Journal Pre-proofs

Research Paper

Numerical investigation on radiative heat loss of a direct absorption solar col-

lector using the discrete ordinates method

Yanlong Zhu, Weiqiang Kong, Jianhua Fan, Gerald Englmair, Yuan Yuan

PII: S1359-4311(23)01589-2
DOI: https://doi.org/10.1016/j.applthermaleng.2023.121560
Reference: ATE 121560

To appear in: Applied Thermal Engineering

Received Date: 12 March 2023

Revised Date: 11 June 2023
Accepted Date: 6 September 2023

Please cite this article as: Y. Zhu, W. Kong, J. Fan, G. Englmair, Y. Yuan, Numerical investigation on radiative
heat loss of a direct absorption solar collector using the discrete ordinates method, Applied Thermal Engineering
(2023), doi: https://doi.org/10.1016/j.applthermaleng.2023.121560

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2023 Published by Elsevier Ltd.

 Numerical investigation on radiative heat loss of a direct
absorption solar collector using the discrete ordinates method

Yanlong Zhu1, 2, Weiqiang Kong2, Jianhua Fan2, Gerald Englmair2, Yuan Yuan1, *
1School of Energy Science and Engineering, Harbin Institute of Technology, 92 West Dazhi Street, Harbin 150001, PR
China

2Department of Civil Engineering and Mechanical Engineering, Technical University of Denmark, Koppels Allé Building
404, 2800 Kgs. Lyngby, Denmark

* … corresponding author ([email protected])

Abstract
Direct-absorption solar collectors have the potential to be highly efficient and exhibit a
uniform temperature distribution. However, the standard numerical model simplifies radiation
and heat loss, which require further consideration at high temperatures. Thus, a novel self-
programming model for photothermal conversion in a direct-absorption solar collector was
developed based on the discrete ordinates method. This novel model contains further details on
radiative transfer and specular reflection boundary conditions. This study compared the results
of the two models at high temperatures. With an inlet temperature of 500℃ and a solar
concentration ratio of 100, the standard model overestimated outlet temperature and conversion
efficiency by 27.6℃ and 18.9%, respectively. This study investigated radiative heat loss at
different temperatures, solar concentration ratios, and incident angles. The results showed that
radiative heat loss will exceed 5% with a fluid temperature exceeding 600℃. Furthermore,
efficiency decreases significantly at incident angles exceeding 30°. The proposed model
expands the range of applications of numerical models of solar collectors to high temperatures.
These results provide a reference for improving the conversion efficiency of solar collectors at
high temperatures, which is regarded as an urgent problem.

Keywords:

Direct absorption; solar collector; high temperature; radiative heat loss; discrete ordinates
method

Nomenclature
Latin characters Greek symbols

heat transfer conductivity,

a β attenuation coefficient, m-1
W/(m·K)

specific heat capacity, modified attenuation

Cp  mi
J/(kg·K) coefficient, m-1

speed of light in vacuum,

c Δ change of variables
2.998×108 m/s

diffusion conductivity,
Di ε emissivity
W/(m·K)

volume radiative coefficient, photothermal conversion

Di,P η
m·Sr efficiency

convective conductivity,
Fi θ polar angle
W/(m·K)

G incident radiation, W/m2 θ1 incident angle

H height of solar collector, m θ2 refracting angles

Planck constant, 6.626×10-34

h κ absorption coefficient, m-1
J·s

convective heat loss Planck-mean absorption

ha 2

coefficient, W/(m ·K) coefficient, m-1

Boltzmann constant,
I radiative intensity, W/(m2·Sr) κB
1.381×10-23 J/K

radiative intensity of black

Ib λ spectral length, m
body, W/(m2·Sr)
 Nusselt number of convective thermal conductivity,
Nua λt
heat loss W/(m·K)

n1 refractive indexes of the air ρ density, kg/m3

n2 refractive indexes of the fluid ̂ specular reflectivity

n̂ unit normal vector of walls ρrg reflectivity of the glass

Pe Peclet number σs scattering coefficient, m-1

qw boundary radiative flux, W/m2 τg transmissivity of the glass

qr radiative heat flux, W/m2 Φ scattering phase function

Ri,P volume radiative source, W/m φ azimuthal angle

r location vector Ω solid angle, Sr

sun’s apparent size,

ra aspect ratio of solar collector Ωs
6.8×10-5

rc solar concentration ratio ω scattering albedo

S radiative heat source, W/m3

transmissivity of the
Scat Subscripts
atmosphere

modified source function,

S mi E, e east
W/(m3·Sr)

direction vector of radiative

ŝ inc incident
transfer
 ŝ ' other direction vectors l left

T temperature, ℃ loss_rad radiative heat loss

u velocity of the fluid, m/s N, n north

x x coordinate, m P position

y y coordinate, m r right

S, s south

Abbreviations W, w west

value of parameters in last

CSP Concentrating solar power *
step

Direct absorption solar

DASC
collector

Surface absorption solar

SASC
collector

1. Introduction
Globally, energy production infrastructure is undergoing a transformation driven by
decarbonization goals[1]. Renewable energy sources, such as solar energy, play a vital role in
this transformation[2]. It is necessary to develop solar energy applications to reduce carbon
emissions and address environmental issues[3]. In addition to solar photovoltaics,
concentrating solar power (CSP) has the potential to decarbonize power generation[4]. By 2022,
the installed capacity of CSP was 7GW, generating a total of 15 TWh of electricity[5].
Although there is a large gap between installations of CSP and solar photovoltaic, only CSP
can be integrated with thermal energy storage to provide a stable and dispatchable output[6].
Thus, there has been a growing interest in CSP for peak shaving of power grids.

Because of its short development time, CSP technologies are less attractive than solar
photovoltaics. 3rd generation CSP [7] utilizes Brayton-sCO2 as a power cycle to address this
issue. Compared to the original Steam-Rankine cycle, the efficiency of the system using this
new power cycle increases from 41.5% to 55%[8]. In this new system, the inlet temperature of
the power block increases from 565℃ to 700℃. Accordingly, the solar collector should
provide an output with a temperature exceeding 720℃[9]. In traditional surface absorption
solar collectors (SASC), solar radiation is converted to thermal energy by an absorbing coating
on the surface. Thermal energy is transferred to the fluid depending on the temperature
difference. Owing to thermal resistance, the temperature of the surface of the SASC is always
higher than that of the fluid[10]. Heat flux in the solar collectors must be limited to avoid
thermal damage, which can decrease solar collector efficiency from 89% to 78.2%[11]. To
provide the required temperature and mass flow rate, the size of the solar collector must be
increased, which increases the cost of construction.

The use of a direct-absorption solar collector (DASC) is a possible approach to address

this size issue. The top surface of a DASC is composed of transparent materials and solar
radiation is directly absorbed by the fluids. Consequently, DASC can provide an output with a
more uniform distribution and lower surface temperature[12]. Novel fluids such as Ag-
plasmonic[13], cuttlefish ink (a biomass material), water[14], TiN-ethylene glycol[15], and
SiC-water[16] have been examined as potential heat transfer fluids. The DASCs in these
experimental studies performed better than SASCs.

Several numerical investigations of DASCs have been conducted. Some researchers

model the solar collector using commercial software, including Ansys® Fluent [12,17] and
Comsol Multiphysics®[18–20]. However, the results obtained using commercial software
cannot separate the different types of heat loss. Therefore, it was difficult to determine the
specific mechanism of heat loss. Additionally, at a system-level economic assessment, the
absorbed ratio of the working fluid to solar radiation was obtained only from specific
experimental results[21]. The values varied for the different working fluids. To consider
additional numerical details and provide a reference for a system-level model, many
researchers have focused on self-programming codes.

Tyagi et al., and Taylor et al. [22,23] proposed a 2D model of a non-concentrating DASC.
The initial temperature was 35℃ and consequently the model ignored the emission and
scattering of the fluid. The results showed that the efficiency of this novel solar collector
increased by 10% compared with that of a traditional SASC. Vital et al. [24] numerically
studied the photothermal conversion efficiencies of DASC using different types of nanofluids.
The Beer-Lambert equation was used in this model and only convective heat loss was
considered. Veeraragavan et al. [25] proposed an analytical model for DASC using the same
radiation treatment as used by Taylor. However, Therminol®VP-1 was used as fluid, with a
temperature of 400℃. The assumption that ignoring fluid emissions does not compromise the
results requires further consideration. Some researchers have considered radiative heat loss in
their models. Sharaf et al. [26] studied the effect of the heat loss coefficient on the performance
of DASC and they demonstrated that a lower heat loss coefficient led to a higher efficiency,
but a less uniform temperature distribution. Cregan et al. [27] introduced reflectivity, R, on the
top surface to balance the effect of radiative heat loss. When R was 0.35, the calculated
efficiency agreed well with that of the experiment. Hewakuruppu et al. [28] solved a 1D
radiative transfer equation (RTE) along the y-axis using a two-flux method, which considered
fluid emissions in two directions. Consequently, the RTE is only suitable for 1D models. The
results revealed that when the optical depth of the fluid was 3, the DASC exhibited good
absorption of solar radiation. The authors also suggested that adopting an infrared-reflective
coating on the top surface could significantly reduce radiative heat loss. Siavashi et al. [29]
developed a similar model based on two-flux approximations. Radiative heat loss at the top
surface was also considered, based on the Stefan-Boltzmann law.

The models mentioned above, which are defined as standard models in the following
section, all adopt a simplification of radiative transfer. For example, solar incidence was treated
as varying only along the y direction; emission and scattering of radiation of the working fluid
were also ignored. Thus, the effect of radiation was simplified as a heat source depending only
on the y coordinates. Radiative heat loss is considered by an effective coefficient, together with
convective heat loss.

As a novel type of solar collector, DASCs differ from traditional solar collectors in terms
of their design, optimization, and mechanisms of heat loss. In addition to absorbing and
scattering incident solar radiation, the working fluid also emits or receives radiation to/from
other positions. The radiation heat loss of this collector includes unabsorbed and reflected solar
radiation and emissions from the fluid into the environment. Therefore, a clear understanding
of the photothermal conversion process is required. Because of the complexity of solving the
RTE, the standard models of DASCs, which were used in previous studies mentioned above,
neglected the emission and scattering of radiation from the working fluid. Therefore, it is
necessary to examine the applicability of this assumption at high temperatures. Second, the
mechanisms of heat loss of the DASC is a composite because of the radiative transfer in the
semitransparent medium. Determining the mechanism of heat loss is helpful for minimizing
heat loss and improving the conversion efficiency of DASCs.

This study solved the RTE of a fluid using the discrete ordinates method in finite-volume
form and proposes a self-programming model using Python script for the photothermal
conversion of DASC. The results calculated using the novel and standard models at low and
high temperatures were used to verify the accuracy and benefits of the novel model. Radiative
heat loss was evaluated for different collector temperatures and solar concentration ratios.
Furthermore, the effect of the angle of incidence on conversion efficiency is discussed. The
proposed photothermal conversion model expands the range of application of the DASC
numerical model from low to high temperature. In addition, this study aimed to evaluate
radiative heat loss with variations in temperature, solar concentration ratio, and angle of
incidence, which are all important parameters in solar thermal applications. The results of this
study provide a reference for designing and adjusting the photothermal conversion performance
of DASCs at high temperatures.

2. Methodology
2.1 Geometry of DASC

The geometry of the DASC in the novel model is the same as that of the standard model.
Fig. 1 presents the details of photothermal conversion in DASCs. It is noteworthy that different
radiative phenomena are marked on the nanoparticles for clarity; however, the radiation fills
the entire fluid. The RTE in the working fluid is defined as [30]:

dI
 sˆ I  r ,sˆ     r  I b  r     r  I  r ,sˆ    s  r  I  r ,sˆ 
ds
(1)
 r 
I  r ,sˆ '    r ,sˆ ' ,sˆ d '
4 4
 s
 where, I  r , sˆ  (W/(m2·Sr)) is radiative intensity; κ (m-1) and σs (m-1) are absorption and
scattering coefficients of the fluid, respectively; Φ is scattering phase function; r is location
vector; ŝ is direction vector of radiative transfer. Eq. (1) identifies that the radiative intensity
is augmented by emission (κIb) and decreased by absorption (κI) and by scattering away from
ŝ (σsI). The last term is the increase in intensity owing to scattered radiation from other
directions ( ŝ ' )[30].

Fig. 1. Diagram of photo-thermal conversion processes in a direct absorption solar collector.

The four terms on the right of eq (1) are marked 1–4 in Figure 1. Photothermal conversion
can be summarized as follows: (1) A fraction of the solar radiation is absorbed or scattered by
the atmosphere, and the rest arrives at the top surface of the DASC; (2) the semi-transparent
top surface absorbs and reflects a small fraction of incident solar radiation; (3) solar radiation
passes through the top of the DASC, and the wall and fluid absorb a fraction of the radiation.
The remainder escapes to the external environment; (4) the emission of radiation by the fluid
and wall are the same as that in (3). The working fluid flows from left to right in Figure 1, with
an attendant increase in temperature.

The main necessary assumptions in this study are as follows:

(1) Fluid flow is incompressible. The flow was laminar.

(2) The inlet temperature in this study ranged from 10℃–700℃. The same fluid was used
throughout to allow comparison at different temperatures. However, no suitable working fluid
such as water, oil, or molten salt can cover this large range. Thus, an assumed working fluid
was used in this study, except in the model validation section. The thermophysical properties
were assumed to be stable in this temperature range, whereas the radiative properties were
obtained through a third-party experiment.

(3) The novel model solves the RTE by considering the entire nanofluid. The emission,
absorption, and scattering caused by nano-particles were characterized based on the radiative
properties of the nano-fluid. Thus, this study did not include the selection of nanofluids except
for model validation.

(4) Internal conduction was not observed at the top surface. No heat loss occurred at the
inlet or outlet.
 (5) The radiative properties of the working fluid depend on the spectrum, whereas those
of the top surface are independent of it.

(6) The emissivities of the inlet and outlet are 0[31].

(7) Reflection by the top surface of the DASC was specular, whereas reflection on the
other surfaces was diffuse. Scattering in the fluid was isotropic.

2.2 Energy conservation

According to the assumptions, energy conservation of the fluid is defined as:

 C p u T    t  T   S (2)

where ρ (kg/m3), Cp (J/(kg·K)), and λt (W/(m·K)) are the density, specific heat capacity,
and thermal conductivity of the working fluid, respectively. u (m/s) is velocity; T (℃) is
temperature; S (W/m3) is the heat source.

The finite volume method (FVM) [32] was used to solve eq. (2). A central difference
scheme [32] was adopted for discretization of the diffuse term, and a first-order upwind scheme
[32] was adopted for the convection term. Thus, the energy conservation of the DASC can be
expressed as:

FeTP  FwTW  FnTP  FsTs  De TE  TP   Dw TP  TW 

(3)
 Dn TN  TP   Ds TP  TS   S xy

Subscripts P, W, S, E, and N represent the centers of the element and adjacent elements,
respectively. The subscripts w, e, s, and n represent the boundaries of the five elements, where
Fi and Di are defined as:

  C p u x x, i  s, n
Fi   (4)
  C p u y y, i  w, e

 t y
 , i  w, e
 x
Di   (5)
 t x , i  s, n
 y


According to Eqs. (3)–(5), the temperature at point P can be calculated using the
temperatures of the other points, defined as [32]:

a pTp  awTW  aeTE  asTS  anTN  S xy (6)

aw  Dw  Fw , ae  De
as  Ds  Fs , an  Dn (7)
a p  aw  ae  as  an  Fe  Fw  Fn  Fs

2.3 Radiative transfer equation

 As described in the introduction, the standard model considers only the absorption of solar
radiation, whereas the proposed model includes a complete radiative transfer process. Thus,
this section describes the RTE in the standard and novel models.

Standard model

The standard model neglects emissions and scattering owing to the complexity of the RTE.
Thus, eq. (1) can be simplified into a 1D equation along the y-axis as follows:

dI
  I (8)
dy

where β (m-1) is the attenuation coefficient of the fluid, and is defined as      s .

Therefore, the radiative heat source in eq. (6) is converted to:

q 
 rc s g  I 0, e    d 
 H  y
S  (9)
y 0

where rc is the solar concentration ratio; H (m) is the height of DASC; λ (m) represents
the spectral length of solar radiation. Ωs (6.8×10-5) in eq. (9) is the solid angle between the earth
and the sun; τg is the transmissivity of the top surface; I0,λ (W/(m2·m·Sr)) is the spectral
radiative intensity of the sun, defined as [33]:

1 S att  2 hc 2 
I 0,  (10)
 5  e hc  B Tsun

1

where Satt (0.73) is the transmissivity of the atmosphere for solar radiation; h (6.626×10-
34 J·s) is Planck constant; c (2.998×108 m/s) is the speed of light in vacuum; κ (1.381×10-23
B
J/K) is the Boltzmann constant.

The heat source calculated using eq. (9) varies only along the y-axis. Thus, the fluid
temperature calculated by energy conservation did not affect the heat source. The heat source
is defined at the beginning of the simulation.

Novel model

Different from the standard model, the RTE in the novel model depends on both location
and radiative transfer direction, which is described using the solid angle Ω. The solid angle is
determined as follows:

d   sin  d  d  (11)

where, φ and θ are azimuthal and polar angles respectively [30].

Eq. (1) contain both differential and integral terms; therefore, it is necessary to convert
the integral term into a series of differential terms. The discrete coordinate method (DOM)
identifies the radiation intensity as a discrete form in 4 π space. The numerical solution for the
RTE can be obtained by solving a set of discrete-direction equations. The DOM uses a form
similar to the finite difference method in the radiation direction; therefore, it has some problems
with the non-conservation of radiation flux[34]. A DOM in the finite-volume form [35], also
known as the finite-angle method, was used in the present model. Similar to the FVM for
convection-diffusion equations, this method integrates the radiative intensity in angles based
on the DOM, thus ensuring the conservation of the radiation flux.

Accordingly, the RTE can be converted as:

dI  r , sˆ 
    r  I  r , sˆ   S  r , sˆ  (12)
ds

where the source is defined as:

 s r 
S  r , sˆ     r  I b  r , sˆ   I  r , sˆ'   sˆ' , sˆ  d '
4 4
(13)

Like the FVM for convection-diffusion equations, the solid angle is discretized into
Nφ×Nθ mesh elements. For both the location and angle elements, integration of eq. (12) can be
converted into:
w,e , s , n

 I li Al 
i
 sˆi  nˆl  d i     mi I i  Smi vi (14)
l w

where mi and Smi are the modifications of the attenuation coefficient and source [35]:

s
 mi      sˆi , sˆi  i
4
(15)
 s N  N
Sm   Ib 
i
   sˆ j , sˆi   j
4 j 1, j i

Assuming that radiation flux is conservative in both the control volume of the location
and angle spaces, the final discretization of the eq. (1) can be expressed as follows:

Ri , P
Ii,P  , i  1, 2,..., N  N
Di , p
Ri , p  S mi xyi  si  nˆw I i ,W y  si  nˆe I i , E y
 si  nˆn I i , N x  si  nˆs I i , S x (16)
Di , p   xyi  si  nˆw y  si  nˆe y
i
m

 si  nˆn x  si  nˆs x

where nˆw , nˆ e , nˆ s , and nˆ n represent the unit vectors along the forward and reverse
directions of the x- and y-axis.
2.4 Boundary conditions and materials properties

The thermal boundary conditions of the present model are defined as:

x  0, T  Tin
T
y  0, 0 (17)
y
T
y  H,  ha Ta  T 
y

where ha (W/(m2·K)) and Ta (K) are the convective heat loss coefficient and ambient
temperature, respectively. The radiative boundary conditions on the top surface are as follows:

y  H , sˆ  nˆw  0 : I   I bw  ˆ r Iˆ  1   g   rg  I sun (18)

where nˆw denotes the normal direction of the wall facing the fluid. ̂ and Iˆ represent
specular reflectivity and specular radiative intensity, respectively. The specular reflectivity can
be calculated as[30]:
2 2
1  n cos  2  n2 cos 1  1  n1 cos 1  n2 cos  2 
r   1     (19)
2  n1 cos  2  n2 cos 1  2  n1 cos 1  n2 cos  2 

where n1 and n2 are the refractive indices of air and fluid, respectively. θ1 and θ2 are the
incident and refracting angles, which are identified as:

n1
sin  2  sin 1 (20)
n2

εg and ρg in eq. (18) emissivity and reflectivity of the top surface, respectively.

In contrast to the upper surface, reflections from the other surfaces were diffuse. There
are two parts to the radiative intensity of these surfaces. One is the emission of the surface and
the other is the reflection of incident radiation by the surface, defined as:

1   
sˆ  nˆw  0, I   I bw 
 
sˆnˆw  0
I sˆ  nˆw d  (21)

Table 1 lists the thermal and radiative properties of the materials used in the model.

Table 1. Thermal and radiative properties of the heat transfer fluid.

Parameter Value Parameter Value

ρ 997 ω 0
 Cp 4180 εb 0.05

λt 0.607 εg 0.067

nf 1.33 εl, εr 0

β f (λ)

The attenuation coefficient of the fluid is obtained by third-party experiment result [36].
Table 2 lists the spectural attenuation coefficient of the fluid.

Table 2. The spctural attenuation coefficient of the fluid used in this study [36].

λ (nm) β (m-1) λ (nm) β (m-1)

200 600 1000 200

400 400 1200 200

500 300 1400 500

600 200 >1400 1000

800 200

2.5 Convection-radiation coupled heat transfer

The radiative properties of the materials in this model are independent of temperature.
However, the emission term in the RTE varies with the temperature. Thus, the results obtained
from energy conservation affect the emission term in the RTE, and the radiative intensity of
the RTE affects the heat source in energy conservation. Accordingly, a semi-implicit method
was introduced to solve the convection-radiation coupled heat transfer problem. This can be
summarized as follows[37]:

(1) Based on the estimated initial temperature distribution of the DASC, the radiative
intensity in eq. (16) was solved. The incident radiation G and boundary radiative heat flux qw
was calculated as follows[34]:

G   Id  (22)
4
 qw   I cos  d  (23)
2

(2) The heat source in the energy conservation was calculated using the G obtained in step
(1):

 
S    qr   4 Eb T      G d 
0 
 dEb T *  
   q  4 
*
 T  T *  (24)
r
 dT 
 
   qr*  16 n 2 T *3 T  T * 

where superscript * represents the value in the last iteration. According to the Stefan-
Boltzmann law, the emission of a black body is proportional to the fourth power of temperature.
The semi-implicit method adopts T*3T to reduce the order of the equations.  in eq. (24) is the
Planck-mean absorption coefficient, defined as:

   I b , d 
 0
(25)
T 4
where n is the refractive index of the medium and σ is the Stefan-Boltzmann constant.

(3) A heat source was used for energy conservation to obtain the new temperature
distribution of the DASC.

(4) Replace the temperature distribution T* using the new temperature distribution
obtained in step (3), and repeat steps (1)–(3) until the tolerance is acceptable.

Table 3 summarizes all the methods used in the novel DASC model.

Table 3. Details of the model of photo-thermal conversion of the DASC.

Items Method or value

Discretization of the geometry Internal-node mesh

Discretization of the energy conservation Finite volume method

Discretization of the diffuse item Central difference

Discretization of the convective item First order upwind

Tolerance of energy solution 10-7

 Discretization of the radiative transfer equation Discrete ordinates method

Tolerance of radiative solution 10-6

Combined radiation and convection Semi-implicit method

Tolerance of coupled solution 10-4

2.6 Parameters used for model validation and comparison

Three non-dimensional parameters, the convective heat loss coefficient, Nua, Peclet
number, Pe, and aspect ratio of the DASC, ra, are defined as:

ha H
Nua  (26)
t

uCp H
Pe  RePr  (27)
t

L
ra  (28)
H

The solar concentration ratio is defined as:

qinc
rc  1
(29)
qinc

1
where qinc and qinc are incident radiation fluxes before and after concentration by the solar
concentrators, respectively.

The angle of incidence was defined as the acute angle between the light and the normal
direction of the collector surface.

Table 4 lists the parameters of DASC used in this study. Besides, the parameters used for
validation can be found in section 3.1.

Table 4. Parameters used in the calculations of photothermal conversion of DASC.

Parameters Values Parameters Values

Nua 0.83272 ra 2080

 rc 1700 H (m) 0.1, 0.2

θin (°) 060 Pe 43068680

The average temperature of the fluid was defined as:

Tin  Tout
Tavg  (30)
2

The radiative heat loss ratio was defined as the percentage of radiative heat loss in the total
solar incidence as follows:

rloss _ rad
rloss _ rad  (31)
S att  s I sun

The effective radiative heat loss was defined as the ratio between radiative heat loss and
the solar incidence without concentration:

rc  rloss _ rad
rc _ loss _ rad  (32)
Satt s I sun

3. Results and discussion

After model validation, the differences in performance between the novel and standard
models at high temperatures were analyzed. A development in radiative heat loss with
variations in the temperature and solar concentration ratios was obtained. Finally, the effect of
the angle of incidence on conversion efficiency of the DASC was examined.

3.1 Model validation

The results of the novel model and the standard model of the DASC were compared with
the experimental results of Lee’s work[36]. As in the experiment, MWCNT-water–water with
a concentration of nanoparticles of 0.003% was selected as the working fluid. The bottom
surface of the solar collector was coated with materials with a 0.95 reflectivity to increase the
transfer path of solar radiation. Accordingly, the heat source in eq. (9) can be modified as
follows:
    H  y 
S  rc s g  ( I 0, e  I 0, e  H e  y )d  (33)
0

Eq. (33) includes the transfer of solar radiation after reaching, and being reflected by, the
bottom surface. Considering the radiative properties of the working fluid and height of the
DASC, the reflected radiative intensity was less than 2% when it reached the top surface. Thus,
this transfer process was ignored. Table 5 lists the parameters used for validation.

Table 5. The parameters used for validation of the models.

 Parameters Values Parameters Values

Nua 0.82372 H 0.01m

ra 45 rc 22

τg 0.913 Tin 10℃

Pe 430–861

The solar simulator provides a square radiative output to the DASC. The aspect ratio was
calculated by dividing the width of the flow channels by the area of the solar simulator output,
which was approximately 40, different from the experimental value. Fig. 2 shows a comparison
of the experimental and simulated DASC temperatures and energy conversion efficiencies. The
Peclet number varied across the range of 430–861.

(a) Development of the outlet temperature with variation of Peclet number.

 (b) Development of the efficiency with variation of Peclet number.

Fig. 2. Validation of two models of DASC utilizing third-party experimental results.

The outlet temperature and energy conversion efficiency of the simulation results agreed
well with those obtained using the standard model. Maximum errors of 0.2℃ and 0.93%,
respectively occurred. With an inlet temperature of the working fluid of 10℃, consideration of
the emission term in eq. (1) in the novel model does not substantially affect the results.

The experimental uncertainty ranges of the temperature and energy conversion

efficiencies were 3% and 5%, respectively[36]. Results obtained using the novel model were
within this range, except for the outlet temperature, which was calculated using a Peclet number
of 574. Here, the deviation between the numerical and experimental results was 2℃. The other
results with different Pe numbers were all within the scope of this study, considering
measurement uncertainty. Thus the two models could predict the photothermal conversion
performance of the DASC when temperature was low.

3.2 Model comparison

In this section, the results of the two models are compared to explore whether the
assumption is still acceptable when fluid temperature is high, and to analyze the errors using
the standard model.
 (a) The distribution of temperature calculated with the presented model.

(b) The distribution of temperature calculated with the standard model.

Fig. 3. Comparison of temperature of the standard and novel models with an inlet temperature of 500℃.

Fig. 3 shows the temperature distributions of DASC calculated with the two models with
a working fluid inlet temperature of 500℃. Differences between results were significant. The
minimum and maximum temperatures calculated with the novel model were 475℃ and 575℃,
respectively, while using the standard model temperatures of 492℃ and 623℃ were obtained.
The average temperature rise at the outlet of the novel model was 63.9℃, while the standard
model average was 91.5℃ (an increase of 27.6℃). The temperature differences resulted in an
overestimation of the conversion efficiency by 18.9% when the standard model was used.

Because the inlet temperature was 500℃, infrared emission of the working fluid was
significant. Convection between the DASC and the environment and radiation of the working
fluid resulted in heat loss. The fluid near the top surface was exposed to a higher radiative heat
loss. Thus, the fluid temperature near the bottom was higher than that near the top (Fig. 3a).
The standard model ignored radiative heat loss. The fluid absorbed solar radiation, as described
in eq. (8). Consequently, outlet temperature near the top surface was higher than that near the
bottom surface of the DASC. These results indicate that it is necessary to use the novel model
to predict the performance of DASC at high temperatures, the standard model may
overestimate fluid temperature and conversion efficiency by neglecting radiative heat loss.

3.3 Radiative heat loss of DASC

A series of parameter variations was applied to both models to evaluate the performance
error of the DASC when using the standard model. With temperature-independent material
properties, the parameters that may affect the performance of DASC are the solar concentration
ratio, inlet temperature, fluid velocity, and DASC size.

(a) variation of the length of DASC.

(b) variation of the outlet temperature.

 Fig. 4. Radiative heat loss ratio with variation of collector length and outlet temperatures.

Fig. 4 shows the sensitivity of radiative heat loss ratio to variations in channel lengths and
outlet temperatures. The Peclet numbers were 17170, 34340, and 68680, corresponding to
velocities of 0.25, 0.5, and 1 m/s, respectively. An increase in the length of the DASC increased
radiative heat loss when other parameters are constant because the energy absorbed by the fluid
increased along the flow channel (Fig. 4(a)). Fluids at higher temperatures emitted more
radiation from the solar collector, increasing their radiative heat loss. With a constant channel
length, the ratio decreased with higher Peclet numbers because the velocity of the fluid
increases with a larger Peclet number. The larger the velocity, the smaller the increase in the
temperature of the fluid. Thus, the radiative heat loss resulting from fluid emission was small
for a large Peclet number.

Radiative heat loss ratio increased linearly when the other parameters were the same (Fig.
4(a)). The gradient of the curve decreased with an increase in Pe. Accordingly, the effect of
the Peclet number and DASC length on this ratio can be simplified into a variation in this ratio
with the difference in inlet and outlet temperatures (Fig. 4 (b)). All 15 data points agreed well
with the fitted curve, reducing the number of parameters affecting the performance of the
DASC.

The effects of different inlet temperatures and solar concentration ratios on the radiative
heat loss ratio were also examined. Nua was 0.82372, and the solar concentration ratio was in
the range 1–700, which included low-, middle-, and high-temperature applications of the
DASC. The inlet temperature varied from 20℃ to 700℃. The increase in temperature of the
fluid was set in the range of 2.5–14℃. The Pe number was calculated according to different
concentration ratios and temperature increases. This is also a general strategy used in
applications for adjusting the output to the desired temperature. The relationship between the
parameters mentioned above can be described as follows:

Lrc  s I sun   C p uH T (34)

The Pe number can then be calculated by:

Lrc  s
Pe  (35)
 t T

The range of Pe number used in this section was 430–68680, corresponding to a fluid
velocity of 0.006–1 m/s.

Based on the results shown in Fig. 4, all parameters, except the height of the DASC, could
be simplified into the solar concentration ratio and average temperature. Within the scope of
these parameters, we randomly generated 100 simulation data points.

Fig. 5 (a) shows the radiative heat loss with variations in the solar concentration ratio and
average temperature. At a constant average temperature, the radiative heat loss changed slightly.
Consequently, the radiative heat loss ratio decreased with increasing solar concentration ratio.
At a constant solar concentration ratio, the radiative heat loss increased with increased average
fluid temperature, which increased the radiative heat loss ratio.
 (a) the ratio of radiative heat loss and solar radiation.

(b) the ratio between radiative and convective heat loss.

Fig. 5. The development of the radiative and convective heat loss with variation of inlet temperature
and solar concentration ratio.

The radiative heat loss ratio was less than 5% when the temperature was low and the solar
concentration ratio was large (Fig. 5(a)). In the area of the temperature and concentration ratio
to the left of the dashed line, the standard model could predict the performance of the DASC
with minimal error. In the area to the right of the dashed line, radiative heat loss exceeded 5%.
Thus, radiative heat loss cannot be ignored and the standard model overestimated the results
compared with the novel model. It is noteworthy that DASC suffer from a high radiative heat
loss ratio at high temperatures, even at high concentration ratios. For temperatures exceeding
600℃, the ratio varied in the range of 5%–33% with different solar concentration ratios.

Figure 5(b) shows a relationship between radiative and convective heat losses. Two dash
lines with values of 0.5 and 1.0 represent that the radiative heat loss is half of and equals to the
convective heat loss, respectivily. There are two situations with a larger radiative heat loss in
Figure 5(b). When temperature is low and solar concentration ratio is large, convective and
radiative (resulting from infra-emisson of the fluid) heat losses are small. However, the
reflection by the top wall and the solar radiation not absorbed by DASC are significant. Thus,
the radiative heat loss is larger than the convective heat loss. For temperatures exceeding 500℃,
radiative heat loss (resulting from infra-emisson of the fluid) becomes significant.

Fig. 6 shows the effect of the average temperature of the fluid on effective radiative heat
loss.

Fig. 6. The development of effective radiative heat loss with variation of average temperature.

The effective radiative heat losses at different inlet and outlet temperatures, velocities,
and solar concentration ratios agreed well with the fitted curves. In addition to the 100
simulated data points, the results in Fig. 4 (b) are also described by the curve. When the
temperature was low, the difference between the calculated effective radiative heat loss and
predicted value was significant, with a maximum difference of 152%. With an increase in
temperature, the difference between the calculated and predicted values decreased. When the
average temperature of the fluid exceeded 600℃, the maximum difference was 23%. This is
because the radiative heat loss of the collector included the unabsorbed fraction of incident
solar radiation and the fraction emitted by the fluid itself. When the average temperature was
low, the radiation emitted by the fluid was small; therefore, the radiation heat loss was
significantly affected by the solar concentration ratio. The difference between the calculated
and predicted values was large. In contrast, when the temperature was high, the radiation
emitted by the fluid itself was significant, and radiation heat loss was less affected by the solar
concentration ratio; thus, the deviation was slight.
 The effective radiative heat loss of the DASC increased with an increase in average
temperature. With an average temperature of 100, 300, and 700℃, the radiative heat loss of
the collector equaled 2.4, 6.95, and 51.2 times the incident radiation intensity. Accordingly, a
solar concentration ratio much larger than this value should be selected to ensure a higher
energy conversion efficiency of the collector.

3.4 Effect of angle of incidence

Because of the changing position of the sun (daily and seasonally), and the focusing
strategy used, and the roughness of the mirror surface, the solar radiation reaching the collector
is not always normal to the surface of the collector, that is, the angle of incidence changes.
Therefore, the energy conversion efficiency and heat loss of the collectors at different angles
of incidence were examined.

(a) The radiative heat loss with incident (b) The radiative heat loss with incident
angles for different attenuation coefficient. angles for different scattering albedo.
 (c) The efficiency with incident angles for (d) The efficiency with incident angles for
different attenuation coefficient. different scattering albedo.

Fig. 7. The photo-thermal conversion performance with variation of incident angles.

Fig. 7 shows the radiative heat loss and efficiency with the variation of solar incident
angles. The angle of incidence exceeded 0° along the direction of fluid flow. fβ is the scaling
of attenuation coefficient selected. The attenuation coefficient of fβ = 1.0 is the data in Table 2.

Figure 7 (a) shows the radiative heat loss incresed with a larger attenuation coefficient.
Because the fluid with a smaller attenuation coefficient did not absorbed solar incidence fully.
Moreover, radiative heat loss decreased with an increase of incident angle, which means a
longer optical length to absorb more solar radiation. For the large attenuation coefficient, the
radiative heat loss almost did not changed with incident angle. Because the solar radiation was
absorbed fully by the fluid even in a small incident angle.

Figure 7 (b) shows the radiative heat loss increased with a larger scattering albedo.
Because the fluid with a larger scattering albedo scattered more solar radiation to enviroment.
For the large scttering albedo, an increase of incident angle results in longer optical length to
scattering more radiation. Thus, the radiative heat loss increased with the incident angles.

In addation to radiative heat loss, specular reflective loss caused by the top glass wall also
affected the photothermal conversion of DASC. According to eq. (19), the reflective loss only
depended on the meidum refrative index and incident angles. Thus, for the same refrative index,
the reflective losses are 0.02, 0.02, 0.021, 0.028, and 0.059 with the incident of 0°, 10°, 30°,
45°, and 60°. As shown in Figure 7 (c) and (d), when the attenuation coefficient factor was 0.4,
the efficiency increased with the incident angle < 45°. In this situation, the reduce in radiative
heat loss was more significant than the rise in reflective loss with large incident angle. However,
for the other attenuation coefficient and incident angles, the situation was being reversed,
resulting in a drop in efficiency.

4. Conclusions
This study proposes a novel photothermal model of a direct-absorption solar collector
based on the discrete ordinates method in a finite-volume form. The novel model included a
solution for modeling radiative transfer in the working fluid, which was missing in the standard
model. Additionally, a specular reflection boundary condition was applied to the upper glass
wall of the novel model.

This study compared the temperature distributions of the standard and novel models.
Changes to radiative heat loss with variations in fluid temperature, velocity, and collector
length was analyzed. Furthermore, the effect of solar angle of incidence on energy conversion
efficiency is discussed. The main conclusions are summarized as follows:

(1) A comparison of the temperature distributions of the two models showed that the
standard model overestimated the outlet temperature and conversion efficiency at high
temperatures. The simplification of the radiative transfer equation used in previous studies is
not suitable for high-temperature applications.
 (2) The radiative heat loss with different velocities and collector lengths can be combined
into a function that depends only on average temperature.

(3) Radiative heat loss becomes the dominant process of heat loss at high temperatures
(above ~540℃, especially when solar concentration ratios are low), and it is necessary to use
the novel model to predict the photothermal performance of the direct absorption solar collector.

(4) According to point (2), the effective radiative heat loss that depends only on the
temperature is defined as the ratio of radiative heat loss to the non-concentrated solar radiation
intensity. This parameter helps determine a suitable solar concentration ratio to ensure a high
conversion efficiency of the solar collector according to the average temperature.

(5) Reflective losses cannot be ignored when the absolute angle of incidence exceeds 30°.
On the one hand, an increase in incident angle results in a larger reflective loss. On the other
hand, the fluid absorbs more solar radiation with a larger incident angle which means a longer
optical length. Thus, the efficiency increases with the increase of incident angle when the
attenuation coefficient is small. The situation will be opposite for large attenuation coefficients.

The novel model proposed in this study expands the temperature range of numerical
research on direct-absorption solar collectors. This study analyzed the effects of key parameters
on radiative heat loss, including temperature, solar concentration ratio, and solar angle of
incidence. These results provide a reference for the design and optimization of direct-
absorption solar collectors at high temperatures.

Future perspectives

This paper discusses the differences between a standard model and a novel model, at
different temperatures, using the discrete ordinate method. Two points should be considered in
future studies:

(1) The angular discretization should be sufficiently fine owing to the angle of incidence
and specular reflection of the upper surface. This resulted in an increase in the time required to
solve the novel model. Thus, it is necessary to study whether this can be reduced by applying
other methods, such as the Monte Carlo method.

(2) As described in the Abstract, to provide a fast evaluation of radiative heat loss for
system calculations, a Pearson correlation coefficient analysis was performed to quantify the
influence of all possible parameters on radiative heat loss.

Acknowledgment
This work was supported by the National Natural Science Foundation of China (Grant
No. 52041601). The Chinese Scholarship Council (CSC) (No. 202106120167) has also partly
funded the research activities – enabling cooperation of the Harbin Institute of Technology
with the Technical University of Denmark. Without their supports, the presented research
wound not have been possible.

References
[1] International Renewable Energy Agency. Innovation outlook thermal energy storage.
2020.

[2] International Energy Agency. Next generation wind and solar power: from cost to value.
2016.

[3] Alsagri AS, Alrobaian AA, Almohaimeed SA. Concentrating solar collectors in
absorption and adsorption cooling cycles: An overview. Energy Conversion and
Management, 2020, 223. https://doi.org/10.1016/j.enconman.2020.113420

[4] Ardani K, Denholm P, Mai T, et al. Solar Futures Study. 2021.

[5] International Energy Agency. World energy outlook 2022. 2022.

[6] Calvet N, Slocum AH, Gil A, et al. Dispatchable solar power using molten salt directly
irradiated from above. Solar Energy, 2021, 220: 217–29.

https://doi.org/10.1016/j.solener.2021.02.058

[7] Mehos M, Turchi C, Vidal J, et al. Concentrating Solar Power Gen3 Demonstration
Roadmap. 2017.

[8] Turchi CS, Ma Z, Neises TW, et al. Thermodynamic study of advanced supercritical
carbon dioxide power cycles for concentrating solar power systems. Journal of Solar
Energy Engineering, Transactions of the ASME, 2013, 135.

https://doi.org/10.1115/1.4024030

[9] Augustine C, Kesseli D, Turchi C. Technoeconomic cost analysis of NREL concentrating

solar power gen3 liquid pathway: preprint. 2021.

[10] Bandarra Filho EP, Mendoza OSH, Beicker CLL, et al. Experimental investigation of a
silver nanoparticle-based direct absorption solar thermal system. Energy Conversion and
Management, 2014, 84: 261–7. https://doi.org/10.1016/j.enconman.2014.04.009

[11] Martinek J, Jape S, Turchi CS. Evaluation of external tubular configurations for a high-
temperature chloride molten salt solar receiver operating above 700 °C. Solar Energy,
2021, 222: 115–28. https://doi.org/10.1016/j.solener.2021.04.054

[12] Zhu Y, Li P, Ruan Z, et al. A model and thermal loss evaluation of a direct-absorption
solar collector under the influence of radiation. Energy Conversion and Management,
2022, 251. https://doi.org/10.1016/j.enconman.2021.114933
[13] Kumar S, Sharma V, Samantaray MR, et al. Experimental investigation of a direct
absorption solar collector using ultra stable gold plasmonic nanofluid under real outdoor
conditions. Renewable Energy, 2020, 162: 1958–69.

https://doi.org/10.1016/j.renene.2020.10.017

[14] Zhu W, Zuo X, Ding Y, et al. Experimental investigation on the photothermal conversion
performance of cuttlefish ink nanofluids for direct absorption solar collectors. Applied
Thermal Engineering, 2023, 221: 119835.

https://doi.org/10.1016/j.applthermaleng.2022.119835

[15] Wang K, He Y, Zheng Z, et al. Experimental optimization of nanofluids based direct

absorption solar collector by optical boundary conditions. Applied Thermal Engineering,
2021, 182: 116076. https://doi.org/10.1016/j.applthermaleng.2020.116076

[16] Hooshmand A, Zahmatkesh I, Karami M, et al. Porous foams and nanofluids for thermal
performance improvement of a direct absorption solar collector: An experimental study.
Environmental Progress and Sustainable Energy, 2021, 40.

https://doi.org/10.1002/ep.13684

[17] Dugaria S, Bortolato M, del Col D. Modelling of a direct absorption solar receiver using
carbon based nanofluids under concentrated solar radiation. Renewable Energy, 2018,
128: 495–508. https://doi.org/10.1016/j.renene.2017.06.029

[18] Karim MA, Arthur O, Yarlagadda PKDV, et al. Performance investigation of high
temperature application of molten solar salt nanofluid in a direct absorption solar collector.
Molecules, 2019, 24. https://doi.org/10.3390/molecules24020285

[19] Qin C, Kim JB, Lee BJ. Performance analysis of a direct-absorption parabolic-trough
solar collector using plasmonic nanofluids. Renewable Energy, 2019, 143: 24–33.
https://doi.org/10.1016/j.renene.2019.04.146

[20] Qin C, Kim JB, Lee J, et al. Comparative analysis of direct-absorption parabolic-trough
solar collectors considering concentric nanofluid segmentation. International Journal of
Energy Research, 2020, 44: 4015–25. https://doi.org/10.1002/er.5165

[21] Ahbabi Saray J, Heyhat MM. Modeling of a direct absorption parabolic trough collector
based on using nanofluid: 4E assessment and water-energy nexus analysis. Energy, 2022,
244. https://doi.org/10.1016/j.energy.2022.123170

[22] Tyagi H, Phelan P, Prasher R. Predicted efficiency of a Low-temperature Nanofluid-based

direct absorption solar collector. Journal of Solar Energy Engineering, Transactions of
the ASME, 2009, 131: 0410041–7. https://doi.org/10.1115/1.3197562

[23] Taylor RA, Phelan PE, Otanicar TP, et al. Applicability of nanofluids in high flux solar
collectors. Journal of Renewable and Sustainable Energy, 2011, 3.

https://doi.org/10.1063/1.3571565
[24] Vital C, Farooq S, Araujo R, et al. Numerical assessment of transition metal notrides
nanofluids for improved performance of direct absorption solar collectors. Applied
Thermal Engineering, 2021, 190: 116799.

https://doi.org/10.1016/j.applthermaleng.2021.116799

[25] Veeraragavan A, Lenert A, Yilbas B, et al. Analytical model for the design of volumetric
solar flow receivers. International Journal of Heat and Mass Transfer, 2012, 55: 556–64.
https://doi.org/10.1016/j.ijheatmasstransfer.2011.11.001

[26] Sharaf OZ, Kyritsis DC, Abu-Nada E. Impact of nanofluids, radiation spectrum, and
hydrodynamics on the performance of direct absorption solar collectors. Energy Convers
Manag, 2018, 156: 706–22. https://doi.org/10.1016/j.enconman.2017.11.056

[27] Cregan V, Myers TG. Modelling the efficiency of a nanofluid direct absorption solar
collector. International Journal of Heat and Mass Transfer, 2015, 90: 505–14.
https://doi.org/10.1016/j.ijheatmasstransfer.2015.06.055

[28] Hewakuruppu YL, Taylor RA, Tyagi H, et al. Limits of selectivity of direct volumetric
solar absorption. Solar Energy, 2015, 114: 206–16.

https://doi.org/10.1016/j.solener.2015.01.043

[29] Siavashi M, Ghasemi K, Yousofvand R, Derakhshan S. Computational analysis of

SWCNH nanofluid-based direct absorption solar collector with a metal sheet. Solar
Energy, 2018, 170: 252–62. https://doi.org/10.1016/j.solener.2018.05.020

[30] Modest MF, Mazumder S. Fundamentals of Thermal Radiation. Radiative Heat Transfer,
Elsevier, 2022, p1–29. https://doi.org/10.1016/b978-0-12-818143-0.00009-2

[31] Ansys. Ansys Fluent Theory Guide. 2022.

[32] Tao WQ. Numerical Heat Transfer, (2nd ed.). Xi’an Jiaotong University Publishing
Company, 2001.

[33] Lee SH, Jang SP. Efficiency of a volumetric receiver using aqueous suspensions of multi-
walled carbon nanotubes for absorbing solar thermal energy. Int J Heat Mass Transf
2015;80:58–71. https://doi.org/10.1016/j.ijheatmasstransfer.2014.08.091

[34] Modest MF, Mazumder S. The Method of Discrete Ordinates (SN-Approximation).

Radiative Heat Transfer, Elsevier, 2022, p563–616.

https://doi.org/10.1016/b978-0-12-818143-0.00024-9

[35] Chai JC, Lee HOS, Patankar S v. Finite volume method for radiation heat transfer. Journal
of Thermophysics Heat Transfer, vol. 8, American Institute of Aeronautics and
Astronautics Inc., 1994, p419–25. https://doi.org/10.2514/3.559

[36] Lee SH, Choi TJ, Jang SP. Thermal efficiency comparison: Surface-based solar receivers
with conventional fluids and volumetric solar receivers with nanofluids. Energy, 2016,
115: 404–17. https://doi.org/10.1016/j.energy.2016.09.024
[37] Modest MF, Mazumder S. Radiation Combined with Conduction and Convection.
Radiative Heat Transfer, Elsevier, 2022, p775–817.

https://doi.org/10.1016/b978-0-12-818143-0.00029-8

Highlights

 A novel photothermal model of direct absorption solar collectors is established.

 The model contains further details of radiative transfer and specular reflection.
 The radiative loss with different temperatures and solar radiation is obtained.
 The radiative heat loss exceeds 5% with fluid temperatures higher than 600℃.
 The efficiency decreases significantly with incident angles larger than 30°.

CRediT authorship contribution statement

Yanlong Zhu: Conceptualization, Methodology, Visualization, Writing- Original draft

preparation. Weiqiang Kong: Methodology, Validation. Jianhua Fan: Methodology,
Reviewing. Gerald Englmair: Writing-Reviewing & Editing, Visualization. Yuan Yuan:
Conceptualization, Methodology, Writing- Reviewing & Editing.