DESIGN, CONSTRUCTION AND TESTING OF A PARABOLIC TROUGH SOLAR COLLECTOR FOR A DEVELOPING-COUNTRY APPLICATION Michael Brooks Ian Mills Department of Mechanical Engineering Mangosuthu Technikon PO Box 12363, Jacobs, 4026 South Africa [email protected]

Thomas Harms Department of Mechanical Engineering University of Stellenbosch Private Bag X1 Matieland, 7602 South Africa [email protected]

ABSTRACT

2.

A parabolic trough solar collector similar in size to smaller-scale commercial modules has been developed for use in a South African solar thermal research programme. An appropriate mix of advanced and less sophisticated technologies was employed during construction. The collector length is 5 m, aperture width is 1.5 m and rim angle is 82.2°. The surface consists of stainless steel sheets covered with SA-85 film. Two receivers were fabricated for comparative testing, including one enclosed in an evacuated glass cover. An optical error analysis was conducted to estimate the intercept factor. The tracking system employs an electric motor, gearbox and rotary encoder for angular feedback. The PSA Algorithm is used to determine the solar vector for tracking purposes. Low-temperature testing of the PTSC was conducted according to the ASHRAE 93-1986 (RA91) standard. Peak efficiencies of 55.2 % and 53.8 % were obtained with the unshielded and glass-shielded receivers respectively.

2.1 Collector structure

1.

INTRODUCTION

Many developing countries with high levels of solar radiation are well placed to implement solar thermal technologies for bulk power and process heat generation. Examples are the proposed South African central receiver power station and concentrating solar power (CSP) projects in India, Egypt, Morocco and Mexico 1. In addition, a small number of PTSCs have been designed, constructed or tested for research purposes in developing regions 2,3,4,5. The aim of this study was to design, construct and test a PTSC module to serve as a test-bed for collector components in an ongoing research program at Mangosuthu Technikon, a South African University of Technology.

DESIGN AND CONSTRUCTION

Factors considered in the construction of the parabolic trough solar collector (PTSC) included stability and accuracy of the parabolic profile, optical error tolerance, method of fabrication, cost, material availability and strength constraints. Güven et al.6 proposed a PTSC design approach that differentiates between developed and developing nations, where design objectives are not limited to maximising thermal efficiency but must also favour cheaper, labourintensive design and production techniques. Elements of this approach, which partitions the PTSC design problem into a macro-level stage dealing with the reflector, receiver and tracking system and a micro-level stage in which the subsystems are integrated, were employed in this study. Specifically, universal error parameters were used in the design of the receiver. Deviations included pre-selecting the rim angle based on parabolic-rib material constraints and selection of the receiver glass envelope diameter based on availability of glass tubing. Like many developing countries, South Africa exhibits a mix of high-technology and less advanced manufacturing capabilities. The construction of the collector reflected this. A torque-tube structure was chosen as it represented an achievable solution in terms of accessibility, accuracy, ease of fabrication and cost. The collector has a length of 5 m, aperture width of 1.5 m and a rim angle of 82.2°. Seven lightweight parabolic ribs were machined from 25 mm thick polypropylene sheet using a computerised high-pressure abrasive water-jet (AWJ) process. These were integrated with an off-the-shelf aluminium tube of outer diameter 101.6 mm and wall thickness 6.4 mm. Standard aluminium flat-bar sections provided additional bracing (Fig. 1). The accuracy of the parabolic ribs and the method of assembly reduced the time and cost of

fabrication and minimised the degree of technical skill required by workshop staff, while providing sufficient structural strength.

2.2 Optical error analysis The optical efficiency of a PTSC is the ratio of solar energy that falls on the surface of the absorber tube, to that which falls on the reflective surface of the collector. It is commonly given as: (1) ηo = Kατ(θi)[(τα)eργ]n The intercept factor γ contains the effects of all optical errors 8, which must be minimised. The intercept factor may be obtained from a closed form expression given by Güven and Bannerot 9: (2) γ =

1 + cosΨrim 2sin Ψrim

Ψ ∫ 0 rim [erf {M} − erf {N}]

dΨ

(1 + cos Ψ )

where: Fig. 1: PTSC during construction with torque-tube, parabolic ribs and bracing members Thermally sagged glass mirror segments are durable but are complex to make, expensive and seldom available for small-scale projects in developing country environments. Highly specular, vapour-deposited aluminium, of the type under development by Alanod and tested by Fend et al.7, was the preferred choice for this PTSC, although as an interim measure SA-85 aluminised acrylic film was used. This was applied to a stainless steel substrate, which was clamped into the profile formed by the parabolic ribs (Fig. 2). The clamping system was designed to allow easy swapping of surface sheets for comparative testing of different materials, ensuring flexibility of the test-bed.

M=

(

)

sin Ψrim (1 + cos Ψ ) 1 − 2d* sin Ψ − πβ* (1 + cos Ψrim ) 2 πσ (1 + cos Ψrim ) *

sin Ψrim (1 + cos Ψ )1 + 2d* sin Ψ  + πβ* (1 + cos Ψrim )   N= * 2 πσ (1 + cos Ψrim )

Estimates of the error parameters σ, β and (dr)y were obtained by modifying developing country values given in the literature (see Table 1): TABLE 1: COMPARISON OF ERROR PARAMETERS Parameters given by Güven et al.6

Reflective surface

Fig. 2: Reflective surface edge clamps

Estimated for this PTSC

#1

#2

σ [rad]

0.0064

0.0087

0.0113

0.0113

β [deg]

0.25

0.50

1.00

0.375

(dr)y [mm]

3.10

6.20

6.20

7.75

D [mm]

24.80

24.80

24.80

30.00

C

28.00

21.00

16.00

15.92

σ* [rad]

0.1792

0.1827

0.1808

0.1798

β* [rad]

0.1222

0.1833

0.2793

0.1111

d*

0.125

0.250

0.250

0.2583

Parameter

Edge clamping Clamping bracket Compressing bolt

Developing country

U.S. Hightech PTSC

Intercept Interceptfactor, factor γ

Fig. 3: Intercept factor γ as a function of β* and d* for σ* = 0.1798

1

50

0.9

45

0.8

40

0.7

35

0.6

30

0.5

25

0.4 0.3 0.2 0.1

The optical error analysis was also used to evaluate the choice of collector rim angle. Eq. (2) was solved for a range of rim angles from 50° to 130°. Fixing β* and σ* according to their values in Table 1, γ was determined as a function of d* for increasing ψrim (Fig. 4).

0 0

20

(dr)y (mm)

5

15

C

3.10 6.20 7.75 9.30 15.50 21.70

10

Ψrim = 82° σ* = 0.1798 β* = 0.2583 10

15

20 25 30 35 Receiver Diameter D (mm)

Concentration Concentration ratio, Ratio C C

A higher rim angle would have had little positive effect on γ and increased the depth of the parabolic ribs, adding weight to the collector and increasing cost. The effect of varying D on γ and C is shown in Fig. 5. By setting D between 27 mm and 32 mm, γ was kept above 80 % and the concentration ratio above 15. A degree of safety was also ensured since the curve for (dr)y = 7.75 mm is relatively flat over the given range of D. An absorber tube of diameter 28.6 mm was therefore chosen based on available copper tubing, giving a final concentration ratio of 16.7.

γ

The solution of Eq. (2) is shown graphically in Fig. 3, which illustrates the degradation of γ with increasing σ*, β* and d*. The graph is drawn for a rim angle of 82°. The intercept factor for the given error parameters was 0.823.

5 40

45

0 50

Receiver diameter, D (mm)

Fig. 5: Intercept factor as a function of receiver diameter for fixed σ* and β* and varying (dr)y 2.3 Receiver design

1 0.9

Two receivers were fabricated, one an unshielded copper tube and the other enclosed in an evacuated glass shield (Fig. 6).

σ* = 0.1798 β* = 0.2583

0.8

Intercept Interceptfactor, factor γγ

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

ψrim 50° 70° 82° 90° 110° 130° 0.2

0.4

0.6

0.8

1

d* d*

Fig. 4: Intercept factor as a function of d* for fixed β* and σ* and six rim angles Clearly, higher rim angles give better intercept factors, but the closeness of the curves from 80° to 130° indicates that for the errors assumed here, the range of optimum rim angles is wide and includes the chosen value of 82°.

Fig. 6: Glass-shielded receiver under test with vacuum hose and gauge

TABLE 2: PTSC KEY FEATURES Collector dimensions

5.0 m x 1.5 m

Rim angle

82.2°

Absorber diameter

28.6 mm

Concentration ratio

16.7

Intercept factor

0.823

PTSC surface reflectance

0.83

Absorptance (coating)

0.88

Emittance (coating)

0.49

Glass shield dimensions

Length: 5380 mm Outer diameter: 44 mm Wall thickness: 2.3 mm

Glass shield transmittance

0.92

Optical efficiency (unshielded receiver)

0.601

3.

PTSC PERFORMANCE RESULTS

The PTSC was tested according to the ASHRAE 93-1986 (RA 91) standard 12. Tests were conducted to determine the collector time constant, thermal efficiency, collector acceptance angle and incidence angle modifier using both the unshielded and glass-shielded receivers. The heat transfer fluid was water circulated at 0.3 m3/h to ensure turbulent conditions in the absorber and a reasonable temperature rise for the expected levels of irradiance. The extensive test program was conducted under controlled conditions at STARlab, Mangosuthu Technikon’s Solar Thermal Applications Research Laboratory 13. At 0.3 m3/h, the cooling time constants were 30.5 s and 28.6 s for the unshielded and glass-shielded receivers respectively. Fig. 7 illustrates the exponential nature of the parameter’s relationship with flow rate. 120 Glass shielded 100 Time constant (s)

A selective coating was applied to both absorber tubes to reduce emittance and increase absorptance. The shield was fabricated by a chemical glass blower using four lengths of Schott Duran borosilicate glass tubing with transmittance of 0.92 and refractive index of 1.473. The annulus was sealed using temperature-resistant o-rings allowing for expansion of the inner absorber. A single vent enabled evacuation of the annulus using a vacuum pump. Table 2 gives a summary of PTSC key features.

Unshielded

80 60 40 20 0 0

200

400

600

800

Flow rate (L/h)

Optical efficiency (glass-shielded receiver)

0.553

2.4 Tracking system Drive hardware consists of a 0.25 kW, 685 rev/min, 8pole AC motor with electromechanical brake and a 463:1 high-reduction helical gearbox. The addition of a Siemens variable speed drive (VSD) enabled trough rotational speeds below 1.5 rev/min. The control hardware consists of a 2500-pulse rotary encoder mounted on the shaft of the PTSC to provide angular feedback information, and a programmable logic controller (PLC). The alignment of the PTSC is along a true north-south axis and tracking is exercised via PLCcontrol of the VSD. Three methods of control were available: manual jogging of the collector, fixed-rate angular corrections based on the sun’s apparent motion (0.25 °/min) and virtual tracking using solar angle information from the implementation of the PSA Algorithm 10. Rotational accuracy of the system was found to be 0.144°. The tracking system is described in detail by Naidoo 11.

Fig. 7: Collector time constant at increasing flow rates The ASHRAE 93-1986 standard requires efficiency tests to be run for a period equal to one time constant or 5 min, whichever is larger 12. The duration of all subsequent thermal efficiency tests was therefore set at 5 min. To determine thermal efficiency, the temperature rise across the receiver (∆t) was measured for a range of inlet temperatures. In this study the temperature was restricted to 85 °C due to limitations in the fluid circulation system. Linear models of thermal efficiency of the type described by Duffie and Beckman 14 and Stine and Harrigan 15 were then imposed on the data according to Eq. (3). The efficiency curves for both receivers are shown in Fig. 8 and the described mathematically in Table 3. Peak values of 55.2 % and 53.8 % were obtained for the unshielded and glass-shielded receivers respectively. (3)

A U F ηg = −  r L R  Ag 

 ∆ t    G bp 

   +  Aa   Ag  

 F η  R o 

0.65 Unshielded data Glass dat a Unshielded best fit Glass best fit

0.60

Best-fit curves for Kατ(θi) are presented in Fig. 9, while the resulting equations are given in Table 3. Fig. 10 shows the PTSC under test.

0.50

1.1

0.45

1.0

0.40 0.35 0.30 0

0.02

0.04

0.06

∆t/Gbp (m2K/W)

Incident angle modifier, Kατ

Thermal efficiency, ηg

0.55

0.9 0.8 0.7 0.6 Unshielded best fit

0.5

Fig. 8: Thermal efficiency data and best-fit curves for shielded and unshielded receivers

Glass best fit

0.4 0

5

10

15

20

25

30

35

40

45

50

55

60

Angle of incidence (deg)

At lower temperatures heat loss is low and the glass adversely affects performance by reducing optical efficiency. At higher temperatures performance is dominated by heat loss, which is prevented more effectively by the shielded receiver.

Fig. 9: Incidence angle modifier curves obtained by regression analysis for shielded and unshielded receivers

The acceptance angle results for the unshielded and glassshielded receivers were 0.43° and 0.52°. In both cases the angular tracking accuracy of the PTSC was exceeded, ensuring that the collector operated within 2 % of its optimal efficiency at all times. The incidence angle modifier, Kατ(θi), is calculated using Eq. (4), where the denominator represents the peak efficiency at normal incidence and the numerator the measured efficiency at a set value of θi. (4) K ατ =

ηg

(A a / A g ) FR [(τα) e ργ ]n

Fig. 10: PTSC under test

TABLE 3: PTSC PERFORMANCE SUMMARY

Time constant at 0.3 m3/h (s) Thermal efficiency

Unshielded receiver

Glass-shielded receiver

30.5 (cooling test)

28.6 (cooling test)

ηg = – 2.0099(∆t/Gbp) + 0.5523 ηo = 0.601

ηg = – 1.0595(∆t/Gbp) + 0.5381 ηo = 0.553

Acceptance angle (deg)

0.43

0.52

Incidence angle modifier

Kατ = – 2.032 x 10-6 (θi)3 + 1.199 x 10-4 (θi)2 – 3.940 x 10-3 (θi) + 1.005

Kατ = 9.360 x 10-7 (θi)3 – 1.616 x 10-4 (θi)2 + 1.061 x 10-3 (θi) + 1.009

4.

CONCLUSION

The performance of a research PTSC, constructed using a mix of advanced and less sophisticated technologies, has been fully characterised for two types of receiver. Although the thermal efficiencies are lower than state-ofthe-art modules, the collector shows good potential as a source of process heat in developing countries. The system would benefit from further optimisation prior to commercial deployment. 5.

NOMENCLATURE

Aa Ar Ag C D (dr) y d* FR Gbp Kατ UL

Collector aperture area Receiver area Gross collector area Geometric concentration ratio Absorber tube diameter Receiver mislocation Universal nonrandom error parameter due to receiver mislocation Heat removal factor Beam in plane irradiance Incidence angle modifier Heat loss coefficient

Greek symbols α Absorptance β Reflector misalignment and tracking error angle β* Universal nonrandom error parameter due to angular errors γ Intercept factor ∆t Receiver inlet temperature minus ambient ηg Thermal efficiency ηo Optical efficiency θi Angle of incidence ρ Reflectance σ Random optical error σ* Universal random error parameter τ Transmittance ψ Angle between parabolic axis and point on mirror surface ψrim Rim angle 6.

ACKNOWLEDGEMENTS

The authors wish to thank Eskom (TESP), USAID, AfricaBlue, Autodesk, the University of KwaZulu-Natal (UKZN) and the Norwegian University of Science and Technology (NTNU) for their support. 7.

REFERENCES

(1) Sargent & Lundy LLC Consulting Group, Assessment of parabolic trough and power tower solar technology cost and performance forecasts, NREL/SR-

550-34440, National Renewable Energy Laboratory, 2003 (2) Thomas, A., Simple structure for parabolic trough concentrator, Energy Conversion Management, 1994, Vol. 35, No. 7, pp. 569-573 (3) Kalogirou, S., Parabolic trough collector system for low temperature steam generation: design and performance characteristics, Applied Energy, 1996, Vol. 55, pp. 1-19 (4) Almanza, R., Lentz, A. and Jiménez, G., Receiver behavior in direct steam generation with parabolic troughs, Solar Energy, 1997, Vol. 61, No. 4, pp. 275-278 (5) Bakos, G. C., Adamopoulos, D., Soursos, M. and Tsagas, N. F., Design and construction of a line-focus parabolic trough solar concentrator for electricity generation, Proceedings of ISES Solar World Congress, Jerusalem, 1999 (6) Güven, H. M., Mistree, F. and Bannerot, R. B., A conceptual basis for the design of parabolic troughs for different design environments, Journal of Solar Energy Engineering, 1986, Vol. 108, pp. 60-66 (7) Fend, T., Jorgensen, G. and Küster, H., Applicability of highly reflective aluminium coil for solar concentrators, Solar Energy, 2000, Vol. 68, No. 4, pp. 361-370 (8) Güven, H. M. and Bannerot, R. B., Determination of error tolerances for the optical design of parabolic troughs for developing countries, Solar Energy, 1986, Vol. 36, No. 6, pp. 535-550 (9) Güven, H. M. and Bannerot, R. B., Derivation of universal error parameters for comprehensive optical analysis of parabolic troughs, Journal of Solar Energy Engineering, 1986, Vol. 108, pp. 275-281 (10) Blanco-Muriel, M., Alarcón-Padilla, D. C., LópezMoratalla, T. and Lara-Coira, M., Computing the solar vector, Solar Energy, 2001, Vol. 70, No. 5, pp. 431-441 (11) Naidoo, P., Intelligent control and tracking of a solar parabolic trough, DTech dissertation, Nelson Mandela Metropolitan University, 2005, In preparation (12) ANSI/ASHRAE 93-1986 (RA 91), Methods of testing to determine the thermal performance of solar collectors, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., 1991 (13) Brooks, M. J., The development and impact of an outdoor solar thermal test facility, In preparation, 2005 (14) Duffie, J. A. and Beckman, W. A., Solar engineering of thermal processes, John Wiley & Sons, Inc., United States, 1991 (15) Stine, W. B. and Harrigan, R. W., Solar energy fundamentals and design with computer applications, John Wiley & Sons, Inc., United States, 1985