Sie sind auf Seite 1von 7

Available online at www.sciencedirect.

com

Solar Energy 86 (2012) 132138 www.elsevier.com/locate/solener

Concentrating sunlight with an immobile primary mirror and immobile receiver: Ray-tracing results
Steven C. Rogers, Connor Barickman, Greg Chavoor, Matt Kinni, Nik Glazar, Peter V. Schwartz
Dept. of Physics, California Polytechnic State University, San Luis Obispo, CA 93407, United States Received 3 November 2008; received in revised form 13 September 2011; accepted 15 September 2011 Available online 6 October 2011 Communicated by: Associate Editor L. Vant-Hull

Abstract Using a combination of custom computer code and commercially available ray-tracing software, we explore variations of concentrator geometries where sunlight is rst incident onto a stationary primary mirror of circular cross section. The reected radiation is incident onto a smaller, secondary moveable mirror, which focuses the radiation onto a stationary target. Simulations for this trough geometry show peak concentrations of 38 solar equivalents. 2011 Elsevier Ltd. All rights reserved.
Keywords: Solar energy; Solar concentrator; Ray tracing; Trough concentrator; Fixed focus

1. Introduction Large solar concentration devices have traditionally consisted of a parabolic primary mirror, which focuses radiation onto a target, such as a heat collecting element (HCE) or photovoltaic cell (PV). In order to keep the target at the focus of the primary mirror, the entire mirror must rotate about either one axis (for trough systems), or two axes (for dish systems). The need to rotate the assembly constitutes a considerable amount of the capital cost for modern Solar Thermal Electric (STE) facilities. The mirrors and HCE are often made of glass and must be delicately held in place with a high degree of angular tolerance, while hot (and possibly pressurized) uids circulate in the HCE and through piping that must pivot as the drive mechanism tracks the sun. A recent NREL (National Renewable Energy Laboratory, Golden, Colorado) study (Stoddard et al., 2006) indicates that 60% of the cost of an STE facility is the solar eld.
Corresponding author. Tel.: +1 805 756 1220.

E-mail address: pschwart@calpoly.edu (P.V. Schwartz). 0038-092X/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.solener.2011.09.012

Half the cost of this solar eld including pylons, metal support structure, mirrors, drive system and piping interconnects may be due to the need to track the large parabolic mirrors (Sargent and Lundy LLC Consulting Group, 2003). Some innovations reduce the extent of hardware necessary to track the sun and most are described in a single text (Winter et al., 1991). Many innovations include an immobile primary mirror of circular cross section that is either a trough or dish (i.e. sphere, or bowl), and a mobile target that tracks the area of concentrated radiation (Rabl, 1985). The most famous spherical concentrator is the radio telescope at Arecibo, where incident radio wave radiation is reected from an immobile spherical bowl about 300 m across onto a movable secondary concentrator and receiver allowing observations of up to 20 from the zenith. Similarly, a trough of circular cross section can be used to concentrate radiation onto a moveable linear secondary target. Whether a trough or dish, the primary mirror has a circular cross section of radius R, rather than a parabola in order to avoid the aberration that increases with radiation of increased o axis incidence (Fig. 1a). A mirror of circular

S.C. Rogers et al. / Solar Energy 86 (2012) 132138

133

Fig. 1. Incident sunlight (shown as parallel lines) on (a) a parabolic mirror, and (b) a circular mirror, at the angles indicated. The parabolic mirror focuses only the axially incident radiation to a point. Radiation reected from the circular mirror is not focused to a point. However, at all incident angles, the area of focused radiation has the same mathematical form albeit in a dierent place.

cross section produces a consistent spherical aberration for all incident angles (Fig. 1b). Radiation that is near perpendicular to the circular surface (incident angle of 0) is reected through a point R/2 from the center of curvature. Radiation at higher incident angles is focused to points at larger radii (closer to the surface of the circular mirror), focusing on the mirror surface itself for incident angles of 60. Incident angles greater than 60 result in multiple reections. For a solar bowl collector all incident radiation is ultimately reected onto a movable target kept parallel to the incident radiation along the radius vector of the bowl, extending from the mirror surface to R/2 (E1-Refaie, 1987; OHair et al., 1986, 1987). Another trough geometry makes use of an immobile faceted (Fresnel) primary mirror that focuses incident radiation onto a line on a cylindrical surface, where the reected radiation is absorbed by a mobile line absorber (Russel and Deplomb, 1975). All the above innovations lower the mass that needs to be moved in solar tracking. However, these innovations may also result in the following problems: additional complication in the tracking mechanism, as well as reduction of solar concentration or collection eciency. We propose a trough design that allows both the primary mirror as well as the absorber to be stationary, using a smaller, moving secondary mirror to redirect the radiation reected o the stationary primary mirror. Consistent with the above alternative concentrator designs, our design greatly reduces the mass that needs to track. This gain comes at the expense of collection eciency, concentration, and simplicity of tracking. While this ray-tracing study presently satises academic

interest, future improvements may enable applications in solar energy or other elds. The design we propose is similar to that of a Gregorian telescope (Fig. 2). In such an optical system, a large primary concave mirror directs radiation onto a smaller secondary concave mirror, which serves to correct for spherical aberration and redirect the radiation back toward the target past the primary mirror. Such a system has been studied before as a way to achieve high solar ux concentrations (Leutz and Ries, 2005). However, in this previous work

Fig. 2. Diagram of a Gregorian telescope.

134

S.C. Rogers et al. / Solar Energy 86 (2012) 132138

(a) Equinox

(b) Summer

(c) Winter

Fig. 3. Incident radiation is reected from a stationary primary mirror and focused onto a stationary target by means of a smaller, mobile secondary mirror. For an eastwest oriented trough the surface is optimized for equinox, summer and winter sunlight can still be focused onto the same target.

the spatial relationship between the target, and primary and secondary mirrors remains xed and the entire system must be oriented with respect to the incident radiation to maintain focus. Our proposed concentrator design has both the primary mirror and target stationary. Focusing is achieved through the movement of the smaller, secondary mirror (Fig. 3). In this paper, we explore the performance of a linear trough geometry. The same technology can equally be applied to dish geometry. 2. Methodology Modeling of the shape and placement of reecting surfaces was made possible with commercial ray-tracing software (LightTools 6.3.0 by Optical Research Associates). We used the Solar Tracking Utility included in LightTools to provide realistic simulations of collector performance at San Luis Obispo, CA (35.27 N Latitude). The simulated sun was a uniformly radiant circle of radius 100 m, located 21,483 m from the collector, subtending an angle of 0.533. Modeling was purely geometric: mirrors were assigned a reectivity of 100% and the HCE, a reectivity of 0%. Troughs are oriented eastwest, facing both perfectly

upward as well as tilted at latitude with respect to the azimuth. Performance at other locations can be inferred from the simulation data gathered for the location mentioned above. We ran simulations for three dates that give the operating extremes of collector performance: Vernal Equinox, Summer Solstice and Winter Solstice. We ran simulations of two dierent HCE types with the same size aperture: a cylindrical absorber with an involute CPC (Winston et al., 2005), and a Flat Plate absorber. Results are reported in terms of eciency (portion of radiation incident on the primary mirror that reaches the target) as well as concentration. The primary mirror has a circular cross section, as opposed to a parabolic cross section. We chose a circular primary mirror shape because the coma (concentrated area of reected radiation) has the same mathematical form for all incident angles. The coma follows a path that is roughly concentric to and located at half the radius of the primary mirror. We chose this trajectory as a starting point for the path of the secondary mirror. Subsequent optimization resulted in slight deviations from a circular path. We explored combinations of model parameters, and present results for the following values:

Fig. 4. LightTools output showing equally spaced, parallel rays (a) reected from the primary mirror to the secondary mirror, onto the target, and (b) an enlarged view of the box in (a), showing the tertiary involute CPC concentrator.

S.C. Rogers et al. / Solar Energy 86 (2012) 132138

135

Fig. 5. LightTools output showing a reduced number of incident rays from a simulated sun (a) reected from the primary mirror to the secondary mirror, onto the target, and (b) an enlarged view of the box in Fig. 5a, showing the tertiary involute CPC concentrator.

(1) Length of the collector, LCOLLECTOR = 20 m. (2) Radius of curvature of the primary mirror, RPRIMARY = 3.0 m. (3) The rim angle of the primary mirror, qRIM = 80. (4) Width of the secondary mirror, WSECONDARY = 0.6 m. (5) Radius of the cylindrical HCE, RHCE = 4.4 mm. (6) Input aperture of the CPC, WCPC = 3.9 cm. (7) Width of Flat Plate HCE, WPLATE = 3.9 cm. The secondary mirror was optimized in Light Tools by using a ve-point symmetrical spline patch. As a starting point, we rst examined reections only in the plane perpendicular to the (eastwest) trough orientation and optimized the shape of the secondary mirror for every angle of solar

incidence from 0 to 40. We chose an upper limit of 40 because it allowed our system to track the sun for most of the usable daylight hours over the course of an entire year. From the 40 resulting secondary mirror shapes, we selected the surface that provided the highest average solar ux to the target. The shape of the secondary mirror is dened by a ve point spline passing through the points (0, 0), (0.1515, 0.0245) and (0.3030, 0.0874) measurements in meters. In comparison, a parabola passing through the middle three points would have slightly more curvature, passing through (0.3030, 0.0978). Using this shape as the secondary mirror, we wrote computer code utilizing an Application Program Interface (API) within LightTools 6.3.0 to further dene the path of the secondary mirror maximizing solar ux on the target.

Fig. 6. Concentration of sunlight onto the target as a function of angular oset at solar noon for an eastwest oriented trough. That is, the direction to the sun is perpendicular to the long axis of the trough but is displaced rotationally from the concentrators axis by the angle indicated in the x axis.

136

S.C. Rogers et al. / Solar Energy 86 (2012) 132138

Fig. 7. Concentration vs. Time of Day for a trough oriented eastwest; for concentrators facing vertically upward during Vernal Equinox, Summer Solstice and Winter Solstice with Flat Plate and CPC.

3. Modeling results An example of the computer output is shown in Figs. 4 and 5. Fig. 4 is a simplied graphic making use of equally spaced, parallel rays. Fig. 5 makes use of our simulated sun, albeit with a reduced number of rays to improve readability. A blowup of the involute CPC tertiary concentrator is shown in Figs. 4b, and 5b. Fig. 6 displays the models angular response for in-plane incident radiation that is

for radiation during solar noon on an eastwest oriented trough. While Fig. 6 is useful in order to understand the geometric response of the model, it does not provide practically useful data. Subsequent gures display data obtained from our model following actual solar trajectories (Figs. 710). Fig. 7 compares the concentration achieved as a function of the Time of Day of both the Flat Plate and CPC absorber types when the collector is oriented straight upward. Fig. 8 displays concentration for the collector

Fig. 8. Concentration vs. Time of Day for a trough oriented eastwest; for concentrators tilted at latitude toward the equator during Vernal Equinox, Summer Solstice and Winter Solstice with Flat Plate and CPC absorbers.

S.C. Rogers et al. / Solar Energy 86 (2012) 132138

137

Fig. 9. Eciency vs. Time of Day for a trough oriented eastwest; for concentrators facing vertically upward during Vernal Equinox, Summer Solstice and Winter Solstice with Flat Plate and CPC absorbers.

tilted at latitude. Fig. 9 compares the eciency of intercepting redirected rays when the collector is facing straight upward. Fig. 10 displays eciency for the collector tilted at latitude. Facing straight upward, the collector yields the greatest number of useful hours and highest concentrations during Summer Solstice. However, it oers poor concentration and eciency during Winter Solstice due to the angle of the sun being greater than the collectors 40 design limit. Concentrators tilted at latitude provide the

highest year round energy. The CPC we chose oers higher concentrations but lower eciency than the Flat Plate HCE. 3.1. Comparison with parabolic trough concentrators: land use eciency and concentration The design proposed here, with an 80 rim angle, achieves a peak geometrical eciency of 32% and concen-

Fig. 10. Eciency vs. Time of Day for a trough oriented eastwest; for concentrators tilted at latitude during Vernal Equinox, Summer Solstice and Winter Solstice with Flat Plate and CPC absorbers.

138

S.C. Rogers et al. / Solar Energy 86 (2012) 132138

trations up to 38 suns. A conventional parabolic trough achieves concentrations of 71 suns (Cohen, 2004). While a single parabolic trough achieves eciencies near 100%, troughs are widely spaced in a solar eld to avoid shadowing when the sun is lower in the sky. Standard land use eciency of a typical parabolic trough eld is about 25.8% (Canada et al., 2005), although the light use eciency is higher when averaged over all incident angles throughout the day. Because our normal oriented solar troughs would not shade each other, we could theoretically utilize 100% of the solar eld. For troughs oriented at latitude, in order to have no shading at Winter Solstice at noon, the necessary spacing between troughs would be: wtanhlat 23:5 sinhlat 1 cos hlat ; where w is the width of the trough, and hlat is the latitude. Under this criterion, at hlat = 35, the land use eciency would be 60%. Because of multiple reections required for our geometry, reection losses would further lower the performance of our design compared to a single parabolic trough. In summary, our results predict solar concentrations and eciencies somewhat below those of conventional parabolic troughs. 4. Conclusion We have modeled the focusing of sunlight from a stationary primary mirror onto a stationary target by means of a smaller, mobile secondary mirror, which tracks the area of concentrated radiation through a complex path. Using this conguration, with a tertiary optical concentrator around the target element and ignoring reection losses from the two or three reections, we have achieved a peak concentration of 38 solar equivalents with greater than 30% peak eciency. Improvements are possible with further optimization of the shape and path of the secondary mirror, as well as by allowing the tertiary mirror to rotate


about the target, which would require a larger slit in the primary. Acknowledgements This work was sponsored in part by the Department of the Navy, Oce of Naval Research, under Award # N00014-07-1-1152, and in part by the Donors of the American Chemical Society Petroleum Research Fund. References
Canada, S., et al., 2005. NREL/CP-550-37077. Cohen, G.E., 2004. In: National Solar Energy Conference: The Parabolic Trough Technology. <www.nrel.gov/csp/troughnet/pdfs/cohen_neva da_aps_projects.pdf>. El-Refaie, M.F., 1987. Applied Energy 28, 163189. Leutz, R., Ries, H., 2005. Hemispherical solar concentrator corrected for aberrations. In: 3rd ISCC Conf. OHair, E.A., Simpson, T.L., Green, R., 1986. Results of operation of the crosbyton solar bowl. In: Ferber, R.R. (Ed.), Proc. 8th ASME Solar Energy Div. Conference, Anaheim, CA, pp. 205209. OHair, E.A., Simpson, T.L., Green, R., Krile, T., Smith, M., Jones, W.B., Vann, W.P., 1987. The crosbyton solar power project: solar bowl subsystems tests and analyses. Department of Energy Report Number: DOE/AL/21557-T12. Rabl, A., 1985. Fixed reectors with tracking receivers. In: Active Solar Collectors and Their Applications. Oxford UP, New York, pp. 186 188. Russel, J.L., Deplomb, E.G., 1975. Line Focused Fixed-Mirror Collector, 1975. USPTO, Patent #3868823. Sargent & Lundy LLC Consulting Group, 2003. Assessment of Parabolic Trough and Power Tower Solar Technology Cost and Performance Forecasts. in: NREL/SR-550-34440. Stoddard, L., et al., 2006. Economic, Energy, and Environmental Benets of Concentrating Solar Power in California. NREL-SR550-39291. Winston, R. et al., 2005. Nonimaging Optics. Elsevier Academic Press, San Diego, pp. 52. Winter, C.J., Sizmann, R.L., Vant Hull, L.L., 1991. Solar Power Plants: Fundamentals: Technology, Systems, Economics. Springer Verlag.