Nonimaging Optics Winston Episode 12 pptx

The losses in the series resistance degrade thecell fill-factor FF, which is one parameter that is related with the cell efficiency by the formula 13.1 where P L is the solar power, V OC

case of the CPC), at least a small gap must be provided Moreover, these thermalaspects are responsible for the difference between thermal and photovoltaics with

respect to the use of optically dense transparent media (n > 1) surrounding the

receiver This is difficult to apply for thermal concentrators, and if excluded, the

concentration limit is reduced by a factor n2for solar thermal systems (which is

n2= 1.52= 2.25 for typical materials)

Another aspect to consider is the electric nature of the photovoltaic cell Thenecessity of extracting the photogenerated current and of interconnecting the dif-ferent cells must be considered at the optical design stage Also, the use of metal-lic mirrors in contact with the cell should be avoided to prevent short circuits.Obviously, these problems do not matter for solar thermal systems

Also, the spectral sensitivity of the receivers is different in both fields For example, the quantum efficiency of the silicon solar cells is very low over 1,200 nm This means that the further infrared radiation of the sun, which usually

is nearly 25% of the sun power, is not useful, contrary to the solar thermal case.These spectral differences may make it so that a good optical material for photo-voltaics may be unsuitable for solar thermal applications due to its poor infraredresponse

Finally, the last aspect to be considered here is the receiver geometry Thisgeometry affects the design importantly, as we have seen in Chapter 5 In the case

of medium to high concentration in photovotaics, the active surface of the solarcell is always flat, typically with a round or squared contour of the active area.Sometimes the cells tessellate to build a strip receiver The inactive face of the cell is used to evacuate the heat by fixing a heat-sink to this face Only low-

concentration systems (typically C < 5) can use bifacial solar cells—that is, cells

that can collect the sunlight on their two faces The bifacial geometry must be ously considered in the concentrator design

obvi-On the contrary, in solar thermal concentration there is a wider variety ofreceiver geometries Flat (monofacial and bifacial) and tubular are the mostcommon ones In the case of a tubular receiver, the conventional cross section iscircular However, the contour can also been considered a surface to design, whichconstitutes an additional degree of freedom As an example, for the specific case

of the parabolic through, Ries and Spirkl (1995) proved that designing the contourusing the caustic of the edge rays allows to noticeably increase the concentrationratio, as shown in Figure 13.1 With 100% ray collection efficiency within the accep-

tance angle ±a, the geometrical concentration is limited to CMAX = 1/sin a For aparabola with rim angle of ±90° and also for 100% collection efficiency, the circu-

lar receiver only achieves C/CMAX= 1/p = 0.32, while the optimum receiver provides

C/CMAX= 0.47

13.2.2 Irradiance Uniformity on the Receiver

The irradiance distribution on the receiver produced by most concentrators is notuniform The effect of this nonuniformity has not been critical in solar thermalsystems, although it has some relevance in the development of Direct Steam Generation (DSG) systems (Goebel et al., 1996/1997) However, the nonuniform illumination of a photovoltaic cell may produce a dramatic decrease of its solar-to-electric power conversion efficiency This is due to the Joule effect power losses in

the cell series resistance R (if the series resistance were null, the efficiency will

always increase with concentration) The losses in the series resistance degrade the

cell fill-factor FF, which is one parameter that is related with the cell efficiency by

the formula

(13.1)

where P L is the solar power, V OC is the open-circuit voltage, I SCis the short-circuit

current, and FF is the fill factor The nonuniform illumination mainly affects the

cell fill factor As an example, Figure 13.2 shows the PSIPICE simulation with adiscretized cell of the effect of illumination nonuniformity on the performance of

a 1 mm ¥ 1 mm Gallium Arsenide (GaAs) solar cell (Álvarez, 2001) This cell wasdesigned for optimum performance at a uniform irradiance of 1,000 suns (1 sun =

1 kW/m2) The simulation considers an average irradiance along the total cell area

of 1,000 kW/m2and a pillbox irradiance pattern with higher peak irradiance on asquared area at the cell center and null irradiance outside said area When thecell is illuminated with a peak irradiance of 5 times the average, the cell conver-sion efficiency reduces approximately as the fill factor does, and then the simula-tion indicates that the efficiency decreases a 100(1 - 0.77/0.85) = 9.4%

Both concentrator and cell designers should be concerned about this problem

at the design stage and try to minimize its negative effects The variation of theirradiance pattern with the mispointing angular error must also be considered.Therefore, two issues must be addressed:

P

OC SC L

Straight lines Caustic of

+ a

Parabolic mirror Receiver

Figure 13.1 The concentration of a solar thermal parabolic trough (on the left) can be imized by designing the contour of the receiver (on the right) The conventional circular receiver providing the same acceptance angle (for 100% ray collection) is shown for com- parison purposes.

max-0 5 10 15 20 25

) and different pillbox local concentration

at the cell FF denotes the cell fill factor.

1 How a given nonuniform irradiance distribution affects the cell efficiency Notethat analytical models rather than numerical simulations verb (as that inFigure 13.2) because the formers give much more information to the design-ers about where and how much the light can be concentrated.

2 How to design concentrators producing the desired irradiance patterns, pendently of the sun position within the concentrator acceptance angle.Discussing issue 1, the effect of the nonuniform illumination depends on both thephysical parameters of the solar cell and on the irradiance distribution No generalanalytical model to quantify this effect is available yet As an illustrative example,

inde-in the case inde-in which the front metal grid contribution to the series resistance R S

is negligible, the cell voltage near the maximum power point can be estimated withthe following approximate formula (Benítez and Miñano, 2003):

(13.2)

where V0is the cell voltage when the series resistance is null, kT/e is the thermal voltage (about 29 mV at usual cell operating temperatures), f(C) is the probability density function of the irradiance C, and I sc,1sun is the cell short circuit current

under 1 sun irradiance The parameter C0coincides approximately with the centration level at which the cell efficiency is maximum The integrand in Eq.(13.2) indicates the relative importance of the different concentration intervals

con-(C, C + dC) on the cell performance degradation, pointing out that the dependence

on irradiance is strongly nonlinear due to the exponential function Consequently,even if a small fraction of power is highly concentrated, it can produce the celldegradation dominating the integral in Eq (13.2)

A general and simple rule of thumb that is generally used says that an diance distribution with a peak concentration doubling the average usually pro-duces an affordable decrease of conversion efficiency This rule is only inaccurate

irra-when the average irradiance is several times higher than the paramenter C0 Forinstance, following with the example of the cells fulfilling Eq (13.2), let us con-sider the two extreme irradiance distributions: (1) the pill-box type distribution(half the cell is not illuminated and the other half is illuminated with double theaverage irrradiance), which has maximum standard deviation; and (2) the distri-bution with a very small area with the peak distribution, and the rest of the area

is illuminated slightly below the average, which is the minimum (null, at the limit)standard deviation case According to Eq (13.2), the voltage difference near themaximum power point between cases (1) and (2) is

(13.3)

For a high concentration GaAs cell, the 10% relative difference between the

effi-ciencies of cases (1) and (2), approximately coincides with V(2) - V(1) = 100 mV This implies <C> ª 4C0 Since presently high concentration GaAs cells have C0ª

500 suns, this limits the 10% guaranteed accuracy of the rule to <C> ª 2,000 suns.

It can be proved (Benítez and Miñano, 2003) that, for a given average centration, the irradiance distribution maximizing Eq (13.2) is uniform However,when the grid series resistance is not negligible, it is clear that the irradiancepattern that produces the maximum conversion efficiency on conventional

e

C C

ˆ

¯

ÊË

concentrator cells is nonuniform (Benítez and Mohedano, 1999) The maximumefficiency of this optimum irradiance pattern is close to that obtained for theuniform illumination in practice Therefore, the efficiency increase is not the maininterest of this optimal nonuniform irradiance distribution but the fact that itguides the concentrator designers to where and how much can they concentratethe sun light.

In order to address now issue (2), let us consider the uniform illumination tribution as the goal If only the uniformity is required for a given concentrationfactor, the solution to this design problem is well known and was described inChapter 7 However, in the photovoltaic application, especially due to the nonlin-ear effects when series connected cell are illuminated differently (see Section13.2.3), the concentrator is desired to have an acceptance angle a substantiallygreater than the sun angular radius aS(typically, a ª ±1°, while aSª ±0.26°) Theuniformity is then required for the sun placed anywhere inside the acceptanceangle, which is a more complex design problem (the solutions in Chapter 7 coin-cide with case in which a = aS), especially because in high-concentration systems

dis-(<C> >1,000), the value a ª ±1° approaches the thermodynamic limit (and thus the

maximum illumination angle of the cell, b, comes closer to ±90°, typically b ª

±60°–75°)

There are two methods in classical optics that potentially can achieve thisinsensitivity to the source position The two methods, used for instance in con-denser designs in projection optics, are the light-pipe homogenizer and the Kohler

illuminator (commonly called integrator) (Cassarly, 2001).

The light pipe homogenizer uses the kaleidoscopic effect created in the ple reflections inside a light pipe, which can be hollow with metallic reflection orsolid with total internal reflections (TIR) This strategy have been proposed severaltimes in photovoltaics (Fevermann and Gordon, 2001; Jenkins, 2001; O’Gallagherand Winston, 2001; Ries, Gordon, and Laxen, 1997), essentially attaching the cell

multi-to the light pipe exit and placing the light pipe entry as the received of a tional concentrator It can potentially achieve (with the proper design of the pipewalls and length) good illumination on a squared light pipe exit with the sun inany position within the acceptance angle For achieving high-illumination angles

conven-b, the design can include a final concentration stage by reducing the light pipecross section near the exit However, this approach has not been proven yet to lead

to practical photovoltaic systems (and no company has commercialized it as aproduct yet)

On the other hand, the integrating concentrator consists of two imaging optical elements (primary and secondary) with positive focal length (that is, producing a real image of an object at infinity, as a magnifying glass does) Thesecondary is placed at the focal plane of the primary and the secondary imagesthe primary on the cell This configuration makes it that the primary images ofthe sun on the secondary aperture, and thus the secondary contour, defines theacceptance angle of the concentrator As the primary is uniformly illuminated bythe sun, the irradiance distribution is also uniform, and the illuminated area will have the contour of the primary and will remain unchanged when the sun moveswithin the acceptance angle (equivalently when the sun image moves within thesecondary aperture) If the primary is tailored in square shape, the cells will beuniformly illuminated in a squared area The squared aperture is usually the preferred contour to tessellate the plane when making the modules, while the

squared illuminated area on the cell is also usually preferred because it fits thecell’s shape.

Integrator optics in PV was first proposed (James, 1989) by Sandia Labs inthe late 1980s, and it was commercialized later by Alpha Solarco Now, high-concentration SMS concentrators (see Section 13.4.3) are including this strategyfor achieving good uniformity and improving tolerances

Sandia Labs’ approach used a Fresnel lens as primary and a single-surfaceimaging lens (called SILO, from SIngLe Optical surface) that encapsulates the cell

as secondary, as illustrated in Figure 13.3

This simple configuration is excellent for getting sufficient acceptance angle aand highly uniform illumination, but it is limited to low concentrations because itcannot get high angles b Imaging secondaries achieving high b (high numericalaperture, in the imaging nomenclature) are, to the present, impractical Classicalsolutions, which would be similar to high-power microscopes objectives, need manylenses and would achieve b ª 60° Another simpler solution that nearly achieves

b = 90° is the RX concentrator (Benítez and Miñano, 1997; Miñano, Benítez, andGonzalez, 1995) (see Figure 13.4) Although the Lens+RX integrator is still notpractical, it is theoretically interesting because shows that the optimum photo-voltaic concentrator performance (squared aperture concentrator, acceptance angle

a several times larger than the sun radius aS, isotropic illumination of the cell (b

= 90°), squared uniform cell irradiance independently of the sun position withinthe acceptance) is nearly attainable

As an example, for a = 1°, this optimum performance concentrator will get a

geometrical concentration C g= 7,387¥, which is the thermodynamic concentrationlimit for that acceptance angle Note that since all rays reaching the cell come fromthe rays within the cone of angular radius a, no rays outside this cone are col-lected by the optimum concentrator

Of course, in practice, the optimum photovoltaic concentrator performancemight be not desired For instance, as already mentioned, the high reflectivity ofnontextured cells for glazing angles may make the isotropical illumination useless

As another example, illuminating a squared area inside the cell is perfect for contacted solar cells (such as SunPower’s or Ammnix’ cells), but it may be not soperfect for front-contacted cells, for which an inactive area is needed to make the

Cell

Figure 13.3 The “Sandia concept 90” proved that the cell could be nearly uniformly minated on a square (with a squared Fresnel lens) for any position of the sun in the accep- tance angle (a) Normal incidence (b) Incidence near the acceptance angle.

illu-front contacts (breaking the squared shape active area restriction) Finally, for

medium concentration systems (let’s say, C g= 100) the aforementioned optimumperformance would imply an ultrawide acceptance angle a = 8.6° It seems logicalthat over a certain acceptance angle, there must be no cost benefits due to therelaxation of accuracies (and 8.6° seems to be over such a threshold) If this is thecase, coming close to the optimum performance seems to be unnecessary for thismedium concentration level

The present challenge in the optical design for high-concentration photovoltaicsystems is concentrators that approach optimum performance and at the sametime are efficient and suitable for low-cost mass production

13.2.3 Dispersion of the Optical Efficiency

Corresponding to Different Receivers

In solar thermal concentrating systems—for instance, in parabolic trough technology—the optical efficiency along the receiving tube does not need to beuniform because the system performance depends on the cumulative solar powercast along the receiver

However, in general this is not the case in photovoltaic concentration systems.The level of degradation of performance due to the dispersion of the solar powercast by the different solar cells depends on the electrical interconnection configu-ration The extreme cases are the all-parallel connected cells, whose performancedegradation is unimportant, and the all-series connected cells, whose performance

is much worse when nonequally illuminated

Consider a set of N solar cells illuminated by nonperfect concentrators that

produce dispersion of the solar power cast by each individual cell of the set Assumethat the cells are identical (thus, when illuminated equally and independently, all

the individual cells present the same fill-factor FF, short-circuit current I SC, and

open-circuit voltage V ) Since I is proportional to the power cast by the cell

with very good approximation, instead of referring to the dispersion of the solar

power cast, we can directly refer to the dispersion of I SCinstead

Let us consider the effect of illuminating the cell set in the two extreme cases(all cells are parallel connected or all cells are series connected) Referring to theparameters of Eq (13.1) applied to the cell sets, if all cells are equally illuminated

(I SC,k = I SC , 1 £ k £ N), we get

(13.4)

where V OC and FF are open-circuit voltage and fill-factor of all the (equally

illu-minated) individual cells Therefore, according to Eq (13.1), both series and allel sets of cells get the same efficiency when uniformly illuminated

par-However, when there is dispersion in the illumination, it is obtained that(Luqve, Lorenzo, and Ruiz, 1980)

us consider the case of the EUCLIDESTMconcentrator, which was developed jointly

by the Spanish Solar Energy Institute of the Technical University of Madrid UPM), the British company BP solar and the Spanish Institute of Technology andRenewable Energies (ITER) The biggest photovoltaic concentration plant in theworld, of 480 kW, was made in Canary Islands with the EUCLIDESTMtechnology.This concentrator tracks the sun with a single north-south axis, and it is composed

(IES-14 units of a parabolic trough 84 m long (composed (IES-140 individual parabolicmirrors, as shown in Figure 13.5) that concentrate the sunlight onto silicon solarcells, with a geometrical concentration of 38¥ The interconnection configurationwas based on 1,380 series connected cells grouped in 138 modules

The structural support of the EUCLIDESTMis given by an equilateral gular beam This beam suffers flexion and torsion due to the system weight andthe wind loads, which causes pointing errors along the concentrator trough Thelosses associated they’re beam deformations are not relevant for the annual energyproduction (because with below 1%; see Arboiro, 1997), but the instantaneouslosses, which can be noticeable, are useful to illustrate the effect of the series con-nection of the cells Figure 13.6 shows the results of a simulation of the instanta-neous performance of the EUCLIDESTMif installed in Madrid for the worst case

trian-hh

series parallel

SC series

SC parallel

I I

,

SC series SC OC series OC series

SC parallel SC OC parallel OC parallel

(the time of the year when the beam deformation and the sun position make the

pointing errors maximum) Figure 13.6a shows the photocurrent I L of the cellsalong the concentrators, as a function of the cell position and for several wind

speed values The photocurrent of each cell I L is proportional to the optical ciency of the mirror that illuminates the cell Figure 13.6b shows the losses intro-duced by the nonlinear effects of the series connection as a function of the windspeed Note that this graph shows only the losses associated with the series con-nection (it does not include the decrease of the average of the optical efficiency).This means if the EUCLIDESTM were solar thermal, its corresponding value inFigure 13.6b would be 100% independently on the wind speed Note that for thecase of no wind, although there is up to ±5% of photocurrent dispersion, only a 2%loss should be expected However, if the wind has a speed of 30 km/h, the minimumphotocurrent reaches the 62% of the nominal value, and the series connection effi-ciency drops to 79% In practice, this low value would be pessimistic because eachone of the 138 modules has a bypass diode, which keeps the series connection effi-ciency at 83% These bypass diodes (not mentioned before) introduce another non-linear effect to provide a way for the current to flow skipping the low photocurrentmodules

effi-This difference between solar thermal and photovoltaics is important becauseeffects to basic concepts not always clearly identified For instance, in solar

thermal it is usual to use the concept of effective sun This consists of analyzing a

real system as perfect but transferring its different imperfections (mirror profileand scattering, tracking errors, structural misalignments, etc.) to the sun shape(Rabl, 1985) This model, which is correct for solar thermal, it is not correct for

Figure 13.5 EUCLIDES TM

photovoltaic concentrator plant installed in Tenerife, Spain (Courtesy of ITER, BP, and IES-UPM)

photovoltaics when the aforementioned nonlinear effects of series connection take

place

Another direct consequence of all this is the definition of the acceptance angle.

Photovoltaic concentrators with series-connected cells cannot be used effectively

for the angular positions of the sun where the optical efficiency for some

individ-ual cells is low, whereas in solar thermal systems this can be allowed because the

optical efficiency is averaged along the receiver For example, consider the

afore-mentioned EUCLIDESIM parabolic trough Let us assume that each mirror

has an angular collection curve that can be approximated by a Gaussian curve

with a 50% transmission angle of ±1° It is easy to obtain from the Gaussian curve

formula that the 90% transmission angle is ±0.4° In order to avoid a significant

system performance degradation, standard deviation of the mirror positioning and

structural deformation errors should be kept below 1° for the thermal case

(speci-fically, if these errors are Gaussian distributed, a 5% degradation implies a

stan-dard deviation of 0.65°) However, in the photovoltaic case, probably a better

criteria would be to keep the mirror positioning and structural deformation errors

below ±0.4° during 95% of the operation time (if these errors are Gaussian

dis-tributed, this implies a standard deviation of 1°/4.4 = 0.23°, to be compared with

0.65° for the solar thermal) This is the reason why in photovoltaic concentrators

the acceptance angle is usually defined at 90% transmission, whereas the 50%

transmission angle is used as the definition of acceptance angle in solar thermal

systems

SOLAR THERMAL APPLICATIONS

Table 13.1 shows various solar thermal concentrator types ordered by their

geo-metrical concentration ratio (Cg) In this regard, it is worth pointing out that the

50 60 70 80 90

10 0

Without bypass diodes

Wind speed v (km/h)

Series connection efficiency (%)

With bypass diodes

concentrator neous performance in the worst case (a) Photocurrent of the cells depending on the posi-

instanta-tion along the concentrator and for several selected wind speed values (b) Losses due to the

nonlinear effects of the series connection.

concept of concentration ratio is frequently misapplied As is clear from the ceding discussion, this ratio divides the entrance aperture area by the area of the

pre-absorber For example, the Cg for a trough of diameter D with tubular absorber

of radius r is D/2pr, and not D/2r The latter definition is sometimes quoted for

parabolic troughs—perhaps because it gives a higher number—but it is not thethermodynamically correct definition We will discuss various concentrator

regimes with Cg as the organizing principle As already noted, we are concerned

with energy efficient concentrators only Systems with low throughput are not ticularly useful for solar applications

par-13.3.1 Stationary Concentrators (Cg < 2)

This category may well be the most important of all because of the practical tages enjoyed by fixed solar systems Recall that the elevation angle of the sunvaries by over 6° over the course of the year, a consequence of the tilt of Earth’saxis of 23° to the plane of the ecliptic (Figure 13.7)

advan-Table 13.1 Implications of availability of solar flux over a range of 1–100,000 suns.

2–100 Nontracking/tracking (linear focus) Power generation (cooling and

heating)

70,000–100,000 Speciality solar furnace Materials, lasers, experiments

S N

Therefore a fixed concentrator would have to accommodate this annual sion in elevation on top of the diurnal east-west variation of nearly 180° Clearly,the diurnal variation indicates a 2D (trough) geometry aligned east-west But what

excur-of the excursion in elevation? At one time, the concept excur-of a fixed concentrator was

considered an oxymoron After all, the Cg of a parabolic trough is 1/p sin q for a

tubular receiver So a parabolic trough aligned east-west that accepts, say, 60° in

elevation would have a concentration ratio Cg = 2/p Therefore, such an

arrange-ment actually deconcentrates To be sure, the geometry of the involute was known

to the ancient Greeks, as was the geometry of the parabola The involute in principle would have a concentration ratio approaching 1 but, in practice, with a

necessary space or gap between absorber and reflector, Cg falls short of 1 With

the advent of nonimaging optics and the recognition of the sine law of tion the situation changed dramatically Fixed solar concentrators were developedand applied to thermal uses where the temperature requirements exceeded that

concentra-of the flat plate collector A number concentra-of CPC collectors for midtemperature use arenow manufactured (Figures 13.8 and 13.9) A particularly attractive use is thermaldriven absorption cooling In this application, rooftop solar collectors convert solarradiation that would otherwise warm up the building into heat at temperatures,say, 115°C to 180°C (Figure 13.10) The low end of the temperature range can drive

a “single effect” absorption chiller, which can be characterized as quasi-static Theyshow a coefficient of performance (COP) of about 0.7–0.8 COP is the ratio of heatremoved by the chiller divided by heat supplied to the chiller Thermodynamically,this ratio can exceed 1 The upper end is sufficient to drive a “double-effect” chillerwith a COP of about 1.1–1.2 The heat drives an absorption chiller that actively

Figure 13.8 CPC “flat plate” collector for midtemperature.

cools the building, a double benefit At the low end, the solar collector’s elementscan be nonevacuated, whereas at the upper end, vacuum insulation is needed.When triple effect machines become commercially available, with driving temper-atures in the 200°C range, solar cooling systems with stationary concentrators will

be even more efficient One can look forward to the day when the COP sates for collector losses so that the amount of heat removed matches the heat collected

compen-13.3.2 Adjustable Concentrators (Cg = 2–10)

Narrowing the acceptance angle from 30° to, say, 6° still allows for stationary ation at least during the day The required frequency of adjustments goes up

oper-rapidly with Cg Thus, Cg = 3 needs only biannual adjustment, whereas Cg = 10

requires almost daily updates Thus, systems of this kind can be characterized asquasi-static Such collectors and systems have been successfully operated forthermal and PV applications on both prototype and commercial scales An indus-trial process heat installation in Israel using seasonally adjusted nonevacuatedCPCs is shown in Figure 13.11 A project for heating in a very cold ambient used

a Cg = 3 nonevacuated CPC This was a school building on a Navajo reservation

in Breadsprings, New Mexico, USA, that is over 2,000 m elevation near the

Figure 13.9 Vacuum tube CPC collector for mid- to high temperature.

13.3 Nonimaging Concentrators for Solar Thermal Applications 331

Figure 13.10 Solar cooling using evacuated CPCs driving double-effect absorption chiller.

Continental Divide, which required only winter heating so that the collectors could

be stationary and oriented for winter use Figure 13.12a shows a schematicdiagram of the collector, and Figure 13.12b shows the collector array on the roof

13.3.3 Tracking Concentrators, One Axis (Cg = 15–70)

In this category, the most widely deployed type of solar concentrator is the bolic trough, which is not a nonimaging solution A good example is the solar powerplant at Kramer Junction, California, USA (Figure 13.13) The modern parabolictrough technology was developed in the United States in the 1970s and early 1980swith extensive testing and demonstration at Sandia Laboratory in New Mexico.The technology was commercialized by the LUZ company in Israel, now succeeded

para-by the Solel Corporation The parabolic mirrors are back-silvered glass for highreflectivity and durability The receiver heat collecting element (HCE) is stainlesssteel tubing coated with a highly selective metallo-ceramic (cermet) material andhoused in an evacuated borosilicate glass tube The glass is chemically treated toreduce surface (Fresnel) reflections The working fluid is heated to 400°C so thatthe power block is a conventional steam turbine The backup fuel to the solar heat

Figure 13.11 CPC array (nonevacuated) for process heat.

input is natural gas This system represents the largest-scale implementation byfar of solar electric power generation.

Nonimaging design approaches in this concentration range have been mainlyfocused on the design of CPC-type secondary concentrators for the parabolic Thishas been done for two reasons: (1) increasing the concentration (which can increasethe conversion efficiency by raising the operation temperature of the absorbertube) and (2) obtaining a more uniform irradiance distribution on the absorber(which reduces the thermomechanical stress in the tube wall, especially in direct-steam generation systems; see Goebel et al., 1996/1997)

(a)

Receiver

Reflector surface

Foam

Exterior housing

Low-iron glass-cover ( 1/8 inch)

(b)

Figure 13.12 CPC (nonevacuated) collectors for winter heating and hot water.