Introduction

Various devices for collecting solar radiation thermally have been devised. At the simplest level, a flat metal plate painted black and placed in the sun will heat up until it reaches a temperature where the heat that it looses to the air around it and also by radiating itself, exactly balances the amount of energy it receives from the sun. This “stagnation temperature” occurs at around 80°C for a simple flat plate. If water, for example, is passed through passages in the plate, then it will stabilize at a lower temperature and the water will extract some of the energy in being usefully heated up. This is the essence of solar thermal energy collection.

Greater levels of sophistication are aimed at reducing the amount of “thermal loss” from the collector surface at a given temperature. This allows energy to be collected more efficiently and at higher temperatures.

Starting with the flat plate, a cover layer of glass helps by cutting down the energy lost by the circulation of cold air. If metal tubes and glass cylinders are used instead of plates, then the space can be completely evacuated, so that air convection losses are completely eliminated. Clever use of coating materials to produce an optically selective surface that absorbs as much as possible of visible solar wavelengths whilst emitting as little as possible of thermal radiation from the plate helps to reduce radiation losses. Various combinations of these measures are used in the production of systems for the production of solar hot water that are used around the world for domestic and industrial applications. In principle, solar collectors of this nature could be used for electricity production. However the low temperatures achievable limit the conversion efficiencies possible to low levels.

Further increasing the temperature at which energy can usefully be recovered requires some method of optically concentrating the radiation so that the size of the absorbing surface and hence its thermal loss, is reduced. The conceptually simplest approach is to employ a series of flat mirrors (called heliostats) that are continuously adjusted to direct solar radiation onto the absorbing surface. This is illustrated conceptually in Figure 1. Large plants called “Central Receiver Systems” or “Power Towers” have been built based on this principle and are discussed in detail in section 2. The concept can also be adapted to linear absorbers and long strips of mirror to create a “Linear Fresnel” concentrator. Figure 1The central receiver concept; a field of plane mirror heliostats all move independently to each keep a beam of solar radiation focussed on a single central receiver (Figure H. Kreetz).

Alternatively, the mathematical properties of a parabola can be exploited. The equation for a parabola in the x y plane is;

y=x2/4f

Rays of light parallel to the y axis of a mirrored parabola will all be reflected and focused at the focal point at a distance f from the vertex.

As illustrated in Figure 2, this effect can be used in a linear arrangement, where a mirrored “trough” with a parabolic cross section will focus solar radiation onto a line focus when it is pointed directly at the sun. The largest Solar Thermal Power plants constructed to date employ this principle. Alternatively a mirrored dish with a parabolic cross section (a paraboloid) will focus solar radiation to a point focus (see figure 3). Dish systems and troughs require continuous adjustment of position (or at least frequent readjustment) to maintain the focus as the sun moves through the sky. Figure 2A parabolic trough concentrator focuses solar radiation onto a linear receiver when faced directly at the sun (figure H. Kreetz)

Figure 3A paraboloidal dish concentrator focuses solar radiation onto a point focus receiver (figure from H. Kreetz).

Similar focusing effects can obviously be achieved with lenses of various kinds, but this has not been employed on the scales needed for solar thermal power systems.

Another alternative which potentially avoids the need to track the sun is to employ “non imaging” concentration. As illustrated in Figure 4, this involves the construction of a mirrored “light funnel” of some kind. Such a device will be able to collect rays into its aperture over a range of incidence angles and cause them to exit via a smaller aperture via multiple reflections. Non imaging concentrators have not found application as the “primary” means of concentration for Solar Thermal Power systems but they are frequently applied as “secondary concentrators” at the focus of central receivers, dishes or troughs, where they serve to further reduce the size of the focal region. Figure 4A non-imaging concentrator concentrates solar radiation without the need to track the sun.

The rays of light from the sun are not exactly parallel. This means that even a perfect optical system will produce an image of finite size, with an intensity distribution that is a maximum in the center and tapers off to zero at the edges. Imperfections in mirror shape and tracking accuracy have the effect of further spreading out the sun image.

Each of these approaches to concentration has a typical ratio of collected radiation intensity to incident solar radiation intensity, termed the “concentration ratio”. The limiting conversion efficiency arises from the second law of thermodynamics. The maximum efficiency for conversion of heat from a constant high temperature source given by;

Maximum conversion efficiency = 1-Tcold/Thot

This is the “Carnot limit”. A simple understanding of why there should be such a limit can be developed by realizing that empirically heat will not spontaneously flow from cold objects to hot ones. If all the heat flow from a hot object could be converted to electricity then this could in turn all be used to heat an even hotter object which we know to be impossible. In real Solar Thermal Power systems, conversion efficiencies around one third or less of the ideal maximum are typically achieved.

Clearly higher concentration ratios give higher efficiency, however they also lead to potentially higher complexity and cost. The ultimate challenge with Solar Thermal Power systems is to produce the desired output as economically as possible. This invariably means a trade off between system efficiency and capital investment drives the design process.

Central Receiver Systems

The central receiver concept was first proposed by scientists in the USSR in the mid 1950s. The first experiment was established in Sant’ Ilario near Genova, Italy, in 1965 by Professor Giovanni Francia. He installed 120 round mirrors each the size of a ‘tea-table’, focusing on a small steam generator on top of a steel frame. The product was superheated steam (500°C, 10 Mpa). Central receivers have the advantage that all the energy conversion takes place at single fixed point. This avoids the need for energy transport networks and allows investment to improve the efficiency and sophistication of the energy conversion process to be made more cost effectively. Associated disadvantages are that they must be built as single large systems, without the modularity benefits of distributed systems like troughs or dishes. The fixed position of the receiver also means that heliostats do not point directly at the sun, so that the amount of collected solar radiation per unit area of mirror is less than with the other technologies. Major investigations during the past twenty years have focused on four heat transfer fluid systems; water/steam, sodium, molten salt and air.

To date power generation has been via conventional steam Rankine cycles at the base of the tower. In early systems, receivers were built to produce superheated steam directly, however thermal problems associated with the unsteady boundary between liquid water and steam in the tubes motivated a move to secondary heat transfer fluids. Liquid sodium provides good heat transfer behaviour, but carries the disadvantage that all the transport piping must be heated above sodium's melting point (slightly below 100°C). The possibility of leaks also provides a significant fire risk. Use of molten salt avoids the fire risk and is thus also well suited to storage in tanks to allow power generation when there is no sun. The nitrate salts used has a melting point of 260°C. They do however lead to corrosion problems if leaks occur. Air is also suggested as a heat transfer fluid, a 3MWth system has been demonstrated at the Plataforma Solar test facility in southern Spain.

Heliostat fields can either surround the tower or be spread out on one side. For a surround field externally radiated circular cross-section receivers are employed, for one sided fields cavity receivers can be used. System designers have developed optimization strategies which determine the best arrangement for a given number of heliostats. These take into account the effects of shading between heliostats, the spread of the field and the optical inefficiency that increases as heliostats are further from the tower.

Two general approaches to heliostat design have been used. The most obvious is a plane structure with rigid mirror facets mounted on it. The structure sits on top of a pedestal with a gearbox arrangement that allows for two axis tracking of the sun.

The other alternative is termed the “stretched membrane” approach. As the name implies a membrane is stretched across a circular frame in a similar manner to a drum skin. The membrane is then covered with mirrors. Thin stainless steel membranes with glass mirrors glued to the surface have been tried as have polymer films which are mirrored themselves. The stretched membrane approach allows the mirrored surface to be curved slightly by the application of a small internal vacuum. In this way heliostats can actually focus their own sun image on the tower receiver rather than just directing a plane beam at it. Figure 5 illustrates an example of a stretched membrane heliostat design. Development trends have suggested that larger heliostats are more cost effective, heliostat modules up to 200m² have been tested. Figure 5Stetched membrane heliostat.

The largest central receiver solar thermal power plant demonstrated to date is the “Solar Two” plant in Southern California. This plant is actually an updated version of the previously operated “Solar One” system. Extra heliostats were added and the receiver converted from direct steam generation to molten salt. Figure 6 shows the Solar Two plant in operation and Figure 7 is a schematic illustrating the operating principles. When the solar field is operating, the molten nitrate salt moves from a cold (288°C) storage tank via the receiver at the top of the tower, where it is heated, to the hot storage tank (566°C). Independent of the solar energy collection process, salt from the hot tank is passed through a heat exchanger where heat is transferred to produce superheated steam, with the salt passing back to the cold storage tank. The steam is used in a conventional steam turbine power plant for electricity generation. Figure 6The Solar One (later Solar Two) central receiver power plant in operation.

Figure 7Schematic of Solar Two operation (figure from Bechtel Group Intenational).

The Solar Two plant has one thousand and eighteen 39.1m² heliostats plus a further one hundred and eight 95m² heliostats. Under nominal conditions, 48 MWth is concentrated onto the receiver which sits at the top of a 91m high tower. Steam is produced in the heat exchangers at 10MPa and 538°C and the net electrical output is 10.4MWe. One of the most recent developments in the Central Receiver area, is the “beam down” concept proposed by the Weizmann Institute in Israel. Rather that converting the energy at the top of the tower, a hyperbolically shaped secondary directs it vertically downwards. At the bottom a further non imaging concentrator concentrates it further before it is captured by a volumetric receiver capable of heating air to very high temperatures needed for a Brayton cycle.

Figure 8 shows one of the SEGS plants and figure 9 illustrates the operating principles schematically. A synthetic heat transfer oil is pumped through the trough array and heated to up to 400°C. This oil is then used to produce steam in heat exchangers before being circulated back to the array. The steam is used in a conventional steam turbine based electricity generating plant. Although some hot oil based energy storage was provided in the first plant, the SEGS systems overall rely on natural gas firing to provide continuous operation when the sun is not available. Figure 8View of SEGS trough based solar thermal power plant in southern California

Figure 9Schematic representation of SEGS plant operation (figure from ABB).

The LUZ troughs are built using a galvanized steel space frame. This frame supports 4mm thick glass mirror segments which are shaped by heating and molding to match the parabolic profile, before silvering. Each mirror facet is supported only at 4 points of attachment. The most recently constructed systems (termed LS-3) are 5.76m wide and 95m long giving a total aperture of 545m². 224 glass mirror segments are used and a concentration ratio of up to 80:1 is achieved.

The trough units are lined up in north south rows and track the sun from East to West during the day. During the winter when midday sun elevation is lower some radiation is lost from the end of each trough, but because they are long, this is only a small fraction of the total. Each trough has its own positioning and local control system.

The receiver units of the LUZ troughs consist of a stainless steel tube 70mm in diameter, covered by a Pyrex glass envelope which is sealed to the tube via metal bellows at the ends. The space between the steel and glass is evacuated to minimize thermal losses. The surface of the stainless steel absorber tubes is coated with a black chrome selective surface which absorbs 94% of the solar radiation incident whilst minimizing the amount that is radiated at thermal wavelengths. The combination of trough and receiver is capable of operating at temperatures in excess of 400°C. However the heat transfer oil becomes chemically unstable and begins to degrade at temperatures above 300°C. Approximately 350m³ of oil circulates in one of the 30MWe plants. Depending on the operating regime this can need replacing at rates of up to 8% per year as a result of chemical breakdown.

The latest SEGS plants each employ an Asea Brown Boveri steam turbine with reheat cycle and multiple extraction. With steam inlet conditions of 10MPa and 370°C, thermal to electric conversion efficiencies of approximately 37% are achieved, giving overall peak solar to electric efficiencies of 24%.

Regular maintenance is required for all solar thermal power systems. For the SEGS plants, routine cleaning and replacement of broken mirror facets and receiver modules forms a major part of the maintenance program.

Research and Development on trough systems has continued since the LUZ plants were completed. A major area of investigation has targeted the replacement of the heat transfer oil with direct generation of steam in the receivers. Direct steam generation would allow collection of energy at higher temperatures and so improve the efficiency of the steam turbine. It would also avoid the need to replace the costly oil and eliminate the inherent fire risk. The challenge is managing the rapidly changing thermal stresses induced in the receiver tube by the unsteady movement of liquid and vapor water when boiling is taking place.

In a variation of linear focusing technology a group based at the university of Sydney has developed a concept known as the “Compact Linear Fresnel Reflector”. This operates rather like a linear version of a central receiver system. Fixed linear receivers are illuminated by a series of long narrow mirrors which track the sun individually. If several receiver rows are installed side by side, then the individual mirror units can switch from one receiver to another depending on the relative optical efficiencies at different times of the day. A demonstration system based on this technology is under construction as an addition to an existing coal fired power station in northern Queensland (Australia).

Paraboloidal Dishes

The first realization of dish concentrator/engine system design was in 1864. The Frenchman Augustin Mouchot built a variety of ‘thermo-piles’ and ‘solar engines’ using conical mirrors. Currently, there are many designs and states of development of paraboloidal dishes. Paraboloidal dish concentrators must point directly at the sun for proper operation. This necessitates two axis tracking mechanisms. There are two approaches to this, “altitude/azimuth” tracking or “polar/equatorial” tracking. In the first, the dish is mounted with a horizontal axis pivot (to allow altitude adjustment) plus a vertical axis pivot (to allow azimuth adjustment). A polar/equatorial system has a main axis of rotation, the angle of which is adjusted on a daily basis so that it is at a right angle to the sun's noon elevation. With the polar axis correctly adjusted, tracking during the day involves following the sun from sunrise to sunset with this single axis of movement. Altitude/azimuth arrangements have the disadvantage that both actuators must work throughout the day and the azimuth adjustment must be very rapid near midday for high sun elevations. Despite this, it is sometimes used for pragmatic structural design reasons.

Dish based solar thermal power systems can be divided into two groups. Those that generate electricity with engines at the focus of each dish, and those that use some mechanism to transport heat from an array of dishes to a single central power generating block. From a research and development perspective, dish systems have a major advantage over central receivers in particular, in that a single dish prototype can be constructed and tested relatively economically. Furthermore, for the dish/engine approach, a single module on its own can demonstrate the operation of the full system. Consequently there have been well over 30 dish prototypes developed to date, both by research institutes and private companies.

The variations in structural design of dish concentrators are very similar to the different types of heliostats used with central receivers.

Possible different approaches include; A space frame structure that supports a number of rigid mirror elements held in a paraboloidal orientation, a single rigid paraboloidal surface covered with mirror elements, a single mirrored stretched membrane that is pulled into a paraboloidal shape on a drum like structure and a space frame supporting a series of small stretched membrane mirror elements.

The high concentration ratios achievable with dish concentrators allow for efficient operation at high temperatures. Stirling cycle engines are well suited to construction at the size needed for operation on single dish systems and they function with good efficiency with receiver temperatures in the range 650°C to 800°C. To achieve good power to weight ratios, working gas pressures in the range 5 – 20 MPa are employed and use of either the high conductivity gases hydrogen or helium gives improved heat transfer. The Advanco corporation and McDonnell Douglas have produced 25 kWe dish Stirling units which have achieved solar to electric conversion efficiencies of close to 30%. This represents the maximum solar to net electric conversion efficiency achieved by any solar energy conversion technology.

Stirling systems array were constructed at the Plataforma Solar de Almeria test site in southern Spain and has been operating since 1991. Each unit consists of a 44.2m² stretched membrane dish fitted with a 9 kWe 2 cylinder kinematic Stirling engine. The membranes are made from a sheet of 0.23 mm thick stainless steel sheet which has beenplastically deformed to form a parabolic shape and is then kept in place by a slight internal vacuum. 1mm thick back silvered glass mirror tiles are bonded to the membrane and elastically deformed to follow its shape. The dishes use a polar equatorial tracking system. The first 3 dishes were installed in 1991 and a further 3 dishes added in 1996. In the period up to 2000 an accumulated total of 30,000 MWh of electricity was generated and fed into the local grid.

The biggest distributed array/central plant solar thermal power system that has been trailed is the “Solarplant 1” system built in Southern California by the “Lajet” consortium in 1983/84. It consisted of 700 dishes with a total collecting area of 30,590 m². The dishes generated steam in their cavity receivers, with 600 of the dishes producing saturated steam and others taking the saturated steam and superheating it to 460°C. The steam was transported through an insulated pipe network to a central steam turbine based generating plant which produced a nominal output of 4.9MWe. The dishes were constructed with multiple stretched membrane mirror elements. The membranes were made from aluminized acrylic film. Unfortunately this film proved not to be very durable, which was a major factor in the plant ceasing operation in 1990.

The Australian National University's large dish prototype (designated Solar Generating System 3 – SG3) is intended for use in distributed systems similar to Solarplant 1. The first prototype was completed in 1994. A similar unit has been built at the Sde Boquer Desert Research Center in Israel. At 400m² aperture, they are the biggest dishes built so far. The SG3 unit in Canberra Australia, has 54 triangular mirror panels supported on a hexagonal aperture space frame structure. Altitude azimuth tracking is employed, with the horizontal axis near the base of the dish so that it can be parked in a horizontal position relatively close to the ground. This helps to reduce wind resistance and so improve storm survivability. The SG3 dish has a cavity receiver based on a single helical winding of stainless steel tubing. This receiver serves as a “once through” boiler which produces superheated steam at 5MPa and 500°C. Although intended for operation in large arrays, the SG3 prototype is currently connected to a small reciprocating steam engine which is capable of generating 45kWe and is connected to the local grid. The thermal efficiency from solar to steam is approximately 85% and in a large steam turbine based system production of electricity at a rate of approximately 100 kWe per dish would be expected.

Research and development on dish systems around the world, looks at improving and developing dish designs to bring the cost down and at improving the reliability and efficiency of conversion systems. A number of groups are investigating using small solar driven gas turbines at the focus of dishes (dish/Brayton systems). These offer the potential for high efficiency operation but with less maintenance than Stirling engines.

Chemical conversion systems are also investigated. High temperature reversible chemical reactions such as methane reforming or ammonia dissociation can be used to store energy in chemical form in products which are stored and transported at low temperature and then recombined to provide the heat for power generation on a continuous basis.

Concentrating Solar Thermal Systems