You are on page 1of 6

Applied Thermal Engineering 102 (2016) 695–700

Contents lists available at ScienceDirect

Applied Thermal Engineering


journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Design and optical analysis of the band-focus Fresnel lens solar


concentrator
Gang Wang a,⇑, Zeshao Chen b, Peng Hu b, Xiaofang Cheng b
a
School of Energy and Power Engineering, Northeast Dianli University, Jilin, Jilin 132012, China
b
University of Science and Technology of China, Hefei, Anhui 230027, China

h i g h l i g h t s

 A new kind of band-focus Fresnel lens solar concentrator was proposed.


 The design principle of band-focus Fresnel lens concentrator was given.
 Optical analysis of this Fresnel lens concentrator showed a good concentrating uniformity.

a r t i c l e i n f o a b s t r a c t

Article history: Solar energy is one of the most promising renewable energies and meaningful for the sustainable devel-
Received 27 November 2015 opment of energy source. In this paper, a new kind of band-focus Fresnel lens solar concentrator was pro-
Revised 5 April 2016 posed. The design principle of this solar concentrator was given and the spectral concentrating
Accepted 8 April 2016
performance was simulated by the means of Monte Carlo Ray Tracing Method (MCRT), which was com-
Available online 8 April 2016
pared with the linear Fresnel lens. The results show that both the spectral concentrating uniformity and
optical efficiency of the band-focus Fresnel lens were better than those of the linear one. Meanwhile sev-
Keywords:
eral characteristic parameters of the band-focus Fresnel lens concentrator were analyzed under different
Solar energy
Fresnel lens
conditions and it can be drawn from the results that a high-ratio band-focus Fresnel lens concentrator
Band-focus could increase the optical efficiency of a concentrating photovoltaic (CPV) system.
CPV Ó 2016 Elsevier Ltd. All rights reserved.
Renewable energy

1. Introduction uniformity, but because of the design limitations, this new lens
must face the incident rays strictly when it is used in PV systems.
Solar energy is one of the most promising renewable energies Otherwise, even a very small deviation could make the concentrat-
and meaningful for the sustainable development of energy source. ing light deviate from solar cells and decrease the total efficiency
As the solar radiation energy flux density is low on the surface of greatly. So a precise two-axis sun-tracker is necessary for this kind
the earth, the spectral concentrating can be used in the solar of Fresnel lens.
energy photovoltaic utilization process to reduce costs. Fresnel This paper proposes a new kind of Fresnel lens concentrator,
lens is a common kind of spectral concentrator, which is used based on the linear Fresnel lens, which could focus incident rays
not only in PV systems [1–3] but also in solar thermal systems to a uniform solar flux band. The optical simulation and analysis
[4–6]. Traditional Fresnel lenses make lights focus to a point or a of the band-focus Fresnel lens were carried out by the means of
line with wedges which are distributed on a plane or a curved sur- Monte Carlo Ray Tracing Method, aiming at the investigation of
face [7,8]. When they are used in PV systems, the energy flux den- the spectral concentrating performance of this new concentrator.
sities on solar cells are very non-uniform and that reduces the
efficiency of solar modules [9,10].
For this problem, Ryu et al. improved the point-focus Fresnel 2. Concentrator design
lens and made rays through the new lens focus to a square which
had the same size as a solar cell [11]. That improves the spectral In order to make incident rays focus to a band which is pre-
sented in Figs. 1 and 2, the design principle was given as follows:
(1) The horizontal length of the Fresnel lens d was set to be odd
⇑ Corresponding author. times of the width of a solar cell w, and the ratio N was equal to
E-mail address: kinggang009@163.com (G. Wang). d/w. The horizontal length was divided to N units. (2) In these N

http://dx.doi.org/10.1016/j.applthermaleng.2016.04.030
1359-4311/Ó 2016 Elsevier Ltd. All rights reserved.
696 G. Wang et al. / Applied Thermal Engineering 102 (2016) 695–700

where sj and lj were the energy flux transmission efficiency and


tooth width of Wedge j, respectively. The ideal optical efficiency
of the lens was expressed as following:
g ¼ s1 s2 ð4Þ
The optical efficiency of lens under the condition which the rays
were not normal incident was analyzed later in this paper.

3. Method and calculation model

Monte Carlo Ray Tracing Method (MCRT) was employed for the
spectral concentrating simulation of the band-focus solar concen-
trator. It was a statistical method of tracking the random process
of a large number of beams. The calculation process of MCRT
was: Assuming that solar radiation energy was carried by a lot of
beams evenly, every beam would experience many optical pro-
Fig. 1. Spectral concentration effect of band-focus Fresnel lens. cesses including reflection, refraction, absorption and scattering.
Whether these optical processes occurred or not were controlled
by random numbers. By tracking the propagation processes of
these beams, the energy flux density distribution on the radiation
absorbing surface could be obtained [13].
Assuming that incident rays distributed evenly on the incident
plane of the band-focus Fresnel lens and this plane was considered
as the emitting surface of sampling beams, the probability model
of the emitting point (x0, y0, z0) was:
8
< x0 ¼ xL  RX  xL =2
>
y0 ¼ f ð5Þ
>
:
z0 ¼ zL  RZ
where RX ; RZ 2 ð0; 1Þ were the random numbers of x axis and y axis
respectively, and xL, zL and f were the x axis length, y axis length and
focal length of the band-focus Fresnel lens respectively. The non-
Fig. 2. Design of the horizontal orientation of band-focus Fresnel lens. parallel angle of solar radiation was 320 , so the incident rays was
considered as a light cone with ha = 160 . In this cone, the solar
units, the center one was made as a plane without wedges. Other energy distribution followed the Lambert law which was that the
units all have the same number of wedges and the angles u of directional radiation intensities were the same. So we could obtain
every wedge which were in the same unit should be the same. the zenith angle and circumferential angle perpendicular to the
If the index of refraction of the lens material was n, the angle u incident light direction as follows [14]:
of a wedge could be calculated as following: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  h ¼ arcsin Rh  sinðh2a Þ ð6Þ
sin c
u ¼ arctan ð1Þ
n  cos c u ¼ 2pRu ð7Þ
where c was the light deviation angle of this wedge, which can be where Rh ; Ru 2 ð0; 1Þ were the random numbers of cone angle and
seen in Fig. 2. When the wedge angles of the Fresnel lens were circumference along the emitting direction. The direction vector of
given, the overall size of the lens would be obtained according to the normal incident light P was:
the size of solar cells and the focal length f. In this paper, the size
P ¼ ðsin h sin u;  cos h; sin h cos uÞ ð8Þ
of solar cells was assumed to be 25 mm  125 mm, f would be
changed according to different conditions. and the direction vector of the refraction light A was:
The ideal optical efficiency was an important characteristics
A ¼ P þ CN ð9Þ
parameter of the band-focus Fresnel lens. When the light was nor-
mal incident, the energy flux transmission efficiency should be where N was the normal direction vector on the incident point of
[12]: the refraction surface, and C was a constant. If the incident angle
was a, the refraction angle was b, C would be:
4n
s1 ¼ ð2Þ
C ¼ n cos b  cos a ð10Þ
ðn þ 1Þ2
As shown in Eq. (2), the energy flux transmission efficiency was When the rays were not normal incident, the direction vector P⁄
only related to the index of refraction of lens material when the could be obtained by using a transformation matrix U:
light was normal incident. P  ¼ UP ð11Þ
The energy flux transmission efficiency of the light output sur-
face of the band-focus Fresnel lens was equal to the weighted aver- where transformation matrix was:
2 3
age of the energy flux transmission efficiencies of all wedges and u11 u12 u13
the center plane [12]: 6 7
U ¼ 4 u21 u22 u23 5 ð12Þ
!,
X
M
u31 u32 u33
s2 ¼ sj lj þ ws1 d ð3Þ
1
G. Wang et al. / Applied Thermal Engineering 102 (2016) 695–700 697

8
> u11 ¼ cos hy cos hz focus Fresnel lens were both better than those of the linear one
>
>
>
> u12 ¼ sin hx sin hy cos hy þ cos hx sin hz under the same condition.
>
>
>
>
>
> u13 ¼ sin hx sin hz  cos hx sin hy cos hz
>
>
>
> ¼  cos hy sin hz 4.2. Research on the relationship of CR, CR0 , f and g
>
< u21
u22 ¼ cos hx cos hz  sin hx sin hy sin hz ð13Þ
>
> For band-focus Fresnel lens with different focal lengths, the
>
> u23 ¼ cos hx sin hy sin hz þ sin hx cos hz
>
> relationships of theoretical spectral concentrating ratios, actual
>
>
> u31
> ¼ sin hy spectral concentrating ratios and ideal optical efficiencies were cal-
>
>
>
> u32 ¼  sin hx cos hy culated, and results are presented in Figs. 6 and 7.
>
>
: As shown in Fig. 6, the maximal theoretical spectral concentrat-
u33 ¼ cos hx cos hy
ing ratio increased with the focal length increasing. Thus a high-
ratio Fresnel lens should have a big focal length.
hx, hy and hz were the three angles between the incident light and
According to Fig. 7, when f was fixed, g decreased as the spectral
three axes, respectively. By this way, incident rays, refraction rays
concentrating ratio increased. And the decreasing rate decreased if
and output rays could be all ascertained, and the energy flux density
f was increased. So it can be drawn from Fig. 7 that a high-ratio
distribution on the absorbing surface could be also calculated.
Fresnel lens could make better use of solar energy and increase
the optical efficiency of the PV system.

4. Spectral concentrating simulation and analysis


4.3. Effect on the optical efficiency under different incident angles

4.1. Concentrating performance simulation


The analysis of the ideal optical efficiency of band-focus Fresnel
lens above was under the condition which the rays were normal
It is assumed that the base thickness of the band-focus Fresnel
incident, and that needed a precise two-axis sun-tracking. For
lens was 2 mm, the horizontal length of the lens d was 225 mm,
the one-axis sun-tracking (The horizontal direction of the lens
the vertical length was 125 mm, the tooth width of every wedge
was along the north–south and the vertical direction was along
was 1 mm, f was 100 mm, theoretical spectral concentrating ratio
the east–west.), assuming that the north–south tracking was pre-
CR was 9, the lens material was PMMA, n was 1.4935; Meanwhile,
cise, there was an incident angle, which was not equal to 0°, in
a linear Fresnel lens model, which had the same material and
the surface perpendicular to the top surface of the Fresnel lens.
parameters except the wedge angles, was established to compare
The incident angle was b1 and the refraction angle under the top
with the band-focus lens. The energy flux density of incident rays
surface was b2, then the incident angle on the wedge surface could
was assumed to be 1000 W/m2 and the number of the incident rays
be expressed as:
was 106. The width and length of the receiving surface were 25 mm
and 125 mm, which was the same as the size of solar cells. The b3 ¼ arccosðcos u cos b2 Þ ð14Þ
absorbility of the incident surface was assumed to be 100%.
and the final exit angle of output ray was:
The light density distribution performance contrast of the two
kinds of Fresnel lenses is presented in Fig. 3. Fig. 3(a) is the concen- b4 ¼ arcsinðn sin b3 Þ ð15Þ
trating performance of the linear Fresnel lens and Fig. 3(b) is the
one of the band-focus lens. It can be seen from Fig. 3 that both Assuming as follows:
the vertical light flux density distributions of the two lenses were 8
> bþ ¼ b1 þ b2
>
>
uniform, but the horizontal light flux density distribution of the < b ¼ b1  b2
band-focus lens was much more uniform than the linear one. The ð16Þ
horizontal light flux density distribution contrast of the two lenses
>
>
> b0þ ¼ b3 þ b4
: 0
is presented in Fig. 4, in which red1 dotted lines represent the linear b ¼ b3  b4
Fresnel lens and black solid lines represent the band-focus one.
the light flux transmission efficiency of the top surface of the lens
The natural logarithm of horizontal light flux density distribu-
was:
tions on the receiving surface under different focal lengths and
" #,
light deviation angles is presented in Fig. 5. It can be seen that dif- 2
tan2 ðb Þ sin ðb Þ
ferent spectral concentrating ratios for the same light deviation T1 ¼ 1  þ 2 ð17Þ
tan2 ðbþ Þ sin2 ðbþ Þ
angle c had close spectral concentrating uniformities. When f
was fixed, the uniformity under a small CR was better than the and the light flux transmission efficiency of every wedge bottom
one under a big CR. Therefore, the band-focus Fresnel lens concen- surface was:
trator could obtain a better spectral concentrating uniformity by " #,
tan2 ðb0 Þ sin ðb0 Þ
2
increasing the focal length and spectral concentrating ratio
properly. T2 ¼ 1  þ 2 ð18Þ
tan2 ðb0þ Þ sin2 ðb0þ Þ
In this paper, when the focal length was 100 mm, the actual
concentrating parameters of the linear and band-focus Fresnel lens Using Eqs. (17) and (18), the theoretical optical efficiencies of
under different theoretical spectral concentrating ratios were cal- the lens under different incident angles could be calculated. For a
culated, and the results are presented in Tables 1 and 2. Here, Imax, band-focus Fresnel lens whose f was equal to 100 mm and CR
Imin and Ia were the maximal, minimal and average light flux den- was equal to 9, the calculation result of the relationship between
sity on the receiving surface respectively, CR0 was the actual spec- b1 and g is presented in Fig. 8.
tral concentrating ratio, and g0 was the actual optical efficiency. According to Fig. 8, when b1 was between 0° and 26.5°, the opti-
Results in Tables 1 and 2 show that with CR increasing, g0 cal efficiency of the lens was high (bigger than 90%) and had small
decreased. And the optical efficiency and uniformity of the band- changes with b1 increasing. When b1 was just past 26.5°, the opti-
cal efficiency decreased sharply, which was caused by the total
1
For interpretation of color in Fig. 4, the reader is referred to the web version of reflection of the light in the Fresnel lens. Different lens materials
this article. had different critical incident angles, past which the optical effi-
698 G. Wang et al. / Applied Thermal Engineering 102 (2016) 695–700

Fig. 3. The concentrating performance contrast: (a) linear Fresnel lens; (b) band-focus Fresnel lens.

Table 1
Results of the concentrating simulation of the linear Fresnel lens.

CR Imax (W m2) Ia (W m2) Imin (W m2) CR0 g0 (%)


3 5955.06 2574.5 676.78 2.627 87.57
5 7454.72 4276.7 745.39 4.364 87.28
7 9860.23 5585.2 833.20 5.699 81.41
9 9959.88 5956.4 840.15 6.078 67.53

Table 2
Results of the concentrating simulation of the band-focus Fresnel lens.

CR Imax (W m2) Ia (W m2) Imax (W m2) CR0 g0 (%)


3 2735.45 2663.6 1498.34 2.72 90.67
5 4389.91 4321.8 2273.08 4.41 88.20
7 5766.12 5688.9 2917.80 5.81 83.00
9 6410.23 6095.6 3150.14 6.22 69.11

Fig. 4. Light flux density distribution contrast.

Fig. 6. The relationship of CR, CR0 and f.

Fig. 5. Horizontal light flux density distributions on the receiving surface under
different f and c. north–south direction increased from 0°, the concentrating light
on the focal plane would deviate slowly. When b1 increased to a
ciency decreased sharply. In Fig. 8, when b1 increased to 60°, the certain value, the output rays of different units of the lens were
optical efficiency decreased to about 43%. already not able to coincide completely, which made the concen-
For another case of the one-axis sun-tracking, which was the trating light turn wider and non-uniform, and the output rays of
east–west tracking was precise and the incident angle b1 in the some units (for example the outermost units) even separate from
G. Wang et al. / Applied Thermal Engineering 102 (2016) 695–700 699

Fig. 10. Horizontal light flux density distributions under different incident angles.
Fig. 7. The relationship of CR, g and f.

significance of this concentrator. Assuming the base thickness of


the lens was 2 mm, the relationship between b1 and concentrating
light deviation quantity (Dd) on solar cells under different focal
lengths was calculated and the result is presented in Fig. 9. For
the band-focus Fresnel lens whose focal length was 500 mm and
CR was 35, the horizontal light flux density distributions of the
receiving surface under different incident angles are shown in
Fig. 10.
According to Figs. 9 and 10, the concentrating light deviation
quantity increased with f increasing. Though the concentrating
light on solar cells would deviate and turn non-uniform, if the inci-
dent angle was controlled to be very small (for instance, b1 was
smaller than 0.5°), the deviation quantity and concentrating light
uniformity changes caused by the north–south tracking error
would be ignored and the normal use of the CPV system would also
be not influenced. That means this band-focus Fresnel lens solar
concentrator had a more simple sun-tracking requirement than
the Fresnel lens concentrator proposed by Ryu [11].

Fig. 8. The relationship between the incident angle and optical efficiency. 5. Conclusions

In this paper, for the solar energy application, the band-focus


Fresnel lens solar concentrator was proposed. The design principle
was given and the concentrating performance investigation was
carried out by MCRT simulation and theoretical analysis. The spec-
tral concentrating simulation results show that both the spectral
concentrating uniformity and optical efficiency of the band-focus
Fresnel lens were better than those of the linear lens. Meanwhile,
the relationships of several key parameters of the band-focus Fres-
nel lens were analyzed and the results indicate that a high-ratio
band-focus Fresnel lens could make use of solar energy more effec-
tively and increase the optical efficiency of the CPV system. The
effect on the optical efficiency under different incident angles
was also studied. The analysis results show that even a precise
one-axis would meet the application requirement of the solar con-
centrator and very a tiny north–south tracking error would not
influence the concentrating performance of this CPV system.

Acknowledgement

Fig. 9. The relationship between the incident angle and concentrating light The authors appreciate the support of the Natural Science Foun-
deviation quantity.
dation of China (Grant No. 51376167).

the concentrating light. As b1 increased even bigger, the output References


rays of the outermost units would disappear because of the total
[1] D. Chemisana, M. Ibanez, J. Barrau, Comparison of Fresnel concentrators for
reflection and the concentrating light also turn more building integrated photovoltaics, Energy Convers. Manage. 50 (2009) 1079–
non-uniform, which the band-focus Fresnel lens lost the original 1084.
700 G. Wang et al. / Applied Thermal Engineering 102 (2016) 695–700

[2] D. Chemisana, M. Ibanez, Linear Fresnel concentrators for building integrated [9] M. Mijatovic, D. Dimitrovski, V. Veselinovic, Fresnel lens-absorber system with
applications, Energy Convers. Manage. 51 (2010) 1476–1480. uniform concentration and normal incoming rays to the absorber, J. Opt. 18
[3] C. Sierra, A.J.V. Zquez, High solar energy concentration with a Fresnel lens, J. (1987) 261–264.
Mater. Sci. 40 (2005) 1339–1343. [10] Y. Xu, M. Li, L. Wang, W. Lin, X. Zhang, M. Xiang, Y. Wang, S. Wei, The effect of
[4] K.E.J. Al-Jumaily, M.K.A. Al-Kaysi, The study of the performance and efficiency concentrated light intensity on output performance of solar cell arrays, Acta
of flat linear Fresnel lens collector with sun tracking system in Iraq, Renew. Phys. Sin. 58 (2009) 8067–8076.
Energy 14 (1998) 41–48. [11] K. Ryu, J.G. Rhee, K.M. Park, J. Kim, Concept and design of modular Fresnel
[5] H. Zhai, Y.J. Dai, J.Y. Wu, Investigation of concentrating solar photovoltaic/ lenses for concentration solar PV system, Sol. Energy 80 (2006) 1580–1587.
thermal system performance based on Fresnel lens, J. Eng. Thermophys. 28 [12] R. Leutz, A. Suzuki, A. Akisawa, T. Kashiwagi, Design of a non-imaging Fresnel
(2007) 725–728. lens for solar concentrators, Sol. Energy 65 (1999) 379–387.
[6] H. Zhai, Y.J. Dai, J.Y. Wu, R.Z. Wang, L.Y. Zhang, Experimental investigation and [13] J.A. Duffie, W.A. Beckman, Solar Engineering of Thermal Processes, Wiley, USA,
analysis on a concentrating solar collector using linear Fresnel lens, Energy 2006.
Convers. Manage. 51 (2010) 48–55. [14] S. Du, X. Xia, Y. Tang, Numerical investigation on effects of non-parallelism of
[7] F. Duerr, Y. Meuret, H. Thienpont, Miniaturization of Fresnel lenses for solar solar rays on concentrating solar power, Acta Energi. Sol. Sin. 27 (2006) 388–
concentration: a quantitative investigation, Appl. Opt. 49 (2010) 2339–2346. 393.
[8] N. Yeh, Analysis of spectrum distribution and optical losses under Fresnel
lenses, Renew. Sust. Energy Rev. 14 (2010) 2926–2935.