, and thermal management of the photovoltaic cell is impractical in this configuration of.

In US 8,000,018 B2 some of the inventors here found a practical solution to increase the concentration acceptance product compared to the previous Kohler approach. It involves dividing the POE and SOE into sectors providing independent Kohler channels, so each SOE sector needs to manage only a correspondingly smaller field of view and provide a correspondingly smaller concentration. Furthermore, the multi-channel approach offers further improvement and robustness due to the principle of superposition: degradations caused by any cause in one of the POE images are less likely than in the single-channel case due to their smaller contribution to the total irradiance produced. (such as the SILO problem mentioned earlier) is even less significant.

The most notable embodiment in US 8,000,018 B2 comprises a 4-fold symmetrical flat Fresnel lens and a 4-fold single surface secondary lens. The device uses a single wavelength to conceptually design each quadrant of the Fresnel lens. The location of the monochromatic focus is indicated as being on the surface of the SOE, or deeper inside the body of the SOE, closer to a certain chord. A polychromatic design that takes into account the different responses of optical elements to the different spectral bands of a multi-junction solar cell, the combined behavior of which is strongly nonlinear, has not been indicated.

This polychromatic optimization has not been applied to other related architectures such as the 9-fold Fresnel-Kohler design also mentioned in US 8,000,018 B2.

Although most of the Fresnel lenses used in this application are flat, better concentration acceptance angle products have been achieved with rotationally symmetric dome lenses [5].

As described in commonly assigned US 2010/0269885, obtaining optimum efficiency from multi-junction solar cells may require extreme care in balancing the irradiation of the different cells.

Most of the same or a very similar problem occurs when generating a white beam from a luminaire with a white light source, reversing the direction of the light.

Contents of the invention

Embodiments of the present invention provide different photovoltaic concentrators that combine high geometric concentration, high acceptance angle, and high irradiance uniformity on solar cells. In all embodiments, the primary optical element and the secondary optical element are each lenticular shaped to form a plurality of segments. Sections of primary optics and sections of secondary optics combine to form a Koehler integrator. Multiple nodes lead to multiple Kohler integrators that collectively focus their incident sunlight onto a common target, such as a multi-junction photovoltaic cell, allowing for polychromatic optimization means to separate the multi-junction solar cells Responses to different spectral bands. Any hot spots are typically in different places for different individual Kohler integrators, and the multiple Kohler integrators further average the hot spots on the target cell.

Embodiments of the present invention provide an optical device comprising: a multi-junction photovoltaic cell, wherein each junction is operable to convert light of a corresponding wavelength band into electricity; a refractive first optical element having a plurality of segments each arranged to focus incident collimated light from a common source; and a second optical element having a plurality of segments each arranged to direct light from a corresponding segment of said first optical element to said on a photovoltaic cell; wherein the acceptance angles of incident light in two of the wavelength bands are within a ratio of 5:4 to 4:5.

The acceptance angle may be in a ratio of 5:4 to 4:5 for incident light of the shortest and longest of said bands, and advantageously for all of three or more bands.

If the cell, the first optical element (projected into a plane perpendicular to the direction of perfect alignment) and the sections in the first optical element (projected similarly) are all square and are all aligned in the same direction, then The ratio of α _p (top) to α _d (bottom) is ideally within the ratio of 5:4 to 4:5, where α _p (top) is measured in a plane parallel to one side of the cell in the band The acceptance angle of the shortest band, α _d (bottom) is the acceptance angle of the longest of the bands measured in the plane containing the diagonal of the cell, and each of the acceptance angles is defined as uniform incidence The angle between collimated light and the direction of perfect alignment at which the light energy directed onto the cell is 90% of the energy directed onto said cell for the same incident collimated light in the direction of perfect alignment .

The first optical element may be a Fresnel lens or other discontinuous surface lens. Each segment of the Fresnel lens may then comprise a Fresnel lens with a different center. Alternatively, the first optical element may then comprise a sheet formed on one face with Fresnel lenses common to all segments and on the other face with individual continuous slope lenses for each segment. Fresnel lenses may be domed.

The CAP may be at least 0.45 for at least two of said bands and preferably simultaneously for all of said bands.

The uniformity in the direction of perfect alignment may be at least 0.5, better at least 0.67, preferably at least 0.8 for all wavelength bands.

One embodiment of the present invention provides an optical device comprising: a primary optical element having a plurality of segments, in one example the number is four; and a secondary optical element having a plurality of segments, in one example being 4 lenticulations of the lens optical surface; where each segment in the primary optics together with a corresponding segment in the secondary optics constitutes one of a plurality of Koehler integrators. The plurality of Kohler integrators are positionally and azimuthally arranged to direct light in the plurality of spectral bands from a common source to a common target.

For example, in the case of solar photovoltaic concentrators, the source is the sun. Whether it is a common source or a common destination, another may be part of or connected to the device. For example, in a solar photovoltaic concentrator, the target may be a photovoltaic cell.

Embodiments of the invention also provide other forms of light collectors and collimators, including light collectors and illuminators with similar optical properties. The common source in case the device is a light collector or the common target in case the device is a luminaire may be located outside the device. The following examples are primarily intended for use as solar concentrators. For illuminators, the source and target are usually interchanged so that the light is highly concentrated at the source behind the "secondary" optics and collimated primarily on its way to the external target in front of the "primary" optics.

Embodiments of the invention also provide methods of designing and fabricating solar concentrators and other optical devices with specified novel properties.

Embodiments of the present invention make it possible to solve the following three problems simultaneously or at least mitigate the consequences of not solving the following three problems simultaneously:

1. For all of the three or more bands at normal (perfect alignment) incidence, the light collection efficiency will be as close to 100% as possible, that is, the light of the three top, middle and bottom junctions is fully collected.

2. The irradiance on the cell is balanced for three or more bands. This is not obtained in monochrome designs without homogenization schemes such as Kohler integrals, where usually the mid-band produces a hotspot in the center.

3. The overall acceptance angle of the three segments will be maximized, which generally requires the acceptance angles of the three segments to be balanced as equally as possible, since the smallest of the three acceptance angles effectively limits the overall acceptance of the device.

Description of drawings

Figure 1 is a perspective view of the primary and secondary lenses of a previously proposed 4-fold Koehler concentrator.

Figure 2 shows a set of three ray traces through the SOE of one embodiment of a solar concentrator for three different wavelengths of light.

Fig. 3A shows the ray trace in the parallel direction of the top-junction band with an incident angle equal to 0.95α.

Figure 3B shows a ray trace similar to that in Figure 3A, but in a diagonal direction and for the bottom knot segment.

Figure 4A is a plot of irradiance at location over the area of the top junction segment for a flat POE design optimized for furthest point use of SOEs in a photovoltaic cell.

Figure 4B is a plot similar to Figure 4A for the bottom knot segment.

Figure 4C is a plot similar to Figure 4A using closest point optimization of SOE.

Figure 4D is a plot similar to Figure 4C for the bottom junction.

Figure 5A is a perspective view of the primary and secondary lenses of a RR dome Fresnel-Kohler concentrator.

FIG. 5B is an enlarged view of the SOE of FIG. 5A showing a ray trace.

Figure 6 is an axial sectional view showing the design of the dome concentrator.

Figure 7A is a 3D plot of the irradiance of the top junction of a concentrator with a domed POE optimized for maximum uniformity.

Figure 7B is a plot similar to Figure 7A for the bottom knot segment.

Figure 7C is a 3D plot of the irradiance of the top segment of a concentrator with a domed POE optimized for maximum CAP.

Figure 7D is a plot similar to Figure 7C for the bottom knot segment.

Figure 8A shows plan and perspective views of a circular dome Fresnel lens with four segments formed by a lens grating on the upper surface and a common non-rotationally symmetric (spiral) Fresnel lens on the underside.

8B is a top perspective view of a square Fresnel lens taken from the circular lens of FIG. 8A.

Figure 9A is a perspective view of the primary and secondary lenses of a 2-section RR Fresnel-Kohler concentrator.

FIG. 9B is an enlarged view of the SOE in FIG. 9A showing a ray trace.

Figure 10 is a perspective view of the primary and secondary lenses of a 9-section RR Fresnel-Kohler concentrator, and two different enlarged views of the SOE with and without ray tracing.

detailed description

A better understanding of the various features and advantages of the invention may be obtained by referring to the following detailed description of the embodiments of the invention and the accompanying drawings which illustrate exemplary embodiments employing the various principles of the invention.

The primary optical element (POE) described in these embodiments is formed as multiple segments, thereby exhibiting multiple symmetries. In embodiments taught herein, the secondary optical element (SOE) has the same multiple symmetry as the corresponding POE. Each section of the POE together with a corresponding section of the SOE constitutes one of a plurality of Kohler integrator sections. The multiple Kohler integrator nodes combine to concentrate incident sunlight onto a common photovoltaic cell.

The solar cells of currently available solar concentrators use three junctions commonly referred to as top (top), middle (middle) and bottom (bottom), which are sensitive to different spectral bands in solar radiation. The semiconductor physics of a junction determines the minimum photon energy (maximum wavelength) of light that the junction can convert into electricity. Typically, the top junction is sensitive from 350 to 690nm, the middle junction is sensitive from 690nm to 900nm, and the bottom junction is sensitive beyond 900nm (the germanium bottom junction can in principle use light down to about 1800nm, while indium gallium arsenide (InGaAs) or nitrogen antimony arsenic InGaAsNSb bottom junctions can use light down to only about 1400nm), the transition between cell bands does not have to be abrupt. When the POE is a mirror, the direction of reflected light does not depend on wavelength, and a monochromatic concentrator design is sufficient to predict full-spectrum performance.

However, in all current embodiments, the POE is refractive, and the variation of the lens material's refractive index with wavelength (often referred to as material dispersion, which is responsible for chromatic aberration in imaging optics) causes light of different wavelengths to be Refract in different directions to reach different points in the SOE. For solar concentrators, these wavelength-dependent deviations of light from different junctions will cause two effects that should be taken into account: (1) there may be three different acceptance angles for different junctions; and (2) the irradiance The degree distribution may also be different for different knots. The consequence of the first effect is that the effective acceptance angle of the concentrator as a whole is the smallest of the three, thereby limiting the CAP of the device. The second effect reduces the overall solar cell efficiency because the three junctions work in series and the lowest illuminated junction limits the current output of the stack. When the irradiance distributions of the bands used by the three junctions differ, the minimum luminance limitation occurs locally, only partially mitigated by the lateral current, even when the total integrated illumination of the cell is the same for the three junctions.

Figure 1 shows one of the earlier embodiments in US 8,000,018 B2, where the POE is a flat Fresnel lens with 4-fold symmetry. Each of the four Fresnel lens segments is a portion of a lens that has rotational symmetry about one of four axes that are not coincident with each other and with the center of the overall optical system. The normally incident ray 11 is divided into four discrete beams to reach the four lobes of the 4-fold symmetric SOE. The focal point 12 of the POE lens section is formed close to the front surface of the SOE. The design ignores dispersion and assumes that tracing a single set of rays is sufficient.

The embodiments in this application are polychromatically optimized to obtain a solution that can achieve high optical efficiency and also correct for the two effects mentioned above. That is, they can be achieved as additional performance goals: (1) the concentrator acceptance angle α given by the smallest of the three junctions is the largest, and (2) the irradiance distribution of the three junctions is very akin.

Although the optimization described here can be applied to a general NxM symmetric design, three specific preferred embodiments are included in the present invention: 2x2 symmetry with flat or domed Fresnel primary lenses, which will be referred to simply as 4-fold; 3x3 symmetric with flat or domed Fresnel lenses, which will be referred to simply as 9-fold; and 2x1 symmetric with flat or domed Fresnel primary, which will be referred to simply as 2-fold.

The performance objective (1) is obtained through POE optimization. To illustrate this optimization, FIG. 2 shows a side view of a cross-section along the diagonal of a quadruple SOE 202 with a triple-junction cell 201 surrounded by light 205 in the top junction band, light 204 in the middle junction band, and a bottom junction cell 201. The band of light 203 is illuminated. The POE (not shown in Figure 2) is assumed to be a Fresnel lens. The focal regions are located at three very different locations 206, 207 and 208, with the top junction point 206 being the shallowest and the bottom junction point 208 being the deepest into the SOE.

Although only a single focal area is mentioned in US 8,000,018 B2, the polychromatic optimization disclosed here will take into account the positions of the three focal points. In the case of flat Fresnel POEs, their positions cannot be independently controlled, so we can specify the position of one focal point and calculate the other two.

A focal point can preferably be designated as the point at which light of the selected color will conceptually be focused by the POE if further refraction of the light by the SOE is not intervened. For example, point 209 in Figure 2 corresponds to such an imaginary focal point for light having a wavelength of 550 nm. The two coordinates (x _m , z _m ) of a point 209 in the tilted coordinate system xz shown in FIG. 2 constitute the two parameters to be changed in order to achieve the performance goal (1). Therefore, the goal is to solve the mathematical problem of finding the maximum value of the bivariate function α(x _m , z _m ). Since the definition of α=min{α(top), α(middle), α(bottom)} is very nonlinear and its derivatives are not continuous, it is useful to visualize the overall shape of the function. For example, in the case of the 4-fold embodiment, the following inequalities are valid around the optimum value.

α _p (top)<α _p (middle)<α _p (bottom)

.α _d (bottom)<α _d (middle)<α _d (top)

where α _p and α _d denote the parallel and diagonal acceptance angles as defined in the Glossary. The first two equation lines show that the parallel acceptance angle of the top (short wavelength) junction is smaller than that of the other two junctions, while for the diagonal orientation, the bottom junction is the limiting junction.

The last two lines of equations show that, for a constant z _m , as x _m increases, the bounded parallel acceptance angle α _p (top) increases while the bounded parallel acceptance angle α _d (bottom) decreases. Thus, for each z _m , there is a compromise solved for the value of x _m where α _p (top) = α _d (bottom), then accepting that the angle α will be the maximum. Thus, we have found that the desired maximum in objective (1) is obtained when the top and bottom acceptance angles are balanced.

The previous calculation can now be done with varying _zm , so we can find the value of _zm at which the uniformly accepted angle αp(top) = _{αd (} bottom) is the maximum, leading to the desired absolute maximum . Note that since α(top)=min{αp(top), _αd (top)} and _α (bottom)=min{αp(bottom), _αd (bottom)}, we get α= _α (top )=α (bottom)<α (middle).

Figure 3A shows the ray traces of the top junction spectral bands for the optimized design along parallel directions at an angle of incidence equal to 0.95α. Figure 3B shows the raytrace along the diagonal direction at the same angle of incidence 0.95α but for the bottom junction spectral band. When ray 301 entering the SOE closest to the top cusp of the SOE reaches the adjacent lobe with the top junction parallel (see FIG. 3A ), and when ray 302 entering the lowest SOE below is with the bottom junction diagonal When the target cell is missed (FIG. 3B), the 10% reduction that would occur in both directions at the angle of incidence α will occur.

Because the refractive index of typical POE lens materials is not very different from the middle junction to the bottom junction, the focal points 207 and 208 in FIG. 2 are relatively close and the middle and bottom acceptance angles are also close. As a result, instead of making the parallel top acceptance angle equal to the diagonal bottom acceptance angle, the parallel top acceptance angle and the diagonal mid-acceptance angle can be made equal. This is especially true in the case of solar cells where there is too much bottom junction photocurrent (as in commercially available GaInP-GaInAs-Ge cells), so that the bottom junction current is less likely to be limiting.

In US 8,000,018 only a monochromatic design is disclosed and no mention is made of different parallel and diagonal acceptance angles. In US 8,000,018 we propose to place a single focus on the surface of the SOE or along the chord joining the edge of the meridional curve passing through the segmented optical axis, which corresponds to _zm = 0 in this specification points in the line.

As will now be shown with an example, monochromatic designs with those focus choices result in very low acceptance angles compared to the present polychromatic optimization, and result in a poor balance between the acceptance angles of the top, middle and bottom junctions.

Two 4-fold devices with flat Fresnel lenses were designed, both with geometric concentration Cg=1024×, POE made of silicone on glass of size 160mm×160mm, with 21.8mm average diameter SOE, gallium indium phosphide-gallium indium arsenide-germanium (GaInP–GaInAs–Ge) triple-junction 5mm × 5mm solar cells made of Savosil glass and a depth-to-POE diagonal ratio of 1.08. One of the devices was designed using the polychromatic optimization just described, while the other device was designed using the procedure disclosed in US8,000,018.

In order to reproduce the monochromatic design of US 8,000,018, the acceptance angle of the selected wavelength is maximized. The chosen wavelength is 550nm, which is the center of the top junction. This wavelength is often used in optical components because the refractive index at this wavelength takes approximately the median value of the distribution. The choice of wavelength at 550nm is not stated in US 8,000,018, but can be easily deduced from column 8, lines 40-50, which state the polychromatic ray tracing analysis implementation of the top junction segment Acceptance angle of ±1.43° and monochromatic analysis of ±1.47°. The proximity of these two values is consistent with the 550nm selection.

Table 1 shows a comparison of the performance parameters of the two designs obtained using ray tracing analysis:

Table 1

α (top) α (medium) α (bottom) alpha CAP x _m (mm) z _m (mm) Color optimization ±0.90° ±0.94° ±0.90° ±0.90° 0.51 11.7 -5.1 US 8,000,018 ±0.92° ±0.66° ±0.63° ±0.66° 0.35 10.3 0

The first significant difference between the results of the two devices is that the focal point of the 550nm light in this polychromatic optimization is 5.1 mm from the line z _m =0 in US 8,000,018. This distance is as large as the battery side, indicating a significant difference between this polychromatic optimized design and that disclosed in US 8,000,018.

The second difference is that although the acceptance angles of the three junctions are excellently balanced in the polychromatic design, the bottom and middle junction acceptance angles in the concentrator of US 8,000,018 are 30% lower than the top junction acceptance angles, so there is no Balance is remarkable.

The third significant difference is that the resulting concentrator acceptance angle and CAP given by α=min{α(top), α(middle), α(bottom)} are also 30% less in the device of US 8,000,018 .

For this comparison, the US 8,000,018 concentrator is designed with the 550nm focus at the line _zm = 0, but as mentioned earlier, in that patent we propose instead to place the focus at the SOE surface. This alternative resulted in even lower acceptance angles and CAPs than shown in the table above.

The POE and SOE materials selected for the previous examples are now of particular interest due to their expected long-term durability. However, they have relatively low refractive indices (silicone has n=1.41 and Savosil has n=1.46 at 550 nm), which makes them especially for low depth-to-POE diagonal ratios (1.08) Achievable acceptance angles and CAP are lower than alternative materials. For example, use PMMA for POE and B270 glass for SOE (which have n=1.49 and n=1.52 respectively) - they are also excellent candidates in terms of durability, at the same depth to POE diagonal ratio In the case of , the same design algorithm results in a 15% higher CAP value due to their higher refractive index.

On the other hand, the performance objective (2) is obtained by optimizing the SOE, and the performance objective (2) is that the irradiance distributions of the three junctions must be very similar. In this case, a single third design parameter is sufficient to obtain excellent results, and that is the point along the cell diagonal that will be imaged by each quadrant to its corresponding POE segment. Because the required SOE imaging is less affected by the dispersion caused by the large field of view imaged by the SOE, the central wavelength can be used for the calculations. The actual optimum location depends on the particular embodiment.

For example, in the case of concentrators with 4-fold or 9-fold flat POEs, the optimum is obtained when the selected cell diagonal point is the end of the diagonal furthest from the POE paired sections, as in Disclosed in US 8,000,018. However, for a POE of a dome, the sweet spot is the end closest to the diagonal of the paired segments. Although this choice is optimal for the balance of irradiance (2), it is not optimal for other criteria such as acceptance angle. For example, Figures 4A and 4B show the irradiance at the extreme bands of the top and bottom junctions, respectively, of a flat POE design with PMMA POE and B270 glass SOE using the furthest point selection, which is the most extreme for target (2). Nice and so they look alike. On the other hand, Fig. 4C and Fig. 4D show the same plot for a flat POE design using closest point selection, where a significantly worse uniformity balance is visible. However, the farthest-point configurations in Figures 4A and 4B have a CAP = 0.59, while the closest-point configurations in Figures 4C and 4D achieve a CAP = 0.63.

Furthermore, the present polychromatic optimization enables much better uniformity over the three knots. For previous monochromatic designs, as shown in Figures 9 and 10 of Kritchman's reference [15], too perfect focus at the nominal frequency sometimes resulted in wildly inhomogeneous irradiance at that frequency, and/or from one band The uniformity varies widely from one to another.

Tables 2 and 3 below show numerical data describing the shape of the polychromatically optimized design just discussed, the performance data of which is given in the previous table 1). For both POE and SOE, a Cartesian coordinate system was used with the origin at the center of the cell and the X and Y axes parallel to each cell side, and the z axis perpendicular to the cell plane. The size of the cells is arbitrary, and the battery active area is 5x5 and the POE is 160x160 in these cells. Due to the 4-fold symmetry, the lens needs to be described only in the quadrants X>0 and Y>0. Both POE and SOE are symmetrical with respect to the plane X=Y.

The surface of the SOE has rotational symmetry with respect to the line joining the cell corner at X=Y=-7.071 and Z=0 and the POE corner at "X=Y=113.2, Z=244.5". The list of SOE sagittal (sag) is given in Table 2:

Table 2

The Fresnel lens segment has an axis of rotational symmetry parallel to the Z axis at X=Y=5.14468, and the vertices describing its polygonal profile are given in Table 3 below:

table 3

Figures 5A and 5B disclose another preferred embodiment consisting of a 4-fold symmetric device, where the POE is a dome-like shape rather than a flat Fresnel lens 501 . The additional degree of freedom to choose the curve of the overall profile of the Fresnel lens makes it possible to control the positions of the two foci of two junctions rather than one junction in polychromatic optimization.

Figure 6 shows the steps in the design of a dome 4-fold concentrator. First, point A on the POE arranged on the optical axis is selected. Then, point C of the SOE on the axis of symmetry is chosen, and the SOE is designed as a Cartesian oval coupled spherical wavefront with origins at A and point E on the near side of the cell passing through C. Next, point D is chosen as the point of the SOE where the meridional tangent to the SOE forms an angle (typically 5° to allow easy release of the SOE part) from the vertical. Later, POE is designed from A to B. A suitable method is described by Kritchman et al, Appl. c starting in ray a and ending with ray a) is focused on D and parallel rays (starting in ray d and ending with ray b) of oblique -α are focused on C, where α is the desired acceptance angle. Kritchman only describes the design of cylindrical lenses, but it is now within the state of the art to generalize Kritchman's method to dome lenses. Kritchman did not consider the possibility of a Kohler configuration, but could still be used to locate the primary focus 209 .

In this example, the Kritchman method is modified into a polychromatic design, where rays focused on C are selected to have a short wavelength in the solar spectrum (e.g., 450 nm, in the top junction) and rays focused on D are selected is of long wavelength (eg, 1000 nm in the bottom junction). A ray impinging on point A within ±α will go to E after being refracted in the POE and SOE. Then, similar to the described polychromatic optimization, the coordinates of points F, C and the abscissa of D are treated as free parameters that define the space in which the acceptance angle is maximized to achieve the performance goal (1).

Due to the less constrained POE design, the dome Fresnel design can achieve a lower depth-to-diameter ratio than the flat Fresnel design (0.7 for the dome Fresnel compared to 0.9 to 1.2 for the flat Fresnel to 0.9) and higher CAP (up to 0.73).

With respect to the balanced irradiance distribution desired in performance goal (2), Figures 7C and 7D show the results for the design of Cg = 1,234 for the extreme bands of the top and bottom junctions of the dome designed according to the previous paragraph, respectively. radioactivity. It shows that the irradiance is significantly less uniform and less similar than the flat Fresnel case in FIGS. 4A-4D . The reason is, because the lens is not flat, so even if its projection on the plane perpendicular to the direction of the sun is square, the angular field seen by the SOE lobe is not square, and the image projected by the SOE is very distorted, and due to The low depth-to-POE diagonal ratio is wavelength-dependent. This design achieves CAP = 0.73. If the abscissas of C and D were reduced, a much more balanced uniformity could be provided, as shown in Figures 7A and 7B, but with a lower CAP=0.55.

Besides higher compactness and CAP, the advantage of dome Fresnel over flat Fresnel is its smaller SOE (which means lower absorption inside the material and lower cost during glass molding ). However, fabrication of dome lenses becomes challenging because the combination of lens convexity and high optical efficiency can result in facets of domed POEs with negative draft angles. One technique is based on the use of PMMA injection molding with movable moulds, as has been developed for rotationally symmetric lenses by the Japanese company DaidoSteel, see Ref. [5]. An alternative is shown in Figure 8, where the Fresnel inner face of the lens is made using a helical profile 81 (which can be demolded, like a bolt, by a combination of turning and pulling), truncated as Square protruding aperture 82. The outer surface has four lobes 83 to produce the desired beam splitting. Construct a helix from a 2D polygonal profile contained in a meridian plane in three steps that are available in many CAD packages: (a) generate a linearly varying helix through concave vertices, (b) also for convex vertices with (a ), and (c) sweep the facet profile along the spiral using the spiral as a track. Of course, it is also possible to use the 4-fold front surface 83 of the POE in combination with a rotationally symmetric inner Fresnel surface.

The two following Tables 4 and 5 show numerical data describing the shape of the polychromatic optimized dome design just discussed, where CAP = 0.73. For both POE and SOE, a Cartesian coordinate system is defined with the origin at the center of the cell and the X and Y axes parallel to each cell side, and the z axis perpendicular to the cell plane. The size of the cells is arbitrary, and the battery active area is 5x5 and the POE is 176x176 in these cells. Due to the 4-fold symmetry, the lens needs to be described only in the quadrants X>0 and Y>0. Both POE and SOE are symmetrical with respect to the plane X=Y.

The surface of the SOE has rotational symmetry about the line joining the cell corner at X=Y=7.071 and Z=0 with the POE vertex at X=Y=0 and Z=200. The SOE sagittal list is given in Table 4:

Table 4

The domed Fresnel lens segment has an axis of rotational symmetry parallel to the Z axis at X = Y = 5.211, and the points describing the inner polygonal profile and the smooth outer profile are given in Table 5 below:

Embodiments so far have had 2x2 symmetric cells, but the polychromatic optimization described can be applied to other more general NxM schemes. Figure 9 shows a 1 x 2 design in which a rectangular POE lens 91 is divided into two sections that concentrate sunlight onto a double lobe SOE lens 92, thereby splitting the beam into two channels that produce two Focus 93 and 94. Devices where N is different from M have the ability to produce different acceptance angles in N and M directions. This is beneficial for setting high acceptance angle orientations parallel to the elevation axis, where mechanical constraints are higher in typically rectangular arrays. The design in Figure 9 has Cg = 312x and acceptance angles of ±1.37° and 1.62° (larger in the direction of the long side of the POE rectangle).

The acceptance angle of the 1x2 concentrator can be optimized similarly to the 2x2 concentrator described above. The "long parallel (longparallel)" direction is defined as parallel to the edge extending first along one section of the primary optical element 91 and then along the other section, and the "short parallel (short parallel)" direction is defined as parallel to the normal Cross-edge optimization can be done in three ways.

(1) As in the 4-fold case, the simplest way is to find the maximum value where α_length(top)=α_length(bottom) and is the largest. In this case, α_short(top) and α_short(bottom) will generally not be equal to one another.

(2) Adjust the two equations α_long(top)=α_long(bottom) and α_short(top)=α_short(bottom) using x _m and z _m . In this case, α_long will not be the largest.

(3) Find the maximum value when α_short(top)=α_short(bottom) and the maximum value. In this case, α_length(top) and α_length(bottom) will generally not be equal to one another. This is the same as swapping long and short (1).

Fig. 10 shows another embodiment of a Koller concentrator in the form of a 3x3 concentrator, to which polychromatic optimization has been applied. The POE is a Fresnel lens 100 comprising nine segments or segments. The Fresnel lens is not completely rotationally symmetric, but includes a symmetrical central section 106, four lateral sections 105 (each of which is symmetrical to each other with respect to the center of the Fresnel lens), and also with respect to the Fresnel lens. The four diagonal segments 104 are symmetrical to each other in the center of the Neel lens. The four side sections and the four diagonal sections can be fabricated as off-center squares of a symmetrical Fresnel lens. SOE lens 101 also includes nine segments that are each aligned with a corresponding segment in POE lens 100 . All nine segment pairs send the sun's rays within the acceptance angle to the cell 102 .

Compared to the 4-fold flat Fresnel design, the 9-fold yields a higher CAP (up to 0.65) in performance goal (1) and an even better balance of uniformity and irradiance in performance goal (2). For example, the 9-fold in Figure 10 achieves a geometric concentration of Cg = 1000 x (ie CAP = 0.65) with an acceptance angle of ±1.18°. The device is attractive for solar cells 102 with particularly high spectral sensitivity, as is expected to happen in future four-junction and five-junction solar.

With their prior US 8,000,018, the inventors have achieved a balance between the top and bottom junctions of no better than 0.7:1 and a CAP of no better than 0.40 for the worst of the three bands. They believe that a balance of 0.75:1 and a CAP of 0.45 will be achievable with an improved design. With these devices, in contrast, a balance between the top junction and bottom junction of at least 0.99:1 and a CAP of at least 0.63 for the worst of the three bands can be obtained even with flat primary optics.

Embodiments of the present invention simultaneously consistently achieve a CAP of greater than 0.45 for all wavebands and a uniformity (ratio of minimum to maximum irradiance on the cell with the sun centered in a perfectly aligned position) of at least 0.5 ). The inventors have found that with proper design a uniformity of at least 2/3 and typically at least 0.8 can be consistently achieved for realistic configurations.

references

[1] W. Cassarly, "Nonimaging Optics: Concentration and Illumination" in Handbook of Optics, Second Edition, pp2.23-2.42 (McGraw-Hill, New York, 2001)

[2] H.Ries, J.Μ.Gordon, Μ.Laxen "High-flux photovoltaic solar concentrators with Kaleidoscope based optical designs (High-flux photovoltaic solar concentrators with Kaleidoscope based optical designs)", Solar Energy, Vol.60 , No1, pp.11-16, (1997)

[3] J.J.O'Ghallagher, R.Winston, Nonimaging Optics: Maximum Efficiency Light Transfer VI (Nonimaging Optics: Maximum Efficiency Light Transfer VI) "Nonimaging Solar Concentration with Near-Uniform Radiation for Photovoltaic Arrays (Nonimaging solar concentrator with near-uniform irradiance for photovoltaic arrays)", Roland Winston, Ed., Proc. SPIE4446, pp.60-64, (2001)

[4] D.G.Jenkings, "High-uniformity solar concentrators for photovoltaic systems" in Nonimaging Optics: Maximum Efficiency Light Transfer VI )", Roland Winston, Ed., Proc. SPIE 4446, pp.52-59, (2001)

[5] http://www.daido.co.jp/en/products/cpv/technology.html

[6] http://www.solfocus.com/

[7] http://www.suncorepv.com/

[8] L.W.James, "Use of imaging refractive secondaries in photovoltaic concentrators", SAND 89p-7029, Albuquerque, New Mexico, (1989)

[9] Chapter 13, "Next Generation Photovoltaics", (Taylor&Francis, CRC Press, 2003)

[10] R. Winston, JC P. Benítez, "Nonimaging Optics", (Elsevier-Academic Press, New York, 2005)

[11]JC P. Benítez, JC González, "RX: anonimaging concentrator", Appl.Opt.34, pp.2226-2235, (1995)

[12] P. Benítez, JC "Ultrahigh-numerical-aperture imaging concentrator", J.Opt.Soc.Am.A, 14, pp.1988-1997, (1997)

[13]JC M.Hernández, P.Benítez, J.Blen, O.Dross, R.Mohedano, A.Santamaría, "Free-form integrator array optics in Nonimaging Optics and Efficient Illumination Systems II" Components (Free-form integrator array optics)", SPIE Proc., R. Winston & T.J. Koshel ed. (2005)

[14] J. Chaves, "Introduction to Nonimaging Optics", CRC Press, 2008, Chapter 17

[15] E.M.Kritchman, A.A.Friesem, G.Yekutieli, "Highly Concentrating Fresnel Lenses", Appl.Opt.18, 2688-2695(1979)

Claims (17)

1. a kind of optics, including：

Multi-junction photovoltaic battery, wherein each knot is operable to the light of corresponding wave band being converted to electric power；

The first optical element is reflected, it has multiple sections, and it is accurate that the plurality of section is each arranged to focus on the incidence from common source Direct light；And

Second optical element, it has multiple sections, and the plurality of section is each arranged to the phase from first optical element The light that should be saved is directed on the photovoltaic cell；

Wherein, ratio respectively with the acceptance angle of the incident light of two wave bands in the corresponding wave band of each knot 5: 4 to 4: 5 It is interior.

2. optics as claimed in claim 1, wherein, respectively with the most short-wave band in the corresponding wave band of each knot and most long The acceptance angle of the incident light of wave band is in 5: 4 to 4: 5 ratio.

3. optics as claimed in claim 1, wherein, respectively with the incidence of all wave bands in the corresponding wave band of each knot The acceptance angle of light is in 5: 4 to 4: 5 ratio.

4. optics as claimed in claim 1, wherein, the battery is square, and first optical element is to vertically Projection in the plane in perfect alignment direction is square, and each section in first optical element is to perpendicular to described Projection in the plane in perfect alignment direction is square.

5. optics as claimed in claim 4, wherein, α_p(top) and α_dThe ratio at (bottom) in 5: 4 to 4: 5 ratio, its Middle α_p(top) be measured in the plane parallel to one side of the battery respectively with it is most short in the corresponding wave band of each knot The acceptance angle of wave band, α_d(bottom) be measured in cornerwise plane comprising the battery respectively with the corresponding ripple of each knot The acceptance angle of most long-wave band in section, and each in the acceptance angle is defined as uniform incident collimated light and perfection is right Angle between quasi- direction, the luminous energy being directed into the acceptance angle on the battery is on the perfect alignment direction Identical incident collimated light be directed into 90% of the energy on the battery.

6. optics as claimed in claim 1, wherein, first optical element is Fresnel Lenses.

7. optics as claimed in claim 6, wherein, each section in the Fresnel Lenses is included with not concentric Fresnel Lenses.

8. optics as claimed in claim 6, wherein, first optical element includes being formed on one face for institute There is the Fresnel Lenses common to section and form the piece of the independent continuous slope lens for each section on the other surface.

9. optics as claimed in claim 6, wherein, the Fresnel Lenses is dome.

10. optics as claimed in claim 1, wherein, for respectively with least two in the corresponding wave band of each knot Wave band, intensity receives product CAP and is at least 0.45.

11. optics as claimed in claim 10, wherein, for respectively with all ripples in the corresponding wave band of each knot Section, the CAP is at least 0.45.

12. optics as claimed in claim 1, wherein, the uniformity on perfect alignment direction for all wave bands at least It is 0.5.

13. optics as claimed in claim 12, wherein, the uniformity on the perfect alignment direction is for all wave bands At least 0.67.

14. optics as claimed in claim 13, wherein, the uniformity on the perfect alignment direction is for all wave bands At least 0.8.

15. optics as claimed in claim 1, wherein, second optical element is the preceding surface of solid transparent body, should The rear surface of solid transparent body is contacted with the photovoltaic cell；And wherein incident collimated light is by first optical element Each saves the focus refractive into the transparent body；And if wherein string is defined within corresponding second light of the transparent body Learn element surface on warp-wise curve end points between, then for respectively with least two ripples in the corresponding wave band of each knot Section, the focus is between the string and the photovoltaic cell.

16. optics as claimed in claim 15, wherein, for respectively with all ripples in the corresponding wave band of each knot Section, the focus is between the string and the photovoltaic cell.

17. optics as claimed in claim 15, wherein, the distance between the string and described focus are at least equal to described The length on the side of photovoltaic cell.