Solar power is a distributed energy source that can be captured at or near its point of use provided that is not behind a mountain or otherwise shaded from direct sunlight during the day. This allows distribution losses and the cost of distribution systems to be avoided.
Lets assume we average 1 kW/m^2 incident solar radiation on an optimally inclined collector in the southern UK. This is a generally accepted value for the insolation received by a collector set perpendicular to the sun's rays at sea level. Using my house as an example we can calculate how useful solar collectors might be. I measured my energy inputs for a year and converted these figures into a continuous annual power requirement of 400 watts of electricity for lighting and appliances and 1.68 kW for heating and hot water. Currently (2007) the best PV solar collectors are 16% efficient (160W/m^2), so I would need a 7.5 m^2 collector for electricity. Now let's look at heating. A good solar water heating installation can reach 60% efficiency (600 W/m^2), so I'd need an 8.4 m^2 solar heating collector to handle my heating and hot water requirements. As a sanity check, a typical house in the UK is reckoned to average 400 watts of electric consumption and around 1.8 kW for heating. If its solar powered it should have collector capacities of 1260 watts for electicity and 5 kW for hot water: my calculations allow for 1200 watts and 5040 watts respectively. My roof is peaked with the ridge oriented east-west and has a south facing area of about 16.5 m^2. I need a total collection area of 15.9 m^2 for my entire energy requirements, so my house could, in theory, be energy self-sufficient from solar energy alone - provided that the line of large oak trees along the south edge of my garden wasn't there, the collectors are kept scrupulously clean and the system is well maintained.
This shows that, depending on its orientation, the entire sun-facing side of a peaked house roof may come close to providing household needs for heating and power in a temperate climate. This is fine for heating: heat is easily enough stored, but cheap, reliable, non-polluting electricity storage for the 6-7 kWh needed to run the house during the night is another problem. However, note that this is probably over-optimistic for winter because no allowance has been made for increased cloud or for the effect of orbital eccentricity, which further reduces winter insolation at high northern lattitudes.
Electricity from PV panels can be stored in lead-acid batteries, though whether these will still be as cheap or available when the oil shortage affects road transport remains to be seen. The currently preferred solution is to use the national grid as the storage system. This involves feeding excess electricity into the national grid during the day and drawing on the grid at night. This is highly cost effective now, assuming the grid operators pay a fair rate for your excess power, but simply does not work once a large proportion of buildings are doing the same.
PV generation can work on a larger scale too. At least one American Internet server farm is powered entirely by solar energy collected from its roof. This installation uses PV cells as roofing material. This is claimed to be cost effective at present and can only get better as PV cells become cheaper and more efficient. However, the server farm descriptions did not say whether the installation stored its nighttime power requirement locally or merely banked it in the national grid.
It is claimed that solar buildings are already cost-effective compared with traditional buildings whose only energy source is drawn from the national grid. This requires that solar heating arrays occupy part of the roof area, PV collectors replace building wall and roof cladding and the collected energy is used in the building.
At present global PV panel production is rising rapidly, but PV panel use is only cost effective where mains electricity is not available. For instance the Eddystone Lighthouse, off the south-west coast of England, replaced its diesel generators with PV panels in 2000 to save money on fuel transport and generator maintenance. According to New Scientist (21 January, 2006), Germany and Japan account for about 66% of the world's installed capacity in mains-connected houses but both governments subsidise these installations. Its worth noting that neither country has much in the way of indigenous energy sources. The same source expects the cost of local PV micro generation to equal that of mains power by 2020, but as the result of price reductions from mass production rather than from increased PV conversion efficiency.
In the UK there is suprisingly little interest in domestic solar heating. My guess is that the lack of interest is a reaction to cowboy installers, aided and abetted by manufacturers pushing hi-tech, expensive prepackaged solutions designed to squeeze as many watts per square meter out of their collector systems regardless of cost or practicality. For instance, Siemens supplies distinctly space age looking collectors: heat pipes with specially coated collector fins mounted inside 150 mm diameter evacuated glass cylinders. These are mounted on standoffs above conventional roofing. The heat pipes plug into a central manifold which is part of a heat transfer circuit that transfers the heat to your hot water cylinder via a copper coil. The collectors look cute, but, with the two thermal conductive junctions (heat pipes → transfer circuit → hot water cylinder) I wonder how good the overall efficiency is. How much does the system efficiency drop when the inevitable dirt film forms on the glass tubes. How easy is it to clean them? What about getting the autumn leaf accumulation out from under the collectors without breaking them? How does the installation affect the cost of roof repair and maintenance? These points are never mentioned.
At the start of this section I calculated that I'd need a 8.4 m^2 solar collector operating at 60% efficiency to meet my hot water requirement, yet I have been quoted for a high tech system with collectors of only 2 m^2: this would need to reach an overall efficiency of 250% to produce the same amount of hot water. Sorry, guys. Wrong answer, especially at several thousand pounds per square metre of collector.
The same economics do not apply to large central solar generators, which suffer from the same distribution losses as contemporary central generating systems. In addition they need a bulk electricity storage facility to keep the grid supplied with power during the night and during eclipses. These installations are best suited to deserts and other cloudless regions. There are a few technologies that are best suited to large central plants, though this does not include huge PV collectors which are best scaled for installation at the point of use.
With the exception of a handful of PV systems (see below) all large scale solar power systems are CST systems. The technique is simple: use large mirrors or Fresnel lenses to collect sunlight and concentrate it on a device that generates electricity from the incoming energy.
The oldest large scale technology is the solar steam generator. In this system concentrated solar energy is used to generate steam to drive a generator.
Solar Troughs were the first system to be tried due to their simplicity. The concentrating mirrors don't need to track the sun and the working fluid's temperature is low enough for common materials to be used in the heat transfer circuit. This approach uses fixed 2D parabolic mirrors to focus solar rays on a collector pipe containing circulating oil and raising its temperature to almost 400°C. The hot oil is used to generate steam for a conventional turbine generator set. There are currently (May, 2007) nine CST installations in the Mojave desert, which have supplied 350,000 people with electricity for some 20 years. The cost is fairly high at $US0.15 per kWh compared with $US0.04 for gas and $US0.07 for nuclear generation. However, this price falls in the peak demand cost range. It will become more competitive in future as fossil fuel costs rise and more experience with CST reduces its equipment and running costs. CST is only suitable for desert areas with large amounts of empty space and reliable sunshine. It is estimated that 39,000 square km of CST plant installed in the southwestern US could supply half America's energy needs.
CST systems can only generate electricity when the incident solar radiation is strong, so this system doesn't work after dark and its output is vastly reduced under cloud, so the provision of bulk energy storage systems will be critical if these systems are to meet a large proportion of the total energy demand. It has been suggested that by replacing the heat transfer medium by molten salt and cycling it through insulated storage vessels the turbo-generators can be run at night. The efficiency should be better than Solar Troughs too because molten salt can be heated to 540°C, which increases turbo-generator efficiency beyond that of systems using oil as the working fluid by increasing the steam temperature at the turbine inlet. Plants that have been built so far include:
All the CST generation schemes described so far contain an operational pitfall: for maximum efficiency CST plants must be built in deserts or other arid regions, yet they use a lot of water. This is used as coolant to condense the exhaust steam from the turbo-generator sets so it can be re-injected into the boilers. Exhaust stream cooling requires 2000-3000 litres of water per MWh generated.
The cooling water requirement can be avoided by focussing the collected sunlight on the hot end of a Stirling engine and air cooling the cold end. The loss of efficiency due to relatively poor cooling is offset by the high efficiency of the Stirling engine in comparison to a steam engine. The Stirling dish arrangement mounts a Stirling engine generator set at the focus of a parabolic reflector and tracks the sun with the whole assembly. In consequence, a plant would consist of a farm of relatively small units. A benefit of this is that the plant can start to supply electricity soon after construction starts and gradually ramp up output as modules come online, while the major drawback is that Stirling dishes have no energy storage ability and so can only produce power when the sun is shining. A single module 50 kW plant has been installed in Australia at Canberra and two 750 MW plants are being built near San Francisco.
An alternative CST approach uses concentrated solar energy to convert carbon dioxide into fuel. Los Alamos Renewable Energy (LARE) simply plans to use intense solar heat to split CO2 into carbon monoxide and oxygen at about 10% energy efficiency. The CO would then be used to feed the Fischer-Tropf synthetic fuel process. The overall energy efficiency is likely to be poor because both stages require high temperatures, though if the waste heat was used to generate electricity the overall efficiency could appoach 48%, comparable with CCGT electricity generation. Gabriele Centi at the University of Messina in Italy suggests using CST and a titanium dioxide catalyst to split water. The resulting hydrogen is separated from the oxygen with a semi-permiable membrane and reacted with CO2 on a platinum nanotube catalyst to make low molecular weight (C2 to C9) hydrocarbons. The overall efficiency is about 1%, though this is better than natural photosynthesis achieves. In terms of land use and total efficiency both schemes should come out ahead of biofuels.
Large scale PV systems, in the sense of systems that supply the grid rather than just the building they are installed on, are more talked about than being built. The main problem is that they are large, expensive to build and suffer from the usual grid losses.
The relatively small 11MW system near Serpa, Portugal, cost 58m euro (£40m) to build. This will produce enough power for 8,000 homes, but these will still be dependent on mains power at night since this plant has no storage capacity. It covers 60-hectares (150-acres) on a south-facing hillside. The collector panels are mounted high enough and with suitable spacing to allow the land underneath to be used for grazing.
The only systems operating in 2009 are six plants ranging from 40 MW to 60MW rated capacity in Germany, Spain and Portugal.
The Solar Chimney is another interesting technology. This is basically a large area solar collector surrounding a large chimney. Ambient air is drawn into the collector, warmed by solar energy and rises up the chimney, where it drives a turbine. A 50 kW prototype was built at Manzanares in Spain where it worked successfully from 1982 to 1987, when the non-weatherproofed plant collapsed in a windstorm because the chimney guy wires had corroded. A 200kW chimney started operation at Jinshawan in Inner Mongolia in December, 2010, with plans to extend it to 277ha of collectors and an output of 27.5MW. There are proposals for the construction of commercial plants of up to 200 MW in India, South Africa, Australia and other countries. The capital cost is high but running costs are very low, the fuel is free and the power station has a long lifetime. Despite having an overall efficiency of only about 3% this technology could become the cheapest method for the large scale generation of solar electricity. An additional benefit is that the collection area can apparently be used as a greenhouse without unduly affecting electrical output.
The Solar Vortex is a new approach to harvesting solar power. Inspired by the dust devils found in all hot deserts, it uses a ring of tangential vanes to create a stable vortex column located in the centre of the ring. A vertical axis turbine at the centre of the vortex drives an electrical generator. A 1m experimental unit has shown that the concept works and it has been calculated that a unit with a 10m diameter turbine should generate up to 50kW of electrical power at a price that is competitive with current (2012) renewable energy technologies and that large scale generation should be possible by installing arrays of these units. However, its not obvious how closely the units can be spaced without interference between them limiting power output. Similarly, there is as yet no indication of how the ratio of solar energy input to electrical output compares with PV systems, solar chimneys and other equivalent systems. Both factors combine to determine the overall output per hectare of desert an installation occupies.