Energy Conversion and Management 184 (2019) 219–234

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Optical efficiency improvement of solar power tower by employing and T

optimizing novel fin-like receivers
⁎
Wen-Qi Wang1, Yu Qiu1, Ming-Jia Li , Feng Cao, Zhan-Bin Liu
Key Laboratory of Thermo-Fluid Science and Engineering of Ministry of Education, School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi
710049, China

A R T I C LE I N FO A B S T R A C T

Keywords: The optical efficiency of the solar receiver is important to the global efficiency of the solar power tower plant.
Solar power tower However, the present commercial cylindrical receiver causes a considerable optical loss because most of the
Novel external receivers with finned structures sunrays reflected by the receiver surface are lost to the environment. To improve optical efficiency, three novel
Monte Carlo ray tracing receivers with vertical finned structures and two novel receivers with horizontal finned structures are proposed
Optical efficiency
in this paper. The optimal configurations of these receivers are obtained after parametric optimization. The
Solar flux
Multi-point aiming strategy
optical efficiency and solar fluxes distributions of the optimal novel receivers are compared with those of the
traditional cylindrical receiver at the noon of spring equinox. The results show that there is an optimal fin
number of 6 for all the receivers with vertical finned structures. The receiver with vertical finned structures and
the flat bottom shows the maximal optical efficiency of 63.9%. It indicates an improvement of 3.2 percentage
points compared with that of the traditional cylindrical receiver. The receiver with horizontal fins with flat
bottom shows the minimal optical efficiency of 60.3% at the noon of the spring equinox. It indicates a decrease
of 0.4 percentage points compared with that of the traditional cylindrical receiver. Additionally, the peak solar
fluxes on the novel receivers are quite close to each other and are nearly double that of the traditional cylindrical
receiver. In the end, a multi-point aiming strategy is applied to the optimal novel receivers, and the results show
that the peak fluxes can be obviously reduced. It is concluded that optical efficiency can be effectively improved
by using the receivers with vertical finned structures.

1. Introduction power tower. So far, four typical receivers including tubular cavity
receiver [20,21], tubular external receiver [22,23], volumetric receiver
The utilization of non-renewable energy resources, including oil, [24,25] and particle-based receiver [26,27] have been widely in-
coal and natural gas, has produced a large amount of carbon dioxide vestigated. Among these receivers, the external receiver is considered as
and caused environment problems [1–3]. In this regard, renewable a promising option for the large-scale utilization of solar energy due to
energy resources, especially solar energy, have attracted much atten- its suitability of a surrounding heliostat field [18]. Most external re-
tion in recent years [4–6]. ceivers, which have been put into operation, are arranged in a cylind-
Concentrating solar power (CSP), one of the solar energy utilization rical configuration, such as the receivers used in Solar Two [28], Ge-
technologies, has a potential for grid-scale dispatchable power gen- masolar [29] and Crescent Dunes [30]. Despite the fact that the
eration when integrated with thermal energy storage [7,8]. Up to now, cylindrical configuration is easier for fabrication and installation, it
four CSP technologies have been developed, which are solar power maximizes the optical loss because most sunrays reflected by the re-
tower [9,10], parabolic dish collector [11,12], parabolic trough col- ceiver surface are lost to the environment [31]. Compared with the
lector [13–15] and linear Fresnel reflector [16,17]. Among these four cylindrical receiver, the cavity receiver can achieve higher optical ef-
technologies, solar power tower is regarded as a promising choice be- ficiency by reabsorbing the reflected sunrays [32]. Qiu and He [33]
cause it can not only provide satisfactory performance but also reduce studied the effect of the reabsorption on the optical efficiency of a
the levelized cost of energy [18,19]. cavity receiver. The results showed that optical efficiency is increased
The solar receiver is one of the most crucial components of the solar by 3.5% at the noon of spring equinox because of the reabsorption.

⁎
Corresponding author.
E-mail address: mjli1990@xjtu.edu.cn (M.-J. Li).
1
The authors contributed equally to this work.

https://doi.org/10.1016/j.enconman.2018.12.029
Received 6 August 2018; Accepted 10 December 2018
Available online 01 February 2019
0196-8904/ © 2018 Elsevier Ltd. All rights reserved.
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

Nomenclature αe effective absorptivity of coating

αs altitude of the sun (rad)
Wh width of the heliostat (m) η optical efficiency (%)
Hh, Hr, Hf height of the heliostat / receiver / fin (m) ηi intercept efficiency (%)
nh number of the heliostat θ angle of the wall (°)
nfa number of the heliostat’s facets β concave and convex angle (°)
nt number of the tube γ angle on the tube (°)
Dw diameter of the wall (m) ρt,d coating diffuse reflectance
De effective diameter of the receiver (m) ρw,2 wall diffuse reflectance
G gap between tubes (m) ρh,1 heliostat reflectivity
do outer diameter of absorber tube (m) ρh,2 heliostat cleanliness
L length of the absorber tube (m) σte,1 altitude tracking error (rad)
Ar area of the receiver panel (m2) σte,2 azimuth tracking error (rad)
Ae area of the tube element (m2) σte tracking error (rad)
ts solar time (h) σse heliostat slope error (rad)
Qt, Qe power absorbed by tubes / element (W) σe effective deviation error of sunrays (rad)
Qh power received by heliostat field (W) σsun sunshape standard deviation (rad)
q solar flux (W·m−2) ωh reflection angle (°)
DNI Direct Normal Irradiance (W·m−2) εh elevation angle of the reflected sunray (°)
Nday ordinal number of a day in a year
Np number of sunrays Subscripts
rk beam radius (m)
rP radius of the beam projection (m) d diffuse parameter
X, Y, Z Cartesian coordinates (m) r,w,t,h receiver / wall / tube / heliostat parameter

Greek symbols

αt, αw absorptivity of coating / wall

Wang and He [20] also investigated the effect of the reabsorption on a structures, which form cavity-like structures. Their results showed that
cavity receiver. Their results showed that the reabsorption can improve effective absorptivity is improved from 0.86 to 0.96 when compared
the optical efficiency at any time. with that of the receiver without finned structure. Moreover, the use of
Therefore, introducing the reabsorption to the external receiver is cavity-like structures can also improve the thermal efficiency of the
an effective way to improve optical efficiency. Friefield and Friedman solar receiver. Christian et al. [37] investigated the thermal efficiency
[34] investigated an external receiver with a star-like cross-section. of external receivers with horizontal finned structures. They found that
Their results showed that this receiver has the potential to improve the this receiver can increase the thermal efficiency by 4.5% compared
optical efficiency. Garbrecht et al. [35] proposed a new external re- with the receiver without finned structure.
ceiver composed of hexagonal pyramid-shaped elements to reabsorb From the above review, it can be summarized that the optical effi-
the reflected sunrays. The results showed that the optical loss is reduced ciency of the external receiver can be improved by reabsorbing the
to 1.3%. Christian et al. [36] studied a receiver with horizontal finned reflected sunrays when finned structures are used. However, most

Fig. 1. Sketch of the solar tower power with a cylindrical receiver.

220
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

previous studies focused on the shape design and performance en- Table 1
hancement of small-scale prototypes, few studies focused on the optical Parameters and assumptions of the heliostat and tower [23,29,33,38].
efficiency improvement of plant-scale receivers. Thus, the receiver de- Parameters Dimensions Parameters Dimensions
sign by introducing the reabsorption in realistic tower plants should be
further conducted, and the performance of the optical efficiency im- Heliostat number nh 2650 Heliostat reflectivity ρh,1 0.90
Heliostat width Wh 11 m Heliostat cleanliness ρh,2 0.97
provement should be further examined. Five novel fin-like receivers are
Heliostat height Hh 10 m Altitude tracking error σte,1 0.46 mrad
first proposed and parametrically optimized through a Monte Carlo ray Facets number nfa 35 Azimuth tracking error σte,2 0.46 mrad
tracing code, which was developed in our previous work [33]. Then, Heliostat shape Spherical Heliostat slope error σse 1.3 mrad
the optical efficiencies and solar fluxes distributions of the optimal
novel receivers and the traditional cylindrical receiver are studied. In
the end, a multi-point aiming strategy is adopted to reduce the over- 2.2. Traditional cylindrical receiver
large peak solar flux of the novel receivers. The main contributions of
this study are summarized as follows. The cylindrical receiver, installed on a tower, has an elevation of
116 m, and the elevation is defined as the height difference between the
(1) Five novel fin-like receivers are proposed to improve the optical heliostat pivot point and the receiver center. The receiver has a dia-
efficiency of the central receiver, and these receivers are para- meter (Dw) of 8.5 m and a height (Hr) of 10 m, as shown in Fig. 3, and its
metrically optimized to obtain their respective optimal geometrical effective diameter (De) is defined as the diameter of the smallest circle
parameters by using a ray tracing code based on Monte Carlo that envelops itself. The cylindrical receiver comprises 18 tube panels,
method. each panel is composed of 31 tubes. The outer diameter (do) and the
(2) The optical efficiency of the optimal novel receivers is investigated length (L) of the tube are 45 mm [38] and 10 m, respectively. The de-
on typical days in a year under both sunny and cloudy conditions. tailed parameters and assumptions of the cylindrical receiver are
The real-time solar flux distributions on the optimal novel receivers summarized in Table 2. For the cylindrical receiver, the receiver panel
are analyzed. area (Ar) is defined in Eq. (1).
(3) A multi-point aiming strategy is adopted to reduce the overlarge
Ar = n wall LDw sin(θ /2) (1)
peak solar fluxes of the novel fin-like receivers. To ensure the multi-
point aiming strategy works in a whole year, the strategy is ex- where nwall is the number of the wall, which is equal to the number of
amined on three typical days including spring equinox, summer panels.
solstice and winter solstice. As shown in Fig. 3, some Cartesian righted-handed coordinate sys-
tems are built to describe the receiver model. The receiver system is
2. Physical model defined as XrYrZr with its origin placed in the center of the receiver. The
tube system is defined as XtYtZt, whose origin is placed in the center of
The Gemasolar solar power tower plant is treated as the physical the tube.
model. The plant is located in Fuentes de Andalucĺa, Spain, at 37°42′N,
5°9′ W [38]. It consists of a surrounding heliostat field and a cylindrical
2.3. Novel fin-like receivers
external receiver mounted on top of a tower as shown in Fig. 1.
To improve the optical efficiency of the solar power tower, three
2.1. Heliostats field novel fin-like receivers with vertical finned structures and two novel
fin-like receivers with horizontal finned structures are proposed. There
The heliostat field is composed of 2650 heliostats as shown in Fig. 2 are two reasons why such finned structures are chosen. First, both the
(a). A single-point aiming strategy is introduced as the basic strategy for vertical finned structures and the horizontal finned structures can re-
the field, where all heliostats aim at a point which is one meter right absorb the reflected sunrays, which may improve the optical efficiency
above the center of the receiver. Each heliostat, which consists of 35 of the central receiver. Second, the vertical finned structures and the
facets, is 11 m in width and 10 m in height with a spherical surface, as horizontal finned structures are easier to be fabricated and installed
shown in Fig. 2 (b). The detailed parameters and assumptions of the when compared with other types of finned structures, such as the wavy
heliostat are given in Table 1. finned structures.

-1000
Even raw
Odd raw
-800
Tower
-600

-400
Xg / m S-N

-200

200

400
Wh
600
-800 -600 -400 -200 0 200 400 600 800
Yg / m W-E
(a) Coordinates of the heliostats (b) A heliostat of the Gemasolar [39]
Fig. 2. Layout of the heliostat field. (See above-mentioned references for further information.)

221
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

Tubee No.
No.1 N
Yr Xt W E

Zt S

Wall
Xr
L (Hr)

Dw Xr
Zr Tube

De

Zr
Tube No.2800
Fig. 3. Schematic diagram of the cylindrical receiver.

Table 2 2.3.1. Receiver with vertical finned structures and flat bottom (Receiver-I)
Parameters and assumptions of Gemasolar plant [33,38]. The receiver with vertical finned structures and flat bottom
Parameters Dimensions Parameters Dimensions
(Receiver-I) is composed of vertical finned panels and flat bottom pa-
nels as shown in Fig. 4.The panel area (Ar) of the receiver is calculated
Receiver elevation 116 m Coating absorptivity αt 0.9 by Eq. (2). The flat bottom panels are arranged around a polygon with a
Receiver height Hr 10 m Coating diffuse reflectance 0.1 representative diameter of Dw. The Dw and the fin number (nfin) will be
ρt,d,
Diameter of the wall 8.5 m Wall absorptivity αw 0.6
optimized later. The rest parameters are the same as those of the tra-
Dw ditional cylindrical receiver.
Tube panels number 18 Wall diffuse reflectance 0.4
ρw,d, θ
Ar = n fin LHf + n wall LDw sin ⎛ ⎞
Tube outer diameter do 45 mm Tube gap in a panel G 2.7 mm ⎝2⎠ (2)
Tube length L 10 m Gap between tubes and 2.7 mm
wall where nwall is the number of the wall, θ = 360°/nfin.

2.3.2. Receiver with vertical finned structures and concave bottom

To obtain the optimal geometric parameters of the novel receivers, a
(Receiver-II)
parametric optimization is applied to all the novel receivers. To avoid
The receiver with vertical finned structures and concave bottom
the increase in the cost of the plant, the novel receiver is desired to have
(Receiver-II) consists of vertical finned panels and concave bottom
a similar cost to that of the traditional cylinder receiver. As a result, in
panels as shown in Fig. 5, where the total panel area (Ar) is calculated
the parametric optimization, the panel area is kept the same with that
by Eq. (3). The concave bottom panels are arranged around the wall
of the cylindrical receiver in Section 2.2. Since the size of the tubes in
with a diameter of Dw and a constant angle β of 120°. The Dw and the fin
each receiver is the same, the novel receivers have the same tube
number (nfin) will be optimized later. The rest parameters are the same
number with that of the traditional cylindrical receiver. The detailed
as those of the traditional cylindrical receiver.
information of the five novel receivers is as follows.
Dw θ β
Ar = n fin LHf + n wall Lsin ⎛ ⎞/sin ⎛ ⎞
2 ⎝2⎠ ⎝2⎠ (3)

Wall Tube
N
Yr
S

Xr
Hf
L

Xr
Zr

Zr De

Fig. 4. Sketch of the receiver with vertical finned structures and flat bottom.

222
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

Wall Tube

N
Yr
S

Hf
Xr
Xr

L
Zr

De
Zr

Fig. 5. Schematic diagram of the receiver with vertical finned structures and concave bottom.

where nwall is the number of the wall, θ = 360°/nfin. panels as shown in Fig. 7, where the total panel area (Ar) is calculated
by Eq. (5). The length of the fins (l) should not be too short because the
2.3.3. Receiver with vertical finned structures and convex bottom (Receiver- short tubes with elbow connections will cause significant pressure loss.
III) As a result, the number of walls or fins (nfin) is constrained, and a
The receiver with vertical finned structures and convex bottom constant value of 4 is used in current work. The horizontal fins in each
(Receiver-III) is composed of vertical finned panels and convex bottom direction are parallel to the vector from the aiming point to the mid-
panels as shown in Fig. 6. The convex bottom panels are arranged point of the edge point of the field and the base point of the tower as
around the wall with a diameter of Dw and a constant angle β of 120°. shown in Fig. 1. The inclination angles of the fins in the north (αN), east
The Dw and the fin number (nfin) will be optimized later, while the rest (αE), south (αS) and west (αW) direction are 74.74°, 71.32°, 66.34° and
parameters are the same as those of the traditional cylindrical receiver. 71.32°, respectively. The length of the horizontal fins (l) will be opti-
Its panel area (Ar) is calculated by Eq. (4). mized later. The rest parameters are the same as those of the traditional
cylindrical receiver.
Dw θ β
Ar = n fin LHf + n wall Lsin ⎛ ⎞/sin ⎛ ⎞ Ar = (n fin Hf + n wall L) l (5)
2 2
⎝ ⎠ 2
⎝ ⎠ (4)
where nwall is the number of the wall.
where nwall is the number of the wall, θ = 360°/nfin.
2.3.5. Receiver with horizontal finned structures and concave bottom
2.3.4. Receiver with horizontal finned structures and flat bottom (Receiver- (Receiver-V)
IV) The receiver with horizontal finned structures and concave bottom
The receiver with horizontal finned structures and flat bottom (Receiver-V) consists of horizontal finned panels and concave bottom
(Receiver-IV) is composed of horizontal finned panels and flat bottom panels as shown in Fig. 8. The concave bottom panels are arranged

Wall Tube

N
Yr
S

Hf
Xr Xr
L

Zr

De
Zr

Fig. 6. Schematic diagram of the receiver with vertical finned structures and convex bottom.

223
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

Wall Tube
N
Yr
W E
S

Xr Xr

L
Zr
De

l
E Zr
Hf
Fig. 7. Schematic diagram of the receiver with horizontal and angled fins.

around the wall with a diameter of Dw and a constant angle β of 3.1. Optical simulation model
120°.The total panel area (Ar) is calculated by Eq. (6). Similar to Re-
ceiver-4, the number of the walls or fins (nfin) is constrained, and a The optical processes in the heliostat field mainly include sunray
constant value of 4 is used in current work. The horizontal fins in each initialization, shading, reflection and blocking as shown in Fig. 1.
direction are parallel to the vector from the aiming point to the mid- Shading is a part of heliostat shadowed by the adjacent heliostats or the
point of the edge point of the field and the base point of the tower as tower while blocking is a part of reflected sunray blocked by other
shown in Fig. 1. The inclination angles of the fins in the north (αN), east heliostats [33]. The optical processes in both the cylindrical receiver
(αE), south (αS) and west (αW) direction are 74.74°, 71.32°, 66.34° and and the novel receiver with finned structures include absorption, re-
71.32°, respectively. The length of the horizontal fins (l) will be opti- flection and reabsorption, as shown in Fig. 9.
mized later. The rest parameters are the same as those of the traditional All optical processes are modeled through a code program called
cylindrical receiver. SPTOPTIC, which is developed by Qiu and He [33] based on Monte
Carlo Ray Tracing (MCRT) model. MCRT is a statistic method in which
L β a large number of sunrays are initialized and traced [40]. In the si-
Ar = ⎛n fin Hf + n wall ⎛ ⎞/sin ⎛ ⎞ ⎞ l
⎜ ⎟

⎝ ⎝2⎠ ⎝ 2 ⎠⎠ (6) mulation, SPTOPTIC works in the following way. Firstly, sunrays are
uniformly initialized on the heliostat [13]. Secondly, the shading, re-
where nwall is the number of the wall. flection and blocking of the sunrays in the heliostat field are simulated.
Finally, the absorption, multiple reflections and reabsorption processes
in the receiver are simulated if the sunrays hit the receiver. In the
3. Optical simulation and validation model, some important assumptions are detailed as follows. The un-
parallel sunrays are assumed to be uniformly distributed within a cone
In this section, an optical simulation model is introduced and vali- with an apex angle of 9.3 mrad [41]. Each heliostat is assumed as a
dated. Moreover, grid system and sunray number independence are also continuous spherical surface by ignoring the gaps between the facets,
tested.

Tube Yr Wall

N
W E
S

Xr
Xr
L

Zr De
Zr
l
E

Fig. 8. Schematic diagram of the receiver with horizontal and angled fins.

224
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

Reflection

Reabsorption
Reflection
Incident rays
Incident rays
(a) cylindrical receiver (b) receiver with finned structure
Fig. 9. Sunray transfer process in cylindrical receiver and finned receiver.

and the radius of each heliostat equals to double that of the corre- computation to ensure the grid independent results and fine resolution
sponding slant range. Both the slope error of the facets and the tracking of the local flux.
errors are assumed to follow the Gaussian distribution law [42,43]. To guarantee the convergence of the simulation, a set of numbers of
More details about the SPTOPTIC model can be found in Ref. [33]. sunrays are used to calculated the maximum solar flux (qmax) and op-
After the simulation, quadrilateral grids are generated on the sur- tical efficiency (η) of the cylindrical receiver at spring equinox noon,
face of the tubes and walls to count the absorbed sunrays, as shown in and the result is presented in Fig. 12. The deviations of qmax and η in
Fig. 10. All tubes have the same grids, and the grids are uniform in both two successive numbers of the sunrays are less than 0.5% when the
circumferential and lengthwise directions on each tube. sunray number Np is larger than 1 × 109 and 1 × 107, respectively.
To characterize the performance, some parameters including the Moreover, it takes the computer with intel(R) Core(TM) i5-4590 CPU
optical efficiency (η), intercept efficiency (ηi), effective absorptivity (αe) processor 457 s to complete the simulation when the number of the
and solar flux (q) on the receiver are defined as Eqs. (7)–(10), respec- sunrays is 1 × 107.
tively. To validate the optical simulation model, the computed solar flux
distribution of the cavity receiver in PS10 plant is compared with that
η = Qt / Q h (7)
in Ref. [22]. The geometric parameters, optical parameters, and simu-
ηi = Qi/ Qref (8) lation conditions in current work are the same with that in Ref. [22].
The PS10 plant consists of a heliostat filed with 624 heliostats, and a
α e = Qt / Q i (9) cavity receiver with an elevation of 93.2 m. The detailed parameters of
q = Qe/ Ae the receiver and heliostats are presented in Table 3. After the solar flux
(10)
distribution on the receiver of PS10 plant is obtained, it is compared
where Qt is the energy absorbed by the tubes. Qh is the energy received with that in Ref. [22], as shown in Fig. 13. It can be seen that the solar
by the heliostat field which is calculated by Eqs. (11) and (12) [44]. Qi flux map obtained by the present model as shown in Fig. 13(a) is
is the energy intercepted by the receiver. Qref is the energy reflected by conformable with that in Ref. [22], as shown in Fig. 13(b). The current
the heliostats without blocking or atmospheric attenuation. Qe is the
energy absorbed by the element of the tube; Ae is the area of the ele-
ment.
Q h = DNI ·Hh Wh nh (11)

2Nday π ⎤ sinα s
DNI = 1376 ⎡1 + 0.033cos ⎛⎜
⎞ ·
⎟

⎢ ⎝ 365 ⎠ ⎥
⎣ ⎦ sinα s + 0.33 (12)
In the above equation, DNI is the Direct Normal Irradiance. Hh and
Wh are the height and width of the heliostat, respectively. nh is the
number of the heliostat. Nday is the day number in a year, which starts
from January 1st. αs is the solar altitude.

3.2. Model validation

To avoid the effect of the grid number on the result, four grid sys-
tems have been checked by examining the local circumferential solar
flux distribution on the cylindrical receiver at spring equinox noon. For
each grid system, all tubes have the same grid (circumfer-
ential × lengthwise), and adequate sunrays of 3.5 × 109 are traced.
The result is shown in Fig. 11, where the examined local flux is located
at Yt = 0 on Tube No.1 in Fig. 3. It is seen that the profile of the solar
flux does not change apparently when the grid of each tube is larger
than 25 (circumferential) × 80 (lengthwise). Finally, a grid system of
35 (circumferential) × 160 (lengthwise) was used in the following Fig. 10. Quadrilateral grids on the surface of tubes and walls.

225
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

2.0 4. Results and discussion

Tube No.1
Local solar flux q / 106 W·m-2
Yt=0 In this section, the geometrical parameters of the novel fin-like re-
1.5 ceivers are firstly optimized. Then, the optical efficiencies and solar
fluxes of the optimal novel receivers are compared with those of the
cylindrical receiver. Finally, a multi-point aiming strategy is adopted to
1.0
reduce the overlarge peak fluxes of the novel receivers.
5 20
15 40 4.1. Parametric optimization of novel receivers with finned structures
0.5 25 80
35 160
The parametric optimization is conducted at the noon of the spring
Grid system for each tube: equinox. For the receivers with vertical finned structures, the optical
0.0
Circumferential Lengthwise performance at different values of the diameter of the wall (Dw) and the
0 60 120 180 240 300 360 fin number (nfin) are studied. Detailed parameters of all parametric
optimization cases are shown in Table 4. For the receiver with hor-
Circumferential angle variable on tube
izontal fins, the optical performance at different lengths of the hor-
Fig. 11. Grid-independence test. izontal fins (l) is studied, and all parametric optimization cases are
listed in Table 5. For each receiver, the study will try to find an optimal
9.0 60.9 geometry by examining the results of its corresponding cases.
106 W·m-2

Grid system for each tube: Fig. 14 shows the variation of the optical efficiency (η) with the
Circumferiential Lengthwise = 35 160 diameter of the wall (Dw) and the fin number (nfin) of the receiver with
7.5 60.8
vertical fins and flat bottom (Receiver-I). It is seen that η decreases with
the increase of Dw when the fin number are 4, 6 and 8, respectively.
Optical efficiency /

6.0 60.7
Maximum flux on tubeqmax /

This is because the fin heights (Hf) decrease with the increase of Dw,
which weakens the reabsorption. However, when nfin is 10, η increases
4.5 60.6 first, then slightly decreases with the ascending Dw and reaches the
maximum value when Dw is near 4 m. This is because the rising Dw will
3.0 qmax 60.5 cause the decrease of Hf and increase of receiver’s effective diameter
(De). The former weakens the reabsorption, while the later increases the
intercept efficiency. As a result, there is an optimal Dw.
1.5 60.4
It is also observed that η increases first, then decreases with rising
nfin at the same Dw, and reaches the maximum value when nfin is 6. This
0.0 60.3 is because the increase of nfin can strengthen the reabsorption. Hf de-
105 106 107 108 109 1010 creases with the increase of nfin, which weakens the reabsorption. As a
Sunray number Np
result, a compromise can be reached between the two parameters, and
Fig. 12. Sunray-independence analysis.
an optimal nfin can be obtained at the same Dw. After the above dis-
cussion, it is found that the Receiver-1 with nfin = 6 and Dw = 1 m can
obtain the largest optical efficiency, which can be chosen as the optimal
Table 3 design.
Parameters and assumptions of PS10 plant.
For the receiver with vertical finned structures and concave bottom
Parameters Dimensions Parameters Dimensions (Receiver-II) and the receiver with vertical finned structure and convex
bottom (Receiver-III), parameter studies of the diameter of the wall
Receiver elevation 93.2 m Number of heliostats 624
(Dw) and the fin number (nfin) are conducted. The results are illustrated
Aperture height 16 m Heliostat width 12.84 m
Aperture width 14.83 m Heliostat height 9.45 m in Figs. 15 and 16, respectively. Similar effects of Dw and nfin on the
Panels height 12 m Reflectivity 0.88 optical efficiency as those described in the analysis of Receiver-I are
Single panel width 5.36 m Total Error on Reflected ray 2.9 mrad observed. In general, both Receiver-II and Receiver-III with nfin = 6 and
Dw = 1 m reach their maximum efficiencies. Therefore, these designs
are chosen as optimal receivers.
calculated peak flux and Ref. [22] are 714.1 kW·m−2 and Fig. 17 shows the variation of the optical efficiency (η) and effective
714.0 kW·m−2, respectively. The deviation is 0.01% that indicates the absorptivity (αe) with the length of the horizontal tube (l) for the re-
present model is reliable. ceiver with horizontal finned structures and flat bottom (Receiver-IV),
and concave bottom (Receiver-V). It can be seen that the optical

12 kW·m2 12 kW·m2
Receiver Height / m
Receiver Height / m

650 650
10 10
550 550
8 450 8 450
6 350 6 350
4 250 4 250
2 150 2 150
50 50
00 0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
Width / m Width / m
Peak flux=714.1 kW·m-2 Peak flux=714.0 kW·m-2
(a) Solar flux obtained by present model (b) Solar flux obtained by DELSOL Ref [22]
Fig. 13. Flux map on PS10 receiver between present work and published data.

226
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

Table 4
Geometric parameters for the receiver with vertical finned structures.
nfin Dw, De (m)

Receiver-I 4 1.0, 13.0 2.0, 12.6 3.0, 12.2 4.0, 11.8 5.0, 11.4 6.0, 10.9
6 1.0, 8.9 2.0, 8.9 3.0, 8.9 4.0, 8.9 5.0, 8.9 6.0, 8.9
8 1.0, 7.0 2.0, 7.2 3.0, 7.5 4.0, 7.7 5.0, 7.9 6.0, 8.2
10 1.0, 5.8 2.0, 6.2 3.0, 6.5 4.0, 7.0 5.0, 7.4 6.0, 7.7

Receiver-II 4 1.0, 12.7 2.0, 12.1 3.0, 11.5 4.0, 10.9 5.0, 10.2 6.0, 9.7
6 1.0, 8.8 2.0, 8.7 3.0, 8.5 4.0, 8.4 5.0, 8.2 6.0, 8.1
8 1.0, 6.9 2.0, 7.0 3.0, 7.1 4.0, 7.3 5.0, 7.3 6.0, 7.6
10 1.0, 5.7 2.0, 6.1 3.0, 6.3 4.0, 6.6 5.0, 6.8 6.0, 7.3

Receiver-III 4 1.0, 12.7 2.0, 12.1 3.0, 11.5 4.0, 10.9 5.0, 10.2 6.0, 9.7
6 1.0, 8.8 2.0, 8.6 3.0, 8.5 4.0, 8.4 5.0, 8.2 6.0, 8.1
8 1.0, 7.0 2.0, 7.0 3.0, 7.1 4.0, 7.2 5.0, 7.4 6.0, 7.4
10 1.0, 5.7 2.0, 6.1 3.0, 6.3 4.0, 6.6 5.0, 6.8 6.0, 7.2

Table 5
Geometric parameters for the receiver with horizontal finned structures.
l, De (m)

Receiver-IV 3.0, 11.3 3.5, 10.0 4.0, 9.2 4.5, 8.9 5.0, 8.8 5.5, 8.9 6.0, 9.0
Receiver-V 3.0, 11.3 3.5, 10.0 4.0, 9.2 4.5, 8.9 5.0, 8.8 5.5, 8.9 6.0, 9.0

Fig. 14. Parametric optimization results for Receiver-I. Fig. 16. Parametric optimization results for Receiver-III.

performances of the Receiver-IV and Receiver-V have the same varia-

tion trend with increasing l. For both of the two receivers, η increase
with the increasing l, but the increasing trend becomes slower. This is
because the rising l can increase the intercept efficiency. Additionally, it
is observed that αe decreases with the rising l. This is because the rising l
causes the decrease of the fin heights (Hf), which weakens the re-
absorption. The Receiver-IV with l = 6 m and Receiver -V with l = 6 m
are selected as their respective optimal one.
After the above parametric optimization, the optical performance of
the five optimal novel receivers and the traditional cylindrical receiver
are compared with each other. Fig. 18(a) shows the comparison of the
optical efficiencies (η). It is observed that the novel receivers with
vertical fins (Receiver-I, II, III) can improve the optical efficiency by
3.2, 3.1 and 3.1 percentage points, respectively, when compared to the
cylindrical receiver. Moreover, the Receiver-I achieves the maximum η
of 63.9%. This because both relatively large intercept efficiency (ηi) and
effective absorptivity (αe) can be achieved by the novel receiver as
shown in Fig. 18(b). For the Receiver-I, its effective absorptivity reaches
94.5% when the coating absorptivity of the tubes is 90%.
Fig. 15. Parametric optimization results for Receiver-II. For the novel receiver with horizontal fins (Receiver-IV, V), it can be

227
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

(a) Receiver-IV (b)Receiver-V

Fig. 17. Parametric optimization results for Receiver-IV and Receiver-V.

seen from Fig. 18 (a) that the Receiver-V has an optical efficiency of time because of the sun blocked by the cloud [45]. To examine the
61.5%, which is higher than that of the cylindrical receiver by 0.8 optical performance of the novel receivers under cloudy conditions, the
percentages points. While the Receiver-IV obtains the minimum η of real variation of DNI with time on a cloudy day, as shown in Fig. 19(b),
60.3%, which is lower than that of the cylindrical receiver by 0.4 is adopted to calculate η of the receivers. The real variation of DNI was
percentage points. This is because although the effective absorptivity measured on the September 19, 2002 at PSA [46], South of Spain, only
(αe) is improved by the reabsorption in the Receiver-IV as shown in 267 km away from the place where Gemasolar is located. The above
Fig. 18 (b), the intercept efficiency (ηi) is decreased due to the decrease real DNI variation on a cloudy day will be used to evaluate the optical
of an effective diameter (De) when the receiver panel area is kept a performances of the novel receivers.
constant among different designs as described in Section 2.3. Since η of Fig. 20 shows the variations of optical efficiencies (η) with local
the receivers vertical fins (Receiver-IV, V) are lower than those of the solar time (ts) on three typical days under sunny condition. It is ob-
receivers with vertical fins (Receiver-I, II, III), Receiver-IV and Re- served that η of the three novel receivers are higher than that of the
ceiver-V will not be discussed in the following sections. While other cylindrical receiver on all the three days. In addition, the difference
three optimal novel receivers with vertical finned structures will be between the η of the optimal receivers and the cylindrical receiver
discussed in a detailed way. reaches the maximum value when ts = 12 on each day. Moreover, the
differences among the η of the three optimal receivers are quite small. It
is also found that η has the maximum values on summer solstices while
4.2. Optical efficiencies under different time
having the minimum values on winter solstice for all the receivers. This
is because the solar altitude is the highest on the summer solstice and
The above parametric optimization process is conducted at the noon
lowest on the winter solstice in the northern hemisphere.
of spring equinox under sunny condition. Therefore, it is still necessary
Fig. 21 shows the optical performances of the receivers on a cloudy
to examine if the optical performance of three optimal novel receivers
day. It can be seen from Fig. 21 (a) that η of the three novel receivers
with vertical finned structures is better than that of the cylindrical re-
with vertical finned structures are higher than that of the cylindrical
ceiver in the whole year under different weather conditions including
receiver during all day. Moreover, the dramatic fluctuation of the DNI
sunny and cloudy days. Therefore, the optical efficiencies of these re-
due to the presence of the cloud has little influence on the optical ef-
ceivers are comparatively studied under different weather conditions
ficiencies of the receivers. This is because the solar energy received by
including sunny and cloudy days. For the sunny days, direct normal
the heliostat field decreases although the solar energy absorbed by the
irradiance (DNI) changes steadily with time in a day, as shown in
tubes decreases due to the presence of the cloudy, as shown in Fig. 21
Fig. 19(a). While for the cloudy days, DNI fluctuates dramatically with

(a) Optical efficiency (b) Intercept efficiency and effective absorptivity

Fig. 18. A comparison of optical performance among different geometries.

228
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

1000 900

750
900

600

DNI / W m-2
DNI / W m-2

800
450
700
300

600
150

500 0
7 8 9 10 11 12 13 14 15 16 17 7 8 9 10 11 12 13 14 15 16 17
time / h time / h
(a) sunny condition, spring equinox (b) cloudy condition, September 19, 2002 [48]
Fig. 19. Variations of DNI under both sunny and cloudy condition.

(b). As a result, the optical efficiencies of the receivers have a small thermal performance and safety. Therefore, in this section, the real-time
change. solar fluxes on three optimal novel receivers with vertical finned
structures are examined and compared with that of the cylindrical re-
4.3. Real-time solar flux distributions ceiver at the noon of the spring equinox.
Fig. 22 shows the non-uniform solar fluxes on the optimal Receiver-
The solar flux on the CSP receiver is of great importance to its I, Receiver-II, Receiver-III and the cylindrical receiver. It is seen that the

65 68
60
64
55
/

60
i,t

i,t

50
Optical efficiency

Optical efficiency

45 Receiver-I:Flat 56
Receiver-II:Concave Receiver-I:Flat
40 Receiver-III:Convex 52 Receiver-II:Concave
35 Cylindrical receiver Receiver-III:Convex
48 Cylindrical receiver
30
44
25
20 40
6 8 10 12 14 16 18 6 8 10 12 14 16 18
The local solar time ts / h The local solar time ts / h
(a) Spring equinox (b) Summer solstice

60

55

50
/ i,t
Optical efficiency

45 Receiver-I:Flat
Receiver-II:Concave
40 Receiver-III:Convex
Cylindrical receiver
35

30

25
7 8 9 10 11 12 13 14 15 16 17
The local solar time ts / h
(c) Winter solstice
Fig. 20. Variations of optical efficiencies on three typical days under sunny condition.

229
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

65 150

Absorbed solar energy Q / MW

/ i,t
Optical efficiency 60 120

55 90
Receiver-I: Flat
Receiver-II: Concave
50 Receiver-III: Convex 60 Flat
Cylindrical receiver Concave
Convex
45 30 Cylindric

40 0
7 8 9 10 11 12 13 14 15 16 17 7 8 9 10 11 12 13 14 15 16 17
The local solar time ts / h The local solar time ts / h
(a) optical efficiency (b) absorbed solar energy
Fig. 21. Optical performances on the cloudy day.

peak fluxes which locate at the northern sides of the above receivers are receiver is about 1.0 MW·m−2 [47–49]. Overlarge flux will cause a local
3.83 × 106 W·m−2, 3.74 × 106 W·m−2, 3.87 × 106 W·m−2 and hot spot and high thermal stress, which can degrade the absorbing
1.97 × 106 W·m−2, respectively. The peak fluxes on the three optimal coating and even cause stress failure [47,50]. Fortunately, the problems
receivers are nearly double that of the cylindrical receiver. This is be- caused by overlarge peak solar flux can be relieved by using a multi-
cause the wall diameters of the novel receivers are much smaller than point aiming strategy [50–54]. Moreover, if the liquid metal is used as
that of the cylindrical receiver. As a result, solar power is concentrated the heat transfer fluid, the peak flux can be maintained as large as
in a smaller area in every novel receiver, which results in a larger solar 10 MW·m−2 due to the high thermal conductivities of the liquid metals.
flux. Moreover, the Receiver-III has the largest peak flux among the Therefore, the current peak flux will be completely acceptable for a
novel receiver, and the peak flux is located in the convex part. This is liquid-metal receiver.
because the tubes in the convex part can receive more reflected sunrays
when compared with those of the Receive-I and receiver-II. In addition,
it is also observed that for all the receivers, the solar flux on each tube 4.4. Peak flux reduction using multi-point aiming
appears to be extremely non-uniform, where flux is quite small at the
two ends but extremely large at the middle part. Usually, it is expected that the peak flux on a receiver should be as
It is well known that an acceptable peak flux for a molten-salt small as possible at the same absorbed thermal power for avoiding
safety issue described above. Thus, a multi-point aiming strategy is

W / m2 W / m2 W / m2 W / m2
3.8E6 3.8E6 3.8E6 3.8E6
5.0
3.4E6 5.0 3.4E6 3.4E6 5.0 3.4E6
5.0
3.0E6 3.0E6 3.0E6 3.0E6
2.5 2.6E6 2.6E6 2.6E6 2.6E6
2.5
2.2E6 2.2E6 2.5 2.2E6 2.5 2.2E6
1.8E6
Yr / m

1.8E6 1.8E6 1.8E6

Yr / m

0
Yr / m

0 1.4E6 1.4E6 1.4E6

Yr / m

1.4E6 0
0
1.0E6 1.0E6 1.0E6 1.0E6
-2.5 0.6E6 -2.5 0.6E6 0.6E6 0.6E6
0.2E6 0.2E6 -2.5 0.2E6 -2.5 0.2E6
-5.0 -5.0
-5.0 -5.0

qmax= 3.83×106 W·m-2, =63.9% qmax= 3.74×106 W·m-2, =63.8%

(a) Receiver-I (b) Receiver-II
W / m2 2 W / m2 W / m2
W/m 2.0E6 2.0E6
3.8E6 3.8E6
5.0 5.0 5.0 1.8E66 5.0 1.8E6
3.4E6 3.4E6
3.0E6 1.6E66 1.6E6
3.0E6
2.6E6 1.4E66 1.4E6
2.5 2.5 2.6E6 2.5
2.2E6 1.2E66 2.5 1.2E6
2.2E6
1.8E6 1.0E66 1.0E6
Yr / m

1.8E6
Yr / m
Yr / m

Yr / m

0 1.4E6 0 0 0.8E66 0 0.8E6

1.4E6
1.0E6 0.6E66 0.6E6
1.0E6
0.6E6 0.4E66 0.4E6
-2.5 -2.5 0.6E6 -2.5 -2.5
0.2E6 0.2E66 0.2E6
0.2E6

-5.0 -5.0 -5.0 -5.0

qmax= 3.87×106 W·m-2, =63.8% qmax= 1.97×106 W·m-2, =60.7%

(c) Receiver-III (d) Cylindrical receiver
Fig. 22. Real-time solar flux distributions on the receivers (single-point aiming).

230
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

adopted to reduce the peak solar fluxes on the optimal novel receivers. Receiver
rk
In this strategy, the heliostat field is divided into odd rows and even rP
rows, as shown in Fig. 2. The heliostats in each row aim at the same P
height level on the receiver. For the odd heliostat rows, the innermost
row aims at the upper level on the receiver, while the outermost row
aims at the bottom level on the receiver. The rest rows aim at the points Ȧh
equally distributed between the upper and bottom levels. For the even
heliostat rows, the innermost row aims at a bottom level, while the
outermost row aims at an upper level. Other rows aim at the points
h
equally distributed between the upper and bottom levels. Fig. 23 il- H
lustrates the flow chart of the multi-point aiming strategy adopted in Heliostat
this study. Fig. 24. The spread process of sunray.
The most important procedure for this strategy is calculating the
upper and bottom aiming points for each row. The first step is to de-
distribution. Although this assumption used is different from the model
termine the radius (rk) of the reflected beam of the heliostats in each
described in the optical simulation model, the beam radiuses obtained
row. The heliostat, which has the largest distance to the tower in each
by the two assumptions are quite close. Therefore, the Gaussian as-
row, is selected as the representative heliostat to determine rk. The
sumption is used to estimate the radius of the beam. ωh is reflection
reflected sunrays are assumed to follow a circular Gaussian distribution
angle at the center of a heliostat. The value of ωh depends on the aiming
as shown in Fig. 24, and its standard deviation can be calculated using
point of each heliostat, for simplification, cosωh is set to be 1.0. σte is
Eq. (13) [55,56]. The flux radius of a heliostat (rk) is estimated using
tracking error of heliostat which is equal to (σte,1 + σte,2)1/2. DHP is the
Eq. (14).
distance between the center of a heliostat and its aiming point. k is the
σe = 2
σsun + 2(1 + cos2 ωh ) σse2 + 4σte2 (13) aiming factor, and it will be adjusted to obtain the satisfactory aiming
points of the innermost and outermost rows.
r k = DHP ·tan(kσe ) ≅ DHP ·kσe k > 0 (14) The second step is to determine the radius of beam’s projection (rp)
as shown in Fig. 24 by Eq. (15).
where σsun is the standard deviation of the unparallel sunrays which is
equal to 2.73 mrad when the sunrays are assumed to follow a Gaussian

Start

Divide the heliostat field into

odd rows and even rows

Chose the innermost heliostats

from the odd rows and even rows

Odd rows Even rows

Odd rows or even rows?

Calculate the aiming point of each heliostat to make the Calculate the aiming point of each heliostat to make the
projected beam tangent to the upper boundary of receiver projected beam tangent to the bottom boundary of receiver

Chose the outermost heliostats

from the odd rows and even rows

Odd rows Even rows

Odd rows or even rows?

Calculate the aiming point of each heliostat to make the Calculate the aiming point of each heliostat to make the
projected beam tangent to the bottom boundary of receiver projected beam tangent to the upper boundary of receiver

The rest aiming point are uniformly distributed between

the upper and bottom aiming points

End

Fig. 23. The flow chart of the multi-point aiming strategy.

231
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

rp = r k /cosεh (15) Table 6

Peak solar flux and optical efficiencies of different receivers. I: Flat receiver, II:
where ɛh is the elevation angle of the reflected sunray. Concave receiver, III: Convex Receiver, IV: Cylindrical Receiver.
Finally, the aiming level of each row is determined by making a
Summer solstice Winter solstice
projection circle of corresponding representative heliostat tangent to
the receiver. As shown in Fig. 24, point P will be chosen as the aiming Receiver type I II III IV I II III IV
point of the illustrated heliostat, and the level of the point will be se-
Solar flux Single- 3.73 3.51 3.59 1.83 3.08 3.09 3.17 1.74
lected as the level of the corresponding row. If the diameter of the
MW·m−2 point
projection circle is larger than the height of the receiver, the aiming Multi- 2.15 2.06 2.17 1.14 1.89 1.83 1.93 1.09
point will be placed on the equator. point
Fig. 25 shows the solar flux distributions on the optimal Receiver-I, η/% Single- 65.0 64.9 64.9 61.8 59.3 59.2 59.2 56.4
Receiver-II, Receiver-III and cylindrical receiver. The peak fluxes of the point
above receivers are 2.20 × 106 W·m−2, 2.18 × 106 W·m−2, 2.29 × 106 Multi- 64.6 64.5 64.5 60.9 59.1 59.0 58.9 55.2
W·m−2 and 1.29 × 106 W·m−2, respectively. They are reduced by point
42.6%, 41.7%, 40.8% and 34.5%, respectively when compared with the
single-point aiming strategy. Moreover, a sacrifice smaller than 1.1
Outlet 1
percentage points in the optical efficiency is observed, which is ac- Outlet 2
ceptable.
The above reduction of the peak solar fluxes is conducted at the
noon of the spring equinox. To examine whether this multi-point
aiming strategy works in a whole year or not, the strategy is applied to
the novel receivers and traditional cylindrical receivers at the noon of
summer solstice and winter solstice. Table 6 presents the peak solar
fluxes of the three optimal novel receivers and the cylindrical receiver
with both single point strategy and multi-point aiming strategy at the
noon of the two typical days. It can be seen that the peak solar fluxes of
the four receivers can be effectively reduced by the multi-point aiming
strategy with only a little sacrifice of optical efficiencies, which means
that this multi-point aiming strategy works in the whole year.

4.5. Manufacturability of the receiver with vertical finned structure Inlet 1 Outlet 1
Inlet 2 Inlet 2 Inlet 1 Outlet 2

The above results show that the receiver with vertical finned (a) Cylindrical receiver (b)Fin-like receiver
structure and flat bottom (Receiver-I) has the best optical performances
Fig. 26. The design of receivers.
among the five novel receivers. In order to apply this receiver to the real

W / m2 W / m2 W / m2 W / m2
2.0E6 2.0E6 2.0E6 2.0E6
5.0 1.8E6 5.0
1.8E6 5.0 5.0 1.8E6 1.8E6
1.6E6 1.6E6 1.6E6 1.6E6
1.4E6 1.4E6 1.4E6 2.5 1.4E6
2.5 1.2E6 2.5
1.2E6 2.5 1.2E6 1.2E6
Yr / m

1.0E6 1.0E6 1.0E6 1.0E6

Yr / m
Yr / m

0
Yr / m

0 0.8E6 0.8E6 0.8E6 0.8E6

0 0
0.6E6 0.6E6 0.6E6 0.6E6
0.4E6 0.4E6 0.4E6 -2.5 0.4E6
-2.5
0.2E6 -2.5 0.2E6 -2.5 0.2E6 0.2E6

-5.0 -5.0
-5.0 -5.0

qmax= 2.20×106 W·m-2, =62.8%, k=1.5 qmax= 2.18×106 W·m-2, =63.4%, k=1.5
(a) Receiver-I (b) Receiver-II
W / m2 W / m2 W / m2 W / m2
2.0E6 2.0E6 1.1E6 1.2E6
5.0
5.0 1.8E6 1.1E6 1.1E6
5.0 1.8E6
1.6E6 5.0 1.0E6 1.0E6
1.6E6
1.4E6 0.9E6 0.9E6
1.4E6 2.5
2.5 2.5 0.8E6 0.8E6
1.2E6 1.2E6 2.5 0.7E6 0.7E6
1.0E6
Yr / m

1.0E6 0.6E6 0.6E6

0
Yr / m

Yr / m

0 0.8E6 0.8E6 0.5E6 0.5E6

0
Yr / m

0.6E6 0.6E6 0 0.4E6 0.4E6

0.4E6 0.4E6 0.3E6 -2.5 0.3E6
-2.5 0.2E6 -2.5 0.2E6 0.2E6
0.2E6 -2.5

-5.0 -5.0
-5.0 -5.0

qmax= 2.29×106 W·m-2, =62.7%, k=1.5 qmax= 1.29×106 W·m-2, =59.6%, k=0.5
(c) Receiver-III (d) Cylindrical receiver
Fig. 25. Solar flux distributions on the receivers (multi-aiming point).

232
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

solar power plant, its manufacturability should be discussed. For the National Natural Science Foundation of China (No. 51721004).
traditional cylindrical receiver, it is constructed by arranging the ver-
tical tube panels into a cylindrical geometry, as shown in Fig. 26 (a), References
where the flow is usually divided into two flow paths. Since the novel
receiver is also composed of vertical tube panels, similarly, it can be [1] Li MJ, Song CX, Tao WQ. A hybrid model for explaining the short-term dynamics of
also manufactured by rearranging the vertical tube panels. Fig. 26 (b) energy efficiency of China’s thermal power plants. Appl Energy 2016;169:738–47.
[2] McGlade C, Ekins P. The geographical distribution of fossil fuels unused when
shows the final design of the receiver with vertical finned structures, limiting global warming to 2 ℃. Nature 2015;517:187–90.
which is very similar to that of the cylindrical receiver. The flow is also [3] Li MJ, Zhu HH, Guo JQ, Wang K, Tao WQ. The development technology and ap-
divided into two flow paths, which separately flows through the bottom plications of supercritical CO2 power cycle in nuclear energy, solar energy and other
energy industries. Appl Therm Eng 2017;126:255–75.
tube panels and the fin tube panels following a serpentine flow pattern. [4] Cheng ZD, He YL, Du BC, Wang K, Liang Q. Geometric optimization on optical
The heat transfer fluid (HTF) first enters the tubes of the tube panel performance of parabolic trough solar collector systems using particle swarm op-
from the inlet of the receiver. Then, the HTF will be heated by the timization algorithm. Appl Energy 2015;148:282–93.
[5] Daabo AM, Al Jubori A, Mahmoud S, Al-Dadah RK. Development of three-dimen-
concentrated solar power before it enters into the header, which is lo- sional optimization of a small-scale radial turbine for solar powered Brayton cycle
cated at the end of each tube panel. After that, the HTF will enter into application. Appl Therm Eng 2017;111:718–33.
the next tube panel from the header until it goes out from the outlet of [6] Li MJ, He YL, Tao WQ. Modeling a hybrid methodology for evaluating and fore-
casting regional energy efficiency in China. Appl Energy 2017;185:1769–77.
the receiver.
[7] Peng X, Root TW, Maravelias CT. Storing solar energy with chemistry: the role of
thermochemical storage in concentrating solar power. Green Chem
5. Conclusions 2017;19:2427–38.
[8] Qiu Y, Li MJ, Wang WQ, Du BC, Wang K. An experimental study on the heat transfer
performance of a prototype molten-salt rod baffle heat exchanger for concentrated
In this paper, five novel fin-like receivers are proposed to improve solar power. Energy 2018;156:63–72.
the optical efficiency of the solar power tower. After parametric opti- [9] He YL, Cui FQ, Cheng ZD, Li ZY, Tao WQ. Numerical simulation of solar radiation
mization, the optical efficiencies and the solar flux distributions on the transmission process for the solar tower power plant: from the heliostat field to the
pressurized volumetric receiver. Appl Therm Eng 2013;61:583–95.
novel receivers are compared with those of the traditional cylindrical [10] Kim J, Kim JS, Stein W. Simplified heat loss model for central tower solar receiver.
receiver. In the end, a multi-point aiming strategy is adopted to reduce Sol Energy 2015;116:314–22.
the overlarge peak solar flux. The main conclusions are summarized as [11] Chang HW, Duan C, Wen K, Liu YT, Xiang C, Wan ZM, et al. Modeling study on the
thermal performance of a modified cavity receiver with glass window and sec-
follows. ondary reflector. Energy Convers Manage 2015;106:1362–9.
[12] Cui FQ, He YL, Cheng ZD, Li YS. Study on combined heat loss of a dish receiver with
(1) An optimal fin number of 6 is found for the novel receivers with quartz glass cover. Appl Energy 2013;112:690–6.
[13] Qiu Y, Li MJ, He YL, Tao WQ. Thermal performance analysis of a parabolic trough
vertical finned structures, under which the optical efficiencies reach solar collector using supercritical CO2 as heat transfer fluid under non-uniform solar
the maximum value. The optimal novel receiver with vertical flux. Appl Therm Eng 2017;115:1255–65.
finned structures and the flat bottom has the maximum optical ef- [14] Bellos E, Tzivanidis C, Antonopoulos KA, Gkinis G. Thermal enhancement of solar
parabolic trough collectors by using nanofluids and converging-diverging absorber
ficiency of 63.9% at the noon of the spring equinox. While the
tube. Renewable Energy 2016;94:213–22.
optimal novel receiver with horizontal finned structures and the flat [15] Wang K, He YL, Cheng ZD. A design method and numerical study for a new type
bottom has the minimum optical efficiency of 60.3% under the parabolic trough solar collector with uniform solar flux distribution. Sci China
same condition, which is higher and lower than that of the tradi- Technol Sci 2014;57:531–40.
[16] Qiu Y, He YL, Wu M, Zheng ZJ. A comprehensive model for optical and thermal
tional cylindrical receiver by 3.2% and 0.4%, respectively. characterization of a linear Fresnel solar reflector with a trapezoidal cavity receiver.
(2) The finned structures can significantly increase the effective ab- Renewable Energy 2016;97:129–44.
sorptivity by reabsorbing the reflected sunrays. The optimal novel [17] Qiu Y, He YL, Cheng ZD, Wang K. Study on optical and thermal performance of a
linear Fresnel solar reflector using molten salt as HTF with MCRT and FVM
receiver with vertical finned structures and the flat bottom has the methods. Appl Energy 2015;146:162–73.
maximum effective absorptivity of 94.5% when the coating ab- [18] Behar O, Khellaf A, Mohammedi K. A review of studies on central receiver solar
sorptivity of the tubes is 90%. thermal power plants. Renew Sustain Energy Rev 2013;23:12–39.
[19] Wang K, He YL. Thermodynamic analysis and optimization of a molten salt solar
(3) The optimal novel receivers with vertical finned structures increase power tower integrated with a recompression supercritical CO2 Brayton cycle based
the peak flux in the meantime increase the optical efficiency. The on integrated modeling. Energy Convers Manage 2017;135:336–50.
optimal receiver with vertical finned structures and the convex [20] Wang K, He YL, Qiu Y, Zhang Y. A novel integrated simulation approach couples
MCRT and Gebhart methods to simulate solar radiation transfer in a solar power
bottom has the maximum solar flux of 3.87 × 106 W·m−2, which is
tower system with a cavity receiver. Renewable Energy 2016;89:93–107.
almost double that of the traditional cylindrical receiver. [21] Zhang L, Fang JB, Wei JJ, Yang GD. Numerical investigation on the thermal per-
(4) The multi-point aiming strategy adopted in the present work can formance of molten salt cavity receivers with different structures. Appl Energy
2017;204:966–78.
effectively decrease the peak fluxes on the optimal novel receivers
[22] Rinaldi F, Binotti M, Giostri A, Manzolini G. Comparison of linear and point focus
by more than 40% with the optical efficiencies sacrifice less than collectors in solar power plants. Energy Procedia 2014;49:1491–500.
1.1%. [23] Astolfi M, Binotti M, Mazzola S, Zanellato L, Manzolini G. Heliostat aiming point
optimization for external tower receiver. Sol Energy 2017;157:1114–29.
[24] He YL, Cheng ZD, Cui FQ, Li ZY, Li D. Numerical investigations on a pressurized
Declaration of interest volumetric receiver: solar concentrating and collecting modelling. Renewable
Energy 2012;44:368–79.
We declare that we have no financial and personal relationships [25] Du S, Li MJ, Ren QL, Liang Q, He YL. Pore-scale numerical simulation of fully
coupled heat transfer process in porous volumetric solar receiver. Energy
with other people or organizations that can inappropriately influence 2017;140:1267–75.
our work; there is no professional or other personal interest of any [26] Ho CK, Mills B, Christian JM. Volumetric particle receivers for increased light
nature or kind in any product, service and/or company that could be trapping and heating. In: ASME international conference on energy sustainability,
Charlotte, North Carolina, USA; 2016. V001T04A016-V001T04A016.
construed as influencing the position presented in, or the review of, the [27] Gobereit B, Amsbeck L, Buck R, Pitz-Paal R, Rögerd Marc, Müller-Steinhagene H.
manuscript entitled. Assessment of a falling solid particle receiver with numerical simulation. Sol Energy
2015;115:505–17.
[28] Pacheco James E. Final test and evaluation results from the solar two project.
Acknowledgement SANDIA Report SAND2002-0120; 2002.
[29] Burgaleta JI, Arias S, Ramirez D. Gemasolar, the first tower thermosolar commercial
The work was supported by the National Natural Science plant with molten salt storage. In: Proc of SolarPACES. Granada (Spain); 2011. p.
20–2.
Foundation of China (No. 51806165) and the Science and Technology
[30] Ho CK. Advances in central receivers for concentrating solar applications. Sol
Planning Project of Xi'an (201809160CX1JC2-02). The authors would Energy 2017;152:38–56.
also like to thank the Foundation for Innovative Research Groups of the [31] Ho CK, Christian JM, Ortega JD, Yellowhair J, Mosquera MJ, Andraka CE.

233
W.-Q. Wang et al. Energy Conversion and Management 184 (2019) 219–234

Reduction of radiative heat losses for solar thermal receivers. SPIE Optics and [45] Ma Z, Yang WW, Li MJ, He YL. High efficient solar parabolic trough receiver re-
Photonics for Sustainable Energy San Diego, CA, USA, 2014; 9175: 917506. actors combined with phase change material for thermochemical reactions. Appl
[32] Daabo AM, Mahmoud S, Al-Dadah RK. The optical efficiency of three different Energy 2018;230:769–83.
geometries of a small scale cavity receiver for concentrated solar applications. Appl [46] Geuder N, Trieb F, Schillings C, Meyer R, Quaschning V. Comparison of different
Energy 2016;179:1081–96. methods for measuring solar irradiation data. In: The 3rd international conference
[33] Qiu Y, He YL, Li PW, Du BC. A comprehensive model for analysis of real-time op- on experiences with automatic weather stations; 2003. p. 19–21.
tical performance of a solar power tower with a multi-tube cavity receiver. Appl [47] He YL, Du BC, Wang K, Qiu Y, Liu ZB. Study on the coupled photon-thermal-stress
Energy 2017;185:589–603. integration method, characteristics with time and failure criterion in the solar
[34] Friefield, JM, Friedman J. Technical Report No. 1: Solar Thermal Power Systems molten salt cavity receiver (in Chinese). Chin Sci Bull 2017;62:4307–20.
Based on Optical Transmission. Rocketdyne Division, Rockwell International; 1974. [48] Flesch J, Niedermeier K, Fritsch A, Musaeva D, Marocco L, Uhlig R. Liquid metals
[35] Garbrecht O, Al-Sibai F, Kneer R, et al. CFD-simulation of a new receiver design for for solar power systems. IOP conference series: materials science and engineering.
a molten salt solar power tower. Sol Energy 2013;90:94–106. IOP Publishing; 2017. vol. 228. p. 012012.
[36] Christian JM, Ortega JD, Ho CK, Yellowhair J. Design and modelling of light- [49] Heinzel A, Hering W, Konys J, Marocco L, Karsten Litfin, Müller G, et al. Liquid
trapping tubular receiver panels. In: Proceedings of the ASME international con- metals as efficient high temperature heat transport fluids. Energy Technol
ference on energy sustainability, Charlotte, North Carolina, USA, 2016; 2017;5:1026–36.
V001T04A005-V001T04A005. [50] Wang K, He YL, Xue XD, Du BC. Multi-objective optimization of the aiming strategy
[37] Christian JM, Ortega JD, Ho CK. Novel tubular receiver panel configurations for for the solar power tower with a cavity receiver by using the non-dominated sorting
increased efficiency of high-temperature solar receivers. In: Proceedings of the genetic algorithm. Appl Energy 2017;205:399–416.
ASME international conference on energy sustainability, San Diego, California, [51] Besarati SM, Goswami DY, Stefanakos EK. Optimal heliostat aiming strategy for
USA; 2015. V001T05A014-V001T05A014. uniform distribution of heat flux on the receiver of a solar power tower plant.
[38] Sánchez-González A, Rodríguez-Sánchez MR, Santana D. Aiming strategy model Energy Convers Manage 2014;84:234–43.
based on allowable flux densities for molten salt central receivers. Sol Energy [52] Qiu Y, Li MJ, Wang K, Liu ZB, Xue XD. Aiming strategy optimization for uniform
2017;157:1130–44. flux distribution in the receiver of a linear Fresnel solar reflector using a multi-
[39] Torresol Energy. Available at http://torresolenergy.com/en/?sfid=4868&_sft_tipo- objective genetic algorithm. Appl Energy 2017;205:1394–407.
de-multimedia=gemasolar-en. [53] He YL, Wang K, Du BC, Zheng ZJ, Liang Q. Non-uniform characteristics of solar flux
[40] He YL, Xiao J, Cheng ZD, Tao YB. A MCRT and FVM coupled simulation method for distribution in the concentrating solar power systems and its corresponding solu-
energy conversion process in parabolic trough solar collector. Renewable Energy tions: a review (in Chinese). Chin Sci Bull 2016;61:3208–37.
2011;36:976–85. [54] He YL, Wang K, Qiu Y, Du BC, Liang Q, Du S. Review of the solar flux distribution in
[41] Yao Z, Wang Z, Lu Z, Wei X. Modeling and simulation of the pioneer 1MW solar concentrated solar power: non-uniform features, challenges, and solutions. Appl
thermal central receiver system in China. Renew Energy 2009;34:2437–46. Therm Eng 2019;149:448–74.
[42] Yu Q, Wang Z, Xu E, Zhang H, Lu Z, Wei X. Modeling and simulation of 1MWe solar [55] Sánchez-González A, Santana D. Solar flux distribution on central receivers: a
tower plant’s solar flux distribution on the central cavity receiver. Simul Model projection method from analytic function. Renewable Energy 2015;74:576–87.
Pract Theory 2012;29:123–36. [56] Vant-Hull LL. The role of “Allowable flux density” in the design and operation of
[43] Badescu V. Theoretical derivation of heliostat tracking errors distribution. Sol molten-salt solar central receivers. Trans-Am Soc Mech Engineers J Solar Energy
Energy 2008;82:1192–7. Eng 2002;124:165–9.
[44] Xu Y, Cui K, Liu D. The development of a software for solar radiation and its ver-
ification by the measurement results on the spot. Energy Technol 2002;23:237–9.

234