Solar Receiver

Abstract

A system and method for maintenance of a reflector with. The system may be incorporated to be an integral part of the dish assembly and may be operated autonomously or remotely. The system may include a controller programmed to activate the system according to set schedule or according to reflectivity drop of the light from the dish. The system includes a foldable arm that is folded when not in use so as not to interfere with the operation of the dish. When in use, the arm unfolds and includes injectors to inject fluids, such as air and/or liquids to clean the surface of the dish. The arm may include water injectors followed by drying air injectors, such as air knife. The system may include waste liquid collection system, such as a gutter and reservoir.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority from U.S. Provisional Patent Application No. 61/917,252, filed on Dec. 17, 2013. This application is also a continuation-in-part of U.S. patent application Ser. No. 13/086,315, filed on Apr. 13, 2011, which claims priority from U.S. Provisional Patent Application No. 61/323,857 filed on Apr. 13, 2010; 61/334,560 filed on May 13, 2010; 61/351,946 filed on Jun. 7, 2010; 61/370,755 filed on Aug. 4, 2010; 61/407,911 filed on Oct. 29, 2010; 61/432,584 filed on Jan. 14, 2011; the entireties of all of which are incorporated herein by reference.

FIELD

[0002] The present application belongs to the field of solar energy systems.

BACKGROUND

Context

[0003] Concentrated solar power (CSP) systems are ones that concentrate incoming solar light before converting it into useful power. The conversion itself can be photovoltaic or thermal, but the common theme is that it is cheaper to collect the light over a large area and into a small power conversion unit (PCU) than it is to build a large power converter.

[0004] There are several methods to concentrate solar power, including lenses, sun-tracking parabolic dish reflectors that position the PCU at the focus of a paraboloid, and central tower systems in which a large number of principally flat tracking mirrors direct the sun onto the top of a tower where the PCU is housed.

[0005] In the case of dish reflectors, the PCU is part of the moving structure since it has to be kept at the focal point of the dish as it tracks the sun. Additionally, since dish reflectors are typically large, they themselves are comprised of smaller reflectors held rigidly together to form the complete optical surface. Typically, these smaller reflectors need to be aligned relative to the dish structure during assembly in the field. Dish reflectors can use photovoltaic (PV) PCUs, Stirling engines, turbines, or heat collectors and steam generators.

State of the Art

[0006] FIG. 1 shows a conventional solar thermal dish design (manufactured by Stirling Energy Systems of Arizona) containing a thermal PCU [10] connected to a dish through a boom [11] (or balance beam), which connects to pivot actuation machinery [12] that is mounted on a pole (also called a pedestal) [13]. The dish and PCU [10] are roughly balanced via the boom [11] with respect to the pedestal [13], reducing the gravity loads on the actuation machinery [12]. The dish is comprised of a carrier truss [14] and reflector tiles [15]. Because of the balance-beam design, such dishes must have a slice [16] cut into them to prevent the pedestal [13] from hitting the dish when it points upwards. The shape of the reflector surface of the dish (made up of the sum of the reflector surfaces of the tiles [15]) is approximately a paraboloid (a parabolic arc revolved around its optical axis) and the aperture of the PCU [10] is located at the focal point of the paraboloid. The pivot actuation machinery [12] is controlled by a sun-tracker that keeps the optical axis of the dish pointed at the sun.

[0007] The reflector tiles [15] themselves are made from thin glass which is warped elastically over a metallic shell and bonded to it. In other systems, thick glass is hot-formed and plastically deformed into the desired shape. In other systems, the thin glass is replaced with a thin Aluminum or Steel sheet with a reflective coating.

[0008] FIG. 2 shows a schematic of a Stirling engine PCU, which is a type of a thermal engine. (In FIG. 1, the PCU was denoted as item 10). Like all thermal engines, this PCU has a hot side [20] and a cold side [21]. The hot side [20] is illuminated by the light reflected from the dish and is traditionally mounted facing it and closest to it. The cold side [21] is therefore traditionally mounted further away from the dish, towards the sun. The cold side [21] is connected to a heat exchanger [22] which rejects heat into the environment and keeps its temperature from rising. An electric generator [23] is powered by the engine and mechanically coupled to it. In most thermal designs, both the heat exchanger [22] and the generator [23] are part of the PCU package [10]. It is the high weight of thermal PCUs that traditionally dictates the balance-beam design for the dish.

[0009] FIG. 3 shows a conventional dish-based photovoltaic system (manufactured by Solar Systems of Australia). Since photovoltaic PCUs are lighter than thermal PCUs, the system does not use a balance beam, and instead the PCU [30] is connected to the dish via a focal support structure [31]. The dish, in turn, is directly connected to the actuation machinery [hidden] and pedestal [32]. The dish is comprised of a carrier truss [33] and reflective tiles [34], same as in the solar thermal dish and the shape of the optical surface of the primary dish is similarly a piece-wise segmented paraboloid. In a PV system, there is no generator, but there is still a need for a heat exchanger, since the PV cells need to be kept cold. In many PV dishes, the heat exchanger is located on the ground and coolant is piped between it and the PCU [30].

[0010] The optical area of both dishes shown above is about 100 m.sup.2. In the PV dish, the concentration factor is about 1000, the area of the aperture of the receiver is 0.1 m.sup.2, and it is built from approximately 1000 PV cells, each only 1 cm on a side, arranged in a "dense array" roughly 30 cm across (shown schematically in FIGS. 4b, 44). Since PV cells produce low voltage (.about.3V), and since the output voltage of the dish must be high (100 --600 V) to keep the current manageable, many of the cells need to be wired in series, a process known as "stringing". Cells on the same string must produce the same amount of current or else the efficiency of the string drops due cell current mismatch, known as a "stringing losses".

[0011] FIG. 4a shows a single photovoltaic cell. The cell has an active area [41] covered with thin conductive lines known as the collection grid [42] that leads to two side contacts [43] commonly known as bus bars. The two bus bars [43] correspond to the "plus" side of the photovoltaic junction, and the back surface of the cell corresponds to the "minus" side. The grid lines [42] are created by metallic deposition and are made tall and thin to minimize shading, but still produce significant shading for light that is arriving at a shallow angle relative to the front surface of the cell. Cells designed for high concentration are typically made by depositing multiple layers of semiconductor materials on a Germanium substrate, but other technologies are equally relevant to this invention. Such "multi-junction" cells are made by companies such as Spectrolab, Emcore, and Solar Junction.

[0012] In traditional dish systems, the carrier truss is non-adjustable and is assembled in the field to the best practical precision. The truss is then placed on top of the pedestal, and the reflector tiles are assembled onto it using adjustable mating mechanisms--typically three adjustment screws at the back of each tile that connect to three points on the truss. At this stage, using an optical reference (e.g. pointing at the moon, or a laser system that bounces off of the reflector tiles) the orientation of each reflector tile is adjusted by turning the screws until it is properly aligned relative to the focal point.

Deficiencies in Existing Art

[0013] The "truss and glass" dish architecture described above is performance limited due to several factors.

[0014] Typical truss-based primary mirror designs weigh around 50 kg/m.sup.2, and their assembly and alignment in the field is very time consuming, taking more than a day for the 100 m.sup.2 reflector described above.

[0015] The truss is made from a very large number of members, which have to be bolted, riveted, or welded together, in the field. The large part count introduces tolerance stack-up errors, and the large number of joints are all sources of stress concentration, fatigue, and structural creep.

[0016] The tile alignment process requires bringing each tile into the correct orientation by tweaking three screws on its back side. The required tolerance is 1 mRad or tighter, and for 100 tiles there are 300 such screws. The person doing the tweaking cannot see the reflection from the tile, and so needs to receive the information from someone else. If the structure creeps over time, alignment has to be re-done. If tiles need to be replaced, they need to be re-aligned--all requiring highly-trained manpower.

[0017] The heavy weight of the dish and in particular its large moment of inertia makes precise tracking difficult, and either increases the cost of the actuation machinery or reduced tracking accuracy, which results in reduced output. In boom-based designs, the mass is distributed in a dumbbell-like way, which is the worst case in terms of rotational moment of inertia, which makes tracking more difficult. Boom flexing adds another oscillation mode and further complicates alignment and tracking Stiffening the boom adds mass to the system. The slice that has to be cut in the dish reduces the optical area, reduces its rigidity, and degrades optical accuracy.

[0018] The optical surface precision of the reflector tiles is also lacking in practice, due to the shell/glass structure having insufficient precision. This problem can be solved with "brute force" precision optical processes such as glass grinding, but in solar power design, cost and fabrication speed are major design considerations, and those processes are prohibitive.

[0019] Finally, optical performance is hindered by accumulated dust and dirt on the optical surfaces. Cleaning frequency is limited since it requires extensive manpower, and since it wastes large quantities of water.

[0020] Dense array photovoltaic receivers also suffer from several performance issues:

[0021] Thermal: The temperature of the PV cells has to be kept low-typically less than 100 C to prevent failure, but preferably near 30 C to prevent performance losses and lifetime issues. The receiver therefore has to reject the generated heat, and at 1000.times. illumination and with tightly packed PV cells this is a difficult problem, since there is very little area the cells can reject heat into.

[0022] Optical: Any lit area covered with the wires or traces used to collect the electricity from the front surface of the cells plus any gaps between the cells, do not produce electricity and thus lead to a corresponding loss in efficiency. Since the cells are small and the current densities large, these effects are much more significant than in non-concentrating PV cells.

[0023] Electrical: The above constraints motivate the use of very thin conductors, which create Ohmic losses and wasted power in the conductor network.

[0024] Uniformity: In the dense array receiver, the cells near the edge of the array receive more cooling and less illumination than those near its center. The non-uniform illumination stems from several reasons, all ultimately stemming from the imaging nature of the paraboloid optics. (In an imaging optical system, if for example one side of the sun is blocked by a cloud, then half the receiver correspondingly goes dark, whereas in a perfect non-imaging optical system, the entire receiver becomes uniformly half-lit.)

[0025] First, the sun is not a uniform source but rather a Gaussian one. Second, the image on the receiver is the sum of many images from the various reflector panels, and since they each have independent deviations, the statistical sum is brightest at their nominal aiming point. Finally, tracking errors move the sun away from the center of the image, so the point of peak illumination moves around and does not coincide with the center of the receiver, thus requiring the field of view of the receiver to extend around the nominal position of the sun and resulting in an image that is even darker in its periphery.

[0026] These effects create a large variance in the electricity production level of the cells, and so result in stringing losses. It is not uncommon for a receiver that uses 40% efficient PV cells to provide only 25% efficiency at the system level.

Other Art

[0027] Many solar system designs have been proposed and implemented over the years. The challenge in them is not simply to make electricity from solar light, but to do so in a way that has acceptable cost, construction time, efficiency, longevity, and environmental impact.

[0028] Other PV dish systems were developed by companies such as Solar Systems of Australia and Zenith Solar of Israel.

[0029] Other Thermal dish-based systems were developed commercially by companies such as Stirling Energy Systems of AZ, Infinia Corporation of WA, and Southwest Solar of AZ, and HeliFocus of Isreal.

[0030] There have been many designs proposed for improving on the truss based design, ranging from flexible structures (e.g. U.S. Pat. No. 4,056,309) to inflatable structures (e.g. U.S. Pat. No. 4,432,342) and even vacuum-pulled membranes (e.g. U.S. Pat. No. 4,352,112). Commercial companies that design dishes include Schlaich Bergermann Solar of Germany.

SUMMARY

[0031] The following summary of the invention is included in order to provide a basic understanding of some aspects and features of the invention. This summary is not an extensive overview of the invention and as such it is not intended to particularly identify key or critical elements of the invention or to delineate the scope of the invention. Its sole purpose is to present some concepts of the invention in a simplified form as a prelude to the more detailed description that is presented below.

[0032] The invention described herein is a dish-based solar power generation system that has several novel features whose utility is to reduce or eliminate the problems outlined above. While each of these features provides independent benefits and can be utilized alone or in combination with other features to enhance prior art systems, they can be made to work in concert with each other to provide a complete system, and so they are described jointly in this specification.

[0033] The primary reflector structure is based on a unique spoke-wheel-like tensile carrier structure that is very lightweight, has a low moment of inertia, high strength and stiffness, and allows for rapid field alignment during assembly. Reflector tiles are mounted to it to create the complete primary reflector.

[0034] In various embodiments of this invention, opposite to the way customary dishes are built, it is the carrier structure that is the entity aligned to form, and the reflector tiles are fixed to it in a non-adjustable way. A very important feature of this structure is that the alignment procedure is based on its geometry and does not rely on its optics, which allows the creation of any revolved optical shape, not just a paraboloid. The structure is named the Alignable Carrier Structure, or ACS.

[0035] Additionally, in embodiments of this invention the primary structure has a central rigid hub which is concentric to the axis of revolution and supports the functioning of several other system components. Like a bicycle wheel, the structure has adjustable tensioners between the spokes and the rim, but in this invention the tensioners are implemented in a kinematic way to enhance precision. The rim is comprised of many individual straight segments, and each tensioner has four lines of force (two spokes and two rim segments) that cross each other at a single point.

[0036] According to embodiments of the invention, the reflectors that complete the dish structure, which in this application are called reflector tiles or panels, are constructed using a unique method based on a special vacuum-formed core that yields very rigid and lightweight bodies, which can have any arbitrary continuous optical shape. The reflector tiles can use any thin reflector material as their front surface.

[0037] According to embodiments of the invention, the reflector structure and reflector panels work in concert to achieve several other advantages. The tiles attach to the primary structure from the back side, which allows them to be replaced without interference with the rest of the dish, without placing people inside the optical path, and without requiring subsequent re-alignment.

[0038] Additionally, according to embodiments of the invention, the primary structure determines the position of the front surface of the tiles which enhances the precision of the formed surface. The composite front surface created is continuous, and so allows for collection of the water used while washing the mirror. The central hub is used to support a built-in rotating washing arm so that the washing process is autonomous.

[0039] Being rigid and coaxial to the optical axis, the front side of the hub may also be used as the anchor point for a kinematic hexapod used to position the secondary optic or PCU. At the front end of the hexapod a precision mechanical mount (called "the fiduciary") may be attached and may be used to co-locate the PCU, a laser guide used to guide and test the alignment of the primary reflector, and an optical tracker used for solar tracking By having all three of these components attach to essentially the same mechanical interface, precision is further enhanced.

[0040] In a photovoltaic embodiment of the invention, the shape of the primary optical reflector surface is not a regular paraboloid as is customarily used, but rather a revolved shape around an axis of revolution (also sometimes called axis of rotation) with a generatrix in the shape of a parabola whose optical axis is parallel but radially offset from the axis of revolution. In this application, this shape is called an offset paraboloid (OP). An OP is naturally constructed with a hole in its center with a radius at least as large as the offset between the optical axis and the axis of revolution.

[0041] The offset paraboloid structure creates a novel narrow ring shaped optical form that has several advantages for a PV receiver.

[0042] In a traditional "dense array" PV dish receiver the PV cells are packed into a 2-dimensional array that is placed at the focal point. This creates differentiation between the cells near the center of the array and those near its edges, with the central cells getting more illumination and less cooling. This non-uniformity is further enhanced due to tracking inaccuracies, as the image of the dish shifts relative to the receiver. Since the cells need to subsequently be stringed into a larger electrical circuit, this non-uniformity creates a large drop in overall receiver efficiency.

[0043] Embodiments of the invention provide a ringed receiver. In the ringed receiver, all the cells are geometrically equivalent, and so receive the same amount of illumination and cooling. Additionally, since the ring is essentially a 1-dimensional array of cells, the radial dimension is free to be uses for cooling and electrical connection routing, two tasks which are very difficult in a two-dimensional array.

[0044] Since each point in the ringed receiver sees its corresponding portion of the offset paraboloid dish with a narrow field of view, a ringed secondary optical element (SOE) can be used to further increase the level of concentration and uniformity of the received light. Several such SOEs are described.

[0045] Several cooling methods are described, including using a conductive heat sink and forced air flow, using a pumped vapor-chamber and a separate condenser, and using liquid coolant.

[0046] The effect of a tracking error on the ringed receiver is always a predictable non-uniformity illumination pattern, with minimum and maximum illumination occurring on diametrically opposed points, along the direction of the tracking error (which is itself not known), and illumination varying smoothly between the minimum to maximum points. Taking advantage of this predictability, in order to reduce stringing losses, the cells are divided into several interleaved groups referred to as circuits, with each circuit having cells distributed essentially uniformly around the ring. Each circuit is wired internally in parallel so in and of itself can accommodate non-uniform illumination. However, since the interleaved circuits are affected in a similar way by this non-uniform illumination effect, the circuits can then be wired in series and only incur minimal stringing losses. The output from the individual cells of any of the circuits is also used as the sensor for providing feedback to the closed-loop tracking system.

[0047] The features just described work in concert to eliminate the problems outlined in the background portion of this specification. For example, the electrical interleaving and cooling method are benefitted by the ring geometry of the receiver, which is itself benefitted by a practical non-paraboloid dish, which is itself benefitted by a non-optical alignment method.

[0048] Some of the design features outlined above have utility outside the scope of a ring-optics photovoltaic dish system. For example, the ACS in combination with the reflector tiles can create any optics of revolution, including a traditional parabolic dish, and thus can be used with other PCU technologies such as thermal PCUs. The reflector tile technology by itself can create reflective optical bodies for other geometries such as for parabolic troughs or rigid heliostats. The ACS can be used for any lightweight approximately round reflector such as a stretched membrane heliostat reflector, or even for non-reflective surface such as thin film solar panels.

[0049] In addition, the ACS is useful when constructing a non-optical receiver or transmitter such as a direction RF antenna, where the reflective surfaces reflect longer wavelength EM radiation, often not for the purpose of producing power but for signal communication.

BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings, which are incorporated in and constitute a part of this specification, exemplify the embodiments of the present invention and, together with the description, serve to explain and illustrate principles of the invention. The drawings are intended to illustrate major features of the exemplary embodiments in a diagrammatic manner. The drawings are not intended to depict every feature of actual embodiments nor relative dimensions of the depicted elements, and are not drawn to scale.

[0051] FIG. 1: Prior art--Traditional solar thermal dish with payload boom

[0052] FIG. 2: Prior art--Schematic of a Stirling engine

[0053] FIG. 3: Prior art--Traditional photovoltaic dish receiver

[0054] FIGS. 4A-4B: Prior art--Dense array photovoltaic receiver and cell

[0055] FIG. 5: Prior art--tensile torque bearing spoked bicycle wheel

[0056] FIG. 6: System overview--Dish system with solar thermal receiver according to an embodiment of the invention

[0057] FIG. 7: System overview--Dish system with solar thermal receiver optical path according to an embodiment of the invention

[0058] FIG. 8: System overview--Offset paraboloid with ring photovoltaic receiver according to an embodiment of the invention

[0059] FIG. 9: System overview--Offset paraboloid with ring PV receiver optical path according to an embodiment of the invention

[0060] FIG. 10: Carrier--Tensile spoked wheel carrier structure according to an embodiment of the invention

[0061] FIGS. 11A-11C: Carrier--Rim and kinematic tensioner node design according to an embodiment of the invention

[0062] FIG. 12: Carrier--Rib, alignment guide ruler, and tile interface according to an embodiment of the invention

[0063] FIG. 13: Carrier--Assembly unit according to an embodiment of the invention

[0064] FIG. 14: Carrier--Hub design with hexapod and fiduciary plate, actuated according to an embodiment of the invention

[0065] FIG. 15: Carrier--Laser alignment guide according to an embodiment of the invention

[0066] FIG. 16: Carrier--Fiduciary reference plate with three spheres according to an embodiment of the invention

[0067] FIGS. 17A-17C: Reflector tile--Core and fabrication according to an embodiment of the invention

[0068] FIGS. 18A-18E: Reflector tile--Structure and fastening according to an embodiment of the invention

[0069] FIGS. 19A-19C: Reflector tile--Core rib patterns according to an embodiment of the invention

[0070] FIG. 20: Reflector tile--Plastic print pattern according to an embodiment of the invention

[0071] FIGS. 21A-21C: Reflector tile--Sheet metal core according to an embodiment of the invention

[0072] FIG. 22: Dish--Hub mounted cleaning arm according to an embodiment of the the invention

[0073] FIG. 23: Dish--Hub mounted condenser/heat exchanger according to an embodiment of the invention

[0074] FIG. 24: Thermal receiver--Hexapod mount according to an embodiment of the invention

[0075] FIG. 25: Ring receiver--Cell placement according to an embodiment of the invention

[0076] FIG. 26: Ring receiver--Interleaved cell circuit according to an embodiment of the invention

[0077] FIG. 27: Ring receiver--Ring receiver with pumped vapor chamber according to an embodiment of the invention

[0078] FIG. 28: Ring receiver--Ring receiver in context of ACS according to an embodiment of the invention

[0079] FIG. 29: Ring receiver--Ring receiver with convective cooling according to an embodiment of the invention

[0080] FIG. 30: Ring receiver--Ring receiver with air-cooled heat sink according to an embodiment of the invention

[0081] FIGS. 31A-31B: Carrier--Flat reflector for heliostats according to an embodiment of the invention

[0082] FIG. 32: Carrier--ACS with stretched membrane design according to an embodiment of the invention

[0083] FIG. 33: Carrier--The ACS with vertical band spokes according to an embodiment of the invention

[0084] FIG. 34: Carrier--The ACS with horizontal band spokes according to an embodiment of the invention

[0085] FIG. 35: Carrier--Rim-based tensioner node according to an embodiment of the invention

[0086] FIG. 36: Carrier--Hub-based tensioner according to an embodiment of the invention

[0087] FIG. 37: Prior Art--A honeycomb core composite panel

[0088] FIG. 38: Prior Art--A compound Parabolic Concentrator (CPC)

[0089] FIG. 39: Carrier--ACS with conical hub according to an embodiment of the invention

[0090] FIGS. 40A-40C: Deployment mechanism, wrapped states according to an embodiment of the invention

[0091] FIGS. 41A-41B: Deployment mechanism, unwrapped state according to an embodiment of the invention

[0092] FIG. 42: Deployment mechanism, deployed state according to an embodiment of the invention

[0093] FIG. 43: Deployment mechanism, ACS with ribs according to an embodiment of the invention

[0094] FIG. 44: Hub tensioners, with radial attachment bands according to an embodiment of the invention

[0095] FIG. 45: Hub tensioners, with no radial attachment bands according to an embodiment of the invention

[0096] FIG. 46: Hub tensioners, showing both crowns according to an embodiment of the invention

[0097] FIG. 47: Rim-side non-adjustable connection node according to an embodiment of the invention

[0098] FIG. 48: Band spoke according to an embodiment of the invention

[0099] FIG. 49: Band spoke with rim-side non-adjustable connection node according to an embodiment of the invention

[0100] FIG. 50: Slotted CPC secondary optical element with PV ring according to an embodiment of the invention

[0101] FIG. 51: Slotted CPC secondary optical element transparent view according to an embodiment of the invention

[0102] FIG. 52: Circular walker single arm according to an embodiment of the invention

[0103] FIG. 53: Circular walker multiple arms flexure according to an embodiment of the invention

[0104] FIG. 54: Circular walker multiple arms hub according to an embodiment of the invention

[0105] FIG. 55: Circular walker one arm according to an embodiment of the invention

[0106] FIG. 56: Flower receiver circular according to an embodiment of the invention

[0107] FIG. 57: Flower receiver single tile according to an embodiment of the invention

[0108] FIG. 58: Automatic washing apparatus--deployed according to an embodiment of the invention

[0109] FIG. 59: Automatic washing apparatus--operating according to an embodiment of the invention

[0110] FIG. 60: Automatic washing apparatus--tube cross section according to an embodiment of the invention

[0111] FIG. 61: Automatic washing apparatus--linked arm according to an embodiment of the invention

[0112] FIG. 62: Automatic washing apparatus--linked arm and water processing according to an embodiment of the invention

[0113] FIG. 63: Focal Plane Flux Sensor--front according to an embodiment of the invention

[0114] FIG. 64: Focal Plane Flux Sensor--back according to an embodiment of the invention

[0115] FIGS. 65 and 66 show an embodiment of the cleaning system of the invention.

[0116] FIG. 67: Hexapod hub according to an embodiment of the invention

DETAILED DESCRIPTION

[System Level and Optical Path]

[0117] FIG. 6 shows an embodiment of this invention using a solar thermal PCU. The system is comprised of a pedestal [60], pivot actuation machinery [66] at the top of the pedestal, a dish reflector [61], a hexapod mount [62], a thermal PCU [63] having a hot end [64], and a heat exchanger [65]. In this embodiment the hot end is shown surrounded by heat absorbing coils in which the thermodynamic fluid flows, but in other embodiments the heat can be transferred directly through the wall of the hot end.

[0118] Note that in this embodiment the thermal engine is oriented so that its hot end is away from the dish (opposite to the conventional mounting scheme), and the light illuminates its entire perimeter, transferring heat inwards. The structure of the dish is discussed further below under the sections "Alignable Carrier Structure" (ACS) and "Reflective Tiles", and the placement and structure of the Thermal PCU is discussed under the section "Thermal PCU"

[0119] The role of the pedestal [60] and actuation machinery [66] is to ensure that the dish reflector [61] is correctly pointing into the sun and tracking it. The role of the dish reflector [61] is to concentrate the solar light towards its focal point. The role of the hexapod [62] is to position the hot end [64] of the PCU [63] at the focal point. The role of the PCU [63] is to convert the light into electricity. The role of the heat exchanger [65] is to cool the cold end of the PCU [63] by exchanging heat with the environment. These roles are largely the same as in a traditional dish-based system.

[0120] FIG. 7 shows the optical path of this embodiment in a meridian cross-section. (one that passes through the center axis of revolution [79]). At each such cross-section, incoming solar light [70] is reflected off of the primary reflector parabolic arc [71] and focused into a focal point [72] which lies on the optical axis of the arc. Since in a paraboloid the optical axis of the generatrix parabolic arc and the center axis of revolution are one and the same, the focal points of all the arcs in the revolved shape coincide and become the single focal point of the paraboloid. The focal point lies inside the hot end [73] of the thermal engine [74]. Roughly speaking, the acceptance full-angle [76] of the system is equal to the length of the hot end [73] of the engine [74] divided by the mean distance from it to the parabolic arc [71]. The heat engine was labeled as [63] in FIG. 6

[0121] FIG. 8 shows an embodiment of this invention using a photovoltaic PCU. The system is comprised of a pedestal [80], pivot actuation machinery [86] at the top of the pedestal, a reflector dish [81], a hexapod mount [82], a photovoltaic PCU [83], and a heat exchanger [84]. The structure of the dish is discussed further below under the sections "Alignable Carrier Structure" (ACS) and "Reflective Tiles". The placement and structure of the PV PCU is discussed under the section "PV PCU".

[0122] The role of the pedestal [80] and actuation machinery [86] is to ensure that the dish reflector [81] is correctly pointing into the sun and tracking it, the role of the dish reflector [81] is to concentrate the solar light towards the PCU [83]. The role of the hexapod [82] is to position the PCU [83] correctly in relation to the dish reflector [81], and the role of the PCU [83] is to convert the light into electricity. These roles are the same as in a traditional dish-based system.

[0123] In this embodiment of the invention, the shape of the primary optical reflector [81] surface is not a regular paraboloid as is customarily used, but rather a revolved shape with a generatrix in the shape of a parabola whose optical axis is parallel but offset in the radial direction from the axis of revolution of the dish. In this application, this shape is called an offset paraboloid and abbreviated as OP.

[0124] FIG. 9 shows the optical path of this embodiment, in a meridian cross-section, showing the axis of revolution [99] of the dish [81]. At each such cross-section, incoming solar light [90] is reflected off of the primary reflector parabolic arc [91] and focused into a focal point [92] which lies on the optical axis [98] of the arc. Thus, light reflected from the complete revolved dish [81] is focused into a ring [97], rather than into a single focal point. A secondary optical element [93] (SOE) has a cross section that is also revolved around the axis of revolution [99] of the dish [81] and has an aperture [94] that encompasses the focal ring [97], further concentrating the light and homogenizing it before directing it to the PV cells [95] that are appropriately arranged in their own ring shape. Thus, light reflected from dish [81] is concentrated onto the ring of PV cells [95]. The surfaces of the SOE [93] as well as the active PV surface of the cells [95] are also surfaces of revolution or close approximation thereof. Roughly speaking, the acceptance full-angle [96] of the system is equal to the width of the aperture [94] of the SOE [93] divided by the mean distance from it to the parabolic arc.

[0125] It is important to note that unlike in a paraboloid dish where the individual focal points of each parabolic arc in a meridian plane coincide into a unique focal point of the entire paraboloid, in an OP the individual focal points do not coincide but rather form a ring. In this embodiment, the ringed shape of the PV PCU [83] matches the ringed focal shape of the focal region of the OP dish [81].

[0126] In the conventional paraboloid optical path shown in FIG. 7 all of the focal points coincide, and the optical path is called "imaging", which means that for a non-point light source such as the sun, the light forms an image of the sun around the focal point. So for example if the sun has a sun spot or is partially obscured by a passing airplane or cloud, there will be a drop in intensity in the corresponding spot in the image. This property is detrimental for photovoltaic receivers since it only affects some of their cells, resulting in stringing losses as described above. The concentration level achieved by the conventional paraboloid dish is equal to the square of the concentration achieved by the parabolic arc of the concentrator. Thus if in a meridian cross section the parabolic arc achieves a concentration factor of 30, then the paraboloid dish will achieve a concentration factor of 30.sup.2=900.

[0127] In contrast, the offset paraboloid (OP) optical path described in FIG. 9, the focal points of the parabolic arcs do not coincide but rather form a ring. This means that if the sun is similarly partially obscured, the corresponding localized drop in intensity is repeated an infinite number of times around the ring, and so is naturally "smeared" across the entire receiver. This is highly advantageous for a photovoltaic receiver. In addition the concentration achieved by the OP dish is comprised of three factors: The concentration achieved by the parabolic arc in the meridian plane, the added concentration achieved by the secondary optical element, also in the meridian plane, and finally the "radial squeezing" concentration that results from the fact that the circumference of the receiver ring is significantly smaller than the mean circumference of the OP dish. For example, if the if the parabolic arc achieves a concentration factor of 30, and the SOE achieves a concentration factor of 3, and the radial squeeze concentration factor is 9, the total concentration of the OP dish is 30.times.3.times.9=810. It is possible to trade off these concentration levels. For example, if the SOE is eliminated completely, but the diameter of the receiver ring is reduced by a factor of 3 thus increasing the radial squeeze concentration factor from 9 to 27, then the total concentration of the SOE-less OP dish remains the same as 30.times.1.times.27=810. The SOE has other advantages such as further homogenization of the light but in general the optical design can be tweaked by using less aggressive or even completely eliminating the SOE.

[0128] In both types of dishes described above, the focal length of the parabolic arc in the generatrix plays an important role, since it determines the length of the physical structure that positions the PCU there. The focal distance is often stated as a multiple of the dish diameter, known as the focal ratio (f/d). Commonly used focal ratios in parabolic dishes are between 0.5 and 1.5 of the diameter of the dish. Longer focal lengths make it difficult to maintain the position of the PCU accurately. In addition, longer focal lengths mean that the image of the sun has more distance to diverge, and so the level of possible concentration drops. Shorter focal lengths mean that the PCU sees the dish with a wider apparent angle, and so is illuminated over a large arc of itself. Additionally, short focal length dishes have more steeply inclined surfaces, and are so less efficient in their use of reflector material. Thus an f/D ratio of 0.5 to 1.5 is a good design choice for dish systems. For non-round or polygonal rims, the mean diameter, or the perimeter/pi, is used as an approximation for a true diameter.

[0129] In heliostats applications (discussed later) the PCU resides on a separate immobile structure and is shared by many reflectors. In such applications, larger f/D ratio, typically over 5, are used, and the reflectors are very close to being flat. Since the direction of illumination in Heliostat systems varies throughout the day, and since the desired concave shape is so close to flat, the precise shape of the reflector is less important and the concavity is often stated as the ratio of the depth of the mirror surface to its diameter, typically no more than a few cm per m.

[0130] As described further below, these components work in concert to eliminate or reduce all of the problems that plague state-of-the-art photovoltaic dish reflectors systems, as enumerated in the background section of this specification. Some of these components have further utility for other types of dish systems, and for other types of solar systems.

[Nomenclature]

[0131] When describing geometries of revolution, there is always a unique axis of revolution, and any plane that contains the axis of revolution is referred to as a "meridian plane", and the two principal directions of the plane are the axial and radial directions. The local circumferential direction is perpendicular to a meridian plane. The shape in a meridian plane that creates the geometry of revolution is called a generatrix. A surface is considered to be a surface of revolution even if it is only partial to a complete surface or revolution spanning a complete revolution of the generatrix about the axis of revolution.

[0132] A paraboloid is a surface of revolution created by a generatrix whose shape is a parabolic arc whose axis coincides with the axis of revolution.

[0133] In this specification, an offset paraboloid (OP) is the surface of revolution created from a generatrix whose shape is a parabolic arc [91] whose optical axis [98] is parallel to and offset radially from the axis of revolution [99]. For example, in the embodiment shown in FIG. 8 and whose meridian plane description is shown in FIG. 9, the optical surface of the dish reflector [81] is an OP, and the aperture [92] of the PCU is a conical ring whose radius is approximately equal to the offset of the axis of the generatrix of the OP. The optical axis of the entire revolved dish is the axis of revolution.

[0134] The angle of acceptance in solar systems is the angle by which light rays can deviate from the nominal direction (parallel to the optical axis of the dish) and still be directed into the aperture of the PCU. In this application the "full angle" is used, which is the angle between the two most extreme light rays that still get directed into the aperture. (as opposed to the "half angle", which is the angle between the nominal ray and an extreme ray. Depending on the symmetricity of the optical path, the full angle value is typically twice that of the half angle value.

[0135] In this specification, when describing portions of the system, the direction towards the sunlight is labeled "front", with associated descriptors such as "in front". Similarly, the opposite direction is labeled "back". Note that inside the solar receiver, which is typically mounted facing towards the primary reflector, the "front" end is the one closer to the primary reflector.

[0136] In this application, the term Photovoltaic (abbreviated PV) device refers to any device that converts light into electricity directly without relying on first converting the light into heat. The prevalent photovoltaic technologies today belong to the group of semiconductor band-gap materials, but other direct conversion methods such as optical rectennas or advanced quantum phenomena such as quantum dots are also considered as photovoltaic in this application. These technologies stand in contrast to thermodynamic technologies that use solar light to heat up a working medium, and then generate power from this heat without reliance on the source of the heat. Some proposed photovoltaic technologies operate at elevated temperatures, but they do not make use of that temperature to produce power, since simply heating them up without exposing them to light will not create any power.

[0137] In this application the term power conversion unit (PCU) refers to the device that converts solar light into electricity. A PCU can be photovoltaic, thermal, or employ a yet-unclassified technology.

[0138] In this application, the term UTS is used in its common capacity to denote Ultimate Tensile Strength for a material, measured in units of stress, equivalent to units of pressure.

[Alignable Carrier Substructure]

[0139] The alignable carrier structure (ACS) replaces the carrier truss in conventional dishes. It differs from a conventional truss in several ways. First and foremost, unlike a conventional truss which is of a fixed shape, the ACS is an adjustable structure, designed to be assembled first, and then tweaked into shape. Second, rather than being built entirely from rigid members like a conventional truss, the ACS mimics the structure of a tensile spoked bicycle wheel, where the radial load carrying members (the spokes) are slender and flexible (bendable) and only loaded in tension. This is achieved by preloading them in tension to a sufficient degree--to a tension higher than the maximum compressive load they would have otherwise experienced. This way, when an external load is applied, the pretension decreases and compressive forces never occur.

[0140] Bicycle wheels can be either free-spinning, in which case the centerlines of the spokes intersect the centerline of the wheel, or torque-carrying, in which case the centerlines of the spokes are tangent to a small circle around the centerline of the wheel and so do not intersect it. Non-torque-bearing spoke geometry are sometimes referred to as radial spokes, and do not have spokes crossing each other. Since the torque loads in the plane of the dish are negligible, in some of the embodiments described herein the spokes may be radial. A spoked bicycle wheel is depicted in FIG. 5, showing the hub [50], rim [51], and spokes [52]. At the connection between the spokes and rim are adjustable tensioners [53]. This image is of a torque carrying (non-radial) wheel.

[0141] FIG. 10 shows an embodiment of the ACS. A central hub has a barrel section [100] and two crowns [101]. The crowns [101] are used to anchor thirty-six pairs of spokes [102] which connect to a rim at thirty-six coupling nodes [106]. In this embodiment the rim is not made of a single ring but is rather composed of thirty-six straight rim segments [103], thereby making it easier to transport and assemble. The tensioner nodes [106] are located at each intersection of two spokes [102] and two rim segments [103], and can adjust the effective length of each of their spokes [102]. Ribs [107] are provided on the back spokes [102] and include reflector guides, as will be described more fully below. In this embodiment the coupling node serves as a tensioning device as explained below, but in other embodiments the tensioning can also occur from the hub, as also detailed below. Also shown are reflector tiles [109] though they are not part of the ACS and will be described later.

[0142] The embodiment of the ACS shown in FIG. 10 differs from a bicycle wheel in several ways. First the rim is not a continuous circle but made out individual straight segments. This is an improvement from a structural point of view since the rim is loaded in compression and straight segments are more resilient against buckling than curved segments. It is an improvement from a logistical point of view since the straight segments are easier to transport than a single monolithic rim hoop. Second, the tensioner nodes contain two spokes and two tensioners each, as opposed to a standard bicycle wheel where each tensioner node has only one spoke and one tensioner. This arrangement eliminates out-of-plane forces on the rim as are present in a traditional bicycle wheel.

[0143] The amount of tension preload (pretension) applied to the spokes is calculated so that when the external load is at its maximum value, all of the spokes remain in tension. This can be seen by considering a single spoke pair. It the tensioning node is pushed backwards in the axial direction, the tensile load on the front spoke increases, while the tensile load on the back spoke decreases. If it is not pretensioned enough, the tensile load on it will decrease past zero and become a compression load, at which point the spoke will buckle since it is too slender to resists compressive loads. Since the most significant source of external loads is wind, and the direction of the wind is not predictable, this can affect both front or back spokes, and so each of them has to be pretensioned to the value load induced on the spoke by the maximum predicted wind in its most unfavorable direction to that spoke. However, when the wind is blowing in the opposite direction, the tensile load on the spoke will equal the sum of the pretension and the wind-induced load, thus equaling to twice the pretension. Thus the UTS of the spoke must be at least equal to twice the pretension load. It is not necessary, however, to preload the spokes all the way to 50% of the UTS, since the designing engineer will typically leave a design margin below the UTS, and also because other considerations (such as stiffness) can drive towards a spoke that has a much higher UTS than the design requires. Even a 10% UTS preload is sufficient to create a tensile structure that can resist significant load, and a 33% UTS preload will leave 33% margin on UTS in the worst-case external load situation, which is a reasonable margin to prevent low-cycle fatigue failure of the spokes. Thus a pretension value of 25% -40% UTS is a good design range, depending on the specific properties of the spoke material.

[0144] In order to decrease the risk of fatigue related failures, in this embodiment the spokes are made of twisted multi-strand wire ropes or cables. Similarly, a weaved roped can be used, or a fiber-composite thin rod. In this specification, the term "braid" is used to include various multi-stranded tensile member structures.

[0145] FIGS. 11a-c shows the detailed composition of a tensioner node according to an embodiment of the invention. A central coupling element [110] in the shape of a short thick-walled tube has two spherical depressions [111] on its ends. The coupling element [110] is held in compression between two rim segments in the form of tubes [112], each of which has spherical protrusions [113] at its ends, thus forming two ball-and-socket compression joints. Two tensioner links [114] are free to rotate around the coupling element [110], and each of them pushes against it with an adjustment screw [115]. The two links [114] are also each connected to a spoke [116]. Turning the adjustment screw [115] causes the tensioner link [114] to move in the direction of the spoke [116], pulling it inwards or outwards. The lines of force therefore intersect (to a very close degree) at the center of the coupling element [110], creating a kinematic joint. This is true even during assembly when the angles between the components are not at their nominal value, since both the rim tubes and the tensioner links can rotate to self-align to the direction of the force. In general, an assembly is kinematic if the numbers of degrees of motion equal the number of constraints. In this case since the forces meet at a single point and therefore do not exert torques, there are only three degrees of motion per tensioner node, and in the complete dish there are three times as many constraints (rim segments and spokes) as there are tensioner nodes. It is also possible to replace the direction of each of the ball-and-socket joints, so that the in the joint the ball is on the coupling element and the socket (spherical depression) is on the rim segment.

[0146] In this embodiment therefore the tensioners are used to adjust the ACS until the rim becomes circular and flat. At each tensioner node, turning the adjustment screws in opposite directions shortens one spoke and elongates the other, thus moving the tensioner node in the axial direction. Turning the adjustment screws in the same direction either elongates or shortens both spokes, moving the tensioner node in the radial direction. The alignment of the ACS is therefore a systematic procedure which can be carried out repeatedly to almost arbitrary precision--as is done in bicycle shops on a regular basis.

[0147] The resultant structure is kinematic, highly rigid, and unlike a traditional dish truss does not have lattice arms that cantilever from the center outwards--the ACS is actually most stiff and precise right at the rim. Just like a bicycle wheel, all the spokes merely pull the rim inwards, and the rim is compressed in the circumferential direction. It is the imbalance in spoke tensions that results from the application of load that provides the rigidity of the rim in respect to the hub, both for in-plane and out-of-plane forces.

[0148] As is the case with a regular bicycle wheel, the spokes [102,116] are pre-tensioned so as to never go slack, and therefore the rim segments [112] are always loaded in compression. The hub barrel [100] is compressed by the combined action of the spokes [102,116], and the hub crowns [101] work principally in tension. When the hub is used to impart torque on the dish in the tip or tilt directions, the magnitude of the forces in the individual spokes vary, but the design is such that the polarity of the load never changes (between tension and compression). These consistent loads reduce fatigue issues and allow for very efficient use of material. (For example, the spokes are slender enough that if they ever experience compression, they will buckle)

[0149] As is common in large diameter sheet metal cylindrical structures, to optimize transportation costs, the barrel can be shipped to the field in three 120.degree. slices which can be nested for packaging and shipping and assembled together to form the a complete cylinder.

[0150] In this embodiment, the back set of spokes has stiffener ribs [104, 117] attached to them. The ribs are not cantilevered from the hub, but are merely attached to the spokes, encompass them, and serve to stiffen the spokes from bending in a radial plane. The back spokes are only visible as they emerge from the ribs at their ends, near the back crown and the tensioners.

[0151] Each back spoke is encapsulated by a rib [117] which stiffens it and contain curved guide rulers [118] welded to them that are later used to position the reflector tiles, which will be brought from the back side and pushed against it. Thus if the coupling elements [110] are brought into their correct locations in space, the guide rulers [118] establish precise radial guides for the optical surface. These continuous radial guides are much more precise than individual support points on a conventional truss.

[0152] It is important to note the dual role of the ribs. In a tensile bicycle wheel, the spokes create a very strong and rigid spatial relationship between the hub and the rim (both in and out-of-plane) but the spokes themselves can be bent by pushing on them sideways in their mid-spans. Since the load on the solar dish will be distributed, the ribs function to stiffen the spokes against such deflections. Additionally, the ribs mediate mechanically between the straight spokes and the curved guide rulers which establish the final optical shape of the dish. The guide rulers are attached to the ribs in-factory during fabrication, and so this process can be done to a very high degree of accuracy--much higher than the accuracy of field assembly.

[0153] It is also important to note that the ACS is constructed, stabilized, and aligned before the tiles are attached. Thus the hub, rim, and spokes together provide the alignable carrier structure. The ribs provide a geometrical fit between the spokes and the desired optical shape and further stiffen the structure for supporting a distributed load, and the tiles merely attach to the ribs and do not contribute in any significant manner to the strength or stiffness of the ACS in bending or translation. If the ACS has radial spokes, the tiles do contribute to stiffness against torsional motion around the optical axis, but such torsional loads are insignificant.

[0154] When the rim is aligned to shape, the guide rulers are necessarily in place and provide thirty-six arcs that are precisely located in space. In the case of a simple paraboloid dish, these arcs are parabolas whose optical axis coincides with the central axis (define as being central and perpendicular to the rim circle). In the case of an OP, the guide rulers are also parabolic arcs, but during fabrication they are attached to the ribs so that their optical axes are properly offset. Thus by controlling the shape of the guide rulers during fabrication, different dish shapes can be produced when the rim is brought into alignment.

[0155] The alignment of the dish is indifferent to which optical shape is being created by the guide rulers, and this is very important if the desired optical shape is not a simple paraboloid. Additionally, the alignment of the dish can take place before any reflector tiles are installed on the ACS. This is also important since reflector tiles can later be replaced without affecting alignment.

[0156] FIG. 12 shows how the reflector tile [120] interfaces with the rib guide ruler [121] in this embodiment. The tile is brought from the back of the dish and pushed forward [122] against the guide ruler. In this tile embodiment, this is done using spring clips shown as [188] in FIG. 18d. Since the guide rulers position the front surface of the tile, the alignment is insensitive to variations in the thickness of the tile--it is the optical surface of the tile that is brought to place.

[0157] The ACS is assembled around its hub on its pedestal. The hub is first oriented vertically so it can spin around the vertical axis without wobbling. At that point, the subassembly shown in FIG. 13, comprised of two spokes [130], a rib [131], and a tensioner node [132] is attached to the hub, followed by a rim segment [not shown, close to perpendicular to the plane of the page], and the hub is then rotated by 10 degrees. This step is repeated thirty-six times, until the ACS is complete. During assembly, the adjustment screws on the tensioners are set for maximum spoke length, and after assembly is complete they are tightened to create the pre-tension in the spokes.

[0158] In this embodiment, the ACS is 8 m in diameter, and the 36 spokes are 1/4'' multi-stranded 7.times.19 steel cable, preloaded in tension at about 500 kgf of force. A typical hub has a diameter of about 10% of the dish diameter, and a height of about 25% of the dish diameter. Other embodiments can have different proportions, diameters, numbers of spokes, cable thicknesses, and material selections. Specifically, the number of rim segments can be reduced very significantly from the value of thirty-six. As this is done, the length of the rim segments increases, and they become more susceptible to buckling. However, even having as few as six segments is still an approximation to a round rim.

[0159] The hub of the ACS is its backbone and most rigid and dense component. The tracking actuation mechanism attaches to the back-side crown of the hub, and the front-side crown attaches to the PCU mount. The rest of the ACS structure is distributed and will not support a concentrated load at any single point. This is a positive trait that indicates that material is used efficiently.

[0160] In addition to aligning the primary reflector to shape, the ACS has to precisely locate the PCU relative to the primary. The connection between the ACS and the PCU must therefore be rigid and precise.

[0161] In the embodiment shown in FIG. 14, this is achieved using a kinematic truss structure known as a hexapod. A hexapod consists of six rods, and attaches to three points on each of two bodies, kinematically constraining the six degrees of freedom that can exist between them. As shown in FIG. 14, the six rods are connected as three pairs to three anchor points 144a-144c on the back of the hexapod which is co-located with the front crown [101] of the hub [100], but each rod from these three pairs form a pairing with a rod from another pair to be pair-connected to anchor points on the front of the hexapod which is co-located with the PCU interface [141], also referred to herein as the fiduciary.

[0162] The six struts [140] connect the front hub crown to a precision mechanical interface [141] called the fiduciary which serves as the defining optical reference for the dish. In other embodiments the hexapod can be replaced by a thin walled-tube, or a thin walled tube with a perforated wall surface for weight reduction. As described later, the hexapod can be actively tilted to perform a tracking function using liner actuators [143].

[0163] The fiduciary is a kinematic coupling. Kinematic couplings are used to connect two objects in a way that a) relieves any stresses due to thermal expansion or misalignments and b) allows them to be disassembled from each other and then reassembled in a highly repeatable way, so each reassembly results in the same spatial relationship between the objects. The kinematic coupling is later used to attach the PCU and other optical elements to the fiduciary interface [141].

[0164] This embodiment uses the kinematic coupling described in U.S. patent application Ser. No. 13/032,607, but other standard kinematic couplings such as a three-groove mount, a six-point mount, or a flat surface with locating pins can also be used.

[0165] In this embodiment, the portion of the kinematic coupling that resides on the fiduciary consists of a triplet of conical depressions [142]. Each cone serves as a seat for a fixed-size sphere that is tangent to it, and so defines a unique point at the center of this sphere. The reference plane is defined as passing through the centers of these spheres, and the reference axis is defined as being perpendicular to this plane and passing through the center of the circle defined by the centers. As per U.S. patent application Ser. No. 13/032,607, the PCU has three beads on it that match the cones.

[0166] In an embodiment that uses the standard three groove mount, the fiduciary has three spheres attached to it, and the PCU has three radial grooves that match the spheres.

[0167] In order to establish a fiduciary-based reference against which to perform the rim alignment, a rotary laser guide (RLG) is fastened to the fiduciary interface. By using the fiduciary interface for both the primary mirror (ACS) alignment guide and as the positioning mechanism for the PCU, there is no need to later align the position of the PCU to the focal point or axis of the ACS. To allow the concurrent mounting of both a PCU and an RLG, the fiduciary interface in the embodiment below contains two triplets of co-located conical depressions--an inner one [157] for the RLG, and an outer one for the PCU.

[0168] FIG. 15 shows an embodiment of the RLG which uses two intersecting laser beams that rotate around the main dish axis. This rotation describes two coaxial cones of light that intersect at a circle, which becomes the reference for the rim. A baseplate [150] is mounted to the inner fiduciary interface [157], leaving the outer one free for mounting the PCU. An inner guide tube [151] is attached to the baseplate and is perpendicular to it. A coaxial outer tube [152] encapsulates the inner tube [151] and rotates around it. A suitable bearing (not shown) may be provided between the inner and outer tubes. A rigid carrier plate [153] carried two lasers [154a,154b], mounted through an adjustment mechanism [156] so that their beams [155a, 155b] are made to intersect at a predetermined point relative to the fiduciary. In this embodiment, many holes are drilled in the carrier plate [153] so as to reduce its weight. As the outer tube [152] rotates, the point of intersection describes a circle to which the rim is subsequently aligned. It is beneficial to have the two laser beams be of different colors.

[0169] In this embodiment the fiduciary has two kinematic couplings and thus two triplets of cones--an inner one and an outer one--with the alignment guide attached to the inner triplet, leaving the outer one for the PCU, so the alignment guide can be mounted even when the PCU (described later) is in place.

[0170] An integral or temporary laser target [159] is placed on each tensioner node as it is being adjusted. The target rests on the center-body of the ACS and on the front spoke, so its position and attitude relative to the rim is fully constrained. An aiming zone [158] is marked on the target. During adjustment, the operator turns the adjustment screws of the tensioner so as to bring both laser beams into the aiming zone.

[0171] FIG. 16 shows another embodiment, in which three tooling balls [160] are placed on a reference plate [161] that attaches to fiduciary interface [162]. The distance between the balls is pre-known and they define a reference axis in the same way that the fiduciary interface does. The alignment of the dish relative to the balls is ascertained either using a laser radar (such as Nikon model MV224) or photogrammetry techniques, using the edges of the rim tubes [163] as the defining features of the inspected geometry. The tooling balls can be made a permanent feature of the fiduciary plate itself, situated permanently in the inner set of conical depressions.

[0172] In another embodiment, alignment is done by placing the dish horizontally and rotating it while comparing the position of the tensioners to a fixed ground base--as is done in bicycle shops when truing a bicycle wheel. Each of the thirty-six data points will have an axial component and a radial component. If the elevation drive does not perfectly align the rim axis with the axis of rotation, there will be a sine-wave embedded in the axial component of the measurement, but it can be removed using statistical fitting. This method does not relate the alignment to the fiduciary, but can serve as a fast "low tech" way of quickly inspecting a dish with no special equipment.

[0173] Another method to confirm that the dish is perfectly aligned is by monitoring the tension of the spokes which should be uniform for all the front spokes, and for all of the back spokes. Since the front spokes are not encapsulated by ribs, their tension can be measured acoustically by plucking them. This can done spoke by spoke with the equivalent of a guitar string tuner, or by imparting a mechanical impulse to the hub and using a microphone and spectrum analyzer to pick up vibration frequencies that indicate an off-tension spoke.

[0174] In another embodiment, as shown in FIG. 14, the three mount points [144a-c] between the hexapod [140] and the hub are actuated along the direction of the optical axis of the dish by short-travel linear actuators [143], resulting in a 3-DOF tip-tilt-piston motion of the fiduciary plate from its nominal position. Only a small actuation travel range is required to tilt the hexapod a few degrees, and so it is used to correct tracking errors in real time. This either enhances the effective tracking accuracy, or alternatively compensates for a much reduced performance tracker. The actuation can be achieved using means such as pistons, bellows, lead screws, voice coils, piezo drives, cams, or eccentric shafts.

[0175] Among the novel features of the ACS as described above are: unlike conventional trusses, the structure itself is adjustable, and its alignment procedure is independent of the optics of the dish. Unlike conventional trusses (including cable-assisted ones) there are no radial members cantilevered or loaded in compression--it is the rim that enables the "magic" by which all radial members are always in tension. The ACS can therefore be fabricated in relatively small parts that can be easily transported to and assembled on location, without losing precision due to tolerance stack ups.

[0176] Among the utilities of the ACS as described so far are: ability to align-to-form in the field even for complex optical shapes, ability to replace reflector tiles without re-alignment, low weight, low moment of inertia, low part count, high strength and rigidity, high geometrical precision, single fiduciary mechanical interface that defines the optical axis for all optical components of the dish.

[0177] The ACS can additionally be used for any lightweight approximately round reflector such as a stretched membrane heliostat reflector, or even for non-reflective surface such as thin film solar panels or Fresnel-lens sheets. In heliostat systems, there is no PCU on the ACS, since it is located on the central tower. In a Fresnel lens embodiment, the tiles held by the ACS are transparent, refracting the light to a focal point behind the plane of the ACS.

[0178] In another embodiment shown in FIG. 39, a ACS is shown with a conical hub [390]. Conical hubs are beneficial in that they nest for transport, and as explained above, they minimize shading if the reflector is not pointed directly at the sun, as in the case of heliostats.

[0179] Another embodiment, also suitable for a heliostat, is shown in FIGS. 31a-b. The reflector tiles [310] are principally flat and located near the front of the hub [311] in order to minimize shading (since heliostat mirrors do not point directly into the sun) and the spokes [312] and ribs [313] are largely behind the tiles.

[0180] In another such embodiment, shown in FIG. 32, the ACS is used to hold a stretched membrane [320] reflector usable with either heliostats in central-tower solar thermal systems or with dish systems. In this embodiment the flatness of the rim [321] controls the periphery of the stretched membrane. The curvature of the membrane is induced from its back side [not shown] either by a mechanical structure or a second membrane and vacuum in between them. (Stretched membrane reflectors are existing art and not in the scope of this specification). Again, to reduce shading in heliostats, the hub [322] is conical in shape.

[0181] In another embodiment of the ACS shown in FIG. 33, the spokes [330] are bands oriented roughly in parallel with the optical axis of the dish, to minimize shading. The reflective surface [331] of the dish in this embodiment is a Fresnel paraboloid, comprised of two paraboloid rings with different focal lengths but sharing the same focal point. Openings [332] in the gap between the rings help reduce pressure differentials across the dish caused by aerodynamic effects. A Fresnel paraboloid can have two or more rings.

[0182] In another embodiment shown in FIG. 34, the spokes [340] are band oriented roughly in parallel with the plane of the dish, and simply wrap around the rim [341]. The rim tensioner mechanisms are not implemented in this embodiment, and preloading of all the spokes after assembly is achieved by moving the two crowns [342a, 342b] away from each other on the hub using threaded rods [343]. Since this embodiment is less precise, it useful as a large tracking structure for flat panel PV receivers.

[0183] Another embodiment of the tensioner node is shown in FIG. 35. A central coupling element [350] in the form of a thick-walled tube is held in compression between two rim segments in the form of tubes [352]. The coupling element [350] has two surfaces at its ends that are appropriately angled with respect to its axis to accommodate the two rib segments [352]. The coupling element [350] also has a fixed bracket [357] attached to it. Two spokes [353] are connected at their ends to a movable bracket [358]. Two adjustment screws [351] adjust the position of the movable bracket [358] with respect to the fixed bracket [357]. Stiffening ribs [354] are attached to the back spokes just like in the embodiment shown in FIG. 11.

[0184] Another embodiment of a tensioner is shown in FIG. 36. In this embodiment, the spoke tensioners reside on the hub rather than on the rim. The drawing depicts the back crown [360] of the hub, with the hub walls removed for clarity. A central column [361] is part of the support structure connecting the back crown [360] to the front crown (not shown). Each spoke [362] is terminated with a swage [363] and goes through a tensioning screw [364] that has a hole drilled through it and is threaded into the crown. By turning the screw [364] clockwise (inwards), the spoke [362] is pulled inwards with it, and is thus tightened. Also shown is the stiffener rib [365]. For clarity, only one spoke is shown. This embodiment is useful if tensioning is automated and motorized, since the tensioning motors all reside next to each other and wiring is thus minimized.

[Reflector Tiles]

[0185] As explained above, in various embodiments of the invention, the ACS provides each reflector tile with two precisely formed and located radial curved guide rulers as mounting points. While other manners of affixing the reflector tiles to the ACS can be used, it is believed that the embodiments described herein provide enhanced precision.

[0186] Traditional composite panels are comprised of a front membrane, a back membrane, and a core. The term membrane signifies that a sheet-like member can only support in-plane stresses. It is stiff in-plane, but flexible out-of-plane. A letter-size sheet of paper and a 1 m.sup.2 sheet of 0.5 mm thick Aluminum are two examples of membranes. The core's role is to connect the two membranes and transfer sheer between them, so that the entire panels can resist out-of-plane bending. Typically, the core itself is rather flimsy and cannot support bending by itself either. Typical cores include honeycombs, general ribbed structures (even including cardboard ribs), and expanding foams. The membranes can be made out of Aluminum, steel, plastic, or glass. A traditional residential door, for example, is made from two very thin plywood membranes, and cardboard ribs in between acting as a core. Surfboards are often made with fiberglass membranes and foam cores. A person's weight will bend the membrane and break the foam, but the composite surfboard can support the person's weight even when the board is supported only at its tips.

[0187] FIG. 37 depicts a standard flat honeycomb core composite panel. The panel is comprised of the honeycomb core, which consists of a large number of ribs [370] oriented in three principal directions, and two membranes [371]. It is evident that the bonding area between the ribs and the membrane is very limited, consisting only of the edges of the ribs.

[0188] Precisely curved composite panels are more difficult to form quickly, since most ribbed structures are naturally flat. Honeycombs in particular, which are the most efficient ribbed structures, are almost exclusively flat, are awkward to handle in small, irregular shapes, and can be difficult to bond to because of the small surface area of the membrane-core joint. Injectable expanding foams get around these limitations and function well in flat or curved panels, but their mass/stiffness ratio is lower, their cost is high, and the curing process is both time consuming and environmentally problematic since it releases large amounts of solvents.

[0189] In some of the embodiment shown here, the core of the reflector tiles is formed using the process of thin-gauge vacuum forming, which is typically used to create objects such as yogurt cups. In this process, a thin plastic sheet is heated to a temperature in which it is easily deformed plastically. It is then placed over a mold containing an array of depressions, and lowered down to seal against it. Vacuum is then applied between the sheet and the mold, and the soft sheet it sucked into the depressions, creating an entire tray of yogurt cups at once. This process is fast, scalable, and cheap.

[0190] In these embodiment, we adapt this process to create a honeycomb-like ribbed core structure which is pre-curved to match the intended curvature of the tile, pre-formed to match the non-regular shape of a slice of a dish, and is able to conform to a high precision optical surface without inducing stresses in itself. (The ideal panel core structure has high stiffness in shear, but low stiffness in bending.)

[0191] FIG. 17a shows the vacuum forming mold that forms the core, which has a front surface [170] that conforms to the intended optical shape of the tile. The back surfaces of the depressions in the mold [171] jointly form a piece-wise surface that defines the back side of the core. As in all vacuum forming molds, small conduits or holes allow the vacuum forming machine to apply vacuum into the depressions, sucking the heated and softened thin plastic sheet being formed into them.

[0192] FIGS. 17b-c depict the formed core, as derived from the mold. It has a front surface [173] that approximates the intended optical shape, but also has a large bonding area (when compared to a conventional honeycomb core). The back of the core [174], comprised of the backs of the formed cups, also features a large bonding area that will attach to the back membrane. In this embodiment, the back surface is curved so as to make the tile thicker in the middle, to optimize it for bending loads when supported by the two guide rulers provided by the ACS. This is evident in the five cups [175] closest to the viewer in the bottom most view of FIG. 17c. The sidewalls [176] of the cups are the members that carry the sheer forces in the completed tile. In this embodiment they are slightly slanted since the depressions in the vacuum mold must be tapered.

[0193] FIGS. 18a-e show the structure and assembly order of the complete tile. As shown in FIG. 18a, The front membrane [180] is first placed (face down) on a precision bending mold [181] that gives it its proper optical shape. The bending mold pulls the membrane down to it using vacuum, ensuring a good surface fit throughout its area. For the curvatures needed for solar applications, the membrane typically only undergoes elastic in-plane deformation during this step. Electrostatic forces or pressure can be used instead of vacuum to constrain the front membrane to the bending mold. It is important to differentiate between the vacuum mold used during the vacuum forming process to form the core and the bending mold which might use vacuum to constrain the front membrane to it.

[0194] As shown in FIG. 18b, the core [182] (which was previously formed to shape) is then placed on the back side of the membrane and bonded to it. Note that since the cups are connected only via the front surface of the core, the core is free to bend in both directions as it is being attached to the front membrane. This is important since the accuracy of the bending mold is higher than the results of the vacuum forming process.

[0195] Next, sidewalls [183] are placed around the tile's perimeter. The sidewalls contain fastening features [184] used to later attach the tile to the ACS. At this point in the process, the backs of the core cups are still exposed and free, and if the tile is removed from the mold, the front surface would still be able to deform and lose its desired shape, by causing the individual cups to "wiggle" as it does so.

[0196] To complete the tile, as shown in FIG. 18c, the back membrane [186] is bonded to the backs [185] of the cups of the core, preventing them from moving relative to each other, and thus making the structure stiff and capturing the geometry of the front membrane that is still held conformant to the mold. The function of the cups is evident in a cross-section of the complete tile (FIG. 18e), as their sidewalls create a web of cross-ribs [187] all throughout the tile.

[0197] In another embodiment, the back membrane is pre-formed around its edges to includes the sidewalls, thus creating a single backshell which encloses the tile, reducing the part count of the complete tile.

[0198] Spring clips [188] are then attached to the sidewalls (FIG. 18d). The clips will be used during dish assembly to mount the tile to the ACS ribs and preload the tile in the forward direction against the ACS rib guide rulers [121, in FIG. 12].

[0199] FIGS. 19a-c shows several embodiments of the core with diff

Claims

1. A system for cleaning a reflector dish, the reflector dish having a base support, the system comprising: an assembly coupled to the base support; at least one foldable arm configured to assume a folded position when not in use and an unfolded position for cleaning; a fluid conduit coupled to the foldable arm and delivering fluids to be injected by the arm onto the reflector dish when the arm assumes its unfolded position.

2. The system of claim 1, wherein the assembly is rotatably coupled to the base support.

3. The system of claim 2, further comprising a motor positioned to rotate the assembly about the base.

4. The system of claim 3, wherein the motor is energized by fluid flow.

5. The system of claim 1, wherein the foldable arms comprise a flexible house and a spring.

6. The system of claim 5, wherein the hose is configured to assume a flat shape when no fluid flows through the hose.

7. The system of claim 6, wherein the spring comprises a flat spring inserted inside the hose.

8. The system of claim 7, wherein the spring is configured to have a spring constant to impart sufficient curling force to fold the hose when no fluid pressure is in the hose, but to yield to fluid pressure to unfold the hose.

9. The system of claim 1, further comprising a gutter configured for collecting fluids sprayed on the reflector dish.

10. The system of claim 9, further comprising a plurality of reservoirs and a manifold coupled to the gutter, the manifold having a plurality of outputs, each configured for delivering fluids from the gutter to one of the plurality of reservoirs.

11. The system of claim 1, further comprising air pressure system configured for delivering air pressure into the foldable arm.

12. The system of claim 1, wherein the assembly is freely-rotatable coupled to the base support and wherein the arm comprises fluid injectors positioned so as to impart rotational motion to the assembly when injecting fluid.

13. The system of claim 1, wherein the arm comprises air injectors configured to generate air knife to dry the reflector dish.

14. The system of claim 1, wherein the arm is rotatably coupled to the assembly.

15. The system of claim 1, wherein the arm comprises liquid injectors at a leading edge thereof and air injectors at the trailing edge thereof.

16. The system of claim 1, further comprising: a liquid reservoir, a pressure conduit coupled to deliver liquid from the liquid reservoir to the art, and a pressure pump configured to pump liquid from the reservoir into the pressure conduit.

17. The system of claim 1, further comprising a plurality of injectors provided on the arm and configured such that collective injection front of the plurality of injectors follows curvature of the dish surface.

18. The system of claim 1, further comprising a plurality injectors provided on the arm and configured such that separation distance between surface of the dish to outlet of each injector is the same for all of the plurality of injectors.