THE LIMITS TO SOLAR THERMAL ELECTRICITY

Ted Trainer

152.2.11.6.101629.8.105

Abstract:  Accessible evidence on the likely output and costs of trough, central receiver and dish technologies is analysed with a view to assessing the capacity of solar thermal systems to meet electrical demand in mid-winter.  It is concluded that Big Dishes using ammonia dissociation for heat storage is likely to be the most cost effective approach, although the rate of electricity delivery at distance and net of all reducing factors would be low.  It is found that for solar thermal systems to meet a large fraction of anticipated global electricity demand in winter would involve prohibitive costs. Problems of variability and storage are discussed and are found to be significant if solar thermal systems are to contribute significantly to overall system supply through protracted periods of low solar radiation. Although solar thermal systems are likely to be the most effective renewable energy sources in future, this study concludes that the general energy and greenhouse problems cannot be solved on the supply side by renewable energy developments; i.e., it will require significant reductions in demand and therefore radical social change.

The major drawback for renewable energy is the inability to store electricity from intermittent sources.  However solar thermal technologies can store heat and use it to generate electricity when it is needed.  Some claim this capacity will enable renewable energy sources to meet all electricity needs. (E.g., Trieb, undated, Czisch, 2004.)  The importance of the issue could hardly be exaggerated.  If solar thermal systems are not able to overcome the gaps left by the intermittency of the other renewable energy sources, then it is not likely that renewable sources can sustain an energy-intensive society.  (On the limits to other renewable energy sources see Trainer 2007, 2008, 2010a, 2010b.)

It is important that attempts should be made to estimate the potential and limits of renewable energy sources even though conclusions might be uncertain at this point in time, rather than wait until the situation becomes clear, because crucial social policy decisions depend on whether or not renewables are likely to be able tocan fuel energy-intensive societies. 

Although sSolar thermal systems are likely to be among the most significant renewable energy contributors.  However they are best suited to the hottest regions and it is not clear how effective they can be in winter, even in the most favourable locations.  T, and there seems to have been no examination of thise question, which . This is the focal concernquestion in this discussion.  Unfortunately iIAt this point in time it does not seem possible at present to arrie to arrive at very confident conclusions,  regarding the crucial issue of winter supply from solar thermal systems.  This is firstly because solar thermal developers are not primarily concerned with this question, being mainly interested in maximising summer and annual output, and therefore they tend not to report winter performance.  Secondly commercial developers (understandably) rarely make publicly available the key data on performance.  (Heller, 2010, Blanco, 2010, Mancini, 2010.)

However the aim of this discussion is to attempt to analyse the situation as best as is possible at this point in time given the limitations in the information available. It is a considerablen advance on the exploration reported in Trainer 2007. 

Solar radiation data indicate that CentralAlice Springs Australia is probably the best global location for solar thermal plant, somewhatsignificantly better than the South Western US. (ASRDHB, 2008, Meteonorm, 2007, RREDC, undated, Odeh, Behmia and Morrison, 2003, Fig. 1, NASA, 2010.)  The NASA solar radiation data source statesgives the Central Australian mid-winter DNI as 26% better than both the SW US and Eastern Shara.following mid winter monthly average DNI for Egypt, South West US and Central Australia, 5.12 kWh/m2. , The source also points out that these are averages and the minimum values can be 4.76, Because various sources indicate that Central Australia has the highest winter value,Consequently the following discussion will focus on the Central Australian situation. If this is problematic then the prospects for winter supply to the NE US and to North Western Europe from North Africa and the Middle East would be less promisinglikely.

Troughs.

Performance in relation to alignment and radiation received.

The ratio of winter to summer output for troughs is considerable, and much greater than for dishes or central receivers.

 This is due to the geometry determining that the angle between sun, reflecting surface and absorber is relatively large much of the time.  

The winter electrical output for the US SEGS VI trough system is reported by NREL (2005) to be about 1/5 of summer output. Odeh, Behnia and Morrison (2003, Fig. 2.) state a ratio closer to 1/7 for thermal energy collected (not electricity produced) at the SEGS site.  Their estimate for heat collection in Central Australia has a summer/winter ratio of 1/3.25.  (Ratios for electricity generated would be higher than for heat collected, due to the negative effect of low DNI on solar-to electricity efficiency; below.) Fig. 2 from Kearney, (1989), shows a ratio of 1/9 for a SEGS plant.  Fig. 3.2 from Bockamp et al. (2003) shows a ratio of 1/4 ¼.  Fig 2 from IEEE (1989 19890shows a ratio of approximately 1/11, with h a winter daily output of c 6.6 W/m2, whereas peak summer output is given as c. 73 Wm2.  Mills (undated, Fig. 3) estimates a ratio of 1/3.4 for troughs at Longreach, Central Australia.  Mills, Morrison and LeLeivre (2003) estimate a ratio of 1/4.8 for a Fresnel arrangement.  Czisch (2001) shows that for Mauritania, Morocco and Portugal, the winter/summer ratio is 1/3.5, 1/4 and 1/27(sic), respectively, with winter collection falling to 12%, 9% and 1%, of peak capacity.  The Solar Advisor Model (SAM) from NREL, (2010) shows a ratio of 1/3.3 for a trough in Southern CaliforniaThe I.E.A.’S Fig. shows a ratio of 1/. 

Solar thermal developers are not concerned with these very low winter figures and concentrate on the obvious potential for significant summer contribution. 

s.

THowever the trough systems that have not been built are not designed to maximise winter performance.  Arranging the troughs on an east-west axis, as distinct from the usual north-south axis, would in general raise the winter/summer ratio for energy entering a trough.  (This is not so for all locations, see RREDC, undated.) Odeh, Behmia and Morrison find that at Alice Springs the summer/ winter difference would in effect be eliminated with east-west troughs.  Trieb (undated) says the North African winter/summer ratio would be c. 1/1.2.  The much reduced difference is evident in Figures 1 and 2 from Odeh, Behnia and Morrison, 2003.

However these high ratios are misleading as even in good solar thermal regions the heat collection performance of east-west troughs in winter is relatively low compared with the summer and annual average performance of north-south troughs.  It is in fact about the same as the winter level for north-south troughs.  This is evident in Figure 1 from Odeh, Behmia and Morrison.  For example in summer at Alice Springs average daily DNI entering a north-south trough reaches almost 800 W/m2 but goes down to 430 W/m2 in winter, yet all through the year the figure for and east-west trough is only a little over 500 W/m2.  The radiation data given by RREDC  point to the same general conclusion, showing that in one of the best US regions although the amount of solar radiation entering an east-west trough in winter is somewhat greater than the amount entering a north-south trough, it is only 4 - 5 kWh/m2/d. 

Consideration of radiation levels, trough layout geometry and threshold effects throws some light on these performance ratiosaverages.  It is evident that the winter performances derive from coercive physical and geographical factors which are not amenable to significant technical improvement.  Hourly Direct Normal Irradiation (DNI) at Alice Springs, Central Australia on a clear winter day (not an average day) is 7.5 kWh/m2/d. The Australian Solar Radiation Data Handbook (ASRDH, 2006) gives the Alice average Alice Springs DNI in winter as about 5.7 kWH/m2/d.  However it does not rise much above 700 W/m2 at any hour of the day.  The values for ideal US sites seem to be c. 5.2 kWh/m2/d.  (NASA, 2020, NREL (SAM), 2010.)

It is important to recognise that these are averages and variation around the average can be quite significant.  Kenaff (19921) reports that variaqionthis is greatest in winter and DNI can go down to 30% lower than average.  The RREDC tables confirm this general trend, for instance indicating that in regions where the monthly average winter insolation is just over 5 kWh/m2/d, the minimum likely is 3 kWh/m2/d.  The NASA data includes similar minima.  For the purpose of this discussion, which is estimating limits, average radiation levels are not as important as those which supply systems might have to cope with when conditions are most unfavourable.  Nevertheless the following discussion will mostly be based on average figures, (which means that the generating capacities needed to meet demand would be greater than those  indicated  below, generally based on averages.) 

A “polar axis” arrangement of troughs, where one end is raised to set the trough parallel to the earth’s axis, also increases winter performance but this would be impractical for large scale power generation at significant distances from the equator, because the troughs would have to be short ifn order to raise  one end is to be raised sufficiently, increasing costs greatly and causing a significant “end loss” effect (angled sunlight reflecting out of the ends of many short troughs.) Even in Egypt and Central Australia troughs would have to be on a c. 23 degree slope.  It would seem therefore that the polar axis possibility can be disregarded as a practical option for large scale power generation in winter at considerable distance from the equator.

            The threshold problem and the effect of low DNI on efficiency.

A critical problem for solar thermal systems is what proportion of collected heat is above the threshold level required for production of sufficient steam pressure to initiate generation?  When DNI is below 700 W/m2 there arises the possibility that  considerablemuch heat cwould be collected but without generating much electricity.  In addition heat must be circulated in cold conditions to prevent the heat transfer salt from freezing, and this requires a significant quantity of energy.  Thus DNI figures can be misleading indicators of electricity output.  In favourable regions daily winter DNI totals appear to be good fractions of annual average totals, e.g., 5.7/7.5 for Central A ustralia, but this does not mean that the ratio for electricity generated would be as high, because muchsome of that 5.7 kWh/m2 wouldill be below the threshold level.

A number of sources provide is clear evidence of a marked effect on the solar                                                                                                                                                                                                                                                                         to electricity efficiency of all three solar thermal technologies as DNI falls.  Fig.14 from Odeh, Behnia and Morrison indicates that trough performance deteriorates as solar radiation falls.  Below 6.1 kWh/m2/d (global horizontal radiation, not DNI), collector efficiency plunges, and they state 6 kWh/m2/d as a limit for effective generation.  (The ARDHB shows that winter average DNI in Australia is less than this, i.e., 5.7 kWh/m2/d.)  Their Figs. 1 and 2 show that when DNI drops 45% (from summer to winter Alice Springs averages) heat collected drops 71%.  This means that at the winter 430 W/m2 the solar-to-heat collected efficiency is only 60% of that at the peak average DNI of 780 W/m2.  Their Fig. 13 shows that if the efficiency of the heat collector in summer is .66 but in winter falls to .43.

Another graph from Odeh, Behnia and Morrison seems to be decisive regarding generation by troughs in winter. Fig. 3 indicates that at Alice Springs 26% of annual DNI is under 550 W/m2, and 18% of it is under 350 W/m2. About 20% of DNI is between 550 W/m2 and 650 W/m2 and after the cosine effect has been taken into account it is likely that much of this would fail to raise absorber temperature sufficiently for significant electricity generation. Unfortunately monthly figures are not given (…and again average figures are not as important as are variations and minima.)  These figures would seem to mean that for at least two months of the year there would be little electrical output from trough systems even at ideal sites.

Figures 5 and 14 from Jones et al. (2001) represent a day on which DNI fluctuated around 500 W/m2 from noon to 6.30 p.m.  From 10.30 a.m. to 12.30 p.m. there was intense sunlight so the system had warmed up by the beginning of the period of interest here.  Output between 1.30 p.m. and 6.30 p.m. was mostly between 8 MW and 12 MW, i.e., one-third of peak output.  Thus it seems that DNI of 500 W/m2 or half the peak level was associated with plant output of about .33 of peak capacity.

Table 3 from Solar Paces (undated) on operating troughs at three sites shows that in one case a 19% fall in DNI was accompanied by a 29% fall in output.  The generalisation is stated that for troughs a 1% fall in DNI results in more than a 1% fall in output.  Another source (Kearney,1989) shows that the SEGS annual average efficiency was 10.7% but the summer peak was 21%, again meaning that winter level would have been well below 10%.

The same effect is evident with respect to Fresnel systems.  Mills, Morrison and LeLivre (2003) show that the summer efficiency of a Fresnel arrangement is  25% higher than the annual average, meaning that in winter the figure is likely to would be considerably lower than the annual average.

Most of the solar radiation entering troughs in winter does so at a considerable angle to the axis of the trough, and thus at any point in time radiation received is reduced by the cosine of the angle between the sun and the vertical.  When this is taken into account for each hour of a sunny day in winter in Central Australia radiation received by the absorber is reduced from 5.7 kWh/m2/d to 3.6 kWh/m2/d.

Table 4 in the Australian Radiation Data Handbook (2006) shows that for Alice Springs, central Australia in June the average daily amount of DNI entering a north-south trough is 4.2 kWh/m2.  However what matters most here is the hourly intensity of this energy.  Table 4.7 shows that the maximum DNI entering a north-south trough in June is 527 W/m2, and the average is 394 W/m2.  The average is just over 500 W/m2 for 6 hours a day.  Given that the sun’s angle would be 47 degrees from normal to the trough at noon (cos 47 = .68 ) and much greater in the morning and afternoon, the average amount of energy per metre reflected to the absorber would seem to be under 300 W/m2, well below the level indicated above as necessary to reach .3 of peak output.

As would be expected from the above performance evidence the ASRDHB tables show the theoretical performance of east-west troughs in winter in Central Australia to be (somewhat) better than for north-south troughs.  However the figures for beam or DNI radiation entering an east-west trough (Table 4.11, p. 80) average 4.9 kWh/m2/d and the average hourly intensity over the 12 daylight hours is only 408 W/m2.  (The intensity does exceed 700 W/m2 for 4 hours a day.)

This evidence indicates that physical and geometrical factors determine that the low output from troughs in winter is not amenable to significant technical improvement.  As DNI falls both the amount of solar energy entering a trough and the efficiency of its conversion fall.  It is therefore important to keep in mind that a site’s generating potential is not well indicated by comparing its winter DNI with average or summer DNI.  Apart from the fact that trough technology is regarded as being relatively mature and marked technical advances are not likely, the foregoing discussion indicates that the low winter performance of troughs is due to geometry and cannot be significantly altered.

Finally, in view of estimated global gas resources, troughs are severely handicapped by their dependence on the use of gas to raise temperatures when solar radiation is low.  About 25% of the electricity delivered by the SEGS system is generated by gas (the maximum permitted by Californian law.)  In a renewable energy world with strict CO2 limits little gas will be used.  In any case little gas will be left late in this century.  The consequences of limited capacity to boost generation are likely to be greatest in winter.

Thus the evidence on performance and on trough geometry seem to show that troughs are not likely to be able to make a major contribution to electricity supply in winter.

  Hayden reports average annual output from the 2.23 million square metre SEGS collection area as 77 MW, or 34 W/m2.  From the above summer/winter figures this suggests that winter output would be in the region of 12 W/m2.  A plant capable of supplying 1000 MW in winter would need 83 times the SEGS VI collection area. 

            Water pre-heating.

A solar thermal plant near Sydney, NSW, some 34 degrees south, has been constructed to pre-heat water for a coal-fired power station. (Mills, Le Lievre and Morrison, 2004.) This is sometimes taken to show that solar thermal systems are viable in the mid latitudes.  However this system delivers heat at about half the temperature required in coal-fired power stations, and therefore does not have to concentrate solar radiation intensely. The absorber is over 1 metre wide (Burbridge, Mills and Morrison, undated) and therefore reflectors can be wide with little curvature and be relatively cheap. Because all energy falling on the collectors helps to pre-heat water the intensity of energy falling on the collectors does not have to reach a certain threshold level before electricity is generated.  For these reasons the capital cost is relatively low. 

These features indicate that this plant is not a good guide to the effectiveness or cost of solar thermal plant at this latitude that would generate electricity at this latitude.  In a world that did not exceed safe greenhouse limits there could be few if any fossil fuel plants, and therefore not mucha need for pre-heating.  Also the performance of the system would be likely to fall markedly in winter in view of the above discussion.

Dishes.

Lovegrove, Zawedsky and Coventry  (2006) claim dishes are in general 50% more efficient than troughs or central receivers.  The advantages of dishes are firstly that they can be pointed directly at the sun all through the day and thus avoid the cosine problem which affects trough and central receiver or tower systems and are for the former are especially serious in winter.  Secondly the high concentration ratios enable much higher temperatures than troughs, which make possible efficient generation of electricity via Stirling engines at the focus of each dish.  Thus the losses involved in transferring hot fluids long distances to a generator are avoided.  It is therefore not surprising that their efficiencies have been reported or anticipated at two the three times those of troughs.  Dishes could therefore be expected to perform considerably better than troughs in winter.  However they are considerably more expensive, currently in the region of $6,000 -10,000/kW according to Mancini, et al, (2003).  Energylan, (undated) put the dollar costs at perhaps 4.5 times as high as troughs.  However this is partly because trough technology is more mature and dishes are closer to being “hand made” at present. Probable future costs are considered below.

It has not been possible to arrive at confident conclusions regarding dishes, mainly because again the relevant performance data has been largely inaccessible.  Evidence on Dish-Stirling systems is more readily available, but there is little evidence regarding the important issue here, the potential use of dishes for heat storage. 

There is some data on the winter performance of dish-Stirling systems.  A year long record of the daily output from a 115 square metre US dish located at Phoenix Arizona (Davenport, et al., 2008) shows that in January (winter) output was 68% of the annual average, which is better than for the typical trough.   A year long output record for the 40 square metre Mod 1 and 2 dish systems at good US sites (NREL, undated) shows that on average annual output corresponded to a continual 24 hour flow of 42 W/m2.  However in January the Mod 1 output corresponded to a continual flow of 18 W/m2, and 22 W/m2 in December, i.e., around 50% of average output.  This  reduction corresponds to evidence fromon Davenport on other dish-Stirling systems (Davenport, undated.)  Heller (2010) states that European dish-Stirling output in winter is around 38% of summer output, not as big a fall as for troughs but a considerable reduction. However the theoretically modelled figures from the NREL (2010) SAM indicate a winter flow of 55 W/m2, 54% of summer output, again raising the question as to whether models represent all the factors affecting actual performance in the field.  Fortunately confidence regarding winter performance of dish-Stirling systems is not crucial for this discussion, which is concerned with systems enabling heat storage.

The Australian National University “Big Dish”.

The main non-Stirling dish initiative is the Australian National University 400 square metre “Big Dish”.  (Lovegrove, Zawedsky and Coventy, 2006.)  Its annual average solar to electricity efficiency has been estimated at just under14%, but it is anticipated that this can be raised to 19% in future. (Uncertainties surrounding this figure are considered below.).The figure is net of operating energy costs and takes into account transient cloud.  However it is not clear what the winter solar to electricity efficiencyperformance would be.  It is an experimental device and has not been used to provide electricity to the grid over extended periods.  The 14% and 19% figures are stated as estimates of output under average annual insolation conditions and therefore solar-electricity efficiency in winter could be expected to be lower in view of the evidence discussed above regarding troughs, and below for dishesissues discussed below.

Heat storage via dishes.

The highest reported efficiencies for solar thermal systems are for dish-Stirling devices, which have Stirling engines at their focus.  These take advantage of the higher temperatures due to concentration of radiation at the focus, and they avoid the energy losses and costs involved in taking heat to a power block.  US and European dish developments seem to have been almost only of this Dish-Stirling kind (Mancini et al., 2003).  The crucial point for the purposes of this discussion is that dish-Stirling systems do not involve storage of heat but if solar thermal systems are to overcome the intermittency problem suffered byset by wind and PV they must includevolve energy storage.  Thus high efficiency Stirling engines would not be used (unless storage is via hydrogen; below), and there would be losses associated with piping the heat from the many dishes to a central generator.  Overall system efficiencies would therefore be expected to be considerably lower than those evident in the above cases.  

The three main strategies open are, taking heat from dishes to a power block, dissociating ammonia at the focus of each dish and pumping this to the power block, and using dish-Stirling devices and storing via hydrogen.

The dish-steam/oil approach.

Because there seems to have been little development of dish systems designed for heat storage it has been difficult to get evidence on their potential performance. 

European (Davenport, 2008) and American (personal communications) dish engineers stress the significant difficulties dishes would suffer compared with troughs if the intention was to collect and store heat.  The higher the temperature the greater the loss at the absorber and in transfer, and dish absorber temperatures are around twice those of troughs.  In a large plant there would be considerable heat loss from long lengths of pipe taking heat from the dishes to the power block, or the need for substantial insulation, affecting dollar and embodied energy costs.  A trough system has to move heat much the same distance but this is mostly done via the absorber pipes which are heated almost all the way.

Long distance transfer of heat to a power block would involve a pumping energy cost and a loss through insuolation, and the embodied energy cost of the insulated pipes and pumps, all of which are avoided in dish-Stirling systems. 

The little information available on this heat transfer and loss issue leaves it unsettled.   Kaneff (1991, Fig. 78) reports that at the White Cliffs 14 dish project when DNI was 700 W/m2 34% of heat energy absorbed was lost between absorber and the nearby engine.  At peak insolation it was 23%. (Table X, see also Figs. 78 and 79.)

In the estimations of the energy loss/efficiency cascades for large scale solar thermal systems given by Kaneff (including a 500 MW plant) and by Lovegrove, Zawadsky and Coventry, (including a 100 MW plant), there is no discussion of the fact that heat would have to be moved long distances, whereas for the Big Dish and the White Cliffs 14 dish project the generator was within a few metres of the dishes.  The distances involved in large scale systems would be great.  If a 1000 MW (winter) plant required 100,000 dishes (see below) then the piping connecting them all to the power block might total in the region of 4000 km, (not that a single solar power station this big would be optimal; the maximum practical size for Central Receiver systems is generally assumed to be 220 MW(p).)  Note again that the output from dishes is around twice the temperature of output from troughs, meaning higher losses and/or the need for greater insulation.  In addition note that the core problem for renewable energy is winter supply and the lower ambient temperatures then would increase the loss.

Both Kaneff and Lovegrove, Zawadsky and Coventry explicitly state that piping losses comparable to those for trough systems are being assuming, i.e., around 4 – 5%. This seems to overlook firstly the fact that unlike the Big Dish and White Cliffs casinstances, very very long distance transfer of heat would be involved in their large scale proposals.  Secondly it is not recognised that in the case of troughs fluids are pumped through absorber pipes almost all the way to the power block and thus are heated and do not require insulation. (Davenport 2008 and Kolb 2010 stresses this point.)  The 4 – 5% loss referred to occurs despite this, i.e., from the heated absorbers, the relatively short connections at the ends of troughs and between absorbers and power block.  A satisfactory analysis for a large scale dish system would therefore have to add what would probably be a major loss and efficiency reduction due to long distance piping.  Consequently it does not seem that the figurebreakdowns given by Kaneff and by Lovegrove, Zawadsky and Coventry can be regarded as complete accounts of losses for heat transfer to a power block in large scale plant.

Another significant issue is to do with the energy required for pumping heat to the power block over these long distances.  Lovegrove, Zawadsky and Coventy include this factor but their reference is to the short distance involved in single Big Dish.  Sargent and Lundy report that for trough systems the pumping task can take 14% of energy produced.

TAnother problematic issue and apparent oversight in these two analyses also seem to have overlooked ac major issue to do with concerns the much greater efficiency of the large generators theywhich are assumed.  Kaneff for instance assumes that a 500 MW system would be able to use a generator that is twice as efficient as that used in the 50 kW  Big Dish.  His breakdown (1991, Table H) shows that this line accounts for the much higher overall system efficiency assumestated for the 500 MW plant.  Similarly the breakdown given by Lovegrove, Zawadsky and Coventry shows that the improvement in system efficiency from the current .139 to the ananticipated .19 is almost entirely due to the difference in generator sizes and efficiencies stated, i.e., going from .27 for a 1 MW size to .35 for a 100 MW size.  However if use of one large and high-efficiency central generator is assumed then there would have to be a large collector field and long distance heat transfer to the power block involving substantial heat losses, which would detract from the overall system efficiency assumed.  For instance the outer reflectors in the proposed S220 central receiver (which would not involve heat piping), likely to average 137 MW as ,distinct from 1000 MW, would be more than 2 km from the power tower. (ZCA, 2010.)

Thus the dilemma for large scale projects is evident.  Either benefit from the greater efficiency of larger generators but suffer losses getting the heat to them from very large fields of dishes, or reduce the transfer distances and losses but suffer the reduced efficiency of smaller generators placed within the field.

These considerations reinforce doubts regarding the viability of dish-steam/oil/salt systems for heat storage, and indicate that the .19 overall solar-electricity efficiency figure given for the Big Dish and used in the following derivations is likely to be a significant overestimate when large scale systems are being considered.

The ammonia dissociation strategy.

The ANU solar thermal group has been experimenting with the use of the high temperature achieved by dishes to transform ammonia into nitrogen and hydrogen which can be stored via processes common in the fertilizer industry, and recombined later to release heat. (Lovegrove, et al., 2004.)  US and European dish developers have expressed concern about the problems involved in dealing with hot ammonia, hydrogen and nitrogen under pressure.  (Personal communications.)  It is fortunate therefore that this technology is reported as being built into a commercial solar thermal plant at Whyalla, South Australia by Wizard Power as this will clarify some of these issues. Its design is not finalised and clear estimates of performance etc. are not available yet.  In addition its designers understandably will not provide technical information on proposals and possibilities.  Although the Whyalla site has a very high summer DNI its winter level is surprisingly low, some 2.5 kWh/m2/d in June.  (ASRDHB, 2006.)  It will therefore not throw precise light on what systems located at Australia’s best winter sites might achieve. 

Lovegrove et al. (2004) estimate that an energy efficiency of .7 might be achieved by the ammonia process, although this seems to be given as the upper end of a possible range under ideal conditions.  (Kaneff, 1992, p.143 states the efficiency at 6.)  It is estimated that some 52% of the solar energy entering the dish would be available after storage.  These figures indicate that in winter with insolation of 5.7 kWh/m2/d electricity corresponding to a 24 hour flow of 31.6 W/m2 would be generated (corresponding to 27 W/m2 continuous flow delivered at distance; below.)  This ignores possible threshold effects (below); i.e., the fact that DNI would be over 700 W/m2 for only about 4 hours a day, and therefore that some or much of the daily 5.7 kWh/m2 would not  be of sufficient intensity to start the generation process. Ity also ignores possible energy losses in long distance pumping of gases to the power block.

The important merits of this approach are that the energy resulting would be at 490 degrees, suitable for efficient electricity generation, and that dissociated ammonia could be stored at ambient temperature.  In other words there would be no large heat loss or need for insulation between dish and generator or in storage (although it seems that a small amount of energy would be needed for pre-heating.)  Kaneff, who  built the White Cliffs system, stresses that no confident conclusions can be stated in advance of practical operating experience.  Theoretical models do not necessarily include all relevant factors and conditions can vary significantly from those predicted from climate table averages etc.

            The effect of low DNI on dish solar-electricity efficiency.

It is not appropriate to assume without caution the average annual .19 solar to electricity generating efficiency anticipated in estimating winter performance of the Big DIsh.  There is evidence that as DNI falls the solar-electricity efficiency of dish-Stirling systems falls significantly (i.e., output falls faster than DNI), as was seen above to be the case with troughs.  The power curves for the 128 square metre Sundish and the SIAC/STM dish (Davenport, 2008) show that output at 700 W/m2 DNI is only about 50% of peak output.  This means thatIn other words solar to electricity efficiency falls to .71 of its value at peak insolation. Winter DNI at the best sites typically totals in the region of 5.7 kWh/m2/day but most of it is usually only a little over 700 W/m2, again meaning that at best dish-Stirling output would be only around half peak output.   Table 3 – 5 from Alpert and Kolb, (1988), shows that when output is 25% below maximum efficiency is 30% below maximum.  The example dish-Stirling cases given in NREL’s SAM package also show lower solar to electrical efficiencies for the lower DNI in winter.  The main factor causing the effect is probably the fact that  generator efficiency is maximised at full load, i.e., when temperatures are at their highest.  Table 3 – 5 from Alpert and Kolb, (199, shows that when load is .25 of full load efficiency is 30% below maximum.

WFrom the values given, winter output from a dish-Stirling system at Alice Springs would theoretically be expected tohere radiation is just above 700 W/m2 for about 7 hours a day would probably correspond to a 24 hour flow of 30+ W, indicating a 33 million square metre collection area for a power plant generating 1000 MW in winter./m2.  However this is considerably above the 18 – 22 W/m2 reported above from actual dish-Stirling performance records, indicating that simple theoretical derivation taking an annual average solar-to electricity efficiency figure and site DNI are likely to be much too high.  TheAnThe important point here is that as Kaeneaff notes there is always a risk that theoretical models and derivations might overlook important factors which make actual performance lower than predicted. 

For instance the derivation of the .19 efficiency figure by Lovegrove, Zawadsky and Coventry does not take into account the possible effect of wind in cooling the absorber, which Kaneff reports was significant at White Cliffs (below).

However this evidence might not help much with the question of most relevance here., i.e., the effect of lower DNI on dish-heat systems.  Again little information on this possible effect  for dish-steam or dish-ammonia systems is available.  However Kaneff (1991) provides impressive useful indicative data from the White Cliffs project.  Fig. 73 shows that when DNI was 800 W/m2 the amount of heat going to the engine room at 500 degrees was 40% of the amount at peak insolation, but at 600 W/m2 it was 0%. Similarly, Table X shows that at peak performance 198 kW were flowing to the engine room at 380 degrees and 60 bar, at 800 W/m2 124 kW were flowing at 360 degrees and 60 bar, but at 600 W/m2 only 107 kW were entering the engine room at 290 degrees and 40 bar.  Fig. 82 gives the marked consequences for electrical output.  WhenAt DNI was of  1000 W/m2 32 kW were generated, at 800 W/m2 22 kW, and at 600 W/m2 only 14 kW, 42% of peak.  Thus as DNI fell 46% output fell 56%, again revealingreflecting a decline in efficiency for dish-heat systems a DNI falls.

At the typical maximum winter DNI in Central Australia, c. 730 W/m2, i.e., 27% below peak, these figures suggest an output that is 40%% below peak, and an decrease in solar-electricity efficiency of 13%. 

There is evidence that the Big Dish exhibits a similar effect.  Figures 3, 4, 10 and 11 from Siangsukone and Lovegrove (2003) show that on day when DNI is c.1000 W/m2 power output averaged about 38-40 kW/m2, but on a day when DNI was around 800 W/m2 output averaged about 20 kW, (reaching 26 kW late in the reported period.)   In other words a 20% fall in DNI from the peak value resulted in a 49% fall in electricity generated.  This is somewhat puzzling as it, which is consider ably greater than the fall evident in the dish-Stirling evidence above.  Extrapolation indicates that at 730 W/m2 output would have been 16 kW, 41% of peak.  This means that solar to electricity efficiency at 730 W/m2 insolation would have been a remarkable 44% lower than at peak insolation.

These will be optimistic estimates because the 13 factors to be discussed below are yet to be taken into account, such as the warm up delay at the start of the day and after the occurrence of cloud transients.  These,  are likely to be more significant in winter., has not been taken into account.   (On the other hand the lower ambient temperature in winter tends to increase turbine efficiency.)

The  main concern here is the effect of lower DNI on the efficiency of the ammonia dissociation process.  Presumably the little evidence available on its efficiency relates to ideal experimental conditions including high temperatures.  However in winter DNI typically reaches only 730 W/m2 for four hours a day, and averages under 700 over 8 hours.  In general thermal energy systems are most efficient when temperatures are at their highest.  Thus it is likely that an ammonia dissociation system designed to operate best with ideal DNI levels would be significantly less energy-efficienct at 700 W/m2.

)(It is conceivable however that a system could be designed with sufficient concentration to achieve ideal temperatures at 730 W/m2 and to deal with the higher energy flows at peak DNI by increasing the rate at which ammonia is pumped through the reactor. However this assumes a considerably larger dish, capable of producing the required temperatures with DNI in the region of 700 W/m2.  An almost 30% increase in area is indicated by the ratio of 700 W/m2 to normal/high operating DNI of c. 900W/m2.  Thus the impact on the budgeting attempted below due to increased dish cost would be significant.)

It is therefore likely that the energy efficiency of the dish-ammonia system would be significantly affected by typically low Central Australian winter DNI, and that the .19 figure assumed here is too high.

It will be assumed that this effect applies to the dish-Ammonia process, although the output and cost findings arrived at below are also explored on the assumption that there is no reduction effect.

            Dish-ammonia conclusions.

To summarise the dish-ammonia strategy, the above figures indicate that if a) the average annual solar-electricity of a Big Dish can be raised to .19, b) the energy efficiency of the ammonia storage system is .7, c) winter DNI is 5.7 kWh/m2/day,  then the average winter electricity output would correspond to a constant 24 hour flow of 31.6 W/m2.

 As will be stressed below this is an average figure and what matters in assessing the limits of renewable supply systems is the fact that around half the time DNI and output are below average.  The climate data sources referred to below state that in the best solar thermal regions winter monthly DNI can actually be 40% below average winter levels.  Frequency distributions have not been accessible so it is not possible to estimate how often it would be say 10%, or 20%  below average, but this general figure indicates that a supply system heavily or wholly dependent on solar thermal sources would at times supply not 31.6 W/m2 but only 19 W/m2. half the energy entering a dish becomes available for generating after the ammonia storage process, b) DNI is 5.7 kWH/m2/d, c) the efficiency of the turbine is .35 as Lovegrove predicts for the Big Dish, and d) reduced winter DNI reduces efficiency by 30%, then electricity would be generated corresponding to a 24 hour continuous flow of 29 W/m2. 

TAgain this 31.6 W/m2 figure is a theoretically derived estimate and appears to be is probably much too high when figures for other systems are considered.  Ias it is some 50% higher than the output per square metre reported above for dish-Stirling systems, which are usually regarded as the most efficient solar thermal systems and involve no heat transfer problems and losses, and do not involve the c. 30% loss in the Ammonia dissociation process. The figure is about the same as the Van Voorthuysen (2006) general estimate for a global solar thermal strategy, which also does not involve the Ammonia process and associated reduction in energy available..  It would seem plausible therefore that  the net average winter output from a dish-Ammonia system is likely to be under 30 W/m2

This output figure also does not take in several of the 13 potential reducing factors discussed below.  Reference to two of these will be noted at this point. If only a 15% transmission loss and a10% winter start up delay loss areis taken into account, then electricity delivered from a dish-ammonia system would be in the region of 22 W/m2, about 37% of annual average output without ammonia storage.  lf so, a plant big enough to deliver 1000 MW in winter from a site where average DNI was 5.7 kWh/m2/d, would need a 46 million square metre collection area.  This would equate to more than 110,000 Big Dishes, and at the futurepredicted cost Lenzen’s review (2009) reports as generally predicted the power station (excluding ammonia processing) could cost $18 billion. (Reasons why the figure might  be much higher are considered below.)

These numbers could be taken to indicate that the dish-ammonia system could deliver at a low but useful and acceptable average rate in winter, although at a high dollar cost (below).  .

The dish-Stirling-hydrogen strategy.

Abbott (2010) argues that the best long term future renewable energy scenario would be based almost entirely on dish-Stirling units producing hydrogen.  The strong points in this argument are that the technologies are relatively simple and do not depend on scarce minerals as PV, nuclear and battery-based technologies do.  It would also provide the 75% of energy demand that does not take the form of electricity, whereas the dish-ammonia strategy solves the electricity storage problem but does not help with the remaining forms required.

However the embodied energy cost and dollar cost implications of this path would be problematic.  The power curve for dish-Stirling units (e.g., Davenport et al. undated, Table 1X) indicate that at 700 W/m2 typical of central Australia in winter output is around 50% of peak output, i.e., .68 kWh/m/d.  If this is reduced 15% to take into account long distance transmission, transients, start up delays, dust on mikrrors, etc. the rate of delivery at distance would correspond to a continual 24 W/m2.

If hydrogen was generated where it was needed the delivery rate would be 16 W/m2 for hydrogen gas and around 12 W/m2 for liquid hydrogen, i.e., after taking into account transmission losses.  If this was used to generate electricity after storage via a .5 efficiency fuel cell the final rate of delivery would be c. 6 W/m2, meaning that a solar thermal plant capable of delivering 1000 MW after storage would need 167 million square metres of collection area, or 420,000 Big Dishes.  At the anticipated future cost estimate derived below, $(A2010)240,000/dish, the cost of such a plant would be an impossible $100 billion, or 27 times the early 2000s cost of a coal-fired plant plus lifetime fuel capable of supplying the same amount of energy.of $163,000 per dish (Lenzen, 2009) the total cost would be $68 billion, or 18 times the early 2000s cost of a coal-fired plant plus lifetime coal fuel.

The hydrogen storage system would also be formidable.  To store sufficient hydrogen in winter to meet the c.15,000 MWh daily demand from a 1000 MW power plant when the sun is not shining, via a hydrogen path with an efficiency of .5, 108 TJ of hydrogen with a volume of 12,000 cubic metres in liquid form would have to be stored under pressure and refrigerated.

Direct hydrogen production.

Hydrogen can be produced by splitting water at high temperature, around 800 degrees, and a practical application of solar thermal to this strategy is being discussed. (Taylor, Davenport and T-Raissi, 2008.)  A theoretical 40% solar to hydrogen efficiency is thought to be achievable.  This is double the figure likely from dish-Stirling devices generating hydrogen from electricity output.  However the hydrogen would be being generated at the distant desert locations and would then have to be transported long distances.  Bossel (2004) explains that it is much moreless energy inefficient to transport electricity than hydrogen, and piping hydrogen from North Africa to Northern Europe would probably cost 65% of the energy in the hydrogen pumped.  Consequently direct generation of hydrogen is not likely to be markedly preferable to dish-Stirling generation, HVDC transmission and generation of hydrogen where it is needed. 

It would seem clear that either hydrogen option would less viable than the ammonia path.

                                                Central receivers.

It is unfortunate thatBecause little or no evidence is publicly available on the performance of central receiver systems little space can be given to them in this discussion. (Only two plants are in commercial operation, in Spain, and the owners will not release performance information.  (Hellyer, 2010, Mancini, 2010.)  However some iidea of their probablempressions regarding their likely performance can be based on the figures given in Sargent and Lundy (2003), Alpert and Kolb, (1988), and Radosevich, (1988) and the NREL (2010) SAM package.9’s report.

The anticipated long term future model is a (nominal) 220 MW peak system with a 280 m high tower and 2.65 million square metres of collection surface, set out over a 2+ km radius.  The anticipated average annual solar to electricity efficiency is .165 (although also given as .173 in some tables.)  This means that peak generating rate would be around 437 MW but its 24 hour average output at a site where DNI is 7.5 kWh/m2/day would be 137 MW.  Gross wWinter output in Central Australia would be 40 W/m2 (before taking into account the thirteen reducing factors discussed below .) 

This is a significantly higher estimate than that of Vvan Voorthuysen, (2006), (an optimist regarding solar thermal systems, arguing that they can supply total world electricity demand.)  He concludes that if 7.5 kWh/m2/day DNI is assumed output would average 32 W/m2.  b

The recent NREL (2010) SAM package gives two example cases for a Southern Californian site (where the NASA radiation source indicates a mid-monthly average DNI of 5.2 kWh/m2day).  The mid-winter outputs represented correspond to continuous flows of 28 W/m2 and 30 W/m2/d. 

The figures given in the latter case indicate an average solar to electricity efficiency of around 16%, which is the figure commonly assumed for future systems.   As with the Big Dish Ammonia approach, these SAM figures are derived from theoretical modelling and again the concern is whether all relevant factors that would detract from performance in the field have been taken into account.  For instance in some of the SAM examples

 Factors tending to reduce delivered power are listed below.  One that could be quite significant and could easily be omitted from modelling of central receivers is the effect of wind in cooling the absorber receiving surface has been taken into account but a speed of 4 – 5 km/h has been assumed.  However the absorber f, which for the 220 MW system envisaged as the long term future standard model would be some 2870 metres above ground level. higher than for troughs.  According to the NASA climate data base average winter wind speeds in the eastern Sahara at 50 m average 10 m/sCentral Australia, SW US, Western Egypt and Mildura are respectively. The effect on a relatively large aarea absorber 280 metres high is therefore likely to be considerable in winter when winds tend to be stronger.  The Central Receiver aa bsorber  has a large area that cannot be protected from winds, whereas the focal point for a dish can be almost completely enclosed.

Similarly the models assume low wind speeds but it is stated in SAM that heliostats would be stowed when winds reached 15 m/s, and it is not clear how often this might happen in the field.  This does not seem to have been taken into account.

This concern is reinforced by the summer to winter output ratio for dish-Stirling systems theoretically estimated in the SAM package, which is .54.  Dishes have significantly higher ratio than troughs and central receiver because they can be pointed directly at the sun at all times, yet Heller reports from observation of actual European dish performance in the field that the summer/winter ratio is .38.

It is unsatisfactory that there is no evidence from detailed commercial operations which confirm the figures given by NREL’s SAM, and there is some reason to suspect that they are somewhat too high.  Nevertheless in the following discussion a winter output  from 5.7 kWH/m2 site (as distinct from 5.2 kWh/m2 at Daggett, CA.) where the example are located) corresponding to a  continuous flow of 32 W/m2 will be assummed.  This corresponds to 27/W/m2 delivered at distance, and is much the same as the figure arrived at above for in the discussion of the dish-Ammonia approach, although the dish-Ammonia figure does not take into account the probably lower than average solar to electricity efficiency, nor any energy cost for the long distance pumping to the power block.  It would seem therefore that Central Receivers would be somewhat superior.

Again note that the 32 W/m2 figure is for average winter monthly DNI and at times it could fall to 60% of this.

Another uncertain reducing factor concerns the difference between summer and winter average angle between sun, reflectors and absorber. In winter troughs suffer marked effects due the geometry of their alignment with the sun throughout the day, causing the quite low performance of troughs in winter.  As the sun travels across the sky in summer its angle with the (N-S) axis of the trough is close to 90 degrees all day, at a good location, but in winter it is around 60 degrees, meaning that a square metre of collector receives half as much (cos 60) solar energy as the DNI value per square metre.  Dishes can be pointed directly at the sun all day in summer and winter, and they have only slight curvature, meaning that this “cosine loss” is much less and only due to the low average angle between sun, reflector and absorber.

With central receivers the situation is more subtle and difficult to estimate.  Radesovich (1988) states the average cosine loss for a central receiver as 21%.  A central receiver field resembles a fresnel arrangement of a dish but a inefficiency is created by the fact that when the sun is at a low angle to the horizon the high absorber is not at a point analogous to the focal point of a parabolic dish.  The angle between it, reflector and absorber is low to very low for all the mirrors on the sun’s side of the tower, and at Mildura at midday in winter one would have to go almost a kilometre south of the tower to find a reflector normal to the sun.  For almost all the reflectors north of this one the angle between sun, reflector and absorber would be low to very low.  The result is that the average angle between sun, reflector and absorber in summer when the sun is higher in the sky all day is somewhat greater than in winter when it is lower all day.  Note that this is for midday and in winter mornings and afternoons the sun is even lower above the horizon.  (NASA tables show that for Algeria the average hourly angle of the sun above the horizon in summer is 46 degrees and in winter only 27 degrees.)  Tables 3 – 6 from Alpert and Kolb (19889) sets out the differences and it is evident that when the angle goes from 90 degrees to 15 degrees the efficiency of the field can fall by one-third.  An attempt to estimate the effect graphically indicates that Aat Mildura, ,35  degrees south in South Eastern Australia, the angle averages 26 degrees over the best 6 hours of the day, and at n attempt to estimate the effect graphically indicates that mid day in winter the difference in energy received at the absorber would be c.15% lower than in summer,  (and it would be lower still in mornings and afternoons as the average angle over the remaining 4 hours of sunlight is under 10 degrees.).

This cosine factor is probably the main contributor to the considerable obsedrved difference between solar-electricity efficiency for central receivers in summer and winter.  (The reference here is to the proportion of daily radiation represented by daily electricity production, not to the conversion efficiency at a point in time or for a level of DNI at a point in time.)  Radosevich (Table 4 - 3) shows that daily solar – electricity conversion in summer was a surprising 4 to 9 times as great as in winter.  Alpert and Kolb report  marked reduction in efficiency with low DNI (c. p. 21.) Over the three years in which NREL experimented with a 10 MW pilot plant the ratio of mid-summer to mid-winter monthly gross output was 4.2, 7.6 and 7.2.  The ratios for net output were up to 25% lower (i.e., 10/1 in one year). Alpert and Kolb, (Table 5 - 1, p. 59.)

Some light can be thrown on these ratios by Fig. 4 –1, p. 43 from Alpert and Kkolb showing daily output in relation to DNI.  For the Solar 1 devicer, DNI must reach almost 8.75 kW/m2 before any power is generated, indicating that output is not proportional to DNI but increases rapidly only after a relatively high value is achieved.  In other words, in winter the lower DNI results in disproportionately lower electrical output.

If reduced to take into account the loss in transmission to the eastern Australian population centres, or from North Africa to northern Europe the delivered winter rate would be around 34 W/m2, somewhat better than for delivery from dish-ammonia systems. 

The ZCA report (2010) claimsed that it would be preferable to avoid the long distance transmission problem costs and losses by locating central receivers closer to population but in less favourable DNI.  The location they choose closest to Australian population centres is Mildura but winter DNI there is given by NASA as 3.9 kWh/m2/day (i.e., below the 4.25 kWh/m2/day ZCA assumes.)  NASA tables also state variation around the average and list the minimum as 3.45 kWh/m2/d.   When other effects are added, such as the limits to do with threshold, warm-up and reduced solar-electricity efficiency, it is unlikely that a significant amount of net power could be produced at such a site in winter. 

The solar to electricity efficiency of the two systems NREL investigated was in the region of 6%.  These were early experimental and pioneering early projects and Sargent and Lundy and the NEEDS report (2008,  ( and Trieb, undated200) anticipate achievement of 16 -18% in future.  However it is not likely that the summer/winter output ratios will alter as these are primarily due to the rather  intractableinsic geometry of the central receiver layout.  As with troughs a configuration that increased winter output could be designed, (e.g., by locating the tower on the sun’s side of all reflectors) but this would reduce annual output, and not raise winter output markedly compared with winter output from a normal layout.

Thus the best central receiver strategy for Australia would seem to be to locate in Central Australia where DNI is around 5.7 kWh/m2/day and incur the possibly 15% loss in transmission to the demand centres.  However in view of the summer/winter output ratios reported above, central receivers would seem to be significantly poorer performers in winter than troughs.  Applying a 6/1 ratio to the average 40 W/m2 derived above suggests a winter average output in the region of 13 W/m2, before taking into account the effect of the 13 factors listed in the next section.      If this (uncertain) figure is more or less valid then the most promising option for in winter supply would seem to be the dish-ammonia strategy.

Factors further reducing solar thermal output.

Accounts and claims are often difficult to evaluate  because it is not clear whether they take in all the factors that affect the output of a solar thermal system.  A full energy accounting would have to includetake into account the following thirteen factors which reduce the net energy that could be delivered. Some of these have been referred to above. 

a)  The embodied energy cost of plant.  The available evidence on this the life cycle embodied energy cost of solar thermal systems is unsettled and unsatisfactory.  There have been few analyses, different approaches make different assumptions, derivations are not transparent and they take into account different components, and they have arrived at adifferent figures.  Lenzen’s 2009 review seems to indicate that no studies have been carried out in the last twenty years.

Dey and Dey and Lenzen  (1999)  reporting Weinriebe, Bonhke and Trieb (2008), Norton (1999), and Vant-Hull (1992-3, 2006), state the embodied cost for trough systems at about 4% of lifetime output (25 year lifetimes are assumed).  Kaneff, (1991), estimates that for a commercial 100 MW Big Dish system the embodied energy cost would be 6% (from cost based estimate; see below.)  Herendeen (1988) states 5% to 16%.  Vant-Hull (1991) reports 4.5%.  However as Lenzen (1999) notes, at least some of these estimates do not include all components.  For instance Lechon, de la Rua and Saez’s figures (2006) indicate a cost of about 2 W/m2 continuous flow for dish-Stirling systems but they do not take into account foundations or components other than the dish.  Lenzen (1999, Table 3), reports on his own findings and that of Kreith et al., on cCentral receiver cost estimates.  These are between is reported as 8.5 – 10.7% of lifetime output., by Lenzen (1999, Table 3), Dey and Lenzen (2000, Table 2), and Lenzen’s review (2009, p. 117.)

Dey and Lenzen (1999, p. 359) say that a “cost-based” estimate (see below) of the White Cliffs collection field was 6% of output.  The breakdown of components for a central receiver system shows that the energy cost of the generator, piping and storage correspond to 70 - 80% of the field cost (Ummel and Wheeler, 2008), which suggests that the White Cliffs total embodied cost would be over 11% of output. 

This aligns with Kaneff’s (1991) estimates embodied energy cost for a proposed 948 dish system.  He lists as the main material items in the ANU Big Dish19 tonnes of steel at 40 GJ/t, 50 tonnes of concrete at 2.5 GJ/t, and 2.9 tonnes of glass. The “materials based” energy cost of these items would add to about 1000 GJ, or 2.5 GJ/m2, corresponding to a continuous flow of 3.165 W/m2 over the plant’s 25 year assumed lifetime.  This would be c. 5.7% of the plant’s probable lifetime output at a site averaging 7 kWh/m2/d, assuming the anticipated Big Dish solar to electricity efficiency of 19%. 

Note again that several important components of a complete system have not been included, such as the turbines, the piping to them and the storage tanks.  Perhaps most important is the possibly 15% of output that would be lost in long distance transmission from desert regions to urban centres.  It is reasonable to deduct this quantity from lifetime delivered output from solar thermal systems (along with the energy costs in constructing and maintaining the transmission lines).

A major issue is to do with the validity of these estimates in view of the fact that they take in only the energy costs of materials included in solar thermal plant.  (See for instance Lenzen’s comments,1999, p. 359, on Vant-Hull, 1991.)  Lenzen and Dey (2000, see also Lenzen and Treloar 2003) point out that such estimates do not take in “upstream” costs, such as the cost of constructing the factories and mines that produced the materials.   Lenzen and Treloar show that when the energy cost of  upstream factors is added to the energy inputs needed to produce the steel the energy cost of steel is doubled.  In the case of PV modules a full accounting is found by Lenzen et al. (2006) to treble the commonly claimed energy “pay back” period.  Lenzen and Munksgaard (2002) find that a similar approach to wind turbine production increases common estimates.  No approach to solar thermal plant along these “full accounting” lines seems to have been carried out.

Estimates of energy costs based on dollar costs of construction or production are thought to be better guides than those based on the energy content of the materials used because the dollar cost is affected by all “upstream” factors.  (However the relationship would seem to be imprecise.  For instance the engineering consultation purchased by a dollar is likely to have a rather different energy cost than the steel purchased by a dollar.)  Cost based estimates tend to be somewhat higher (e.g. by some 10%, in Dey and Lenzen 2000, Table 2), but not in the region of twice as high as the above findings re steel, PV and wind would indicate.   Again it is apparent that further clarifying analyses are needed.

Dey and Lenzen (1999) state the relation between dollar cost and energy cost as 12.3 MJ/$.  The estimated long term future cost of the Big Dish is c. $(A2000)163,000.  (Lenzen, 2009, see also Luzzi, 2000.) These figures indicate a “cost-based” embodied energy estimate of 1.92 TJ, which is 11% of the 25 year lifetime output of a dish operating at .19 efficiency, at a 7 kWh/m2 annual average DNI site.  (Note that if exchange rates and inflation are taken into account the current dollar cost figure would be some 50% higher, again complicating conclusions.)

(Of concern re cost-based embodied energy calculations is the fact that the estimated future cost is expected to fall to one third of the present cost reported by Lenzen 2009 and Luzzi 2000 is three times the estimated future cost, yet there would be little or no difference in the amount of materials built into dishes in future.  There would be a reduction in labour cost but an increase in factory infrastructure..  In other words it would seem that a materials based estimate made at a long term future date would be about the same as it is now, providing the energy cost of materials is the same, but a cost- based estimate would fall to one-third of what it is now.)

Another issue usually not clarified in embodied energy calculations concerns the way plant lifetime output is calculated.  From examination of the few available studies it seems that theoretical estimates of gross plant output have been used, whereas the appropriate figure would be actual, recorded net electrical energy delivered to consumers, when all of the factors in the list being discussed here have had their impact on this quantity, such as operations and management energy, start up delays, water pumping to the site, dust on mirrors, cooling by winds, down time due to high winds and to breakdowns and maintenance, and losses in long distance transmission. Several of these are not included in the theory-based estimates given for instance by Lovegrove, Zawadsky and Coventry, and by Kaneff.  None of the other studies reviewed, including most if not all of those Lenzen refers to, states how lifetime energy output was determined.  I, and it is conceivable that the effect of all the possible reducing factors combined might add to the equivalent of more than 25% of gross output, possibly increasing the embodied energy cost figure 33% over limited theoretical estimates based on gross output.  In addition the issue of lifetime energy delivered at long distance has been mentioned. 

Below it will be assumed that the embodied energy cost of solar thermal plant need not be deducted from gross winter output, on the grounds that these costs can be paid from surplus summer output.  The extent to which this will be possible depends on the amount of energy-intensive production that can be economically carried out only in summer.  In principle steel, cement, aluminium and fertilizers might be mostly produced then, but it is not clear what the cost implications would be.  It would not only involve large scale plant sitting idle for half or most of the year, but the output capacity of that plant would have to be up to three times the average size required to keep a constant 12 month output up to demand.  The associated costs should be accounted to the solar thermal system, but this will not be done in the estimates attempted below.

More importantly, it is not likely that sufficient functions could be found to make full use of the surplus solar thermal power available in summer.  A system capable of meeting demand in winter would be providing approximately twice as much in summer.  Relatively little of this would be taken up in the production of the steel, cement and glass required to produce the solar thermal plant, and the production of  items such as steel, cement, glass and fertilizer for the whole economy do not require around half of total energy production.

In other words it seems that some fraction of embodied energy costs should be deducted from winter output to arrive at a realistic figure.  This will not be done in the following derivations.

b) The embodied energy cost of the ammonia heat storage system.  Without public access to the relevant information it is not possible to assess this factorelement confidently.  ItThis would include the reactors in the dishes, large volumes of piping for storage of the dissociated Aammonia and for transporting it to the turbines, and piping for the return of the hydrogen and nitrogen to the dishes, and the associated pumps. 

Kreetz and Lovegrove (2002) report that in their experimental reactor a flow of 9 grams per second of ammonia corresponded to an energy flow of 993 W.  The average daily output from a Big Dish would be 532 kWh.  To store half of this would involve a flow corresponding to 22 kW for 12 hours, which indicates the need for 8.6 tonnes of ammonia.  An uncertain estimate would be that tTo store this might require 12 m of the 1 metre diameter standard gas pipe it is reported that Wizard Power intends to use at Whyalla, and the same amount would probably be needed to contain the dissociated gases returning from the power block’s reactor.  The embodied energy content of this steel would be around 190 GJ, which is 1.3% of the dish’s lifetime output.  Again the figures can not be no more than indicative but they suggest that the process would not add markedly to total embodied energy cost.

However it is possible that the storage pipe would have to be replaced several times in the 25 year plant lifetime, due to the effects of corrosive Aammonia, high pressure  and the embrittlement of metals which hydrogen causes.  If the pipe had to be replaced each 8 years the lifetime embodied energy cost would treble.

These estimates suggest that when transmission lines and ammonia plant are included the full embodied energy cost of a solar thermal dish-ammonia system could be above 15% of lifetime output delivered at long distance. 

c)    Plant operating and management energy costs would have to be deducted from gross output.  Kreith and Goswani (2007) state these at 8% for a dish-steam system. Presumably pumping ammonia from the many dishes to the turbine would require energy comparable to that needed to pump oil in trough systems. Sargent and Lundy (2003, Section 4 – 3) expect plant operating energy cost to fall to c. 10% of output in future. Jones et al. (2001, Figs. 5, 15) report a 10% cost for SEGS VI. 

It is usually not always clear whether stated O and M figures include all factors in addition to pumping absorbing fluid, such as reflector washing, vehicle use, lighting and keeping heat transfer salt from solidifying on winter nights... 

Because a plant capable of 1000 MW output in winter mightwould be more thanaround twice as big as one capable of this at average or summer DNI, its operating costs, for instance those associated with washing mirrors or pumping heat longer distances, would be correspondingly higher.

Lovegrove, Zawadsky and Coventry include plumping energy cost in their derivation of the .19 solar- electricity figure, but as has been noted this does not take into account the long distances there would be between dishes and power block in the very large systems required to enablewith large and efficient generators.

Unlike embodied energy costs, the full operations and management costs must be paid from output continually, so must be deducted from gross winter monthly output.

d)   The embodied energy cost of the long transmission lines, transformers etc. also have to be taken into account along with their lifetime operations and management cost, e.g., vehicles, vessels for Mediterranean cable service, etc.  For Southern European supply from North Africa Czisch (2004) estimates the dollar cost for these lines would addat perhaps 30%.3 toof plant cost.  Thus lines to Nnorth WesterEastern Europe might add 50%.

It could be argued that in attempting to estimate net delivered winter supply from solar thermal plant, embodied and operating energy costs should not be subtracted from gross winter output, because these can be thought of as being paid from surplus summer output.  The scope for this approach to energy accounting is not unconstrained.  Operating energy-intensive plant only in the summer and/or only when there is full sun,, e.g., for cement, steel, ammonia production, furnaces and manufacturing, would impose costs in the form form of inefficiencies (e.g., warm ups) and of plant lying idle for much of the year.  Some functions can be carried out on a somewhat intermittent basis, such as freezer boosting, but this is not so for processes such as steel production.

e)  “Transients.”  When clouds pass over a dish it can take 5 or 10 minutes for output to rise to previous levels.  For trough systems this can reduce daily dish output an average 10%. (Sargent and Lundy, 2003.)   As has been noted, it is possible that a high daily DNI total could be made up of many short sunny periods separated by cloud, resulting in many warm up delays, and little or no generation of electricity. This factor has been estimated as costing 8% of output and has been included in Lovegrove’sin the derivation of the anticipated .19 annual solar to electricity efficiency for the Big Dish. 

However the winter value for loss due to transients is likely to be considerably higher than the average because of the greater occurrence of cloud, and this has not been taken into account herein the above derivation.  (See below on cloud occurrence in winter.)

f)  Down time for repairs would need to be accounted, although some of these could be carried out at night.  The modelling given for the Big Dish assumes down time will reduce output 6% (which has been assumed in deriving the .19 annual solar-electricity efficiency estimate.)  Further evidence on this factor would be useful as it is only c. one-third the figure for coal-fired generation, and a lower proportion of the typical nuclear reactor figure.

g)  The start up delay is typically an hour on a summer day.  Figures 3 and 10 from Siangsukone and Lovegrove (2003) show that on a morning when DNI was 1000 W/m2 it took an hour for output to rise to peak after cloud, even when output in the previous hour was at 75% of peak; i.e., the warmn up referred to in this case was not from a cold start.  Figures 4 and 11 show that on a day when DNI averaged 800 W/m2 output rose to half its peak for the day in 15 minutes, but then took 50 minutes to reach the 26 kW peak for the recorded period (rising smoothly over that period.)  The peak for the day was more or less reached only at 11.50 a.m., suggesting that on a day when DNI averaged 800 W/m2 the system would not have operated well for more than a few hours. 

Similar evidence is given by Kaneff (1991), Kaneff (1992), Mancini (2007), Broesolme et al. (undated, Fig 6), and Brokman and Kearney, (2002.)   In winter warm up delay is likely to be longer than the 1 hour observed for summer, and given that DNI does not rise much above 700 W/m2 for 7 hours the delay might reduce energy available for generating by c. 15%.

h)   Storage loses some heat energy, although little is lost from presently operating systems. Sargent and Lundy state this as .9% for troughs.  (2003, Table 4.3.l.  , seSee also Lovegrove, Zawadsky and Coventry, 2006.)  However this figure refers to the present c. 5 -7 hr storage and the need to store for periods between 16 hrs and several days, considered below, would increase losses.

i).  i)  Turbine cooling. Solar thermal systems are most likely to be located in desert regions.  From the data given by Solar Paces (undated, 5-43) the turbines of the equivalent of a 1000 MW solar thermal plant would use 18.5 billion litres of water p.a.  ZCA says the use is 340 litres per MWh, and around 17% of the figure for coal-fired generation.  (ZCA, 2010, p. 60.)  The USGS (2010) states that electricity generation actually accounts for half of total US water use, more than agriculture.  The most efficient cooling is by evaporation, meaning that the water cannot be condensed and reused. In some situations sea water can be used but not in Central Australia, nor in North Africa because as cloud occurrence increases with proximity to the sea plant would not be located there. 

Evaporative cooling is reported as costing around 10% of the energy generated by troughs, and takes 2.4 cubic metres of water per MW.   The figures are claimed to be lower for Central Receivers. (Solar PacesACES, undated.) This is a significant problem for large scale solar thermal visions as the best locations are in deserts.  However air cooling is possible and ZCA (2010) claims the energy cost for central receivers is under 2%, although the dollar cost increases slightly.

j) Mirror washing. This factor is mainly of concern regarding water demand in desert areas, rather than energy cost.  Hayden (2004, p. 189) reports that SEGS  reflectors are washed every five days, and subject to high pressure washing every 21 days.  A system capable of delivering 1000 MW in winter might involve washing some 40+ million square metres of collection area, and 470,000 km of travel p.a. for special purpose vehicles carrying water tanks.  Washing 400 square metre big dishes would be more problematic than washing 5 to 10 metre wide troughs.  A Central Receiver example in the NREL SAM package states that 63 washes p.a. would be needed, at .7 litres of water per square metre of reflector.    For the SS220 that would require 117 million litres p.a.

k

j)  Loss in heat transfer.  Reference was made above to the concern expressed by

European and US dish-Stirling researchers regarding the losses likely in moving heat from absorbers to the power lock.  Losses for trough systems are given by as % of output.  Note that with troughs heat transfer fluids pass through absorbers most of the way to the power block, but this would not  beis not the case with a field of dishes.  Never the less Lovegrove and Kenaff assume a loss of 4 to 5% for the Big Dish and White Cliffs ventures.  However these involved negligible transfer distances as generators were very close to dishes.  Kaneff points out that generating efficiency increases markedly with size so it is desirable to have a large unit fed by many dishes, but this would mean long distance transfer of heat to the generator.  This would seem to confirm the antipathy of the European and US dish-Stirling researchers consulted to the use, suspecting that use of dishes to collect heat is not a good idea.  (Heller, 2010,  Mancini, 2007.)  For instance if 1050,000 dishes were needed (see below) to provide 1000 MW about 42000 km of piping might bewould b e needed to move heat from the 25 metre diameter dishes to the central power block.

ll)  Loss in long distance transmission.  For transmission via High Voltage DC lines from North Africa or the Middle East to Europe, or from the South West of the US to the North Eastern cities, a considerable loss of energy would occur.  Mackay (2008) and Czisch (2004) say this could be 15%.   Breyer and Knies (2009) concur, stateing 3% per 1000 km.  The figures in the NEEDS report (2009) align with the foregoing estinates, stateing an 8% loss from Spain to Germany and a 13% loss from Algeria to Germany.  However Ummel and Wheeler (2008) arrive at a higher estimate, 12% per 1000 km, plus .2% for the two substations required at the start and end of the line.  (Others state .6% for each of these.)  They also point out that each line could only carry 3 GW.  The loss  figure assumed below is 15%, mainly because if North American or Northern European supply is to come from desert regions distances would be several thousand km..

m)  Cooling  by the wind.  Kenaff (1991, Fig. 74a) reports that in a region with 881 W/m2 DNI an

increase in wind speed from 2 to 4-5 m/s reduced energy absorbed by dishes by 9%.  No reference to this factor seems not to have been taken into account for the Big Dishh, or in the published literature on central receivers.  Dishes would be more prone to this effect than troughs due to their higher absorber temperatures, greater exposure, and greater height from the ground, and higher temperature. 

Central receivers would  be most prone to the effect given the 280 m height of the proposed 220 MW unit and the impossibility of protecting relatively large absorber areas from the wind.  Alpert and Kolb report that for a central received an increase in wind speed from 2 to 12 m/s reduces receiver efficiency 9.5% (Table 3 – 2, p. 27),k and that a 12 m/s wind doubles the heat loss from the absorber that occurs with no wind. (Table 3 – 4, p. 33.)  The NREL SAM examples state that 4 - 5 m/s wind speeds are assumed, but it is not clear whether these are at ground level or at absorber height. Reference has been made to NASA climate data indicating much higher average speeds at 50 m in North Africa.

Again Kaneff’s practical experience in building the White Cliffs project leads him to stress the difference between the conclusions modelling might lead to and what happens in the real world when all unanticipated factors, variations around assumed means, breakdowns, and mistaken assumptions have had their impact in determining actual performance.  It is likely that the above factors would combine to reduce winter output delivered at distance from central receivers of Big Dishes plus Ammonia systemswhat appears to be the best option discussed above, the Big Dish with ammonia storage strategy, to well below the252 W/m2 derived above.

Dollar costs.

Evidence and claims regarding the likely long term future costs of solar thermal technologies are scarce and vary considerably, therefore and estimates cannot be taken with confidence.  Conclusions can only beare typically educated guesses and are typically not accompanied by numerical arguments providing detailed or itemised derivations that can be assessedverified.  Predictions also tend to assume cost “learning curves” observed in other (selected) engineering fields.  However, but that term might best be confined to improvements in an established technology brought about by increased production scale, plant size, and technical advance, whereas dish and central receiver CR technologies (unlike troughs) aremight best be regarded as no not yet established on a preferable path.  For instance or at the anticipated scales. (Eg., central receivers in use are a small fraction of the greatly increased 220 MW scale anticipated and this will set engineering challenges that have not been dealt with to dateso far.  “Learning curves” for such devices cannot begin until particular devices are built, monitored, adapted etc. over relatively long periods..)  Technologies that are in a pioneering/experimental phase often incur large cost overruns before the best strategies are established.  

It is often assumed that scaling up to mass production will have a large marked reducing effect on unit price, but the NEEDS report (2009, Fig. 3.8) does not anticipate a marked effect.. 

A significant concern problem for those assuming cost reductions is set by recent trends for wind turbines as these run sharply against the conventional wisdom.  In the eEarly 2000s the commonly stated cost was c. $1,500 per kW of capacity.  Wind might be regarded as a “”mature” technology now enjoying the “learning curve” benefits of a rapidly increasing production scale.  However in recent years turbine costs have risen not fallen, and ABARE (2010) reports the average cost or units built in Australia as a remarkable $2,900/kW, including a 30% increase in onethe last year.

The following discussion deals only with capital costs and does not include operations and management costs, which are significant additions to total lifetime cost.  Lenzen (1999) says for large plants these costs would add to almost 20% of plant capital cost.(ABARE, 2010.)

Solar thermal systems are typically located in deserts a long way from demand and the costs of long distance transmission lines should be added.  Transmission lines from the Sahara to Europe under the Mediterranean Sea would probably add one-third to plant cost, according to Czisch (2004).  DESERTEC proposals refer only to supply over relatively short distances, such as from Morocco to Spain and from Egypt to Turkey.  Supply to Sweden or the UK would be considerably more costly.

Easily overlooked is the fact that all these cost figures refer to present materials, construction and energy costs, and in future materials and energy inputs are likely to be considerably more expensive than they are now.  Given the fact thatway all inputs into production involve energy it would not be possible to estimate the total multiplier effect on solar thermal plant cost that might be brought about by significant increase in energy costs. (Unlike nuclear and PV sources, solar thermal requires little more than common materials such as steel, cement, glass and aluminium, so scaling up would not demand large quantities of rare materials.  However these common materials are energy-intensive and in the coming era of probably high energy prices multiplying effects are likely to result in high plant construction costs.)  Effects are likely to be compounded along chains of inputs to production, i.e., it is likely that  increased energy costs will increase costs of all factors of production, which in turn will increase costs of the labour and services feeding into the production of materials and structures in a multiplicative way.

It is important not to directly compare the usually stated capital cost per kW of fossil fuel and nuclear plant with that of wind and solar plant, for which output is intermittent.  Wind is commonly seen as having a capital cost of around $1500/kW and the stated coal cost is about the same.  However coal plant plus fuel (early 2000s price) over plant lifetime would have cost approximately $(A)3,700 million, (although more recently costs of electricity generating plant in general appear to have risen significantly.)  The solar thermal plant cost figures below are for the capacity needed to produce 1 KW at peak output, but over a year the 24 hour average output from a coal plant is around 80% of its peak output, whereas for a solar thermal plant it is around 15% of its peak capacity.  Thus taking the above coal power figure and the Sargent and Lundy estimate, the “near future” capital cost per gross kW delivered on average (or as a continuous flow, as distinct from peak output) from a solar thermal trough plant would be about 8 times as great as for coal including fuel.  The average continuous output flow is what matters, and this is why this discussion focuses on a watt per square metre of collector.

In addition to these sources of uncertainty,

To summarise, tTthe following evidence indicates that estimates of present and future costs vary considerably, the grounds for expecting significant falls are uncertain, and it is conceivable that despite mass production and “learning” effects future costs will not be significantly lower than at present, due mainly to potential energy price rises.  The most optimistic estimatfigures anticipates a fall to one-third of present Big Ddish costs by 2050, and this will be used below, despite the lack of convincing supporting evidence.  However the NEEDS report (2009, p.31.) expects costs to only halve by 2050, and ABARE predicts only a 34% fall in solar thermal cost between 2015 and 2030.  EPRI (2009) actually reports a rise in solar thermal electricity cost from $175/MW to $225/MW in the year to 2009, a 30% increase. 

Trough costs.

According to Sargent and Lundy (2003) the “near term future” cost of solar thermal trough systems will be $(US2003)4,589/kW, or $(A2003)6,556 (using the early 2000s exchange rate.)  This figure includes heat storage, which reduces required generator capacity and cost.  Their long term future (2020) cost prediction is $(US2003)3,220/kW, which when adjusted for inflation and exchange rate difference becomes $(A2010)4,250/kW.

According to Sargent and Lundy (2003) the “near term future” cost of solar thermal trough systems is $(US)4,589/kW, or $(A)6,556 (taking the early 2000s exchange rate.)  This figure includes heat storage, which reduces required generator capacity and cost.  Their long term future, (2020) cost prediction is $(US)3,220/kw.

However, NREL (2005) states that the 2003 cost for the SEGS systems wais $(US2003)7,700/kW which would have corresponded to $(A2003)11,000/kW.  Viebahn, Kronshage and Trieb, (2004) state e5300/kW.  (20, Table 2 – 3, p. 12.)  They expect costs to halve  by 2050. (Fig 3 – 7.)  The example case given in the SAM modelling packge (NREL, 2010) states $(US 2010)8,243/kW.  If, following Lenzen, 20% should be added for lifetime O and M cost, the total would be $9,892/kW. 

ABARE predicts only a 34% fall in solar thermal cost between 2015 and 2030.  EPRI (2009) actually reports a rise in solar thermal electricity cost from $175/MW to $225/MW, a 30% increase in the year to 2009.  It is noteworthy that the recent NEEDS (2008) estimate is in the region of $(A)17,000 per kW. 

A coal plant plus fuel (early 2000s price) over plant lifetime would cost approximately $(A)3,700 million, although more recently costs of electricity generating plant in general appear to have risen significantly.  The above solar thermal plant cost figures are for peak outputs but the average output from a coal plant is c. .8 of peak whereas for a solar thermal plant it is around .2 of peak capacity.  Thus taking the above coal power figure and the Sargent and Lundy estimate, the “near future” capital cost per gross kW delivered on average (as distinct from peak) from a solar thermal plant would be about 12 times as great as for coal including fuel, (indicating  so possibly 6 times as great as in 2010now.) 

In addition solar thermal systems are typically located in deserts a long way from demand and the costs of long distance transmission lines should be added.  Transmission lines from the Sahara to Europe under the Mediterranean Sea would probably add one-third to plant cost, according to Czisch (2004). 

Thus it is not at all clear what should be assumed regarding future costs for solar thermal systems.   Indeed there is evidence from recent trends for wind and solar thermal construction which suggest that costs will not be lower.  Note that the figures discussed are for the annual average output, and thus do not indicate plant sizes and costs that would be required to enable solar thermal plant to produce as much power as a coal-fired plant in winter.

Easily overlooked is the fact that all these cost figures refer to present materials, construction and energy costs, and in future materials and energy inputs are likely to be considerably more expensive than they are now.  Given the way all inputs into production involve energy it would not be possible to estimate the total effect on solar thermal plant cost that might be brought about  by significant increase in energy costs.

c

            Dish costs.

Mancini, et al., (2003) put dish-Stirling costs in the region of $6,000 -10,000/kW.  According to Energylan, (undated) the dollar costs is perhaps 4.5 times as high as for troughs.  However this is partly because trough technology is more mature and dishes are closer to being “hand made” at present. The figure given by the NREL (2010) SAM package for one of the examples is much lower, c. $(US2010)3,014/ kW.  The discrepancy is unsatisfactory, but fortunately is not crucial to this discussion which does not focus on dish-Stirling systems.

Sargent and Lundy do not discuss dishes.  Luzzi (2000) states that the cost of a Big Dish would be $440,000 but in future could fall to one-third of this figure. Luzzi does not provide any derivation or support for the prediction.  Lenzen’s review (2009, Fig. 8.3.4, p. 119) indicates that the estimated future cost would be is around 37% of the initialpresent figure, which would be i.e., $(A2000)163,000.  Lenzen’s references suggest that there is little does not indicate that there is much evidence available on the issue other than from Luzzi., (Thebut the source for Lenzen’s this figure seems to be Luzzi) CHECK say if not muxch ev.  This figure (Sargent and Lundy do not discuss dishes.)  This equates to around $5,140 per peak kW.hcorresponds to around $(A2010)240,000.  Given the 76 kWE peak output assumed for a 400 square metre dish at .19% solar-electricity efficiency, and taking inflation into account between 2000 and 2010, a figure orf $(A2010)3,156/kW is arrived at (not including the cost of the Ammonia dissociation plant.)

This is a quite low figure in view of the information on some other systems.  The figure for the 30 MW SEGS VI system, costing $(US1988)116 million and with 188,000 m2 of collecting area, is $3,867/kW, corresponding to $(A2010)7,657/kW.  Manci and Heller (above) estimate present dish-Stirling capital cost at $(US2003)6,000 -10,000.  The NREL Sam package states $(US2010)8,243 for a trough.

If the Luzzi figure is taken, along with a net winter delivered figure of 25 W/m2, a plant delivering 1000 MW at distance in winter via ammonia storage would require 100,000 dishes, and would cost $A(2010)24 billion   If this cost is assumed and if 110,000 dishes are needed to equate to the winter output of a 1000 MW power station using ammonia storage and located in Central Australia the cost for the collection field and power generation plant, but, excluding the ammonia handling plant and piping, and operations and management costs. (O and M costs are likely to add a sum equal to almost 20% of capital cost according to Lenzen, 1999.), would be $18 billion.  

An important reason why this is probably a significant underestimate is that the systems assumed have heat storage enabling much smaller turbines than would be needed to deal with peak radiation, because then they can run at the much lower constant rate averaged over 24 hours.Where heat is stored turbines can be around one third the capacity that would be needed to produce at peak daily radiation, because they can run at a constant lower level all day.   However this effect would not be available regarding the aammonia reactors, because these would need to be large enough to operate at peak daily radiation, i.e., to process and store all energy available at midday..

To this must be added the cost of either insulated piping connecting every dish to the a central power block or to storage tanks,  (which could be over 4000 km for plant generating a 1000 MW plant in winter; see below) to a central power block, or to the power block viaof the ammonia heat storage system, which would also have to be added to the above cost for the dishes alone.  If 100,000 dishes of 25 metre diameter are spaced at 50 metres centre to centre are to be connected to a power block then connect every dish to the central turbineover 4000 km of pipe might be needed.  This might double as storage for the ammonia, hydrogen and nitrogen gases.  There would be problems regarding probable lifetime in view of corrosion and pressure effects, and the associated impact of replacement costs on total plant lifetime costs.. 

The same length of pipe will be needed to bring the ammonia back to the dishes.  It is not clear whether separate pipes would be needed for the dissociated Nitrogen and Hydrogen as it has been suggested that these might be able to coexist in the one space.  If separate pipes are needed then the pipe distance would triple.

The future cost figure derived from Lenzen $(A2010)240,000 per Big Dish, will be used below, although it cannot be regarded as a precise or confident estimate.

Therefore it would seem that little can be said with confidence regarding the likely future cost of Big Dish-ammonis units.  The figure assumed below for working purposes is $(A)163,000 per dish, excluding the ammonia processing and storage components.  This is the lowest of the figures indicated by the (unsatisfactory) evidence.In the early 2000s the cost of a coal-fired power station plus fuel for its lifetime was in the vicinity of $3.7 billion.

Central receiver costs.

Unfortunately estimates of the probable future Central Receiver cost also vary considerably. 

Sargent and Lundy (2003) state a cost of around $9,090/kW for central receivers but expect this to fall to $2,6843,591 by 2020 (presumably in $US2003, which corresponds to $(A2010)64,605,162, taking into account exchange rates and inflation).  Viebahn, Kronshage and Trieb, (2004) state e10,140.  The NEEDS study (2008) states e10,241/kW.  Based on an updating of the information given by the NEEDS study

Nicholson and Lang (2010) derive 2010 costs of $(A)16,400/kW and $(A)25,700/kW for the two central receiver systems currently being  built,  based on information iven y the NE.  (They also report a 260% increase in projected expected costs for one of these projects.)  The NREL (2010) SAM package provides a 100 MW example for Southern California costing $(US2010)6,575/kW.  (This confuses estimation somewhat as it is not clear whether this is present cost or probable future cost.  The solar-to-electricity efficiency assumed .16 is around that given by various analyses, such as Sargent and Lundy 2003, and ZCA 2009, as the expected future figure.). 

Again the variation prohibits confident expectations.Again the variation prohibits confident expectations and most of thesethe figures are for present cost whereas the above dish-Ammonia  figure is for estimated future cost.  Most useful might be the Sargent and Lundy figure as it is the only one clearly given for future cost.  It is in the region of twice1.35 times the figure for future Big Dish cost from Luzzi.  Therefore even when the cost of the aAmmonia dissociation plant is added the Big Dish strategy would seem likely to remain significantly lower. 

It should be kept in mind that all these figures to do with the three technologies are for peak output, which would be in the region of 165W/m2 and if electricity delivered at distance in winter is in the region of 40W/m2 (and it was seen above that it might be 13 W/m2), the cost of a plant big enough to deliver 1000 W/m2 in winter would be more than 4 times as great.  

Reducing factors.

Sargent and Lundy do not discuss most of the factors listed above under the discussion of troughs as tending to reduce central receiver output.  One that is likely to be significantly different for central receivers compared with troughs is the effect of wind in cooling the receiving surface, which for the 220 MW system would be some 270 m higher than for troughs.  According to the NASSA climate data base average winter wind speeds in Central Australia, SW US, Western Egypt and Mildura are respectively.  The effect is therefore likely to be considerable.

Another uncertain effect concerns the summer compared with winter average angle between sun, reflectors and absorber.   In winter troughs suffer marked effects due the geometry of their alignment with the sun throughout the day, causing the quite low performance of troughs in winter.  As the sun travels across the sky in summer its angle with the (N-S) axis of the trough is close to 90 degrees at a good location, but in winter it is around 60 degrees, meaning that a square metre of collector receives considerably less solar energy than the DNI value per square metre.  Dishes can be pointed directly at the sun all day in summer and winter, and they  have only slight curvature, meaning that this “cosine loss” is much less and only due to the low average angle between sun, reflector and absorber.

With central receivers the situation is more subtle and difficult to estimate.  A central receiver field resembles a fresnel arrangement of a dish but a minor inefficiency is created by the fact that when the sun is at a low angle the high absorber is not at a point analogous to the focal point of a parabolic dish.  When the sun is at a low angle to the horizon the angle between it, reflector and absorber is low to very low for all the mirrors on the sun’s side of the tower, and at Mildura in winter one would have to go metres south of the tower to find a reflector normal to the sun.  The result is that the average angle between sun, reflector and absorber in summer when the sun is higher in the sky all day is somewhat greater than in winter when it is lower all day.  (For Algeria the average hourly angle of the sun above the horizon in summer is 46 degrees and in winter only 27 degrees.)  An attempt to estimate the effect graphically indicates that at Mildura at mid day in winter the difference in energy received at the absorber would be c.15% lower than in summer, and the figure would be higher in mornings and afternoons as the sun’s angle with the horizon fell further.

The question of future central receiver costs would seem to be no clearer than for troughs and dishes as discussed above.  Sargent and Lundy anticipate a fall from $9,090 to $3591  by 2020.

Where; algeria may 37% cloud. Later

????you should take S and L 3220 for troughs long term??

Cost.  S nand L $9,090 to $3591 in 2020.

Cr conclusions

            Some other issues.

If solar thermal plants are to play a major balancing role in an electricity supply system containingwith much solar and wind capacity a major cost saving often claimed for solar thermal systems would not be available.  The ability to store heat from peak mid day collection and to use it to generate at a much lower constant rate, perhaps .2 of peak capacity, means that much smaller and cheaper generators can be used. The power block can make up around 40% of a solar thermal system’s cost so the saving in capital costs, energy costs and operations and management is considerable.  However if at times the solar thermal component of a renewable supply system is to be called upon to plug gaps left by intermittent wind and sun, then there will be times when it must generate at closer to its peak collection rate.  That is, generators large enough to cope with peak insolation would have to be paid for. (or many more solar thermal systems than are needed to meet normal demand must be available.)

Another issue is the need to wash reflectors frequently and the resulting water, vehicle and energy demand, given that solar thermal plant is intended for use in desert areas.  Hayden (2004, p. 189) reports that SEGS  reflectors are washed every five days, and subject to high pressure washing every 21 days.  A dish system capable of delivering 1000 MW in winter might involve washing some 40+6 million square metres of collection area.  This might involve 470,000 km of travel p.a. for special purpose vehicles carrying water tanks.  Washing 400 square metre big dishes would seem to be more problematic than washing 5 to 10 metre wide troughs.

A related issue concerns the amount of water needed for cooling the turbines.  Solar thermal systems are most likely to be located in desert regions.  According to one estimate (Solar Paces, undated, 5-43) the turbines of the equivalent of a 1000 MW solar thermal plant would use 18.5 billion litres p.a.  The USGS (2010) states that electricity generation actually accounts for half of total US water use, more than agriculture.  The most efficient cooling is by evaporation, meaning that the water cannot be condensed and reused. In some situations sea water can be used but not in Central Australia, nor in North Africa because cloud occurrence increases with proximity to the sea.  Air cooling is feasible, but at an energy penalty variously stated as 7 – 10% of output for troughs.

There would also be a problem regarding the need for solar thermal turbines to rapidly ramp up to high levels of output from stored heat, in order to meet more of the demand when sun and wind energies fall suddenly, i.e., assuming no use of gas  in the long term future.  Thermal generators can not be brought up to full output quickly.  Existing thermal power plant copes with the normal daily ramp from night time to morning demand, but that is around a 30+ – 50% increase over some 6 – 7 hours whereas input from an entire integrated regional wind system can drop from almost peak to almost zero in less than that time.

A supply system in the form of troughs capable of meeting demand in winter would produce far more energy in summer than was needed, probably four times as much in view of the  above summer/winter ratios noted above, and would have to dump much of it.  Some of the excess would go into compensating for the lower wind capacity in summer, and some functions such as cement and fertilizer production might be carried out only in summer but the scope for activities of this kind would be limited, and capital costs of plant idle except in summer would be high.  Some processes, such as electric furnace production, are not suited to operating intermittently.

            Output/cost conclusions.

The main conclusions regarding dishes and Central Receivers indicated by the foregoing figures are as follows.  (Due to the clearly lower performance of troughs in winter they will not be considered further.)

For both the Big Dish-Ammonia system and Central Receivers, the net flow of electrical energy delivered at distance in winter, deducting for transmission and O and M losses appears to be c. 27 W/m2.  Therefore taking into account the remaining factors in the above list of additional reducing factors would probably reduce this figure significantly.  In the derivations below a net winter flow delivered at distance of 25 W/m2 will be assumed.

Given that the above exploration of future costs suggested that Central Receivers would cost significantly more, possibly in the region of double the cost of Big Dishes (without Ammonia processing), it would seem that in terms of output per dollar of capital cost the Big Dish would be preferable.

            Exploring a global renewable energy budget.

We are now in a position to explore whether or not it would be possi ble to afford the quantity of plant needed for solar thermal systems to underwrite a global renewable energy world. 

The following derivation is a simplified indicative version of the form developed in Trainer (2010a).  World primary energy demand is likely to be at least 1000 EJ/y by 2050, and final energy demand in 2050 is likely to be 700 EJ/y (Moriarty and Honnery, 2009, p. 31.)  Let us assume that electricity remains at 25% of final demand, transport remains 33%, and that 60% of transport energy (i.e., 20% of final energy) can be provided via electricity.  This would mean 314 EJ/y of electricity would be required.

Because of the problems set by intermittency wind is not likely to be able to provide more than 25% of electricity demand (Lenzen, 2009) except in unusually favourable regions, and PV is not likely to provide more than about 30%.  If solar thermal systems were to provide the remaining 45% of the 314 EJ/y electricity demand, i.e., 141 EJ/y, their monthly supply task would be 11.8 EJ.

To meet this demand in winter via dishes delivering at distance 252 W/m2, i.e., 24.71.8  GJ/dish/month, 478516 million dishes would be needed.  At an estimated future cost of $(A2010)240,000163,000 per dish the total cost would be $ 11584 trillion, or $4.83.4 trillion p.a. assuming a 25 year plant lifetime.  This is more than 117.5 times the early 2000spresent total world annual investment in energy of all forms.  (Birol, 2003, IEA, 20103.).  (Adding the O and M and Ammonia processing costs would increase the figure significantly.)

The magnitude of the sum is indicated by other available figures.  For instance the Executive Summary for the 2010 World Energy Outlook of the International Energy Authority states that renewable electricity investment should rise dramatically in future, to a cumulative $ 5.7 trillion between 2010 and 2035, but this 25 year total, $28 billion p.a., corresponds to only 5% of an annual $4.8 trillion.  Pfuger (2010) reports energy investment at .5% of OECD GDP, and thus in the region of $200 billion p.a., or only 4.5% of $4.8 trillion.

Note that this would be the investment cost for supplying 45% of the 55% of total energy demand that is electrical (when 60% of transport is included), i.e., only 25% of total energy demand. Wind and PV have been assumed to supply the rest of the electricity required, leaving another 45% of total energy to be provided in non-electrical form.  However almost all renewable energy sources only provide electricity, biomass being the major exception. 

If highly optimistic assumptions are made re future global biomass production,(e.g., 1.5 billion ha of biomass energy plantations harvested at 6 tonne/ha yielding ethanol at 7 GJ/t net, Fulton, 2005) biomass might provide 64 EJ/y in the form of ethanol, i.e., 9% of the total world final energy assumed above to be required in 2050.  Let us assume that this meets the remaining 40% of transport energy demand.  If it is assumed that another 10% of total energy demand can easily be supplied as low-temperature heat via solar devices (which is not plausible for Europe or Northern America in winter), there would still be a shortfall of 24% of the 700 EJ/y total energy demand required, i.e., 168 EJ/y, and this would be required in non-electrical form.

If this 168 EJ/y was to be supplied in the form of hydrogen generated by additional wind, PV and solar thermal devices these would have to generate at least 336 EJ/y, given that the general energy efficiency of the hydrogen path would be less thanat best be 50%.  Allocating one-third of this task to solar thermal, i.e., 9.3 EJ/month, would almost double the required number of dishes estimated above.  The task for wind would then be the above 25% of the 314 EJ/y needed for direct electricity and plus electric transport, plus one-third of the 336 EJ/y remainder task, i.e., a total of 190 EJ/y.  This would require around 3150 times the mid-2000s total world installed wind capacity (Coppin, 2008), setting formidable problems to do withre finding suitable sites, and transmission losses.  (Trieb, 2004, estimates total potential European on and offshore sites as enabling only 4 EJ/y.)

The possible contributions of hydroelectricity, nuclear energy and coal with carbon capture and storage have not been included here, to enable a simplified indication of the magnitude of the problem.

If the efficiency of the ammonia path is not affected by the reduced DNI in winter as trough, dish-Stirling and dish-steam systems are, then the task for solar thermal above would require 350 million dishes, not 494 and their cost would be 5 times present world annual total energy investment, not 7 times.

The magnitude of the task might be reduced by assuming maximum conversion of the economy to electricity, thereby minimising the use of hydrogen.  The lack of detailed data on the uses, temperatures and forms of energy (noted by Ayres, 2009, p. 96) makes it difficult to estimate how many functions could be converted.

            The intermittency problem.

As has been noted the significant merit of solar thermal technologies is their capacity to store energy as heat and therefore to be able to generate when there is no sunshine.  The intent in current designs is to build in c. 12 - 16 hour storage to enable 24 hour operation.  This sets two questions, firstly whether storage capacity can enable constant supply from a solar thermal plant despite the intermittency of solar radiation, and secondly whether this capacity can enable solar thermal systems to overcome the gaps left in renewable energy supply by the variability of wind, sun and seasons.  In other words, can they provide the very large scale heat storage that a totally renewable supply system would need in order to maintain electricity supply for several calm and cloudy days in a row?

Focusing on mean DNI levels as in the above discussion can be misleading.  What matters is the variation of DNI around the winter monthly mean, and the minimum that can occur, as distinct from the average winter monthly value.is variation in DNI about the mean, and therefore the minimum levels that will occur, not the average levels.  For instance at the best Australian sites winter DNI averages around 5.7 kWh/m2/d but in particular months it can be 40% below this value. (See ASRDHB 2006, RREDC undated, Kaneff and Hagen 1991. NASA climate data also provide minima as well as averages.)  A solar thermal system intended to guarantee constant supply would have to be big enough to cope with the periods of minimum DNI, just as a conventional power system has to be capable of meeting periods of peak demand.  For conventional coal/nuclear systems this typically requires construction of 30% toup to 50% more plant than would meet average demand.   For solar thermal systems this would require corresponding increase in storage provision. 

In other words despite their capacity to store energy solar thermal systems also have to deal with significant problems of intermittency and integration and these are obscured if DNI averages are attended to.  Lenzen’s review (2009) concluded that the integration problems set by the intermittency of wind limit its contribution to supplying c. 20+% of electricity demand.  There does not seem to be any evidence on the limit to the contribution which a solar thermal component could make in view of theits intermittency problem set for it by typical variations in climatic conditions.

The following information on the variability of solar radiation in winter at ideal sites shows that the problem of intermittency for solar thermal systems is likely to be significant.  Kaneff (1991) reports that in Central Australia there are months when insolation averages 30% below the mean for that month.  For instance the White Cliffs monthly average DNI for June was 3.87 kWh/m/day, but in some years it can be 2.67 kWh/m/day, (compared with the summer month average of 8.2 kWh/m/day.)  NASA radiation source shows that at Mildura winter DNI can be 25%  below the mean.

Data from the US Dish site at Daggett (Davenport, 2008) show that it is an excellent site as on a peak day DNI can total over 11 kWh/m2/d.  However in winter months considerable gaps occur.  The 10 lowest days in the December reported averaged only 2.45 kWh/m2/d with a total of only 5 hours in 10 these days over 700 W/m2.  For January the average for the 12 lowest days was 3.4 kWh/m2/d with a total of only 6 hours over 700 W/m2.  For February the average for the 9 lowest days was 2.1 kWh/m2/d with only 4 hours over 700 W/m2 in these 9 days.  The sequences of consecutive days of low output were as follows. In December there was one run of 2 days, and another of 5 days.  In January there was a run of 4 days and another of 5 days, separated by 2 days.  In February there was a 4 day run in which there was a total of only 4 hours over 700 W/m2.

The DNI data for a year at Phoenix Arizona are even less favourable than for Daggett.  In January the 13 lowest days had an average 2.35 kWh/m2 radiation and these days included only 5 hours over 700 W/m2.  There was one 5 day run of low radiation and twice there were 2 day runs.  In February there were 6 poor days, averaging 1.76 kWh/m2/d and including only 5 hours over 700 W/m2, and a run of 4 days of low radiation. In December there were 10 days averaging 1.5 kWh/m2/d, including 4 hours with radiation over 700 W/m2, 2 runs of 2 consecutive low days and one sequence of 3 days.  At the Mod 2 site over a 19 day period there was output on only 2 days, totalling 25 kWh, less than 2% of the level that peak output would have been for that period.

Australian climate data aligns with the above data from the U.S.  At Alice Springs each of the three winter months averages 5 to 8 “cloudy” days and only 17.1 “clear” days. (Bureau of Meteorology website.)  NASA radiation data shows 21% cloud cover for central Australia in July.  The ASRDHB provides tables on the probability of sequences of cloudy days at Australian sites. (Table 7, for each location.)  At Alice Springs the probability of a 5, 7 or 9 day run in which average daily global radiation in winter is under 4.86 kWh/m2/d (global radiation) is 100% in all cases.  In June there is a 30% chance of a 4 day run averaging under 3.75 kWh/m2, and in each of the 4 winter months there is a 25% chance of a 4 day run averaging under 3.75 kWh/m2/d.  (The significance of these low values should be judged in relation to the above evidence that in Central Australia winter DNI averages around 5.7 kWh/m2/d and yet barely exceeds 700W/m2, a level that reduces dish-Sstirling output to 50% of peak.)

Kaneff (1991, Ttable XV-A) reports the following runs of “completely cloudy” days for three of the years in the 1980s when the White Cliffs (Central Australia) project was under way (each figures stands for the number of days in athe month).  Year one, 7, 17,10.  Year two, 6, 11, 13, 8, 9, 11, 7.  Year three, 7, 10, 7, 11, 6, 7.  In other words in every one of 7 months in one year at least 6 of the days were completely cloudy, and in all years there were at least two periods when there was complete cloud for one-third or more of the month.  In two of the years there was no sun for about half the three winter months.  To put it mildly, sequences like these would set a formidable challenge for any expectation that solar thermal storage capacity could overcome variability problems.  (Kenaff and Hagen state that Alice Springs is a more favourable site than White Cliffs although both are located in Central Australia.)  Bureau of Meteorology data for Alice Springs shows that in the winter of 2006 there was a four day run of cloud, followed by two reasonably sunny days and another run of four days cloud.

NASA climate data for North Africa similarly indicate considerable cloud cover in winter.  In May Algeria has cloud 37% of the time.

As has been explained these figures for days on which DNI is low will probably be misleadingly high indicators of electricity generation on those days, because much of the radiation will have been at levels too low for generation, or will have accumulated in several short periods between cloud.  In the latter case the warm up delays could conceivably determine that little or no electricity is generated during a day when clouds come and go.

Note also that if storage is going to maintain supply through lengthy cloudy periods, then not only must it be correspondingly large but so must heat producinggenerating capacity in order to fill storage in the reduced non-cloudy proportion of the winter month that is available for generating.

To summarise, despite their storage capacity there remains a significant problem of intermittency and solar thermal systems suffer a kind of “peak supply” problem similar to that experienced with coal-fired plant; i.e., the need to build a large amount of generating or storing capacity that will only be drawn on rarely.  This means that the estimation of total system costs must not be based on the cost of the number of solar thermal plants corresponding to average demand but on the number of plants and the amount of storage needed to meet demand from storage through periods when radiation is minimal.

This is why “levelised cost” calculations can be misleading.  They take in onlyindicate the amount of plant needed to meet demand under average generating conditions, whereas the amount actually needed is that which will meet demand under the least favourable conditions.

These considerations would have a marked effect on the number of dishes and the investment costs arrived at above, as that exercise only estimated numbers to meet demand under average DNI conditions, and assumed that no need for surplus capacity. (whereas even for coal-fired plant which suffers no intermittency problem of intermittency in the availability of fuel, plant required is 50% greater than the amount that would meet average demand.)

Some solar thermal systems currently operating have the capacity to store for up to 7.5 hours.  The standard provision in future is expected to be 17 hour storage and this is being built into some systems under construction.  However the above radiation data sources show that runs of four days in a row with little or no sunshine are not uncommon in Central Australia.   NASA says that medical stand- alone solar energy supply systems assume a need to cope with 6 consecutive “black days.”  If electricity from a 1000 MW(e) solar thermal plant was to be despatched from stored heat for 4 cloudy days, some 28350,000 MWh of heat would have to be stored and storage capacity would have to be 10.53 times that built into plants presently operating.

Sargent and Lundy (2003, 4 – 11) expect that by 2020 storage capacity cost will have fallen by 60% of the present cost to be $11.7/kWh(th), and 23% of plant cost.  Although this would correspond to around $(A2010)21/kWh(th), it is a quite low estimate in comparison with some others. Foran (2009) estimates storage embodied energy cost at 1/7 plant cost (apparently using Lenzen’s 1999 data). The NEEDS report states storage cost at Andersol to be e115/kWh(e), in the region of three times as high as the figure from Sargent and Lundy.  The NEEDS estimate would put the cost of storing 280,000 MWH(th) in the vicinity of e9,660 million.  Again the variation in the estimates prohibits confident conclusions but if the NEEDS figure is valid the four day storage provision would cost more than half as much as a Big Dish-Aamonia system capable of delivering 1000 MW in winter.  If Foran’s figure is valid the storage capacity would cost almost twice as much as the plant.

 Sargent and Lundy (200, 4 – 11)expect that storage capacity in 2020 will be $11.7/kWh(th, and 23% of plant cost.  The first figure indicates that sufficient storage to provide four days output at 1000 MW would cost twice as much as a 1000 MW coal fired plant.  (Note that their cost figure is 40% of the present cost they state, and again there is reason to be cautious a bout such predictions.)

Heat losses from the present 5 -7.5 hr storage systems are low, around 1% per day, but if 13 times the amount of heat must be stored for 13 times as long losses would be much greater.  (However the surface/volume ratio of tanks decreases with volume, and storing via ammonia would not involve significant heat losses.) 

            The total energy supply system integration problem.

Some have claimed that the capacity of solar thermal systems to store heat will enable this sector to resolve the problems set by the intermittency of wind and PV sectors.  For this to be so solar thermal systems would not only need to have sufficient storage to enable a constant flow of electricity from the solar thermal sector through several cloudy winter days, but it would also need the capacity to compensate for gaps in wind and PV supply.  Following is an indication of the magnitude of the second task.

Let us assume that wind, PV and Solar thermal each take 1/3 of the load involved in meeting a probable 2050 Australian electricity demand of 1400 PJ, i.e., c. 15 GW each out of a total 44 GW average demand.  There would be times when the first two were providing little electricity, such as on calm nights.  If solar thermal plant was to meet total demand through a 4 day calm and cloudy period, it would have to supply almost all 44 GW for that period.  In other words the solar thermal system would need the storage capacity to provide upwards of 3 times as much power as usual, for 133 times as long as is presently provided in solar thermal power stations.  That is each solar thermal plant would need approximately 450 times the storage capacity currently being built.  If in addition 60% of transport is to be run on electricity the storage task would be doubled.  (On the reasons why pumped storage and vehicle batteries are not likely make a significant difference to the general electricity storage problem see Trainer, 2008.)  There would be times when a greater than 4 day problem would arise. 

It is therefore most unlikely that a solar thermal component of a renewable energy supply system could compensate for much of the shortfall in the variable supply from other renewable sources.

Conclusions.

Because of the inaccessiunavailability of basic performance data this analysis has not been able to come to precise or confident conclusions.   The most promising option would seem to be using Big Dishes with an Ammonia storage capacity.  However the evidence accessed suggests that although solar thermal systems will be valuable contributors they will not be able to make a large contribution in winter to meeting commonly predicted demand levels in winter, let alone to solve the problem set by the variability of other renewable sources.

The most promising option would seem to be using Big Dishes with an ammonia storage capacity.  The cost of sufficient collection and generation capacity to deliver 1000 MW at distance, ignoring many relevant factors including the cost of the ammonia processing plant, would seem to be around $(A)18 billion.  The few and uncertain figures available indicate that providing storage capacity to maintain output over four cloudy days might add e1 – 3 billion per plant.

This discussion aligns with previous attempts to assess the potential and the limits of renewable energy sources.  (Trainer, 2007, 2008, 2010a, 2010b.)  These conclude that no combination of nuclear, hydro, geo-sequestration and renewable energy sources can sustain energy-intensive societies within safe greenhouse gas emission levels.  The least certain element in those discussions has been the assessment of the potential of solar thermal systems.  The present discussion comes to somewhat more confident conclusions than the previous analyses, indicating that solar thermal systems will not be capable of enabling a wholly renewable energy world energy supply.

would seem to indicate that they will not be capable of making the difference. 

If this conclusion is sound then global sustainability problems, centring especially on greenhouse emissions and peak oil, cannot be solved on the supply side.  Only dramatic reductions in demand can solve them, that is unprecedented transition from consumer-capitalist society to ways which enable a satisfactory quality of life on very low per capita resource and energy flows.  At present the failure to consider this path is supported by the common belief that renewable energy sources will obviate the need to take it seriously.  The existence, viability and attractiveness of the radically alternative path required is argued in Trainer 2010b, and inat The Simpler Way website, Trainer 2006.)

ABARE  http://www.abare,gov.au/publications_html/energy/energy10/energy_proj.pdf

 ABARE, (2010),  List of major electricity generation projects.  April.

http://www.abare.gov.au/publications_html/energy/energy_10/EG10_AprListing.xls

Alpert, J. L., and G. Kolb, (1988),  Performance of the Solar One Power Plant As  Simulated  by the SOLENERGY Computer Code, Sandia National Laboratoreis, Alberquerque.

Australian Solar Radiation Data Handbook, (ARDHB, 2006), ANZ Solar Energy Society, April, Energy Partners.

Ayres, R. U., (2009), The Economic Growth Engine, Cheltenham, Elgar.

Birol, F., (2003), “World energy investment outlook to 2030”, IEA, Exploration and Production:  The Oil & Gas Review, Volume 2.

Blanco, J., (2010), Head of Environmental Applications of Solar Energy, Platforma Sollar de, Almera, Spain.  (Personal communication.)

Bockamp, S., T. Gristop, M. Fruth, and M. Ewert, (2003), Solar thermal power generation, http://solarec-egypt.com/resources/Solar+Thermal+PowerGen-2003.pdf

Bossel, U., (2004), “The hydrogen illusion; why electrons are a better energy carrier”, Cogeneration and On-Site Power Production, March – April, pp. 55 – 59.

Breyer, C and G. Knies, (2009), “Global energy supply of concentrating solar power”, Proceedings of  Solar PACES, Berkeley, Sept, pp. 15 – 18.

Brockman, G. and D. Kearney, (2002), “The status and prospects of concentrated solar power technology”, International Conference on Expanding the Marked for Concentrating Solar Power, Berlin, 19 – 20 June.

Broesolme, H., H. Mannstein, C. Schillings and F. Trieb, (undated), Assessment of solar electric potential in north Africa  based on satellite data and a geographic information system, (Duplicated manuscript.)

Burbridge, D, D. Mills and G,. Morrison (undated) “The Stanwell solar thermal project”, http://www.solarpaces.org/CSP-Technology/docs/solar_dish.pdf

Coppin, P., (2008), “Wind energy”, in P. Newman, Ed., Transitions, CSIRO Publishing, Canberra.

Czisch, G., 2001, Global Renewable energy potential; approaches to its use, http://www.iset.uni-kassel.de/abt/w3-w/folien/magdeb0030901/

Czisch, G., (2004), Least-cost European/Transeuropean electricity supply entirely with renewable energies, www.iset.uni-kassel.de/abt/w3-w/project/Eur-Transeur-El-Sup.pdf

Davenport, R., (2008), Personal communications.

Davenport, R., et al., (undated), Operation of second-generation dish/Stirling power systems,  Science Applications International, Corp, San Diego.

Dey, C., and M. Lenzen, (1999), Greenhouse gas analysis of electricity generating systems, ANZSES, Solar 2000 Conference, University of Queensland, 29^th Nov. – 1st Dec., Conference Proceedings, pp. 658 – 668.

Energylan, Undated, “Overview of Solar Thermal Technologies”,  www.energylan.sandia.gov/sunlab/PDFs/solar-overview.pdf

EPRI (2009), Program on technology innovation; Integrated generation technology options; Technical Update, Nov.  http://my.epri.com/portal/server.pt?Product_id=000000000001019539

Foran, B., (2009), Solar Thermal Chapter,  Powerful Choices; The Transition to a Low Carbon Economy, http//:lwa.gov.au/products/pn30178

Fulton, L., (2005), Biofuels For Transport; An International Perspective, International Energy Agency. (No source.)

Grasse, W., and M. Geyer, (2000), “Solar power and chemical engineering systems”, Solar Paces Annual Report, IEA.

Hagen, D., and S. Kaneff, (1991), Application of Solar Thermal Technology in Reducing Greenhouse Gas Emissions, Canberra, Australian Government Department of Arts, Sport, Environment, Tourism and Territories.

Hayden, H. C., (2004), The Solar Fraud, Pueblo West, Co, Vales Lake Publishing.

Heller, P., (2010), Personal communication.

Herendeen, R. A., (1988), “Net energy considerations”, in Economic Analysis of Solar Energy Systems, M. I.T. Press, Cambridge.

IEEE Power Engineering Review, SEGS, (1989), “Solar electric generating stations,” Aug., http://ieeexplore.ieee.org/iel5/39/2848/00087383.pdf?arnumber=87383

Jones, S., Pitz-Paal, A. R., Blair, N. and Cable, R., (2001), “TRNSYS modelling of the SEGS VI parabolic trough solar electric generating system”, Proceedings of Solar Forum 2001; Solar Energy; The Power to Choose, April 21 - 25, Washington.

Kaneff, S., (1991), Solar Thermal Process Heat and Electricity Generation Performance and Costs for the ANU Big Dish Technology, Energy Research Centre, Research School of Physical and Sciences and Engineering, Australian National University, Canberra.

Kaneff, S., and D. Hagen, (1991), Application of Solar Thermal Technologies in Reducing Greehouse Gas Emissions, Department of Arts, Sport, Environment, Journalism and Territories, Federal Government, Canberra.

Kaneff, S, (1992),  Mass Utilization of Thermal Energy, Canberra, Energy Research Centre.

Kaneff, S., and D. Hagen, (1991), Application of Solar Thermal Technologies in Reducing Greehouse Gas Emissions, Department of Arts, Sport, Environment, Journalism and Territories, Federal Government, Canberra.

Kolb, G., (2010), Sandia Solar Thermal Research, NREL, Personal communication.

Kreith, F., P. Norton and D. Brown, (1990), “A comparison of CO2 emissions from fossil and from solar power plants in the United States”, Energy, 15 (12),  1181 – 1198. 

Kreith, F., and D. Yogi Goswami, Eds., (2007), Handbook of Energy Efficiency and Renewable Energy, Taylor & Francis, London.

Kreetz, H. and K. Lovegrove, (2002), “Exegy analysis of an ammonia synthesis reactor in a solar thermochemical power system”, Solar Energy, 73, 3, Sept, 187 – 194.

Lechon, Y., C. De la Rua and R. Saez, (2006), Life cycle environmental impacts of electricity production by solar thermal technology in Spain,  SolarPACES2006, B5-S5.

Lenzen, M., (1999) “Greenhouse gas analysis of solar-thermal electricity generation”, Solar Energy, 65, 6, pp. 353 – 368.

Lenzen, M., (2009), Current state of development of electricity-generating technologies – a literature review. Integrated Life Cycle Analysis, Dept. of Physics, University of Sydney.

Lenzen, M., (1999), “Greenhouse gas analysis of solar-thermal electricity generation”, Solar Energy, 65, 6, pp. 353 – 368.

Lenzen, M., and C. Dey, (2000), “Truncation error in embodied energy analysis of basic iron and steel products”, Energy, 25, pp;. 577 - 585.

Lenzen, M., C. Dey, C. Hardy and M. Bilek, (2006) Life-Cycle Energy Balance and Greenhouse Gas Emissions of Nuclear Energy in Australia. Report to the Prime Minister's Uranium Mining, Processing and Nuclear Energy Review (UMPNER), Internet site http://www.isa.org.usyd.edu.au/publications/documents/ISA_Nuclear_Report.pdf, Sydney, Australia, ISA, University of Sydney.

Lenzen, M.,  and J. Munksgaard, (2002), “Energy and CO2 life-cycle analysis of wind turbines – review and applications.”, Renwewable Energy,  26, 339 -362.

Lenzen, M and G. Treloar, (2003), “Differential convergence of life-cycle inventories toward upstream production layers, implications for life-cycle assessment”, Journal of Industrial Ecology,  6, 3-4.

Lovegrove, K., A. Luzzi, I. Solidiani and H. Kreetz, (2004), “Developing ammonia based thermochemical energy storage for dish power plants”, Solar Energy, 76, 1 – 3, Jan-Mar., 331 – 337.

Lovegrove, K., A. Zawadski,  and J. Coventy, (2006), “Taking the ANU Big Dish to commercialization”, Proceedings of ANZSES Annual Conference, Solar 2006.

Lovegrove, K., A. Zawadski,  and J. Coventy, (2007), “Paraboidal Solar Dish Concentrators for Multi-Megawatt Power Generation”, Solar World Congress, Beijing, Sept. 18 – 22.

Luzzi, A. C., (2000), “Showcase project; 2 MWe solar thermal demonstration power plant, Proceedings of the 10^TH Solar PACES Int. Symposium on Solar Thermal Concentrating Technologies, Sydney.

Mackay, D., (2008), Sustainable Energy – Without the Hot Air, Cavendish  Laboratory. http://www.withouthotair.com/download.html

Mancini, T., (2007), Personal Communications.

Mancini, T., P.Heller, B. Butler, B. Osborn, (2003), “Dish-Stirling systems; An overview of development and status”, Journal of Solar Energy Engineering, 125, 2, May, pp 135 – 151.

Meteonorm, (2008), Maps of Global Horizontal Radiation and Temperature, http://www.meteonorm.com/pages/en/downloads/maps.php

Mills, D., G. Morrison and P. LeLeivre, (2003), “Multitower Line Focus Fresnel Array”, Proceedings of ISEG, Mona Kai Resort, Hawaii.

Mills, D, P. Le Lievre,  and G. L. Morrison, (2004), “Lower temperature approach for very large solar power plants,”  www.solarheatpower.com.

Morrison, G., and Litwak, A., (1988), Condensed Solar Radiation Data Base for Australia, Paper 1988/FMT/1 March.

Moriarty, P., Honery, D., (2009),  “What energy levels can the earth sustain?”. Energy Policy, 37, 2469 – 2472.

NEEDS, (New Energy Externalities Developments in Sustainability), (2008), Final Report concentrating Solar Power Plants. http://needs-poroject.org/RSIa/RSIa.pdf

Nicholson, M and P. Lang, (2010), Zero Carbon Australia – A Strategic Energy Plan: A Critique. http://bravenewclimate.com/2010/08/12/zca2020-critique/

NREL, (2005), Report – Concentrating Solar Power, http://renewablesg.org/docs/Web/AppendixE#.pdf

NREL, (2010), Spolar Advisor Model (SAM),

Norton, B., (1999), “Renewable energy – What is the true cost?”, Power Engineering Journal, Feb, 6 – 12.

Odeh, S.D., Behnia, M. and Morrison, G.L., (2003), “Performance Evaluation of Solar Thermal Electric Generation Systems”, Energy Conversion and Management, V44, 2425-2443.

Radosevich, L., (1988), Final Report on the Power Production Pase of the 10 MWe Solar Thermal Concentrating Receiver Pilot Plant, SAND 87 – 8022, NREL, Sandia National Laboratories, Alberquerque.

RREDC  (undated), website, (http://rredc.nrel/gov/solar/old_data/nsrdb/redbook/atlas/serve.cgi

Sargent and Lundy, 2003, Assessment of Parabolic Trough and Power Tower Solar Technology Cost and Performance Forecasts, NREL.

Siangsukone, P., and K. Lovegrove, (2003), “Modelling of a 400m2 steam based paraboidal dish concentrator for solar therm al power production”, ANZSES 2003 Destination Renewables, Nov.

SEGS, (1989), “Solar electric generating stations,” IEEE Power Engineering Review, Aug., http://ieeexplore.ieee.org/iel5/39/2848/00087383.pdf?arnumber=87383

NREL, (2010), Solar Advisor Model 9SAM),

SolarPaces, (undated), “Solar Parabolic Trough”, http”//www.Solar Paces.org/solar_trough.pdf

Taylor, R., R. Davenport, and A. T-Raissi, (2008), “Solar concentrator options for a thermochemical hydrogen production process”, American Solar Energy Society Solar 2008 Conference, San Diego, CA, 4-8 May.

Trainer, T., (2006), The Simpler Way website, http://ssis.arts.unsw.edu.au/tsw/

Trainer, T., (2007), Renew able Energy Cannot Sustain Consumer Society, Dordrecht, Springer.

Trainer, T.,  (2008),  “Renewable energy  -- Cannot sustain an energy-intensive society”, http://ssis.arts.unsw.edu.au/tsw/REcant.html

Trainer, T., (2010a), “Can renewables etc. solve the greenhouse problem? The negative case”, Energy Policy, 38, 8, August, 4107 - 4114. http://dx.doi.org/10.1016/j.enpol.2010.03.037

Trainer, T., (2010b), The Transition to a Sustainable and Just World, Sydney, Envirobook.

Trieb, F., (undated),  Trans-Mediterranean  Interconnection for  Concentrating Solar Power; Final Report, German Aerospace Center (DLR), Institute of Technical Thermodynamics, Section Systems Analysis and Technology Assessment.

Ummel, K and D. Wheeler, (2008), “Desert power; The economics of solar thermal electricity for Europe, North Africa and the Middle East”, Centre for Global Development, Dec.

USGS, (2010), Thermo-electric power water use,  Water science for Schools, http://ga.water.usgs.gov/edu/wupt.html

Van Voorthuysen, E. H., (2006), “Large Scale Concentrating Solar Power (CSP) Technology; Solar Electricity For The Whole World”, in V. Badescu, R. B. Cathcart and R. D. Schuling, Macro Engineering, Dordrecht, Springer, 2006.

Vant-Hull, L. L, (1991), “Solar thermal electricity – An environmentally benign and viable alternative,” in Proceedings of the World Clean Energy Conference”, Geneva, 4-7  Nov.,, Cercle Mondial, POB 928 Zurich.

Vant-Hull, L.L.. (1992 – 1993), “Solar thermal electricity, an environmentally benign and viable alternative”.  Perspectives in Energy, Volume 2, pp. 157 – 166.

Vant-Hull, L. L., (2006), “Energy return on investment or solar thermal plants”, Solar Today, May/June, pp. 13 – 16.

Viebahn, P., S. Kronshage and F. Trieb, (2004), New Energy Externalities Developments for Sustainability, (NEEDS), Project  502687.

Weinrebe, G, M. Bonhke, and  F. Trieb, (2008), “Life cycle assessment of an 80 MW SEGS plant and a 30 MW Phoebus power tower”, in Proceedings, Solar 98: Renewable Energy For The Americas, ASME International Solar Energy Conference, Alberquerque NM, 13 – 18 June.

____________

Where NEEDS says a site must haved 2000/y

ABARE  http://www.abare,gov.au/publications_html/energy/energy10/energy_proj.pdf

Nicholson, M and P. Lang, (2010), Zero Carbon Australia – A Strategic Energy Plan: A Critique.

Oswald, Coelingh, Sharman, Davey and Coppin

 ABARE, (2010),  List of major electricity generation projects.  – April. 2010

http://www.abare.gov.au/publications_html/energy/energy_10/EG10_AprListing.xls

Alpert, J. L., and G. Kolb, (1988),  Performance of the Solar One Power Plant As  Simulated  by the SOLENERGY Computer Code, Sandia National Laboratoreis, Alberquerque.

Australian Solar Radiation Data Handbook, (ARDHB, 2006), ANZ Solar Energy Society, April, Energy Partners.

Birol, F., (2003), “World energy investment outlook to 2030”, IEA, Exploration and Production:  The Oil & Gas Review, Volume 2.

Blanco, J., 2010, Head of Environmental Applications of Solar Energy, Platforma Sollar de, Almera, Spain.  Personal communication.

Bockamp, S., T. Gristop, M. Fruth, and M. Ewert, (2003), Solar thermal power generation, http://solarec-egypt.com/resources/Solar+Thermal+PowerGen-2003.pdf

Bossel, U., (2004), “'The hydrogen illusion; why electrons are a better energy carrier”', Cogeneration and On-Site Power Production, March – April, pp. 55 – 59.

 Breyer, C and G. Knies, (2009), “‘Global energy supply of concentrating solar power”, Proceedings of  Solar PACES, Berkeley, Sept, pp. 15 – 18.

Brockman, G. and D. Kearney, (2002), “The status and prospects of concentrated solar power technology”, International Conference on Expanding the Marked for Concentrating Solar Power, Berlin, 19 – 20 June.

Broesolme, H., H. Mannstein, C. Schillings and F. Trieb, (undated), Assessment of solar electric potential in north Africa  based on satellite data and a geographic information system, (Duplicated manuscript.)

Burbridge, D, D. Mills and G,. Morrison (undated) “The Stanwell solar thermal project”, http://www.solarpaces.org/CSP-Technology/docs/solar_dish.pdf

Coppin,

Czisch, G., 2001, Global Renewable energy potential; approaches to its use, http://www.iset.uni-kassel.de/abt/w3-w/folien/magdeb0030901/

Czisch, G., (2004), Least-cost European/Transeuropean electricity supply entirely with renewable energies, www.iset.uni-kassel.de/abt/w3-w/project/Eur-Transeur-El-Sup.pdf

Energylan, Undated, “Overview of Solar Thermal Technologies”,  www.energylan.sandia.gov/sunlab/PDFs/solar-overview.pdf

Davenport, R.,  (2008), Personal communications.

Davenport, R., et al., (undated), Operation of second-generation dish/Stirling power systems,  Science Applications International, Corp, San Diego.

Dey, C., and M. Lenzen, (1999), Greenhouse gas analysis of electricity generating systems, ANZSES, Solar 2000 Conference, University of Queensland, 29^th Nov. – 1st Dec., Conference Proceedings, pp. 658 – 668.

Energylan, Undated, “Overview of Solar Thermal Technologies”,  www.energylan.sandia.gov/sunlab/PDFs/solar-overview.pdf

EPRI (2009), Program on technology innovation; Integrated generation technology options; Technical Update, Nov.  http://my.epri.com/portal/server.pt?Product_id=000000000001019539

Foran, B., (2009),

Fulton, L., (2005), Biofuels For Transport; An International Perspective, International Energy Agency. (No source.)

Grasse, W., and M. Geyer, (2000), “Solar power and chemical engineering systems”, Solar Paces Annual Report, IEA.

Hagen, D., and S. Kaneff, (1991), Application of Solar Thermal Technology in Reducing Greenhouse Gas Emissions, Canberra, Australian Government Department of Arts, Sport, Environment, Tourism and Territories.

Hayden, H. C., (2004), The Solar Fraud, Pueblo West, Co, Vales Lake Publishing.

Heller, Pp., (2010), Personal communication.

Herendeen, R. A., (1988), “Net energy considerations”, in Economic Analysis of Solar Energy Systems, M. I.T. Press, Cambridge.

Jones, S., Pitz-Paal, A. R., Blair, N. and Cable, R., (2001), “TRNSYS modelling of the SEGS VI parabolic trough solar electric generating system”, Proceedings of Solar Forum 2001; Solar Energy; The Power to Choose, April 21 - 25, Washington.

Kaneff,S., (1991), Solar Thermal Process Heat and Electricity Generation Performance and Costs for the  ANU Big Dish Technology, Research School of Physical Science and Engineering, Australian National University, Canberra.

Kaneff, S., and D. Hagen, (1991), Application of Solar Thermal Technologies in Reducing Greehouse Gas Emissions, Department of Arts, Sport, Environment, Journalism and Territories, Federal Government, Canberra.

Kaneff, S, (1992),  Mass Utilization of Thermal Energy, Canberra, Energy Research Centre.

Kaneff, S., and D. Hagen, (1991), Application of Solar Thermal Technologies in Reducing Greehouse Gas Emissions, Department of Arts, Sport, Environment, Journalism and Territories, Federal Government, Canberra.

Kearney, D., (1989), Solar Electric Concentrating Stations (SEGS), Editorial, IEEE Power Engineering Review. 

Kreith, F., P. Norton and D. Brown, (1990), “A comparison of CO2 emissions from fossil and from solar power plants in the United States”, Energy, 15 (12),  1181 – 1198. 

Kreith, F., and D. Yogi Goswami, Eds., (2007), Handbook of Energy Efficiency and Renewable Energy, Taylor & Francis, London.

Kreetz, H. and K. Lovegrove, (2002), “Exegy analysis of an ammonia synthesis reactor in a solar thermochemical power system”, Solar Energy, 73, 3, Sept, 187 – 194.

Lechon, Y., C. De la Rua and R. Saez, (2006), Life cycle environmental impacts of electricity production by solar thermal technology in Spain,  SolarPACES2006, B5-S5.

Lenzen, M., (1999) “Greenhouse gas analysis of solar-thermal electricity generation”, Solar Energy, 65, 6, pp. 353 – 368.

Lenzen, M.,  and J. Munksgaard, (2002), “Energy and CO2 life-cycle analysis of wind turbines – review and applications.”, Renwewable Energy,  26, 339 -362.

Lenzen, M., (2009), Current state of development of electricity-generating technologies – a literature review. Integrated Life Cycle Analysis, Dept. of Physics, University of Sydney.

Lenzen, M.., and C. Dey, (1999), “Greenhouse gas analysis of solar-thermal electricity generation”, Solar Energy, 65, 6, pp. 353 – 368.

Lenzen, M., and C. Dey, (2000), “Truncation error in embodied energy analysis of basic iron and steel products”, Energy, 25, pp;. 577 - 585.

Lenzen, M., C. Dey, C. Hardy and M. Bilek, (2006) Life-Cycle Energy Balance and Greenhouse Gas Emissions of Nuclear Energy in Australia. Report to the Prime Minister's Uranium Mining, Processing and Nuclear Energy Review (UMPNER), Internet site http://www.isa.org.usyd.edu.au/publications/documents/ISA_Nuclear_Report.pdf, Sydney, Australia, ISA, University of Sydney.

Lenzen, M and G. Treloar, (2003), “Differential convergence of life-cycle inventories toward upstream production layers, implications for life-cycle assessment”, Journal of Industrial Ecology,  6, 3-4.

Lovegrove, K., A., Luzzi, I. Solidiani and H. Kreetz, (2004), “Developing ammonia based thermochemical energy storage for dish power plants”, Solar Energy, 76, 1 – 3, Jan-Mar., 331 – 337.

Lovegrove, K., A. Zawadski,  and J. Coventy, (2006), “Taking the ANU Big Dish to commercialization”, Proceedings of ANZSES Aannual Cconference, Solar 2006.

Lovegrove, K., A. Zawadski,  and J. Coventy, (2007), “Paraboidal Solar Dish Concentrators for Multi-Megawatt Power Generation”, Solar World Congress, Beijing, Sept. 18 – 22.

Luzzi, A. C., (2000), “Showcase project; 2 MWe solar thermal demonstration power plant, Proceedings of the 10^TH Solar PACES Int. Symposium on Solar Thermal Concentrating Technologies, Sydney.

Mackay, D., (2008), Sustainable Energy – Without the Hot Air, Cavendish  Laboratory. http://www.withouthotair.com/download.html

Mancini, T., (2007), Personal Communications.

Mancini, T., P.Heller, B. Butler, B. Osborn, (2003), “Dish-Stirling systems; An overview of development and status”, Journal of Solar Energy Engineering, 125, 2, May, pp 135 – 151.

Meteonorm, (2008), Maps of Global Horizontal Radiation and Temperature, http://www.meteonorm.com/pages/en/downloads/maps.php

Mills, D., (undated), p2

Mills, D., G. Morrison and P. LeLeivre, (2003), “Multitower Line Focus Fresnel Array”, Proceedings of ISEG, Mona Kai Resort, Hawaii.

Mills, D, P. Le Lievre,  and G. L. Morrison, (2004), “Lower temperature approach for very large solar power plants,”  www.solarheatpower.com.

Morrison, G., and Litwak, A., (1988), Condensed Solar Radiation Data Base for Australia, Paper 1988/FMT/1 March.

Moriarty, P., Honery, D., (2009),  “What energy levels can the earth sustain?”. Energy Policy, 37, 2469 – 2472.

NEEDS, (New Energy Externalities Developments in Sustainability), (2008), Final Report concentrating Solar Power Plants. http://needs-poroject.org/RSIa/RSIa.pdf

Nicholson, M and P. Lang, (2010), Zero Carbon Australia – A Strategic Energy Plan: A Critique.

NREL, (2005), Report –Concentrating Solar Power, http://renewablesg.org/docs/Web/AppendixE#.pdf

NREL, (2010), Solar Advisor Model (SAM),

Norton, B., (1999), “Renewable energy – What is the true cost?”, Power Engineering Journal, Feb, 6 – 12.

Odeh, S.D., Behnia, M. and Morrison, G.L., (2003), “Performance Evaluation of Solar Thermal Electric Generation Systems”, Energy Conversion and Management V44, 2425-2443.

Radosevich, L. (1988), Final Report on the Power Production Pase of the 10 MWe Solar Thermal Concentrating Receiver Pilot Plant, SAND 87 – 8022, NREL, Sandia National Laboratories, Alberquerque.

RREDC  (undated), website, (http://rredc.nrel/gov/solar/old_data/nsrdb/redbook/atlas/serve.cgi

Sargent and Lundy, 2003, Assessment of Parabolic Trough and Power Tower Solar Technology Cost and Performance Forecasts, NREL.

Siangsukone, P., and K. Lovegrove, (2003), “Modelling of a 400m2 steam based paraboidal dish concentrator for solar therm al power production”, ANZSES 2003 Destination Renewables, Nov.

SEGS, (1989), “Solar electric generating stations,”  IEEE Power Engineering Review, Aug., http://ieeexplore.ieee.org/iel5/39/2848/00087383.pdf?arnumber=87383

SolarPaces, (undated), “Solar Parabolic Trough”, http”//www.Solar Paces.org/solar_trough.pdf

Taylor, R., R. Davenport, and A. T-Raissi, (2008), “Solar concentrator options for a thermochemical hydrogen production process”, American Solar Energy Society Solar 2008 Conference, San Diego, CA, 4-8 May.

Trainer, T., (2006), The Simpler Way website,    http://ssis.arts.unsw.edu.au/tsw/

Trainer, T., (2007), Renew able Energy Cannot Sustain Consumer Society, Dordrecht, Springer.

Trainer, F. E. (T.),  (2008),  “Renewable energy  -- Cannot sustain an energy-intensive society”, http://ssis.arts.unsw.edu.au/tsw/REcant.html

Trainer, T., (2010aa), “Can renewables etc. solve the greenhouse problem? The negative case”, Energy Policy, 38, 8, August, 4107 - 4114. http://dx.doi.org/10.1016/j.enpol.2010.03.037

Trainer, T., (2010a),

Trainer, T., (2010b), The Transition to a Sustainable and Just World, Sydney, Envirobook.

Trieb, F., (undated),  Trans-Mediterranean  Interconnection for  Concentrating Solar Power; Final Report, German Aerospace Center (DLR), Institute of Technical Thermodynamics, Section Systems Analysis and Technology Assessment.

Ummel, K and D. Wheeler, (2008), “Desert power; The economics of solar thermal electricity for Europe, North Africa and the Middle East”, Centre for Global Development, Dec.

USGS, (2010), Thermo-electric power water use,  Water science for Schools, http://ga.water.usgs.gov/edu/wupt.html

Van Voorthuysen, E. H., (2006), “Large Scale Concentrating Solar Power (CSP) Technology; Solar Electricity For The Whole World”, in V. Badescu, R. B. Cathcart and R. D. Schuling, Macro Engineering, Dordrecht, Springer, 2006.

Vant-Hull, L. L, (1991), “Solar thermal electricity – An environmentally benign and viable alternative,” in Proceedings of the World Clean Energy Conference”, Geneva, 4-7  Nov,, Cercle Mondial, POB 928 Zurich.

Vant-Hull, L.L.. (1992 – 1993), “Solar thermal electricity, an environmentally benign and viable alternative”.  Perspectives in Energy, , Volume 2, pp. 157 – 166.

Vant-Hull, L. L., (2006), “Energy return on investment or solar thermal plants”, Solar Today, May/June, pp. 13 – 16.

Viebahn, P., S. Kronshage and F. Trieb, (2004), New Energy Externalities Developments for Sustainability, (NEEDS), Project  502687.

Weinrebe, G, M. Bonhke, and  F. Trieb, (2008), “Life cycle assessment of an 80 MW SEGS plant and a 30 MW Phoebus power tower”, in Proceedings, Solar 98: Renewable Energy For The Americas, ASME International Solar Energy Conference, Alberquerque NM, 13 – 18 June.

The cosine problem.

Another uncertain reducing factor concerns the difference between summer and winter average angle between sun, reflectors and absorber. In winter troughs suffer marked effects due the geometry of their alignment with the sun throughout the day, causing the quite low performance of troughs in winter.  As the sun travels across the sky in summer its angle with the (N-S) axis of the trough is close to 90 degrees all day, at a good location, but in winter it is around 60 degrees, meaning that a square metre of collector receives half as much (cos 60) solar energy as the DNI value per square metre.  Dishes can be pointed directly at the sun all day in summer and winter, and they have only slight curvature, meaning that this “cosine loss” is much less.

With central receivers the situation is more subtle and difficult to estimate.  Radesovich (1988) states the average cosine loss for a central receiver as 21%.  A central receiver field resembles a fresnel arrangement of a dish but an inefficiency is created by the fact that when the sun is at a low angle to the horizon the high absorber is not at a point analogous to the focal point of a parabolic dish.  All central receivers built to data have been small compared with the standard 220 MW unit assumed for the future, and this cosine problem is considerably more acute when the field is large and the height of the tower is low relative to its radius.  For a 2+ km radius field and a 280 metre high tower in winter the angle between the sun, reflector and absorber is low to very low for all the mirrors on the sun’s side of the tower.  At Mildura in winter one would have to go almost a kilometre south of the tower to find a reflector normal to the sun, even at midday.  For almost all the reflectors north of this one the angle between sun, reflector and absorber would be low to very low.  The result is that the average angle between sun, reflector and absorber in summer when the sun is higher in the sky all day is somewhat greater than in winter when it is lower all day.  Note again that this is true of midday and in winter mornings and afternoons the sun is even lower above the horizon and most angles are bigger.  (NASA tables show that for Algeria the average hourly angle of the sun above the horizon in summer is 46 degrees but in winter only 27 degrees.)  Tables 3 – 6 from Alpert and Kolb (1988) sets out the differences and it is evident that when the angle goes from 90 degrees to 15 degrees the output of the field can fall by one-third.

My crude graphical estimation of reflector angles, weighted by numbers of reflectors at various positions in the field, indicates that given winter angles the energy received at the absorber would be 73% of the value for the summer angles (i.e., assuming the same DNI in each case.)  This roughly corresponds to the above figure from Radosevich, i.e., a 21% average loss due to the cosine effect.  But note that this is for mid day and at other times of the day the sun is lower and the average angle would be worse.

The little evidence to hand on the winter/summer ratio for Central Receivers reports some surprisingly large and confusing differences, and it is disappointing that a clear and confident conclusion is not possible on this crucial issue.  The cosine factor is probably the main contributor to the difference.  Radosevich (Table 4 - 3) shows that daily solar – electricity conversion in summer was a surprising 4 to 9 times as great as in winter.  Over the three years in which NREL experimented with a 10 MW pilot plant the ratio of mid-summer to mid-winter monthly gross output was 4.2/1, 7.6/1 and 7.2/1.  The ratios for net output were up to 25% lower (i.e., 10/1 in one year). (Alpert and Kolb, 1988,Table 5 - 1, p. 59.)  (Attempts to determine by personal correspondence whether these ratios were misleading or generally valid indicators, have not been successful.  Similarly it has not been possible to get confirmation that the SAM data has taken this factor into account or assumes the same cosine effect in summer and winter.)

Some light can be thrown on these ratios by Fig. 4 –1, p. 43 from Alpert and Kolb showing daily output in relation to DNI.  For the Solar 1 device, DNI had to reach almost 875 W/m2 before any power was generated, indicating that output is not proportional to DNI but increases rapidly only after a relatively high value is achieved.  In other words, in winter the lower DNI results in disproportionately lower electrical output. Alpert and Kolb report marked reduction in efficiency with low DNI (c. p. 21.)

It is noteworthy that in one of these receiver cases given in NREL’s SAM package (2010) a 36% fall in DNI is accompanied by a 58% fall in net energy produced, again testifying to a significantly falling solar-to electricity efficiency with falling DNI, and supporting the conclusion that the above simple derivation of a 40 W/m2 winter flow from the basic 220 MW central receiver characteristics could be too high.

   Other Central Receiver issues.