Power from the desert sky

5 August 2002

The prospect of a landscape dominated by clusters of 1 km high towers is turning into reality in an Australian desert, in what will in the truest sense be a landmark project for solar power. Leonard Sanford

The operating principle of the solar tower (until recently known as a solar chimney, a term abandoned because of its connection with pollution) is far from complicated. Put a chimney in the middle of a greenhouse, wait for the sun to shine and superheat the air inside, and let the updraught of hot air power a turbine.

But in terms of scale, and impact, and importance for the future of the renewables business, the idea is far from commonplace. The next solar tower to be built, the second in history and the first commercial scale enterprise of its kind, will be on a gigantic scale - 1000 m high and centred on a glass canopy covering an area of nearly 40 square kilometres. In fact its value as a unique landmark is so great that its operators expect it to become a tourist attraction (and tourism revenue earner) in its own right.

To create energy conversion at levels sufficient for significant power production requires an unprecedentedly high tower and a vast collector area. To produce 200 MW at a practical efficiency in a high solar gain area requires a tower in the order of 1 km high. It is such a tower, currently going through the planning and permitting stages, that is to be constructed near Mildura in Victoria, Australia. When completed in the year 2005 it will be the world's tallest structure.

The owner of the project is EnviroMission Ltd, a recently formed public company based in Melbourne, Australia. It is also the exclusive licensee in Australia for solar towers of this design, created by German structural engineers Schlaich Bergermann und Partner, a company renowned for its large scale structures including the Munich Olympic stadium and the Hong Kong Ting Kau bridge.

Basic principles

Air is heated under the collector roof mainly by convection associated with heat stored in the ground and rises up the tower, drawing in air through openings located around the chimney base (Figure 1)(Figures currently not available). This in turn draws in colder air from around the perimeter. The calculations for the Mildura tower suggest that air under the canopy will be heated to around 35 °C above ambient, giving rise to 15 m/s wind speeds up the tower. The projected configuration is for a ring of 32 pressure-staged turbines arranged around the tower base, although an alternative plan has a series of larger turbines or even a single large turbine situated in the tower itself. Figure 2 shows a six and a 36 turbine arrangement. The design is based on experience gained from operating a 50 kW prototype built in Manzanares, Spain.

Theoretically, a power plant with a 7 km diameter collector operated in an area with an annual global solar radiation of 2300 kWh/m2 will generate between 700 and 800 GWh per year. To achieve MW outputs requires a canopy of several square km and a tower as high as possible to maximise the temperature difference between base and summit.

Solar towers can convert only a small proportion of the solar heat collected into electricity (1 to 2 per cent) and thus have a poor efficiency level. But they are robust, cheap and simple to construct and attract low maintenance costs. The building materials needed, mainly concrete and glass are available everywhere, while the technology is not restricted to hi-tech countries. Because of this the designers envisage breeding populations of stations in desert regions, using stone and sand available locally; the first tower provides energy for construction of the second, the first and second for the third and fourth, and so on.


To prove the design, first developed by structural engineer Jorg Schlaich of Schlaich Bergermann, a 50 kW, 194 m high, 10 m wide prototype (see panel opposite and Figure 6) was built at Manzanares in Spain as a Schlaich Bergermann/Spanish government joint project. It started operating in 1982. By 1989 it had proved its creator's point, operating without significant problems for seven years, and was closed down.

This tower used a collector area of 46 000 m2 (diameter 240 m) set at a height of 2m. The roof material was mainly a translucent membrane covering an area of 40 000 m2, together with 6000 m2 of glass.

The project proved that the design works, at least at the constructed scale, and provided data for design modifications to support larger scale projects.

Scaling up from Manzanares

Manzanares was an experimental facility to demonstrate the principle, validate theroetical model assumptions and prove laboratory test results. Following the project a detailed study of the transferability of Manzanares results to larger, commercially viable units, was carried out. This study was accompanied by various additional laboratory and wind tunnel testing to verify the assumptions, although generally the key thermodynamic data did not change very much.

The study showed that the transfer of heat into the working medium at the various boundaries depends on temperature and throughflow rate, as well as on surface roughnesses and local turbulence. At the experimental facility, the heat transfer coefficients within a differential temperature range of up to 17 K and in the speed range of 0 to 12 m/s (10 to 12 m/s at no load and 7 to 8 m/s at full load) were verified. For calculation of heat transfers in scaled-up plants, no fundamental modifications are needed. Although collector surfaces increase substantially in size, the maximum speed within the collector rises only to 9 or 10 m/s for a 5 MW plant, and to 14 or 15 m/s for a 100 MW plant. The temperature rise in the collector increases from 17 K (Manzanares) to around 25 K for a 100 MW plant. This strongly supports the idea that it is possible to transfer the computer model reliably to large-scale plants; and that they should have a comparable level of efficiency to that demonstrated in Manzanares.


As already noted, solar chimneys are relatively inefficient, but cheap to build and robust. They do need very large collector areas, and economically viable operation is confined to regions with high levels of solar irradiation. They are in fact ideal for locating in deserts. Land use is not then a significant factor; there is a supply of raw building material in the form of sand and rock, and environmental impact is insignificant.

Technical case

Mainly as a result of the Manzanares small scale solar tower, the subsequent development of this technolgy has had the equivalent of A$35 million and 20 years of R & D invested in it. A technical review by engineering specialists Sinclair Knight Merz found the proposed power station's design concepts and construction methods well proven, and established that it could be built in Australia.

Economic case for Mildura

The economic case depends on the renewable energy credit (REC) incentive paid by retailers under legislation connected with the government's mandated renewable energy target (9500 GWh annually by 2010). The selling price of the tower's output will be based on the average peak pool electricity price paid to generators plus the REC paid by retailers. Further value is expected to be added to the internal rate of return through the emerging synthetic carbon trading instrument, allowing the trading of emissions credits. At present most Australian electricity is generated in black and brown coal fired power stations, which account for more than 35 per cent of the nation's greenhouse gas emissions. This project on its own should account for a reduction in annual CO2 emissions of 900 000 tonnes.

The plant is at its most efficient on the hottest days. In the Austalian market that fact is a distinct advantage because it coincides with peak demand caused by airconditioning usage and therefore the highest prices.


In arid regions dust and sand inevitably settle on the collector roof and reduce its efficiency, although this effect can be reduced by choosing areas with exposed rock rather than sand. Manzanares experience showed that the collector is very insensitive to dust and with good design the occasional desert rainstorm provides a sufficient degree of self cleaning. Manzanares experience also suggests a long lifespan for the collector, as long as 60 years with good maintenance.

The tower itself should last 100 years, in a dry climate and if constructed of concrete. The carbonisation process (by which CaOH in the cement is gradually converted into CaCO3 by the action of CO2 diffusing into the structure and as a result subjects the reinforcing steel to the risk of corrosion) cannot take place unless moisture is present.

Why Australia?

• Because the government supports green projects;

• it possesses high solar radiation levels;

• owners can rely on geological stability;

• suitable national building and engineering standards exist;

• suitable sites near the grid exist;

• there are energy industry predictions of a market shortfall in renewables after 2007.

Components of the plant

The collector

The glass or polymer glazing, supported by a lightweight steel frame, is arranged to admit short wave solar radiation while blocking the longer wave re-radiation from the ground. Near the tower base double glazing may be employed to trap re-radiation from the higher temperature air. The roof slopes slowly upwards towards the tower base, to help with a low-friction transition from horizontal to vertical flow. A flat collector of this kind can convert up to 70 per cent of irradiated solar energy into heat, dependent on air flow. A typical average figure would be more like 50 per cent.

Such a roof has a very long life span, probably 60 years or more. In its glazed form it has proved very resilient in the face of storms and weather damage generally. The ground provides natural energy storage, which can be enhanced considerably by the installation of heat storage water vessels arranged in a closed loop (Figure 3). With such vessels the supply curve is flattened a great deal. The curves in Figure 4 show the projected result. It turns out that output over a 24 hour period is a direct function of the depth of the water layer.

The decision to install storage depends on the energy pattern needed locally. If there is, for example, an evening peak in the energy consumption curve and peak electricity pays well enough, the operator should go for a system with storage.

If there is no peak price advantage or if the peak is in the middle of the day (owing, for example, to air conditioner use) a system without storage might be better. The decision depends very strongly on the local situation. Schlaich Bergermann found that, in general, systems with storage generate electricity a little more cheaply than systems without. But initial investment costs increase with a storage system and consequently financing would become more difficult.

The tower

The chimney (Figure 5) is the engine of the plant and takes the form of a low loss pressure tube with friction losses similar to a hydroelectric pressure tube or penstock. The best shape would culminate in a diffuser, ie a tube flared at the top, to minimise dynamic flow losses, but such a shape is much more expensive to build. Practical designs therefore take the form of a simple cylinder. A suitable chimney diameter for the 1000 m Mildura tower would be 130 m, with a wall thickness of 1 metre at the base reducing to only 25 cm at the top if the design adopted includes spoked wheel reinforcement placed on the inside to stiffen the structure.

Air upthrust is approximately proportional to the air temperature rise DT in the collector and the volume of the chimney. In a large chimney DT would be around 35 °C producing an updraught velocity of around 15 m/s.

The efficiency of the chimney hc ie the conversion of heat into kinetic energy, is given by the following equation:

hc= Hcg/CpTo

where Hc is chimney height, Cp is the specific heat of the air, and g is the gravitational term. To is the ambient temperature at ground level.

It turns out that flow speed and temperature rise in the collector have very little influence, so the plant can make good use even of the small temperature rises created on cold winter days or at night by the ground storage effect.

Although simplified, this equation exhibits one of the basic characteristics of the solar chimney, that its efficiency is fundamentally dependent only on its height. Compared to the collector and the turbines the chimney's efficiency is very low, hence the importance of size. The tower should be as tall as possible; at 1000 m its efficiency rises to more than 3 per cent.

Such a height is not expected to present any technical difficulty as 600 m towers have already been built; the Mildura construction is larger, but much simpler and has far fewer demands placed on it compared to an inhabited building. Reinforced concrete is the favoured material, but guyed sheet metal or a clad cable-net are also possible. All have been used in cooling tower construction.

In general the efficiency of the plant is more or less directly proportional to the height of the tower. But one can generate the same power with a chimney of lower height but larger collector area and vice versa. No optimum has been discovered for the relationship between chimney height and collector diameter for any given power output. The final dimensioning of the plant is an iterative process taking local cost figures for the tower and the collector into consideration.


The turbines would be purpose designed of the low-head high-mass-flow kind, such as a Kaplan, used in hydro plants. In this application they would have to accommodate only minor pressure variations.

Turbines suitable for a solar chimney are not velocity-staged, like a free running open air direct wind turbine, but are built as a cased pressure-staged turbogenerator like a hydro unit. The energy yield of such a turbine is about eight times that of a speed-stepped open air turbine. Air speed through the turbine is roughly constant, the output being proportional to volume flow and pressure drop. To maintain this at an optimum the control system controls blade pitch in response to varying airspeed and airflow.

First approximation equations derived from the Bernoulli equation describing flow through the turbine show that there is an optimum position corresponding to a pressure drop of about two thirds of the inlet pressure. This corresponds to the mpp (maximum power point) condition of a hydroelectric power station; but conditions differ in that pressure gradient is related to flow speed. Interpreting this in terms of dimensions leads to the conclusion that the power output of the solar tower is proportional to the product of its height and collector area. In physical terms, therefore, there is no optimum: that must be determined by considering the economic case, specifically the local cost of building and materials. Turbine choice (multiple or single), heat storage, and glazing options all significantly affect the construction cost, providing a number of options for the operator/owner along the cost versus efficiency curve.

The Manzanares experimental project


Table 1. Typical figures for solar tower scaling characteristics

Linkedin Linkedin   
Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.