See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269230328

Optical simulation of a solar parabolic collector using ray-tracing software

TracePro

Conference Paper · April 2014

DOI: 10.2495/CCCS20130251

CITATIONS READS
5 3,249

All content following this page was uploaded by Sasa R. Pavlovic on 10 March 2015.

The user has requested enhancement of the downloaded file.

Optical Simulation of a Solar Parabolic
Collector using Ray-Tracing Software
TracePro
*
Sasa Pavlovic, *Velimir Stefanovic ,**Milorad Bojic
*
Department for Energetics and process technique, Faculty of
Mechanical Engineering University of Nis, 18000 Nis, Serbia

**
Faculty of Engineering University of Kragujevac, Sestre Janjic 6,
34000 Kragujevac, Serbia

Abstract
The Parabolic solar concentrator has a diameter of 5 m. It consist of thin shell,
composed of 24 segments placed of two rings. It consist of inner and outer
petals. This segments are glass-fiber reinforced resin sandwich with mirror
applied to its surface. The objective of TracePro is to model complex optical
systems for solar power and to analyse their performance. The optical system is
organized in stages in a global coordinates system: stages are sections of the
optical geometry that are successively hit by rays in their pathway from the sun
to the final receiver. Several methods can be used to model a parabolic
concentrator in TrcePro. This paper presents optical simulations of the Parabolic
solar collector unit using the ray-tracing software TracePro. Segments of
parabolic collectors are made of Polymethyl methacrylate PMMA. A receiver
model is added in the ray-tracing simulations in order to study the influence of
cavity physical parameters. The software is capable of incorporating user –
defined sun shape and optical error and can calculate the amount of power
incident upon the receiver.
Keywords: Parabolic dish collector, ray-tracing simulation, TracePro
software, Monte Carlo method

1. Introduction
Many areas in the world abound in free solar energy while in some other areas
wind and other types of renewable sources of energy impose themselves as a
logical choice Marković [1]. By means of the adequate equipment the energy of
the sun irradiation can be converted into thermal and electrical energy. The
critical energy situation forces the mankind to reconsider thoroughly all
possibilities given by renewable energy sources, and our present knowledge and
technology. Depending on the degree of the heating of the working fluid we can
differ between the low-temperature (T<100oC), middle-temperature
(100oC<T<400oC) and high-temperature conversion (400oC<T<4000oC). For
low-temperature solar energy conversion one uses flat collectors with water and
air, for middle-temperature conversion one uses vacuum colletors and collectors
with concentrators, and for high-temperature one uses solar furnaces and CSP
plants Pavlović [2] and Fernández-García et al. [3]. Middle and high-temperature
systems are applicable for refrigeration systems, industrial processes and
polygenerations systems [4]. The optical efficiency representing the ratio of the
thermal power introduced into the cavity receiver to the solar power intercepted
by the concentrator.
Two mains categories of codes for solar flux calculation can be distinguished:
the first category of codes MIRVAL, FIATLUX and SOLTRACE, HELIOS,
CIRCE etc. are designed to detailed analysis of the optical performances and
give a detailed description of the reflected power from a concentrator. The
second category of codes (UHC, DELSOL, HFLCAL) is dedicated to global
system optimization [5]. This paper presents a performance analysis which
utilizes ray-tracing models of a parabolic dish collector. For instance, the optical
assessment of the concentrators was performed by Chemisana et al.6 with
commercial software OPTICAD, by Thibert et al.7 with commercial software
ASAP, by Mouzouris et al.8 with commercial software Optics Works Studio, and
by Bader et al.9 via Monte-Carlo ray tracing technique [6-9]

2. Solar Dish technology

Solar dish systems convert the thermal energy of solar radiation to mechanical
energy and then to electrical energy similar to the way that convectional power
plants convert thermal energy from combustion of a fossil fuel to electricity.
Solar dish systems use a mirror array, in the shape of a parabolic dish, to reflect
and concentrate incoming direct solar irradiation to a receiver, in order to
achieve the temperatures required to efficiently convert to heat and work. In
order to achieve maximum efficiency, these types of concentrators are mounted
on a structure with two axis tracking system, so that the dish concentrator can
track the sun. The mechanical is converted to electrical power by an electric
generator or alternator. In modern high performance Stirling engines, working
gas temperatures over 7000 C and pressure of 200 bar, are used. The Dish
Stirling systems have demonstrated the highest efficiency of any solar power
generation system by converting nearly 31.25% Taggart [10] of direct normal
incident solar radiation into electricity after accounting for parasitic power
losses. The present parabolic collector was modeled completely in SolidWorks.
The mirrors on a parabolic concentrator need to be exactly aligned in order to
maximize solar energy output from the reflective surfaces to the heat collecting
element (HCE). To achieve maximum efficiency, several parameters of the
system can be varied, including the positions of the mirror and the receiver. The
solar flux distribution on the absorber is a key parameter of the dish solar
collector system. To know these flux peaks, it’s necessary to introduce in the
ray-tracing code the local slope errors and not the total concentrator error. The
behavior of each of these rays of light is then calculated, taking into account
several factors which reduce the total power reaching the receiver. Atmospheric
attenuation causes the power of a ray to decrease as it moves through air and is
modeled in this simulation by a mathematical function. The ideal shape of the
collector is 2 rings of parabola with slightly different focal lengths defined by
Eq.1.
Z  An  X 2  Y 2   Bn (1)
-1
Where An in [m ] and Bn in [m] are two coefficients characterizing each parabola
ring.
In this paper two more important factors affecting the efficiency of a solar
parabolic collectors were added to the existing model. The first was sun shape,
which refers to the fact that the sun is not a point source and does not emit
parallel rays of light. In this model, the sun shape was modeled by a Gaussian
function describing the probability that a ray will come from a certain location
on the sun and therefore make a certain angle with the curved surface normal.
The second factor that was considered in this paper was optical error.
This can be divided into two parts, slope and specular error. Slope error is
associated with the macroscopic shape of the parabolic mirror and occurs if the
reflective mirror is not perfectly parabolically, but slightly roughness with small
imperfections of parabolic mirror surface. Specular error is related to the
microscopic roughness of the heliostat surfaces, parabolic reflectors surface,
parabolic trough, etc.

Fig.1: Schematic of truncated Fig.2: 3D drawing of parabolic dish

parabola used to construct a collector ( D 5 m, f  1.5 m)
truncated circular paraboloid
The two sources of error were first combined using the following equation:

 tot   slope
2
  specular
2
  sun
2
(2)
The  total was then used to define a single Gaussian distribution describing the
probability that a reflected rays leaves a surface mirror with a certain angle of
deviation from the predicted ray.
3. The ray – tracing software TracePro

TracePro optical modeling software was used to assess the ray tracing analysis
of the three proposed solar collection geometries (parabolic trough, CPC
collector, solar dish, Fresnel lens collector, ring – array concentrator) to estimate
the radiation flow distribution arriving to the photorreactor. TracePro is the first
optical analysis program with the industry standard solid modeling engine,
ACIS®, at its core. TracePro performs stray light analysis, illumination analysis,
and optical systems analysis. TracePro accounts for absorption, specular
reflection and refraction, scattering and aperture diffraction of light. Results can
be examined in spatial Irradiance plots, angular radiance plots, contour maps,
Candela plots, or ray histories in tabular form. Monte Carlo ray tracing is the
process of using the principles of geometrical optics as a statistical method to get
a complete and statistically viable analysis an optical system. A ray of light
travels through the system and the path of the ray is defined by the origin of the
ray and the reflection and reflaction along the path to the exit aperture. In the
process of tracing one ray, the starting point is randomized within a certain start
area or volume, the source of the rays. Depending on the properties of the source,
the direction of the ray can be either the same for all rays or randomized in a
pattern defined by the source. It could for example be a Gaussian source or a
completely diffuse source. At the point of intersection , the surface normal is
determined in order to calculate the direction of the ray after interaction with the
system. The ray could be reflected, refracted or absorbed. When the new
direction of the rays has been calculated, the next intersection is calculated. The
process continues until either the ray is lost from the system, completely
absorbed, or intersects with the target that detescts the ray.

3.1 Parabolic collector model in TracePro software

Several methods can be used to model a parabolic concentrator in TracePro:

- The easiest is to define the whole parabolic dish with an 5 m diameter and
a 1.5 m focal length. In this way, only a total slope error can be defined:

- A second method is to define the 2 parabolic rings of the concentrator

using Eq.1. This can be the best way to model the ideal shape of the
concentrator. Furthermore, we can attribute a total slope error for each
parabolic ring,

- The method that we use consists to divide the parabola in small elements
and then, to apply at each element a normal vector depending of the local
slope error.
The advantage of TracePro is that the Monte Carlo method gives reliable results
when enough rays are used. The disadvantage is that it is computationally
expensive.
 Fig.3: 3D geometrical models of solar receivers for paraboloidal collectors

flux peak (12 facets)

40 120
flux peak (24 facets)
Radiation solar power (KW)

110

(% of incoming solar energy)

19 cm intecepted solar energy
30
100

20 90

80
10
70

0 60
4.4 4.45 4.5 4.55 4.6 4.65 4.7
distance Z0 (f conc.= 1.5 m)

Fig.4: SOLTRACE results for two models of Fig.5: Trapezoidal mirror

solar concentrators – Radiation solar power petal of the parabolic
reflector
Figure 4 shows TracePro results to a concentrator total power mirror reflectivity
of 94 % and DNI Direct Normal Insulation of 1 KW/m2  1 sun. The efficiency
of the studied solar parabolic conentrator is the ratio of the solar concentrated
flux density and the solar flux density qconc/incid.

3.2 Monte Carlo Ray – trace model

For solar models, a custom “sun” has to be created within each ray-tracing
program. In TracePro, a circular object was created and then set to emit rays. The
ray-trace simulations used a deterministic modeling approach meaning that the
same results are achieved for every ray trace with identical parameters. A Monte
Carlo method is available in TracePro , but was not feasible for this study due to
the modeling approach taken to simulate a sun source. The scatter model is a
Lambertian model limited by a small scatter cone angle of ~ 0.55° which limits
the transmission of rays through the surface unless a very large number of rays
are traced. The sun source was made to move to its estimated position in the sky
according to the day of the year, latitude, and the hour relative to solar noon.
Althoug the Monte Carlo encopasses a broad range of numerical integration tools
based on random sampling, in the case of concentrating collector analysis there
is a convenient physical analogy: “bundles” of photons are ray-traced as they
enter the collector from randomly-sampled locations over the aperture and
multiple randomized surface-interaction events in until they are either absorbed
at a surface or escape from the collector aperture.
In Monte Carlo ray-tracing method, the average radiation flux on the target q
and the area weighted mean solar concentration Cmean on a circular disk are
defined:
Nt arg et

  q r m
2
m  rm21 
(3)
q m 1

 Rt2arg et

The collection efficiency  c , the flux non-uniformity  f , and the average solar
concentration ratio C are defined as:
Nt arg et

 q A m 1
m m
(4)
c 
G  R  R 
0
2
2
2
1

q is an average radiation flux on the target.  c and the facet size a are
simultaneously maximized while retaining acceptable values of C .
qm is the radiative flux intercepted by the mth discrete target element, G0 is the
incident direct solar flux –direct normal irradiance (DNI). Analytical geometry
and the Monte Carlo ray tracing method are used to analyze the optical
performance of the concentrator and to identify the set of geometric concentrator
parameters that allow for flux characteristics suitable for medium-temperature
applications.
nideal
trapezoidal reflective
mirror petal
nreal

R
T
Y
nideal Z
X Soltrace facet
nreal f
Simulated element

Fig.7: Slope errors in X.Y, Radial and Tangential direction

For each facet an algorithm developed in Matlab calculates a „mean“ normal
vector nreal (fig.7) given by Eq. 4. Then, values are exported to TracePro for
optical simulations Butler [11]. The facet coodinates (X0, Y0, Z0) and the ideal
normal vector coordinates are:
 
 X0   X 
   0 

F Y0  and nideal  Y0  (5)

   
  2

 Z 0  An X 0  Y0  Bn 
2
  1 
 2B 
 n
Moreover, each facet is paraboloidal trapezoid and the focal length given by Eq.
6. is equal to the distance between the parabolic dish focal length and the facet
localization:
fF  X 2
0  Y02   f CONC .  Z 0 
2
 6)
Table 1: Geometrical parameters of the parabolic concentrator
Diameter of opening of the parabola 5 [m]
Depth of the parabola 0.26 [m]
Focal distance f 1.5 [m]
Area [m2]
Ideal area of the concentrator Aidel  22,73 m2
The cross section of the opening parabola Aproj  19,63 m2
A sheltered area of the concentrator Ashadow  0,008 m2
The effective area of the concentrator Aef= AprojAshadow  19,622 m2

Fig.8: Geometrical parameters of Fig.9: Ray – tracing simulation of the

parabolic concentrator parabolic concentrator
The optical performance of the parabolic dish collector is quantified using the
intercept factor of the system. In the raytracing models there are difficulties in
obtaining an accurate intercept factor. The intercept factor was calculatedas
shown in Eq.7.
 Power from the collector hitting the receiver
I .F .  7
Total power from the collctor

For reflective surface the following properties apply:

Poly(methyl methacrylate)(PMMA) is a transparent thermoplastic, often used
as a lightweight or shatter-resistant alternative to glass

Table 2: Material proprerty of pmma - poly (methyl methacrylate (optical mirror

materials):
Characteristics of reflective materials
Specific density 1.19 g/cm3
Apsorption 2%
Light penetration 92-93 %
Refractive index 1.49 %
Transformation temperature 100 0C
Linear expansion coefficient 6.8 x 10-5 cm/cm0C

Coating layer: Alanod Miro 2 (Rspec = 90 %. Rdiff = 5%, (BRDF A = 0.01751,

BRDF B = 0.1, BRDF g = 0.0), and A =5%). A, B and g are scatter coefficient in
the Soltrace ABg model.
Bidirectional reflectance distribution function (BDRF) is a four – dimensional
function that defines how light is reflected at an opaque surface [7]
Many phenomenological and analytic models have been proposed including the
Lambertian reflectance model frequently assumed in computer graphics. Some
useful features of recent models include:
 accomodating anisotropic reflection
 editable using a small number of intuitive parameters
 accounting for Fresnel effects at grazing angles
 being well-suited to Monte Carlo methods
The Monte Carlo ray-tracing method is applied to compute the radiation flux
qm .The only source of radiation is the incident solar flux in the concentrator
aperture plane. The angular distribution of the launched rays is modeled by the
“pillbox-sunshape”,which represents the ideal solar disk with uniform brightness
in half-cone angle sun . The angular distribution of the launched rays can be
obtained using Kuiper’s distribution Zhiqiang [12].

  2
   
  1   1   
    sun  
  
I     I  0  , 0     sun (8)
 2  2 
 sun  1   
  3 

 I cs   ,  sun    
2
where I   is the total intensity of the incident solar radiation,  is the angle
between the direction towards the center of the sun disk and the direction
towards the actual point on the sun disk,  sun is the half-cone angle subtended
the sun disk 4.65 mrad,  is a constant, and  is the limb darkening parameter.

4. Results and discussion

Fig.10: Total irradiance map for apsorbed solar flux in the focus

Fig. 11: 3D Irradiance map of apsorbed solar flux in the focus

The figure 12 represents the effective solar flux density received by the parabolic
concentrator. As shown in this figure and table 3 the flux density can reach ine
the simulation conditions 544 W/m2 between 11h and 13h. The relation between
incident and direct solar fluxes is given by the following expression:

S
 incid  (9)
cos  
Figure 13 represents temperature distribution of the receiver aperture. The
temperature reaches an average values of 380 0C after 23 min which represents
the heating time of the receiver.

Table 3: Solar flux density and declination

t(h) p (W/m2) (0) s (W/m2) incid (W/m2)
8 49.3 69 39 118
9 252.6 57 202 370
10 405.4 46 324 466
11 520.8 37 416 521
12 576.1 32 461 544
13 546.1 33 437 521
14 490.3 38 392 497
15 364.8 48 291 434
16 206.3 59 165 320
17 52.2 71 42 129

800 ?incid (W /m2) 2

incid W/m 
700
600
500
400
300
200
100
t (h)
0
7 8 9 10 11 12 13 14 15 16 17 18
Fig.12: Solar flux density directly received of the parabolic concentrator
 450
T c (°C)
450

350

300

250

200

150

100

50

t (h)
0
12:00 12:20 12:40 13:00 13:20 13:40 14:00 14:20

Fig.13: Temperature profile of the center of cavity receiver

5. Conclusions
This paper presents optical simulations of the parabolic dish collector using the
ray - tracing software TracePro. A receiver model is added in the ray-tracing
simulations in order to study the influence of cavity physical parameters. A
receiver model is added in the ray-tracing simulations in order to study the
influence of cavity physical parameters. This type of cavity allows to obtain an
homogeneous solar flux distribution on the absorber and toincrease the cavity
receiver efficiency. The advantage of TracePro is that the Monte Carlo method
gives reliable results when enough rays are used. The positions have a ray trace
applied at different solar incident angles and varying additional RMS slope error
values applied to the reflective surfaces. Future works will focus on the study of
the influence of cavity physical parameters in order to design a new solar
receiver. Using the ray-tracing program TracePro and parabola slope error data, a
parabola and receiver simulation are made to estimate and study the solar flux
distribution in different planes. The next step in this study is develop of
mathematical, physical and numerical models of solar parablic concentrator
coupled by new solar receiver modified cavity receiver.

Acknowledgements
This paper is done within the research framework of research projects: III42006
– Research and development of energy and environmentally highly effective
polygeneration systems based on renewable energy resources and TR 33015 –
Research and development of the first Serbian zero energy efficient house
financed and supported by the Republic of Serbia Ministry for Education,
Science and Technological Development. We would like to thank to the Ministry
of Education, Science and Technological Development of Republic of Serbia for
the financial support for conducting the research.
 Corresponding Author

Dipl.-Ing. Sasa Pavlovic – Research assistant at the Faculty of Mechanical

Engineering in Nis, Department for Energetics and process technique
E-mail: [email protected]. Mob. +381643662513

References
[1] D. Marković, B. Čeperković, A. Vlajčić, S. Ressl, The white book of the
electric power industry of Serbia, PE Electric power industry of Serbia,
Belgrade, Vol. 3. pp 34-45, 2011.
[2] T. Pavlović, B. Čabrić, 2007, Physics and techniques of solar energy,
Građevinska knjiga, Belgrade, (in Serbian)..
[3] A. Fernández-García, E. Zarza, L. Valenzuela, M. Pérez, 2010, Parabolic-
trough solar collectors and their applications, Renewable and Sustainable
Energy Reviews, 14, , pp. 1695-1721
[4] V Stefanovic, M Bojic, S Pavlovic, N Apostolovic, I Nikolic, Z Djordjevic,
and D Catic, Prototype Receiver for Medium Temperature Convesion of
Solar Radiation to Heat, Proceedings of the Institution of Mechanical
Engineers, Part A, Journal of Power and Energy [PIA], . 225, 8, DOI:
10.1177/0957650911416566 (2011).
[5] Vicente J.Romero, Circe2/Dekgen2, A software package for
FacilitatedOptical analysis of 3D distributed solar energy concentrators,
SANDIA REPORT, Sandia National laboratories, 1994
[6] D. Chemisana, M. Ib_a~nez, and J. Barrau, Energy Conversion.
Management. 50, 1079 (2009).
[7] T. Thibert, M.-L. Hellin, J. Loicq, E. Mazy, L. Jacques, D. Verstraeten, J.-M.
Gillis, F. Languy, C. Emmerechts, E. Beeckman, S. Habraken, and J.-H.
Lecat, AIP Conf. Proc. 1477, 225 (2012).
[8] M. Mouzouris, L. W. Roberts, and M. J. Brooks, R&D J. S Afr. Inst. Mech.
Eng. 27, 10 (2011).
[9] R. Bader, A. Pedretti, and A. Steinfeld, J. Solar. Energy Engineering. 134,
021002-1-8 (2012)
[10] Taggart S. 2008, CSP: Dish projects inch forward. Renawable Energy
Focus 9(4):52-58
[11] B. L. Butler, R. B. Pettit, 1977, Optical Evaluation Technoques for
Reflecting Solar Concentrators, Proc. SPIE 0114, Optics Applied to Solar
Energy Conversion, 43 (October 31, 1977); doi:10.1117/12.955619.
[12] Zhiqiang Liu, Justin Lapp, Wojciech Lipinski, 2012, Optical design of a flat
solar concentrator, Solar Energy, Vol. 86, pp. 1962-1966.

View publication stats