Vous êtes sur la page 1sur 6

LOWER TEMPERATURE APPROACH FOR

VERY LARGE SOLAR POWER PLANTS


D. Mills* , P. Le Lièvre**, and G.L. Morrison***

*School of Physics, University of Sydney, NSW Australia 2006.


E-mail: d.mills@physics.usyd.edu.au

**Solar Heat and Power Pty. Ltd. (SHP), 2 Chifley Square, Sydney NSW 2000, Australia
E-mail: peter@solarheatpower.com

***School of Mechanical and Manufacturing Engineering, University of New South Wales,


NSW Australia 2052
E-mail: g.morrison@unsw.edu.au

Abstract - Recently, low cost Compact Linear Fresnel Reflector (CLFR) solar thermal technology has
been developed and is being installed at a project in Australia to provide thermal energy for preheating
feed water in a coal fired generation station. The current array design generates saturated steam at
270°C, which heats feed water via a heat exchanger. The array can be upgraded for use as a boiler with
a conventional high temperature turbine by using lower emittance selective coatings and slower flow
rates. However, while a high temperature turbine is a more efficient conversion device than a low
temperature turbine, high temperature operation carries with it some disadvantages, namely increased
thermal losses from the array and a requirement for more stringent materials and component
specifications, with consequent array cost increases. This paper makes performance and cost
comparisons which suggest that the cost/performance benefits due to lower array operating temperature
may be more significant than the efficiency benefit from using a higher temperature turbine. A 100%
solar system concept of 240MW using thermal storage with 12 hour storage and commercially available
turbines with a 250°C inlet temperature is described for use at a particular Australian site.

1. Introduction

In this paper, the general design of a 240 MWe solar thermal electric power plant is discussed for a site in
inland Australia. The design uses the low cost Compact Linear Fresnel Reflector (CLFR) array system
previously reported conceptually [1] and later described more practically as being constructed for a coal
saver solar preheating project of 35 MWe [2,3], which has recently been increased to 40 MWe. While
preheating is a highly cost effective niche market because purchase of a turbine and new electrical
infrastructure is avoided, it is limited in absolute terms to a small percentage of the centralised electricity
market. However, the CLFR design is suited to all markets currently served by parabolic trough
technology, and this includes stand alone storage applications such as the Andasol 50 MWe trough plant
using molten salt storage [4].

Fig. 1. Australian Bureau of


Meteorology map showing two NSW
microclimates with greater than 9
sunshine hours average per day which
are suitable for large solar plants and
near to the grid. The two NSW sites
together are at least 80 times larger
than the ground area required for the
production of Australia’s entire
current electricity output.

> 9 hr/d
annual
In Australia, microclimates of high solar intensity have been identified near existing grid lines in New
South Wales and South Australia (Fig. 1). These areas are currently being investigated as future stand
alone solar power plant sites, and detailed solar radiation studies are underway at one site with the aim of a
240 MWe storage installation being installed later in the decade.

The authors have previously described some of the cost advantages of the CLFR array system [3] of the
current trough technology, but have not discussed the general issue of optimal stand alone overall plant
design. Because the collector cost is now very low, a considerable cost reduction in the system can be
effected by the use of thermal energy storage, which allows a smaller power block relative to a given
collector area. Our projections suggest that a capacity factor in excess of 50% will show competitive costs
in excellent sites such as inland NSW, the U.S. Southwest, and North Africa. In a 100% solar design, the
cost will remain close to that of wind power because
1) the excess cost of storage and power block is compensated by much increased beam insolation at
such sites (25-30% greater than coastal sites).
2) in a 100% solar system, the power block cost can be reduced through the complete avoidance of
fossil fuel equipment (chimneys, boiler, pollution abatement equipment, fuel handling equipment).

The most common approach to the design of a line focus solar plant is to use a parabolic trough system to
supply of heat at between 320°C and 400°C to both the main boiler and superheater of a conventional
turbogenerator. Future designs may circulate molten salt up to 500°C. Higher array temperatures facilitate
the use of more efficient high temperature turbines. However, in line focus systems, thermal losses can rise
rapidly with array operating temperature, partially cancelling out improvements in thermal conversion
efficiency. In addition, the traditional path of using a superheated turbine requires more highly efficient and
durable selective coatings, thicker-walled tubing for steam pressure containment, and the use of oil or
molten salt instead of water as an array heat transfer fluid if operating above the water triple point.

Some designs are hybridised with fossil fuel and utilise a situation where lower cost fossil is used in off-
solar hours, not only to increase the array capacity factor, but to lower the overall cost of delivered energy.
The improved turbine efficiency is presumed to outweigh the reduction in thermal array thermal efficiency
due to higher array operation temperatures and the higher cost of solar/fossil turbines. However, as array
costs drop the advantages of this approach are becoming less apparent, and incentives in some countries
such as Spain dictate the use of very high solar fractions which effectively necessitate the use of storage.

Recently, the National Renewable Energy Laboratory published an extensive overview and detailed
technical report [5] on cost reductions associated with future trough technology out to 2020. In this present
paper, the authors use the detailed NREL cost estimates for trough-based systems in 2010 as a basis for a
preliminary comparison of the cost of low temperature and high temperature operation of CLFR plants.
Later, it is hoped that more detailed project based comparisons will become available which will test further
the conclusions in this paper.

2. Array Costs

In [5], cost projections were made for trough systems by two independent groups, SunLab and Sargent &
Lundy. In this paper, we will use the report 2010 estimates for trough collectors. For the CLFR, costs for
the first large plant will be used, assumed to be nominally complete by 2009, but here taken as a 2010 cost.

2.1 CLFR Array Costs


For the purposes of this paper, CLFR array costs are based on the current costs (see Fig. 2) for the 40 MWe
coal save plant array currently being constructed at the Liddell power plant in New South Wales, Australia.
[3] which operates at a design outlet steam temperature of 270°C under saturated steam conditions. The
present cost includes the array, receivers and connecting pipes and is assumed to be reduced by 17% due to
much larger volume usage of reflector and steel and improved assembly efficiency, 5% due to the use of
lower iron glass in the reflector, and 5% due to the use of improved air stable selective absorber coatings
and design in the cavity receiver. The resultant cost, converted from Australian Dollars into US dollars at
the nominal rate of $A =US$0.75 is US$1046 per kWe or US$102 per m2.

Fig, 2. Stage 1 array under


construction at Liddell power
station.

2.2 Trough Array Costs


Trough array costs used are based upon 2010 estimates from [5] for trough collectors operating at 500°C
delivery temperature. These are lowest for the Sunlab estimate of $3416 per kWe ; the estimates of Sargent
& Lundy are slightly higher at $3562.

3. Power Block + BOP Costs

Estimates from [5] are used for the high temperature power block costs. The low temperature turbine costs
are based upon approximate estimates [6] directly supplied by JSC “Atomstroyexport” (Russia) for the
supply of a 240 MWe VVER steam turbine and steam separator and control equipment of about US$18
million for a single turbine. It is assumed in this paper that an additional 1/3 will be added to the
turbogenerator price to cover delivery and installation. The power block costs include the steam turbine and
generator, steam turbine and generator auxiliaries, feed water and condensate systems. Balance of Plant
(BOP) costs are taken from [5]. BOP costs include general balance-of-plant equipment, condenser and
cooling tower system, water treatment system, fire protection, piping, compressed air systems, closed
cooling water system, plant control system, electrical equipment, and cranes and hoists. Table 1 gives the
combined power block and BOP costs:

Table 1. Total Power Block and BOP Costs in $/kWe for a 240 MWe Solar Power Station

HT Power Block Cost, 500°C 226


LT VVER Power Block Cost 100
NREL BOP Estimated Cost 163

4. Storage Costs

Storage in two tank systems using hot oil and molten salt have been demonstrated, but single tank
thermocline systems may be substantially cheaper [5]. The year 2010 projection for a direct thermocline
storage uses Hitec XL (ternary) salt HTF in both the solar field and the thermal storage system, removing
the need for a heat exchanger between the solar field and storage system. Use of the Hitec allows higher
outlet temperatures (450°C), increasing the power cycle efficiency and further reducing the cost of the
thermal storage. SunLab [5] estimate the storage system cost at $7.9/kWht for 500°C operation, but
demonstration of this salt at 500°C has not yet been performed.

The concept of Underground Thermal Energy Storage (UTES), which we will refer to in this paper as
‘cavern storage’, appears to have been first advanced by R&D Associates in 1977, as quoted in 1983 report
from SERI [7,8]. It involves storage of water under pressure in deep metal lined caverns where the
pressure is contained by the surrounding rock and the overburden weight. The SERI report was a study of
different storage options prepared by the Solar Energy Research Institute for the U.S. Department of
Energy (DOE) in the early 1980’s. These were developed in a joint effort between the Office of Advanced
Conservative Technologies, Division of Physical and Chemical Energy Storage (PACES) and the office of
Solar Applications for Industry, Division of Solar Thermal Energy Systems. Fourteen organizations were
involved in deriving the comparative rankings, which indicated quite definitively that UTES for a large
system is the cheapest storage method. More recently, Tanner (2003) [9], produced, at the suggestion of
one of the authors, a thesis report on cavern storage applied to the case of the CLFR. This study
investigated the current costs of a steel lined cavern at depths of 200m and 400m and indicates that cavern
storage is much cheaper than other currently proposed storage methods at an installed cost of US$16.2
million for a plant providing 54% capacity factor. For the trough field, which has a higher daytimje
capacity than the CLFR, the storage cost is slightly smaller at

Oil storage is well proven, having been used for many years on the first LS2 trough plant in California, and
can be used instead of cavern storage for operation below 304°C. Oil storage was examined in a recent
NREL report [10] on organic rankine cycle trough systems. In this report, thermal storage using mineral oil
(Caloria) in a two tank system approaches an asymptotic cost of $14/kWht. The temperature range of
operation for an efficient ORC system in [10] is about 100°C. The VVER turbine has a difference between
inlet and outlet temperature of 250°C-171°C = 79°C. As a first approximation suitable for this general
discussion, we will assume that the storage temperature range in the two cases is the same ratio, 0.79, so
that the cost can be scaled simply with a result of (14/2.42) x (100/79) x $92 per kWhe = $674 kWhe. The
resulting options are given in Table 2; these include heat exchange cost where applicable.
Table 2. Storage Options Considered for Solar Thermal Plants
Medium $/kWe Reference
VP-1/Solar Salt 958 [5]
HitecXL at 450°C 425 [5]
Cavern at 300°C 92 [9]
Caloria Two-Tank 674 [10]

5. Total Plant Costs for 240MWe 12 Hour storage Systems

In [5], the receiver thermal loss factor is 0.81 for the 500°C 2010 case and a 40% turbine efficiency is used.
For the 500°C case the temperature across the field is assumed to rise by approximately 200°C, so that the
average fluid temperature will be 400°C. The loss factor will be reduced by operating at 300°C, which
corresponds to an average field fluid temperature of 275°C; the new emissivity loss factor (ELF) -
neglecting the change of emissivity with temperature – will vary as the 4th power of the absolute
temperature difference, so that making a simplifying mathematical assumption that the fields have these
average temperatures throughout and that the emissivity is constant with temperature,

ELF = 1-[(547-300)**4 /(673-300)**4 x 0.19 ] = 0.963

The header pipe loss factor HLF is 967 at 500°C outlet temperature and at 300°C outlet temperature will
be

HLF = 1-[(275-30)/(400-30) x 0.033] = 0.978

These thermal losses may be summed, so that (1- ELF) + (1- HLF) = 0.059 and the total thermal loss factor
at 300°C is consequently 0.94. Thermal efficiency of the power block at 300°C is lower, increasing the
array size by the factor of 0.4/0.315 = 1.27, but lowered heat loss decreases the array size by .81/.94 =
0.861 for a net size increase of 1.27 x 0.861= 1.09 In practice, the collector may be able to be constructed
more cheaply at the lower temperature, but this reduction is not included.

Table 3. Total system costs for trough systems in US$ per kWe in 2010,
for a capacity factor of 56.2%.
Trough System 400°C Trough 500°C Trough 300°C Trough 300°C Trough
(current) Hitec Caloria Cavern
Array 3531 2644 2882 2882
PowerBlock+ BOP 499 389 281 281
Storage 958 383 674 63
Total 4856 3416 3837 3226

It is clear in Table 3 that trough systems can obtain the most positive cost/benefit using VVER + Cavern
storage, but the VVER + Caloria option is less favourable than the 500°C Hitec salt option. While it may
seem that Caloria should be ignored, it is an option which contains little risk and could be installed at a
premium of only 12% compared to a 2010 500°C molten salt system. Operation at 500°C with salt in long
absorber tubes has not yet been demonstrated. Table 4 shows similar results for the CLFR; much reduced
collector costs attained in the CLFR system are the predominant factor in cost reduction. In this case, the
use of Caloria now incurs a 34% penalty over the Cavern storage option, but is still lower in absolute cost
than any trough option. The cost of a Caloria CLFR system is more expensive than a 500°C CLFR, and the
Cavern storage option is cheapest, as with the trough systems.

Table 4. Total system costs for CLFR systems in US$ per kWe
in 2010 for a capacity factor of 54.3%

CLFR System 300°C CLFR 300°C CLFR


Caloria Cavern
Array 1435 1435
Power Block + BOP 281 281
Storage 674 68
Total 2390 1784

Cavern storage has not been proven but can be costed due to the commonality with modern mining
technology, and it is an open question whether the technical risk is higher than that of 500°C molten salt
trough system operation. Certainly, the environmental impact of a Cavern storage accident will be
minimal, as water is the storage medium and any release of HTF will not pollute the environment.

In a recent regional power plan discussion document [11] the cost of a 400MWe pulverised coal plant was
found to be $1468/kWe in the North West USA. In a previous paper [12], it is shown that the added price
premium of electricity necessary for the CLFR with cavern storage to maintain a similar return on capital
invested in such a coal plant is US$0.011 per kWhe. The capacity factors in this paper can be increased with
increased storage, while the cost of electricity per kWh can be reduced. The optimal capacity factor is,
however, site dependent and depends upon the premium charged for peaking sales, because as the capacity
factor is reduced, there is a greater opportunity to indulge in ‘peak lopping’, giving a higher return per
kWhe.

6. Conclusions

This paper is not a detailed project study and does not include an in-depth plant financial analysis. Rather,
it is a preliminary look at the issue of high temperature vs low temperature operation in the design of stand
alone solar thermal electricity plants. The early results suggest that the area deserves increased study.

The advantages gained by low temperature operation derive from a combination of cheap low temperature
turbines developed for the nuclear industry and an inexpensive storage concept which suits that particular
temperature range. Should both options be accessible as proposed, then this may be the most cost-effective
solar thermal electricity development path. The CLFR/Cavern 2010 proposal, at approximately US$1784
per kWe, offers costs well below 2020 estimates for both troughs (2225 – 3220 $/kWe) and towers (2270 –
3591 $/kWe) contained in the NREL [5] report. Perhaps more intriguingly, the US$1784 per kWe, figure,
without the use of any Green support mechanisms, is comparable to the cost of some current conventional
pulverised coal-fired generation in the USA [12].

Cavern storage cannot be taken higher than about 360°C and still has some developmental uncertainty
ahead of it, but the concept is very simple and two reports have now identified it as the lowest cost storage
concept. Recent discussions the authors have had with geologists and mining companies suggest the
concept can be widely applied; suitable rock structures are common. If suitable geological structures are
not available, Caloria oil storage is a low technical risk option available for a cost premium.
Environmentally, however, cavern storage would be safer than either salt or oil solutions.

References

[1] Mills D.R and Morrison G.L.(1999). “Compact linear Fresnel reflector solar thermal powerplants”,
Solar Energy, 68, pp 263 – 283.
[2] Hu E., Mills, D.R., Morrison, G.L. and LeLievre P. (2003), “ Solar power boosting of fossil fuelled
power plants”, Proc. International Solar Energy Congress, Goteborg, Sweden, 2003.
[3] Mills D.R.,Morrison G.L and Le Lievre,P.(2004), “ Multi-Tower Line Focus Fresnel Array Project”,
Submitted to ASME Journal of Solar Energy Engineering.
[4] Andasol (2002), Aringhoff R. Geyer M., Herrmann U., Kistner R., Nava P., Osuna R. "AndaSol -
50MW Solar Plants with 9 Hour Storage for Southern Spain", Proc. 11th SolarPACES International
Symposium, 4-6 Sept, Zurich, Switzerland, pp. 37-42, Paul Scherrer Institut.
[5] NREL (2003), Assessment of Parabolic Trough and Power Tower Solar Technology Cost and
Performance Forecasts. Edited by Sargent & Lundy LLC Consulting Group Chicago, Illinois. National
Renewable Energy Laboratory Report NREL/SR-550-3444, 1617 Cole Boulevard, Golden, Colorado
80401-3393, USA, October.
[6] VVER (2003), Estimate of cost by JSC “Atomstroyexport” (Russia) for a single 240 MW VVER
turbine and generator block.
[7] Copeland R.J. and Ullman J.(1983), Comparative Ranking of Thermal Storage Systems, Volume I – for
Water/Steam, organic Fluid, and Air/Brayton Solar Thermal Collector-Receivers, SERI/TR 631-1283,
Solar Energy Research Institute, Golden, Colorado.
[8] Dubberly et al (1983), Comparative Ranking of Thermal Storage Systems, Volume II – Cost and
Performance of Thermal Storage Concepts in Solar Thermal Systems, Phase I, SERI/TR 631-1283, Solar
Energy Research Institute, Golden, Colorado.
[9] Tanner A.R. (2003) Application of Underground Thermal Energy Storage for Solar Thermal Power
Generation in New South Wales. Engineering Thesis, School of Aerospace, Mechanical and Mechatronic
Engineering, University of Sydney, November.
[10] NREL (2002), Modular Trough Power Plant Cycle and Systems Analysis. Edited by Hank Price and
Vahab Hassani, National Renewable Energy Laboratory Report NREL/TP-550-31240, 1617 Cole
Boulevard, Golden, Colorado 80401-3393, USA, January.
[11] NorthWest Council (2002), typical costing of a community power plant, 400Mwe size, available at
http://www.nwcouncil.org/energy/powerplan/grac/052202/coalfireplants.htm
[12] Mills D.R., Morrison G.L and Le Lievre P.(2004), “Design of a 240 MWe Solar Thermal
Power Plant”, Accepted for publication in Proc. Eurosun 2004 Conference.

Vous aimerez peut-être aussi