Available online at www.sciencedirect.

com

ScienceDirect
Solar Energy 105 (2014) 48–57
www.elsevier.com/locate/solener

Experimental and theoretical study of wind loads and mechanical

performance analysis of heliostats
Chuncheng Zang ⇑, Bo Gong, Zhifeng Wang
Key Laboratory of Solar Thermal Energy and Photovoltaic System, Institute of Electrical Engineering Chinese Academy of Sciences, Beijing 100190, China

Received 27 December 2013; received in revised form 19 March 2014; accepted 3 April 2014
Available online 4 May 2014

Communicated by: Associate Editor Robert Pitz-Paal

Abstract

Wind load greatly affects the mechanical performance and tracking accuracy of heliostats. Therefore, predicting wind loads accurately
is important for developing heliostats with good performance. Quantifying an accurate wind load shape factor is the key to predict wind
loads. In this paper, the wind load shape factor is obtained through measuring wind pressure distribution on the heliostat surface, and
then is used to analyze the mechanical parameters of the heliostat support structure, such as stress distribution and directional deforma-
tion. The said mechanical parameters are compared with those generated by the modeled wind load which is calculated according to the
related codes. The comparison results show that the theoretical method of modeling wind loads can be used to design and research the
heliostat.
Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Heliostat; Wind load; Mechanical performance

1. Introduction moment coefficient. Li and Wang and their coworkers (Li

and Gong, 2007; Wang et al., 2007; Gong et al., 2008,
T-shape heliostats with 100 m2 or larger reflective areas 2013; Wang et al., 2008, 2009, 2007) tested a 95 m2 helio-
are very sensitive to the wind load that greatly affects the stat developed by the Institute of Electrical Engineering
mechanical performance and tracking accuracy of helio- Chinese Academy of Sciences in the boundary layer wind
stats. Therefore, accurate prediction of the wind load is tunnel of Hunan University. They measured the wind pres-
important to the structural design of the heliostat. A mul- sure on the reflective surface and the mechanical stresses in
titude of studies on wind load analysis for heliostats has the pedestal for 130 cases of different wind directions and
been performed. Generally, wind tunnel test is an effective elevations. The test data yielded the static and dynamic
way to study wind load on the heliostat. Peterka and stresses for the heliostat, the wind load shape factor and
Derickson (1992) developed a simplified design method the wind flutter factor, which provided reference data for
for defining wind loads on flat heliostats and parabolic dish wind-resistance design of heliostats. Huss et al. (2011)
collectors through wind tunnel test. The method developed a novel methodology for evaluating the dynamic
generalized the wind load data such as the drag coefficient, effect of wind loads on Ivanpah LH-2 heliostats. They
lift coefficient, azimuthal moment coefficient and hinge employed an aero-elastic heliostat model for wind tunnel
test to measure the total wind forces and moments which
⇑ Corresponding author. Tel.: +86 10 82547050; fax: +86 10 62587946. reflect the total fluctuating forces including the background
E-mail address: zangchch@mail.iee.ac.cn (C. Zang). loading as well as the inertial loading. Therefore, the peak

http://dx.doi.org/10.1016/j.solener.2014.04.003
0038-092X/Ó 2014 Elsevier Ltd. All rights reserved.
 C. Zang et al. / Solar Energy 105 (2014) 48–57 49

Nomenclature

A reflective area of the heliostat (m2) uz wind pressure height variation factor
Cpi wind pressure coefficient v wind velocity (m/s)
F wind load acting on the heliostat structure (N) q air density
Pi wind pressure on the testing surface (N/m2) t influence coefficient of wind pressure fluctuation
i test points number uz structural vibration mode coefficient
bz wind flutter factor x0 reference wind pressure (N/m2)
f peak wind fluctuation factor xz wind pressure (N/m2)
us wind load shape factor

dynamic loading could be obtained. Wu et al. (2010) stud- tubes with different sections. The mechanical performance
ied the effects of gap sizes between the facets on the wind of the support structure stands for the wind-resisting capa-
load acting on the heliostat using the fluent software. The bilities of heliostats.
results showed that the wind load increased slightly with Fig. 2 displays 100 heliostats deployed in curve form of 15
increasing gap size and the flow pattern through the gap rows. The tested heliostat is located in the twelfth row and at
resembled a jet flow which reduced the static pressure on the west edge of the field. The heliostat is marked by a red1
the leeward surface of the facets and, consequently, circle in Fig. 2. In the flat open area that is on the west of the
increased the drag force. Zang et al. (2013) studied numer- field and 58.7 m from the investigated heliostat, the ane-
ical simulation of wind velocity fluctuation on the surface mometers are respectively mounted at 3.5 m, 6.5 m, 10 m
of heliostats and analyzed the wind-induced dynamic and 12.5 m above the ground to characterize the wind veloc-
response. Besides above methods, wind load test on an ities and directions approaching the heliostat from the
actual heliostat is of feasibility to analyze the structure. southwest to west to northwest direction. There are no
Sment et al. (2013) measured the boundary winds over objects between the tested heliostat and the anemometers
the heliostat field and analyzed wind velocities and turbu- so that the wind heading to the heliostat is not blocked.
lence between rows within the field, so as to characterize Due to the approximately westerly wind approaching, the
and understand some differences in the impacts of dynamic other heliostats would not affect the tested heliostat too
wind loads on heliostat strain and cyclic fatigue between much. In addition, 34 wind pressure sensors are symmetri-
perimeter and inner-field heliostats. cally mounted on the front and back of the facets, which
This paper introduces the 100 m2 heliostat for DAHAN are marked by the red spots2 in Fig. 1.
solar power tower demonstration system in China and pre-
sents methods of wind load prediction and analysis of the 2.2. Mechanical performance of the heliostat
mechanical performance of the heliostat support structure
based on the wind load experiments on a full-scale heliostat. 2.2.1. Theoretical method of wind load calculation
The main load acting on the heliostat is the wind loading
2. Structural design and analysis of the heliostat which greatly affects the mechanical performance, optical
performance and cost of the heliostat. The heliostat is a
DAHAN solar power tower demonstration station is kind of low-rise structure, close to the ground, so it works
located at Yanqing County of Beijing, China. The heliostat in a complex wind field. Wind load on heliostats is able to
field for DAHAN was built in 2009 and covers 10,000 m2 be predicted using the below equations according to the
reflective areas. The heliostat is described as below. standards (Ministry of Construction of People’s Republic
of China, 2002) for building structures:

2.1. Introduction of the measured heliostat F ¼ Axz ð1Þ

xz ¼ bz us uz x0 ð2Þ
Fig. 1 shows the front view and the back view of the 2
heliostat with 100 m2 reflective areas. The heliostat mainly x0 ¼ qv =2 ð3Þ
consists of four parts: reflector, support structure, drive The design values of wind speed v are determined accord-
mechanism and control devices. The reflector is composed ing to the environmental conditions under which heliostat
of 64 facets with the size of 1.25 m  1.25 m. Each facet is works. For DAHAN heliostat, the operational wind speed
made up of three layers: silver glass, glue and tempered
glass. The facet surface is formed to curve by adjusting 1
For interpretation of color in Fig. 2, the reader is referred to the web
the bolts mounted on the back of the facet. The support version of this article.
structure is composed of torque tube, pedestal and the truss 2
For interpretation of color in Fig. 1, the reader is referred to the web
supporting the reflector. The truss is made of rectangular version of this article.
50 C. Zang et al. / Solar Energy 105 (2014) 48–57

F15 F16 F17 B17 B16

F14 B15 B14
B13 B12
F10 F11 F12 F13 B11 B10
F9 B9
B8 B7 B6 B5
F5 F6 F7 F8
B4 B3 B2 B1

F1 F2 F3 F4

(a) Wind pressure sensors on the (b) Wind pressure sensors on the
front surface of the facets back surface of the facets
Fig. 1. DAHAN heliostat and wind pressure sensors layout.

high rise buildings. DAHAN solar tower demonstration

station is located in the rural areas. So the category B is
applicable to DAHAN heliostats in this paper.The
influence of wind fluctuation on the structure has been
considered when bz is calculated through Eq. (4).
ftuz
bz ¼ 1 þ ð4Þ
lz
The coefficient f is related to x0 T 21 and its value is deter-
mined according to Table 7.4.3 in Ministry of
Construction of People’s Republic of China (2002), section
of which is shown in Table 2. The parameter T1 is the basic
Fig. 2. Heliostat deployment of DAHAN solar tower demonstration natural vibration period and is calculated through finite
system. element analysis on dynamic performance of heliostat
model. The dynamic performance analysis is presented in
Table 1 the following Section 2.2.4. The coefficient t is related to
Wind pressure height variation factor uz. the geometrical characteristics of structures and terrain
Altitudes (m) Terrain roughness categories roughness categories. Based on the ratio of height to width
of the heliostat and category B of the terrain roughness, the
A B C D
value of coefficient t is determined by Table 7.4.4-3 in
5 1.17 1.0 0.74 0.62
Ministry of Construction of People’s Republic of China
10 1.38 1.0 0.74 0.62
15 1.52 1.14 0.74 0.62 (2002), section of which is shown in Table 3. The coefficient
20 1.63 1.25 0.84 0.62 uz is determined by dynamic calculation method and its
.. .. .. .. ..
. . . . . value is less than 1.0.The factor us is mainly related to
the shape and dimensions of structures. It is usually
obtained through wind tunnel tests or on-site experiments.
is up to 13 m/s. The heliostat can resist the wind load under Besides, the factor us might be calculated approximately
any orientations when wind speed is not larger than 20 m/s, according to Ministry of Construction of People’s
and survive 32 m/s wind gust in stow position. Republic of China (2002). In this paper, us is able to be
Factors bz, us and uz are calculated according to obtained by measuring the wind pressure on the heliostat
Ministry of Construction of People’s Republic of China surface and by calculating as well. Based on the methods
(2002). The factor uz is chosen from Table 1 according to of wind load prediction, the mechanical performance of
Ministry of Construction of People’s Republic of China the support structure was analyzed in this paper.
(2002). It is related to the terrain roughness categories.
There are four categories involving A, B, C and D for 2.2.2. Stress analysis of the support structure
the determination of uz. The category A stands for the The mechanical performance of the support structure
areas close to beaches, islands or deserts. B refers to the varies with the change in heliostat orientations during
rural areas, hills, small towns or suburbs where there are sun tracking process. Therefore, the support structure
sparse buildings. C is for the city areas with dense building was analyzed under different orientations. Fig. 3 shows
groups. D is for the city areas where there are dense and the coordinate system for the heliostat and defines the
 C. Zang et al. / Solar Energy 105 (2014) 48–57 51

Table 2
Peak wind fluctuation factor f.
x0 T 21 (kNs2/m2) 0.01 0.02 0.04 0.06 0.08 0.10 0.20 0.40 0.60
Steel structures 1.47 1.57 1.69 1.77 1.83 1.88 2.04 2.24 2.36
x0 T 21 (kNs2/m2) 0.80 1.00 2.00 4.00 6.00 8.00 10.00 20.00 30.00
Steel structures 2.46 2.53 2.80 3.09 3.28 3.42 3.54 3.91 4.14

Table 3 using the finite element software VSAP (Visual Structural

Influence coefficient of wind pressure fluctuation t. Analysis Program). The calculated wind loads were applied
H/B Terrain roughness categories Total altitude (m) on the center of each facets of the heliostat. The heliostat
630 50 100 ... base was fixed as the boundary constraint conditions.
60.5 B 0.42 0.41 0.33 ... Fig. 4 shows the stress distribution on the support structure
1.0 B 0.46 0.46 0.32 ... at the said orientations. The wind speed v is 32 m/s for the
2.0 B 0.48 0.50 0.47 ... stress analysis in stow position and 20 m/s in other posi-
tions. It can be seen that the stress on the components clos-
ing to the ends of the torque tube reaches the maximum and
the maximum values are respectively around 62.95 MPa at
0° elevation, 97.047 MPa at 30°, 131.542 MPa at 60° and
131.473 MPa at 85° elevation. Basically, the maximum
stress increases with the elevation when the azimuth does
not change too much. If the calculated stress is larger than
the allowable value of yield strength of the material, the
related components should be optimized.

2.2.3. Deformation analysis of the support structure

Fig. 5 shows the deformation of the structure at nearly
vertical orientation when the wind is heading to the back
of the reflector at the speed of 13 m/s. The peak displace-
ment of the fringe node A is obtained in Y direction.
Table 4 lists the displacement values of nodes A, B, C
and D under the action of wind loads and gravity. It can
Fig. 3. Definition for azimuth and elevation angle. be seen from the values that wind load has great influence
on the structure deformation. The maximum error of the
focal image caused by wind load is estimated based on
elevation and azimuth angle. The symbol a stands for the
the displacement of node A in Y direction. Displacements
elevation angle, which is the angle between the horizontal
in X and Z are relatively small and can be neglected at this
plane and the reflective surface of the heliostat. The azi-
position.
muth angle is denoted by the symbol b, which stands for
Fig. 6 and Eq. (5) present the approximate calculation of
the rotating angle of the normal of the heliostat surface
the focal image error. The point H is the ideal location of
to the wind direction. Positive values mean that the rota-
focal image when the heliostat has no deformation. HH0
tion is on clockwise.This paper presents the analysis of
is the deflection of the focal image when the deformation
the stress distribution at four heliostat orientations. One
happens. PP0 is the displacement of point P. M1M2 is the
is nearly vertical position which is at 85° elevation angle
focal length of heliostat. PQ is the size of the reflector.
and 175° azimuth angle. The second is at stow position
which is 0° elevation angle and 165° azimuth angle. The qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 4PP0
other two orientations are respectively at the azimuth of HH0  M1M22 þ ðM2H  M1OÞ  ð5Þ
PQ
166° and 149° and corresponding elevation of 30° and
60°. The loads acting on the support structure mainly con- In the position as Fig. 5 shows, the reflector’s normal
tain the wind force and the gravity of facets. In this section, error caused by the gravity of the structure is nearly
the wind force is calculated according to the Eqs. (1)–(4), 1.95 Mrad while the error from wind and gravity is
which involves the dynamic wind load factor. That is, the 5.76 Mrad, which means the error caused by wind load is
wind force contains not only the mean wind but also the around 5.42 Mrad.
wind fluctuation. The wind speed v is at 10 m above the
ground. The influence of the change in heights on wind 2.2.4. Dynamic analysis of the heliostat
speed is determined by the factor uz. The wind heads to The objective of dynamic analysis of the heliostat is to
the back of the heliostat. The stress distribution resulting research its natural frequency and obtain its basic vibration
from gravity forces and calculated wind loads was analyzed period and vibration modes. The basic vibration period is
52 C. Zang et al. / Solar Energy 105 (2014) 48–57

(a) Elevation of 0° and (b) Elevation of 30° and

azimuth of -165° azimuth of -166°

(c) Elevation of 60° (d) Elevation of 85° and

and azimuth of -149° azimuth of -175°
Fig. 4. Theoretical stress distribution of the heliostat structure at different orientations.

are respectively 0°, 15°, 30°, 45°, 60°, 75° and 90° in eleva-
tion angles. Fig. 7 shows the change in the natural period
of the structure with orientations at the first ten orders.
The change in elevation hardly affects the frequency under
all orders except for the third and the eighth order. The
natural period is longer at lower mode than that at higher
mode. The basic natural period (0.23 s) is relatively short,
which implies that the structure is stiff and not much sensi-
tive to the pulse excitation of certain wind forces.

3. Wind load measurement and determination of wind load

shape factor

The method of wind load calculation has been intro-

duced above, which is based on the codes for building
structures. Actually, the heliostat is a special structure that
is complex and a bit different with buildings. Therefore, to
use the generalized load coefficient for calculating maximal
Fig. 5. Deformation of the support structure at elevation of 85° and stresses in such a complex structure is a very rough
azimuth of 175°. approach. The calculation of wind load shape factor us in
the codes might be not completely appropriate to helio-
used to determine the peak wind fluctuation factor f. The stats, but could be a reference to estimate the wind load
relation of the two parameters has been shown in Table 1. on heliostats. The factor us is generally determined through
The mechanical performance of the support structure wind tunnel test or wind loading measurement on full scale
varies with the change in heliostat orientation during the heliostats. In this paper, the wind pressure on the reflector
sun tracking process. This paper presents the vibration was measured under different orientations of the heliostat.
characteristics of the structure at seven orientations which The wind pressure coefficient was calculated according to
 C. Zang et al. / Solar Energy 105 (2014) 48–57 53

Table 4
Displacement of nodes at elevation of 85° and azimuth of 175°.
Loads Direction Displacement of nodes (mm)
A B C D
Gravity and wind X 1.65 1.58 0.55 0.39
Y 22.91 22.31 1.34 1.96
Z 2.29 3.56 4.17 5.31
Gravity X 0.03 0.03 0.03 0.03
Y 6.21 6.20 3.84 3.84
Z 1.77 1.77 2.68 2.68

analyzed in this paper. The wind pressure on the reflector

was measured using CY2000FAIP wind pressure sensors.
This kind of sensor is able to work under the temperatures
ranged from 20 °C to 70 °C, and its temperature stability
is ±0.3% FS/10 °C. It can measure the pressure ranged
from 1000 Pa to 1000 Pa. Being supplied 24 V power,
the sensors output current signal of 4–20 mA. Fig. 1 shows
the test points position. There are 34 wind pressure sensors
mounted symmetrically on the front and back of the mir-
rors. Points F1, F2, . . ., F17 are on the front, and other
points B1, B2, . . ., B17 are on the back. The sensors at F1
and B1 compose a pair to measure the pressure on the same
point. It is the same for F2 and B2, F3 and B3, and so on.
The positive measured values stand for pressure and nega-
M1 M2
tive values are for suction. Considering that the wind pres-
Heliostat Beam target sure changes with the heliostat orientations, the wind
Fig. 6. Effect of deformation on beam quality. pressure was measured at different orientations. During
the test process, the heliostat faced to the east at the begin-
ning and rotated in azimuth direction to the west on clock-
0.24
0° wise. At every rotation angle of 30° in azimuth, the
0.22 15° heliostat stopped rotating in azimuth but changed the ele-
0.20 30° vation angle from 0° to 85° to be tested. The test time
45° length was 10 min for each orientation. All wind pressure
0.18
Natural periods (s)

60° sensors and anemometers were connected with the same

0.16 data acquisition equipment. The acquisition frequency rate
75°
0.14 90° was set at 20 Hz.
As stated in Section 2.1, the anemometer is able to
0.12
define the wind velocity and direction. The origin of the
0.10 wind direction (0°) was set at the north and 90° was at
0.08 the east. So the wind direction value increases on clockwise.
In this paper, the incoming wind was from southwest to
0.06
northwest. The wind velocity varies with time. The time
0.04 history of wind velocity can be obtained and used to ana-
0 2 4 6 8 10
lyze the mean wind velocity and the peak wind velocity
Order of modes
during the period of the measurement. Different heliostat
Fig. 7. Natural period of the support structure. positions were tested at different times. Therefore, accord-
ing to the test time serials, the relation between heliostat
these measured data, which were used to calculate the fac- positions and wind velocity could be found. Fig. 8 shows
tor us. the wind velocity time history and wind direction time his-
tory during the measurement period from 11 o’clock to
3.1. Wind loads test methods 14 o’clock at 10 m high. Wind velocity and direction here
are both in the horizontal direction. It can be seen from
To obtain the factor us for the heliostat, the wind veloc- Fig. 8 that the range of velocities is around from 2.5 m/s
ity and direction and the wind pressure were measured and to 15 m/s and directions approximately fluctuate from
54 C. Zang et al. / Solar Energy 105 (2014) 48–57

350 350

Wind direction (m/s)

Wind velocity (m/s)
300 300

250 250

200 200
11 11.5 12 12.5 13 13.5 14 11 11.5 12 12.5 13 13.5 14
Time (h) Time (h)
(a) Wind velocity time history (b) Wind direction time history
Fig. 8. Wind conditions measured at 10 m above the ground.

Table 5
position. So the mean wind velocity and direction refer to
Mean wind velocities and directions at 10 m high for different heliostat the average values in ten minutes.
positions.
Elevation angle (°) 0 30 60 85 3.2. Wind pressure coefficient analysis
Azimuth angle (°) 165 166 149 175
Mean wind velocities (m/s) 8.77 9.67 8.86 8.30 Wind load analysis is necessary for wind-resisting design
Mean wind directions (°) 254.82 256.09 238.55 265
of the heliostat structure. Generally, to obtain the wind
load shape factor us or the wind pressure coefficient is an
220° to 325°. The mean wind velocities and directions at effective method to predict the wind loads. The wind pres-
10 m high are listed in Table 5 for the said four heliostat sure coefficient is described in Eq. (6), and it is the ratio of
positions. The test time length is 10 min for each heliostat the real wind pressure to the air flow onto the testing sur-

9000 3
9000
2
Height of Reflector (mm)

Height of Reflector (mm)

8000 8000
2.5
7000 7000 1.5
6000 2 6000
5000 5000 1
4000 1.5 4000

3000 3000 0.5

1
2000 2000

1000 1000 0
0.5
2000 4000 6000 8000 2000 4000 6000 8000
Width of Reflector (mm) Width of Reflector (mm)
(a) Elevation angle of 0°and (b) Elevation angle of 30°and
azimuth angle of -165° azimuth angle of -166°

9000 -1
Height of Reflector (mm)

0
8000
Height of Reflector (mm)

8000
7000 -1.5

6000 -0.5 6000

-2
5000

4000 -1 4000
-2.5
3000

2000 -1.5 2000 -3

1000
2000 4000 6000 8000 2000 4000 6000 8000
Width of Reflector (mm) Width of Reflector (mm)
(c) Elevation angle of 60° and (d) Elevation angle of 85°and
azimuth angle of -149° azimuth angle of -175°
Fig. 9. Wind pressure coefficient distribution at different orientations.
 C. Zang et al. / Solar Energy 105 (2014) 48–57 55

face. Based on the wind pressure coefficient, the wind load using Eq. (7), the mechanical performances of the heliostat
shape factor is obtained using Eq. (7). support structure are analyzed. Fig. 10 shows the stress dis-
Pi tribution of the support structure at different orientations
C pi ¼ ð6Þ when the heliostat faces to the east and the wind heads
0:5qv2
to the back of the reflector. It can be compared with
Pn
C pi Ai Fig. 4 to see the differences clearly. The maximum stresses
us ¼ i¼1 ð7Þ have got the change because the empirically determined
A
shape coefficients are substituted into the former numerical
The test data Pi were analyzed and the average wind pres- model. For better understand, theoretical us and experi-
sure coefficients Cpi were obtained according to Eq. (6). mental us are defined here. Theoretical us means the value
Mean wind speed v was tested at different heights. Ai is of us is calculated according to the former numerical
the corresponding area of test point i. A is the total area model. Experimental us refers to the value of us that is
of the test zone. Fig. 9 shows the coefficient Cpi distribution obtained through wind pressure measurement. The experi-
on the reflector surface when the orientations are 165°, mental us distribute on the surface of the heliostat as differ-
166°, 149° and 175° for azimuth and 0°, 30°, 60° and ent values at different points. But the theoretical us has the
85° for elevation. It can be seen from Fig. 9 that the wind same value or two different values on the surface. That is
force acting on the reflector is obviously asymmetrical when the main factor causing the differences between Figs. 4
the wind is heading to the back of the reflector in some and 10. In addition, the wind direction is random and
directions. As the consequence, the wind pressure increases changes all the time. The average wind direction is chosen
gradually from the left to the right of the reflector. as the reference to calculate azimuth, which is another fac-
tor that affects the accuracy of the mechanical performance
3.3. Mechanical performance of the support structure analysis.Besides the orientations shown in Figs. 4 and 10 in
which the wind heads to the back of the heliostat, the stress
Based on the wind pressure coefficient distribution analysis of the structure were also done under the orienta-
shown in Fig. 9 and the wind load shape factors calculated tions when the wind headed to the front of the heliostat.

(a) Elevation of 0° and (b) Elevation of 30° and

azimuth of -165° azimuth of -166°

(c) Elevation of 60° and (d) Elevation of 85° and

azimuth of -149° azimuth of -175°
Fig. 10. Experimental stress distribution of the heliostat structure at different orientations.
56 C. Zang et al. / Solar Energy 105 (2014) 48–57

Table 6
Maximum stress comparison.
Elevation (°) 0 30 60 85 0 30 60 85
Azimuth (°) 165 166 149 175 22 20 24 23
Maximum stress with theoretical us (MPa) 62.95 97.05 131.54 131.47 60.92 78.54 129.62 147.51
Maximum stress with experimental us (MPa) 67.62 82.91 88.22 115.73 67.91 67.20 92.25 145.65
Relative error (%) 7.4 14.6 32.9 12.0 11.5 14.4 28.8 1.3

Table 7
Displacement values comparison at elevation of 85° and azimuth of 175°.
Displacement of nodes (mm)
A B C D E F G
Theoretical us
X 1.65 1.58 0.55 0.39 0.58 0.60 0.59
Y 22.91 22.31 1.34 1.96 9.12 8.80 8.03
Z 2.29 3.56 4.17 5.31 0.87 1.62 0.07
Experimental us
X 1.28 1.30 0.75 0.65 0.48 0.50 0.49
Y 21.45 16.36 0.28 2.23 8.77 6.80 6.96
Z 1.56 2.53 3.29 3.98 0.93 1.53 0.07

The azimuth angles are respectively 22°, 24°, 20°and mance analysis of the 100 m2 DAHAN heliostat at different
23°. Table 6 presents the comparison results of the max- orientations under wind loads. To verify the method of
imum stress between the cases of theoretical us and exper- wind load calculation according to the codes, it is the key
imental us. Under the similar azimuth angles, the to look into the wind load shape factor which was obtained
maximum stress grows with the increase of the elevation. through measuring the wind pressures on the full scale
The theoretical maximum stresses agree with the experi- heliostat. The results show that the stress and displacement
mental stresses except of the case of 60° in elevation. The of heliostat under the case of theoretical us agree with those
relative error ranges from 1.3% to 32.9%. In most cases, under the case of experimental us. In most cases, the theo-
the theoretical maximum stresses are a little larger than retical maximum stresses are a little larger than the exper-
the experimental ones, but it is to the contrary in the case imental ones, but it is smaller in the case of 0° elevation
of 0° elevation angle. In general, the conclusion could be angle. Generally speaking, the method of wind load calcu-
drawn that the method of wind load calculation according lation according to the codes (Ministry of Construction of
to the codes (Ministry of Construction of People’s People’s Republic of China, 2002) is feasible and can be a
Republic of China, 2002) is feasible and can be a reference reference to predict wind loads on heliostats. To guarantee
to estimate wind loads for heliostats design. To guarantee the reliability of the heliostat, the wind load values calcu-
the reliability of the heliostat, the wind load values calcu- lated using the codes might be increased by 8–12% or more
lated using the codes might be increased by 8–12% or more in the case of the stow position.
in the case of 0° elevation angle.
In addition, the deformation of the support structure is Acknowledgments
presented under the orientation of 175° for azimuth and
85° for elevation when the heliostat faced to the east and This work is supported by the National Natural Science
the wind headed to the back of the reflector at the speed Foundation of China (No. 51308524) and the National
of 13 m/s. Owing to the non-symmetry of the real wind Basic Research Program of China (No. 2010CB227106).
loads acting on the reflector, the deformation values appear We would like to show our thanks to Prof. Zhengnong
left-right asymmetry. Table 7 shows the comparison of Li and his team for their great contributions to the wind
deformation values between the cases of theoretical us load measurement. Also we would like to express our sin-
and experimental us. The nodes A  G has been marked cere appreciation to the reviewers for their insightful com-
in Fig. 5. It can be seen obviously that the displacement ments, which have greatly aided us in improving the quality
values under the case of theoretical us are a little larger than of the paper.
those under experimental us. The accuracy of wind load
prediction mainly depends on the value of us. References

Peterka, J.A., Derickson, R.G., 1992. Wind Load Design Methods for
4. Conclusions Ground-Based Heliostats and Parabolic Dish Collectors, Technical
Report, SAND92-7009.
Wind load greatly affects the mechanical performance of Li, Z.N., Gong, B., 2007. Wind tunnel test of heliostat, Technical Report,
the heliostat. This paper presents the mechanical perfor- IEE-CAS.
 C. Zang et al. / Solar Energy 105 (2014) 48–57 57

Wang, Z.F., Wu, Z.Y., Liu, X.B., Li, Z.N., 2007. Wind dynamics testing Huss, S., Traeger, Y.D., Shvets, Z., Rojansky, M., Stoyanoff, S., Garber,
on Dahan heliostat. In: Process of ISES Solar World Congress, vol. 3, J., 2011. Evaluating effects of wind loads in heliostat design. In: 17th
pp. 1934–1938. International Conference of Concentrating Solar Power and Chemical
Gong, B., Li, Z.N., Wang, Y.G., Li, Q.S., 2008. Wind tunnel test study on Energy Systems (SolarPACES), Granada, Spain, September 20–23.
the wind load shape coefficient of heliostat. J. Hunan Univ. (Nat. Sci.) Wu, Z.Y., Gong, B., Wang, Z.F., Li, Z.N., Zang, C.C., 2010. An
35 (9), 6–9. experimental and numerical study of the gap effect on wind load on
Gong, B., Wang, Z.F., Li, Z.N., Zang, C.C., Wu, Z.Y., 2013. Fluctuating heliostat. Renew. Energy 35 (4), 797–806.
wind pressure characteristics of heliostats. Renew. Energy 50, 307–316. Zang, C.C., Christian, J.M., Yuan, J.K., Sment, J., Moya, A.C., Ho, C.K.,
Wang, Y.G., Li, Z.N., Gong, B., Li, Q.S., 2008. Time-domain analysis on Wang, Z.F., 2013. Numerical simulation of wind loads and wind
wind-induced dynamic response of heliostat. J. Vib. Eng. 21 (5), 458– induced dynamic response of heliostats. In: 19th International Con-
464. ference of Concentrating Solar Power and Chemical Energy Systems
Wang, Y.G., Li, Z.N., Gong, B., Li, Q.S., 2009. Reconstruction and (SolarPACES), Las Vegas, USA, September 17–20.
prediction of wind pressure on heliostat. Acta Aerodynamica Sin. 27 Sment, J., Ho, C.K., 2013. Wind patterns over a heliostat field. In: 19th
(5), 586–591. International Conference of Concentrating Solar Power and Chemical
Wang, Y.G., Li, Z.N., Gong, B., 2007. Distribution and fluctuation Energy Systems (SolarPACES), Las Vegas, USA, September 17–20.
characteristics of wind pressure on heliostat. J. Nat. Disasters 16 (6), Ministry of Construction of People’s Republic of China, 2002. Load Code
187–194. for the Design of Building Structures (GB 50009-2001), Beijing, pp. 28–
48.