Progress in the development of heliostats

Over the last several decades, heliostat designs have primarily used conven­tional glass and steel, pedestal-mounted elevation-azimuth designs, but alternatives include ‘ganged heliostats’, carousel heliostats on tracks, stretched membrane reflectors, inflatable enclosures, etc. Various examples are shown in Fig. 17.1. The US Department of Energy (DOE) studies con­ducted by Sandia National Laboratories Albuquerque (SNLA) resulted in the base line glass-steel/elevation-azimuth (el/az) heliostat being about 150 m2 in area; the most specific version of this base line is the 148 m2 Advanced Thermal Systems (ATS) heliostat (Kolb et al., 2007), which was used to develop the DOE installed cost of $211/m2 in 2010 dollars for sustained, high volume production. As with many el/az designs, a linear actuator is used for elevation, and a multi-stage gear system is used for the azimuth. Information on these heliostats is available from various sources (Kolb, 2006; Kolb et al., 2007; Falcone, 1986; Jones, 2006; Dietrich et al., 1982; Winter et al., 1990), including numerous websites maintained by solar central receiver companies, such as eSolar, BrightSource, Abengoa, etc.

Starting with initial heliostat efforts in the early 1970s up to today, there has been a general tendency to increase the heliostat size from about 12 m2 to approximately 150-200 m2, and even up to 320 m2, with several counter­examples of much smaller heliostats, primarily in the past several years.

The tendency to favor larger heliostats during this period has apparently been based in part on the assumed advantages of ‘economies of scale’. This trend is seen in various design studies and analyses (Kolb, 2006; Kolb et al., 2007; Falcone, 1986; Jones, 2006; Dietrich, et al., 1982; Winter et al., 1990). This trend has also been seen in other solar power systems, such as the 320 m2 (or, 334 m2, depending on version) Amonix concentrating photovol­taic system (www. amonix. com). This system was also proposed by Arizona Public Service for modification to a heliostat, but that plan was not com­pleted. Another expected benefit with larger heliostats was that the fixed costs for a heliostat could be spread over a larger area, thus reducing the cost per unit area. Other factors may have played a role in this general trend, such as availability of custom drive units potentially offering high performance and low cost, or relaxing design criteria to achieve lower costs by increasing the reflector area to the maximum allowable for a given drive unit. These studies covered primarily specific designs, and cost consider­ations for these designs. Only recently have intrinsic cost vs. size consider­ations been available in the literature (Kolb et al. 2007, and an earlier discussion in the Sandia Heliostat Handbook, 1982). Figure 17.2 shows the trend of heliostat size as a function of area, compiled from Kolb (2006), Kolb et al. (2007), Falcone (1986), Jones (2006), Dietrich et al. (1982), Winter

Martin Mariettta Solar One and Solar Two (39.9 m2)

 

CSIRO National Solar Energy Centre Solar Towers, 4.5 m2

 

SAIC 50 m2 Stretched Membrane

 

image428

image429image430image431image432

image433

17.1 Representative heliostat designs and sizes (Kolb et al., 2007).

image434

Year

17.2 Heliostat size trend 1970 to 2010.

et al. (1990), Blackmon (2008), California Energy Commission (2007) and various manufacturers websites (BrightSource, eSolar, CSIRO); these pro­grams are summarized in Table 17.1. Note that this summary captures selected prototypes and several commercial developments. There are undoubtedly other prototypes that have been constructed by various groups in this period, but this selection is representative of the range of sizes and the general trend over the last 35 years or so.

Figure 17.2 and Table 17.1 illustrate substantial variations in design approaches and costs. This variation in size is remarkable; of 32 designs only six are less than about 15 m2 and sixteen are approximately 50 m2 or above and nine of these are about 100 m2 or above. If the assumptions of econo­mies of scale, and, in effect, relatively high fixed costs on a per heliostat basis are true, then the larger heliostats would be the preferred choice. This fixed cost per heliostat aspect may have been at least partially supported by relatively high electronic costs needed for each heliostat during the early period of heliostat development. If, however, the fixed costs associated with a heliostat are not a substantial fraction of the total cost, then the reverse would be true. That appears to be the case, especially with far lower elec­tronic costs available today. Finally, costs are typically based on a relatively well-established, if not fully commercialized, central receiver industry. Actu­ally, the initial costs to form this industry would be higher, and thus can pose a major impediment to market entry and commercialization. It remains to be seen if lower initial costs can be achieved by aggregating costs over high production volumes through large power purchase agreements.

Year

(approximate)

Program

Prime contractor or location

Size (m2)

1970

Trombe Heliostats

France

45

1973-1974

National Science Foundation

University of Houston/McDonnell Douglas

13.4

1975-1977

Pilot Plant

Boeing

48

System Research

Martin Marietta

41

Experiment

Honeywell

40

University of Houston/McDonnell Douglas

31.4

University of Houston/McDonnell Douglas

37.5

1977-1979

Central Receiver Test Facility (SNLA)

Martin Marietta

37.2

1978-1979

Pilot Plant Prototypes

Martin Marietta

39.9

University of Houston/McDonnell Douglas

44.5

1979-1981

Second Generation

Boeing

43.7

Heliostat

Martin Marietta

57.4

University of Houston/McDonnell Douglas

56.9

Arco (Northrup)

57.8

1980-1981

Pilot Plant (Solar One)

Martin Marietta

39.9

1981-1986

Large Area Heliostat

University of Houston/McDonnell Douglas

90

Arco

95

Arco

150

Solar Power Eng. Co.

200

1984-1986

Stressed Membrane (SM)

Solar Kinetics Inc

150

Science Applications

150

1990

Stressed Membrane (SM)

Science Applications International Corp.

100

mid-1990s

SM, but with Glass

Solar 2/Spain

150

mid-1990s

USISTF High Concentration Solar Central Receiver

University of Houston/McDonnell Douglas/НІТек Services

9.2

mid-1990s

ASM-150

Steinmuller (Germany)

150

1995-2000

Gher S. A. Hellas 01

Gher S. A. (Spain)

19.2

2006

Existing Amonix PV Tracker Converted to a Heliostat

APS (Proposed)

320

2006-2007

PS-10 and PS-20

Planta Solar (Abengoa, Spain)

121

2006-2008

Carpe Diem HelioCA 16

DLR

16

2006-2008

SHP (Australia)

DLR-Julich, Germany

8

2007

CSIRO (Australia)

CSIRO National Solar Energy Centre Solar Towers

4.5

2009

BrightSource

Solar Energy Development Center, Rotem, Negev, Israel

14.4

2010

eSolar Sierra SunTower

eSolar, Inc (Five 1 sq. meter)

5

 

Подпись: © Woodhead Publishing Limited, 2012

An example of one type of size growth is seen in the study conducted at McDonnell Douglas (Dietrich et al., 1982). There it was concluded that the cost of that heliostat was reduced by increasing the area, and keeping the same structure and drive unit, with some relatively minor changes. This conclusion was based primarily on determining that the angle of attack of the high wind condition for horizontal stow could be reduced from 10 to 6.5°, and that it was cheaper to replace damaged heliostats, rather than design for a 25-year life with a worst case wind condition. This approach, however, did not address the fundamentals of heliostat loads vs size, and it did not keep the design load conditions constant. Other examples are noted, including a study to increase the size of the stretched membrane (SM) heliostat from 50 to 150 m2 to decrease cost (Kolb et al., 2007); it was con­cluded that this did not reduce cost and the effort was not continued. Note that the ATS design is based on strength, not stiffness. If strength, not stiff­ness, is used, then gravity bending or sagging becomes an issue with larger heliostats. This would ‘make large heliostats less costly on an optics – corrected basis than they will be in reality’ (Kolb et al., 2007).

Sandia determined in 2006 that ‘The ATS heliostat is the low cost baseline in the U. S.’ (Kolb et al., 2007). Their reported cost for this heliostat is $126.49/m2 for 50,000 units per year and $164/m2 for 5,000 units per year. These costs are presumed to be for the 5,000th and 50,000th units per year, respectively, with no further reduction due to learning curve effects. The DOE requires that solar generating cost be determined using the System Advisor Model (SAM 3.0). They state that the current DOE baseline helio­stat is 150 m2 with a cost of $211/m2 (DOE, 2009). This cost is based on the ATS heliostat. However, the detailed cost breakdown and production quan­tity are not available; it may be presumed that the cost differences com­pared to the ATS heliostat are at least partly attributable to broader considerations, such as market entry conditions, with lower production rates, as well as initial startup costs, inflation escalation from 2006 to 2010, and perhaps additional costs being included, such as for site preparation, permits, and various financial cost factors.

Updated: August 22, 2015 — 1:09 am