Solar Energy 166 (2018) 50–58

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Reducing convective heat losses in solar dish cavity receivers through a T

modified air-curtain system
⁎
Song Yanga, , Jun Wanga, Peter D. Lunda,b, Siyu Wanga, Chuan Jianga
a
Key Laboratory of Solar Energy Science and Technology in Jiangsu Province, Southeast University, School of Energy and Environment, No. 2 Si Pai Lou, Nanjing 210096,
China
b
Aalto University School of Science, P.O. Box 15100, FI-00076 Aalto (Espoo), Finland

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

Keywords: Here we propose a forced air circulation system to reduce the convective heat loss across the aperture of a dish
Air circulation system concentrator. The function of the proposed system was validated with computational fluid dynamics (CFD)
Convective losses simulations. Compared to a concentrator without such a system, the modified solar dish cavity receiver could
Concentrator clearly reduce the convective heat loss and its fluctuation in the receiver. In the best case studied, the convective
Parabolic dish
heat losses could be suppressed by up to 58%. In addition, two types of air circulation modes (clockwise and
anticlockwise) were compared, showing that the anticlockwise mode yields better performance.

1. Introduction
The receiver of a concentrating solar thermal (CST) system trans-
forms intensive irradiation into high-temperature heat for generating
electricity or driving thermochemical reactions (Blanco and Santigosa,
2016). The higher the temperature of the heat, the higher is its exergy
and usefulness for the thermodynamic cycles. But at the same time, the
total heat losses of the CST will increase, which will decrease the cavity
efficiency. The heat losses are dominated by radiation and convection.
The structure of the receiver will not only affect these losses, but also
the reflection losses of the incoming solar radiation. Often a cavity type
of receiver is used in solar tower or dish systems, which significantly
reduces both reflected sunlight and emission losses from the absorber
surface. As the aperture needs to let concentrated sunlight to pass
across, its size cannot be indefinitely reduced to minimize the heat and
reflection losses. Some trade-off rules are therefore usually applied, e.g.
the size of the aperture is set to the diameter of the maximal circle
within which at least 90% of the incident rays can enter the cavity
(Yang et al., 2018).
Convective losses form a major loss component, but they are very
difficult to measure or simulate. Past research in this field include as-
sessing the effects of geometry, scale, temperature, inclination, buoyant
or wind forces, etc., on convective heat exchange in the cavity receiver
using theoretical, numerical (CFD), experimental and combined ap-
proaches (Baïri et al., 2014; Wu et al., 2010). One of the first studies on
forced convection research in a cavity structure already originates from
the 1960s (Fox, 1965) Clausing (1981) has illustrated the physics of
convective heat losses from a cubical cavity receiver and established a
correlation between the parameters of the geometrical structures,
gravity, temperature, inclination angle, and the Nusselt number. Hogan
(1994) developed a numerical model for calculating the heat losses
from solar reflux receivers, including estimations on the shares of each
loss mechanism. Leibfried and Ortjohann (1995) studied convective
losses of a spherical and a hemispherical cavity over a range of title
angles, and presented to well-validated algorithms the convective heat
losses. McDonald (19950 executed a series of trials on exploring heat
loss mechanisms from an open cavity with and without the effect of
wind. Reddy and Kumar (2009); Reddy et al. (2016) modified the cavity
receiver of a solar dish concentrator by coupling it with a CPC to reduce
heat losses, and also studied the impacts of wind speed and direction on
convective heat losses. Paitoonsurikarn and Lovegrove (2002);
Paitoonsurikarn and Lovegrove (2006) and Taumoefolau et al. (2004)
reported on simulations and experiments on convective losses from
cavity receivers with small sizes relevant to solar dish systems. Two
important conclusions can be drawn from the studies above: Though
the infrared radiation is the dominant mechanism of the total heat loss,
convective losses can be of the same magnitude in some specific con-
ditions depending on the receiver orientation and temperature; Con-
vective heat losses are the only loss mechanism that significantly de-
pends on the receiver orientation and inclination of the cavity. For solar
dish configurations with tracking systems, the influence of the con-
vective losses could therefore be an uncertainty factor for the design
process. Therefore, better understanding of the mechanisms of con-
vective loss in enclosures would be important, but also to find effective

⁎
Corresponding author.
E-mail address: 230169637@seu.edu.cn (S. Yang).

https://doi.org/10.1016/j.solener.2018.03.027
Received 3 January 2018; Received in revised form 8 March 2018; Accepted 11 March 2018
Available online 21 March 2018
0038-092X/ © 2018 Elsevier Ltd. All rights reserved.
S. Yang et al.
Nomenclature
CFD computational fluid dynamics
CST concentrating solar thermal
FVM finite volume method
k turbulent kinetic energy
PJ pressure jump
PV photovoltaic cell
q radial flux/heat flux
q average heat loss
T temperature
T∗ normalized temperature ( T − Tmin
Tmax − Tmin ) fan
v velocity of air flow
solutions to avoid its fluctuation as this would impact the overall
thermal performance of the CST system.
Previous studies include suggestions on modified solar dish cavity
receivers to reduce the convection heat losses, e.g. combining cone,
CPC, and trumpet reflectors (Reddy and Kumar, 2009) with plate fins
attached to the inner aperture surface (Ngo et al., 2015). However, the
outcomes have been modest and the suppression of convective losses
was limited. Another commonly proposed option with quartz glass
cover (Cui et al., 2013) could considerably reduce natural convection
and IR losses from the inner part, but would also partially intercept the
incident sunlight as well.
In this paper a modified design of the cavity receiver is proposed to
reduce the convective heat losses, based on a special forced-air circulation
system. The idea was inspired by air-curtain technology used in cooling
and refrigeration industries (Foster et al., 2007). Actually, a kind of air-
curtain concept was used by Zhang et al. (2015) in a solar dish system for
reducing heat losses. Hughes et al. (2016) reported that optimal designs of
air-curtain systems depend on the velocity and direction of the air jet as
well as on the inclination of the cavity. Flesch et al. (2016) took into ac-
count the wind effects in evaluating the merits of air-curtain for a solar
tower receiver. In this paper we propose a modified version of the air-
curtain configuration for a dish cavity receiver. It contains a special air
channel with an inserted fan instead of a conventional air nozzle. This
helps to form a closed forced-air circulation system and reduce the con-
vective losses, which increased the receiver performance considerably. We
introduced improvements to the dish receiver using the modified air-
curtain, also making use of previous methodologies for analyzing forced-
air circulation processes (Kolb, 2012; Tan et al., 2009). We presented here
the new concept and a thermal analysis based on CFD simulations to verify
the improved performance in a range of different inclination conditions.
Improvements were obtained both in absolute scale and reducing the
Fig. 1. The sketch of modified model receiver with a special air channel (left: 3D model, middle: cross-section, right: plane). All units are in mm.
51
Solar Energy 166 (2018) 50–58
V∗ normalized velocity ( v − v min
v max − v min )
σ standard deviation
ε dissipation rate
η efficiency
θ inclination
Subscripts
cav cavity
conv convective
ds differential element of section area in air channel
fan configuration
input input
fluctuations of convective heat losses through the opening of the receiver.
2. Methodology
The basic analysis tool used here is a CFD model employing the
Fluent 17.0® and ICEM® software. We used as basis the cavity receiver/
reactor model by the Australian National University (ANU), which was
modified and validated for our specific case (Paitoonsurikarn and
Lovegrove, 2006; Taumoefolau et al., 2004). ICEM® was employed in
the modeling and meshing work. Grid dependency was investigated and
the final computational grid used consisted of approximately 3 × 106
unstructured cells. The 3D CFD calculation of the convection through
the aperture of the cavity receiver was simulated with Fluent 17.0®. The
SIMPLE scheme was applied to couple the pressure-velocity fields based
on the finite volume method (FVM).
The ANU’s original receiver model was simplified and geometrically
considered as an ideal cylinder with a hollowed cylindrical cavity zoom
(Paitoonsurikarn and Lovegrove, 2002; Taumoefolau et al., 2004). The
modified receiver incorporated a U-shaped channel for air circulation
(Fig. 1). The air channel in Fig. 1 consists of four flat rectangle en-
closures connected with four quarter circular passages creating a loop.
To allow comparison to reference cases, we used the original ANU
parameters as basis for the new design (Table 1). For simulating the free
atmosphere around the receiver, a large enclosure of at least 10-times
the receiver size was used around the receiver for guaranteeing that the
influence from the finite boundaries is minimized.
The physical model is based on the Navier-Stokes equations, which
are suitable for describing turbulent flow of an incompressible viscous
liquid with constant properties. The governing equations used follow
the Reynolds averaged Navier-Stokes (RANS) approach (Mavriplis,
1995). Enhanced Wall Treatment (EWT) is applied here to obtain the
S. Yang et al. Solar Energy 166 (2018) 50–58

Table 1 be < 10−4 and < 10−7, respectively.

Main parameters of the modified cavity receiver. An isothermal boundary condition is applied to the inner walls of
the receiver and the outer walls are assumed to be adiabatic. The cavity
Names Values (mm)
wall temperature is set to 723 K and the surroundings (reference tem-
Receiver height (H1) 320 perature) to 300 K, which correspond to the experimental average
Receiver diameter (D) 290 temperatures. A pressure inlet with 0 Pa gauge pressure is applied to the
Cavity height (H2) 150
faces of the enclosure. For simulating the effects of the forced-air cir-
Cavity radius (r1) 35
Channel height (H3) 148 culation, a ‘fan body’ is inserted in the middle of the air channel and
Channel length (L) 210 modeled as a constant pressure-jump (PJ). Its magnitude varies from
Channel width (W) 68 −30 Pa (anticlockwise1 flow) to 40 Pa (clockwise flow). The channel
Channel thickness (T1) 4 walls are also assumed to be adiabatic. The inclination (θ) of the cavity
Channel radius (r2) 35
is defined as the actual angle between the horizontal direction and the
Top edge thickness (T2) 2
rotation axis of the cylindrical receiver, ranging from −80° (cavity

Fig. 2. (a) Convective heat losses from the open cavity with different inclinations [−80°, 80°] at Tcav = 723 K based on Taumoefolau’s experimental data and our simulated results. (b)
Comparison of numerical results to previous studies with inclination from 0° to 80° at Tcav = 723 K.

near-wall turbulent quantities, mean velocities, and temperatures. This
method combines a two-layer model with so-called enhanced wall
functions. The relevant equations are described in more detail in
Appendix A. The steady solutions are calculated. As the governing
equations are solved by iteration, we have set as a limit for the number
of iteration the following conditions: the scaled residuals for the con-
tinuity, momentum equations, and the energy, k-ε equations should
aperture facing upwards) to 80° (cavity aperture facing downwards).
This is simulated by redirecting the gravity vector to the desired di-
rection.
1
In this paper, the anticlockwise (or clockwise) is based on the right-side view.

52
S. Yang et al.
3. Results and discussion
3.1. Validation of the CFD model
As the first step, we validated our model against the experimental
data and numerical results given by Paitoonsurikarn and Lovegrove
(2006); Taumoefolau et al. (2004), which corresponds to a natural
convection case (PJ = 0) with Tcav = 723 K, and varying inclination
[−80°, 80°]. Fig. 2 shows the results of the comparison indicating a
good agreement. From Fig. 2a, the error between the experiments and
our simulations is less than 12% in the range which is relevant to our
case (−80° to 0°). The maximum error found was ∼20% at θ = 60°, but
as the absolute heat loss value at that point is low and the uncertainty of
the experimental data is larger, the inaccuracy is still acceptable for our
study. Fig. 2b presents an excellent consistency of the simulated results
with different inclination angles to previous studies.
3.2. Improved receiver with forced air circulation system
Two main factors impact the total amount of the convective losses
from the receiver. One is the ability to transfer mass, and hence energy,
across the aperture through the combined effect of wind and buoyancy
due to cold external air. The second comes from heating the air inside
the cavity (Clausing, 1981). Previous studies have shown that, for up-
ward facing cavities, the internal resistance to heat transfer is often
dominant and the inner temperature will be close to ambient tem-
perature. Therefore, the convective loss is dominated by energy (mass)
transfer across the aperture for typical cavity receiver designs.
The forced air circulation system (Fig. 1) was introduced to reduce
the convection loss through the aperture outwards. Using a circle
channel installed with a fan, the air-curtain created could cut off the
energy transfer thus storing heat inside the cavity, and also decreasing
the losses with variable inclinations.
3.2.1. Effect of forced air flow on heat losses
Next we adjusted the speed of the forced air flow to analyze its effect
on the convective losses. Fig. 3 summarizes the results for a constant
inclination (θ = 0°). Except for the special case PJ = 5 Pa, both in the
clockwise and anticlockwise cases the losses are initially reduced when
increasing the forced flow. The effect is at highest at PJ = −20 Pa and
30 Pa. At PJ = −20 Pa the convective loss is almost 55% less than at
PJ = 0. But beyond the peaks, the benefits start to deteriorate when
Fig. 3. Savings in convective heat losses from forced air-flow as a function of the magnitude of the air-curtain (described here with PJ, zero value corresponds to no air-curtain) at
Tcav = 723 K and θ = 0°. Positive and negative values represent clockwise and anticlockwise flow, respectively.
53
Solar Energy 166 (2018) 50–58
continuing to enhance the air circulation. Fig. 4 demonstrates the effect
of the flow on the velocity and temperature profiles corresponding to
different values of PJ. The forced airflow constrains the heated air to
the cavity so that more thermal energy from the inner wall can be
obtained. The effectiveness of cutting off the flow through the aperture
to surroundings by the air-curtain depends on the value of PJ. In the
clockwise cases (PJ > 0), the hot air is blown in front of the aperture
from the upper slot while cold air is pumped from the lower slot. As
Fig. 4 g shows, a non-enclosed air circulation emerges when PJ is small.
The heated air inside the cavity can be partially sealed, but the effect is
limited (Fig. 4b). When further increasing PJ from 10 Pa to 20 Pa, a
closed airflow is formed (Fig. 4i), and the heat losses are clearly re-
duced (Fig. 4d). In the anticlockwise cases (PJ < 0), cold air is blown
in front of the aperture from the lower slot while hot air is pumped from
the upper slot. This mode improves the heat loss reduction mechanisms
due to the natural convective flow pattern in and around the cavity. In
the buoyancy dominant condition (PJ = 0), cold air enters from the
bottom of the aperture, being heated up while flowing through the
cavity along inner walls, and finally exits from the top (Fig. 4f).
Therefore, heated air is more easily constrained in the anticlockwise
cases than clockwise as confirmed by Fig. 4b and d versus Fig. 4c and e.
It’s worth noting that the benefits from the air-curtain may be lost
when increasring the forced airflow too much. Fig. 5 depicts the profile
of pathlines throughout the cavity (the relative values of temperature
are shown in colors). Although the airflow can seal hot air inside effi-
ciently, a small portion of it is still spilled due to the diffusing effect.
Since the spillage cannot be avoided by increasing |PJ|, the sealing
effect cannot be improved when PJ < −20 Pa. However, at the same
time, the turbulence inside the cavity will grow. As a consequence, the
convective heat transfer of the inside area close to the aperture will be
enhanced. The anti-clockwise eddy in the front of the cavity expands,
when changing PJ from −20 Pa to −30 Pa. Moreover, in this region the
temperature drops, which indicates that more heat is carried out
through the unavoidable spillage of airflow. On the other hand, a
stronger forced airflow will require more energy input, which is not
cost-efficient.
In Fig. 6, we show the temperature distributions for three cross
sections in case of θ = −15° when PJ = 0 and −20 Pa. The contour
maps indicate that a considerable improvement is achieved in reducing
both convective heat losses across the opening and temperature gra-
dients throughout the inner areas of cavity. The temperature distribu-
tions in the bottom and middle sections clearly show how the forced
S. Yang et al. Solar Energy 166 (2018) 50–58

Fig. 4. Comparison of temperature and velocity contour images with and without the air-curtain effect (θ = −15°). Left: (a–e) show the contours of the temperature with different values
of PJ (0, ± 10, ± 20 Pas). Right: (f–j) velocity contours related to a–e.

flow even outs temperature gradients. The temperature differences
observed in the top section are caused by the lower temperature of the
incoming air than that inside the cavity, which would reduce the
homogeneity of the spatial temperature distribution near the opening.
The reduced convective heat losses from forced flow improve the
thermal efficiency; the reduced temperature differences could also re-
duce the thermal stresses in the receiver.
3.2.2. Boosted stability of convective heat losses with varying inclinations
Next we investigated the effect of the forced flow with varying

Fig. 5. Profiles of pathlines, colored with local values of relative temperature, throughout the cavity (θ = −30°) in cases of (a): PJ = −20 Pa, (b): PJ = −30 Pa.

54
S. Yang et al. Solar Energy 166 (2018) 50–58

Fig. 6. Contour maps of the temperature in three different cross sections: bottom (left), middle (center), top (right) planes inside the cavity. Natural flow: PJ = 0 Pa (above) and air-forced
flow: PJ = −20 Pa (below) conditions shown.

inclination. The convective heat losses significantly vary with the tilt
angle (Leibfried and Ortjohann, 1995), but this would be in particular
pronounced with the so-called 2-stage dish concentrator (Wang et al.,
2017; Yang et al., 2018) which always points upwards. In Fig. 7 we
show the convective losses as a function of forced flow (PJ) and in-
clination angle. The positive effect of the air-curtain is not observed
until the flow is beyond 10 Pa. An adverse effect may even appear when
PJ is low. For instance, the peak value of the heat losses is found at
around PJ = 5 Pa (but not at 0 Pa) when θ = −30°. The clockwise and
anticlockwise flows are not symmetrical in terms of the positive effect.
For that same |PJ| value, the anticlockwise circulation yields lower
convective losses than the clockwise one. For example, the peak of the
convective losses found at PJ = −5 Pa is almost a third lower than that
at PJ = 5 Pa. A 20% lower flow is needed to reach the optimal
operation point with the anticlockwise circulation than with the
clockwise one. These results clearly indicate that an anticlockwise flow
is more effective to reduce the convective losses of the receiver. The
positive effects of the air-curtain at θ < 0 are clearly more pronounced.
This is meaningful for up-tilted receivers and further demonstrated in
Figs. 8 and 9, which show the numerical values. Inversely, we also see
that the variance of the heat loss decreased with θ when θ > 0. This
indicates that the air-curtain for receivers, which are down-titled, is less
effective. This is mainly because the convective zone can shrink rapidly
as θ increases in cases of a down-tilted receiver, i.e. the convective
losses through cavity are reduced to a low value even without the air-
curtain. For example, the convective heat losses for θ = 30° and PJ = 0
is 50.4 W and for PJ = −20 Pa 40.2 W, respectively, i.e. the benefit
from the air-curtain is much less than in case θ < 0.

Fig. 7. Convective heat as a function of inclination and pressure-jump (forced flow).

55
S. Yang et al.
Fig. 8. Influence of inclination on the convective heat loss with different forced flows (PJ < 0) in case of anticlockwise flow.
3.3. Estimation of required auxiliary power for forced flow
To realize the air cushion through forced flow, external power will
be necessary to run a fan. This could be realized e.g. by using a pho-
tovoltaics module (Green and Emery, 1993). The amount of required
power input is given by, the following equation:
∫ |PJ|vds
qinput =
ηPV ηfan
where ηPV and ηfan represent the efficiencies of the PV cell and the fan in
the air circulation system, v and ds correspond to the velocity of air flow
and the differential element of section area in air channel, respectively.
For |PJ| = 40 Pa, the mass flow corresponds to 0.66 g/s. To obtain this
value, only 0.72 W of power is needed, which could be realized with a
very small PV module adding a negligible cost.
Fig. 9. Influence of inclination on the convective heat loss with different forced flows (PJ > 0) in case of clockwise flow.
56
Solar Energy 166 (2018) 50–58
4. Conclusions
The air-curtain concept, which is usually applied in air conditioning
and refrigeration, was here applied to a solar dish receiver to suppress
convective heat losses. Through a forced airflow, the heat losses across
the aperture could clearly be reduced. Also, the variation of heat losses
at varying inclinations could be decreased.
The improvements from the forced-air circulation systems were
clearly higher with the anticlockwise that with the clockwise flow
mode. In the optimal case at PJ = −20 Pa and θ = −60°, more than
58% of the convective losses could be reduced compared to natural air
circulation conditions.
Another advantage from the novel configuration was the smaller
variance of convective heat losses of the receiver with changing in-
clination. The swing of convective losses from the average value qconv
does not exceed ± 5.9 W which equals to a relative deviation
of ± 13.9%, when θ varies between [−80°, 80°] and PJ = −20 Pa. By
contrast, the old configuration without an air circulation system shows
S. Yang et al. Solar Energy 166 (2018) 50–58

a large swing with a relative deviation up to 84.3%. Acknowledgements

The extra power required for driving the air circulation system is
very small compared to the amount of reduced heat losses and could in This work was supported by the National Science Foundation of
practice be ignored. China (No. 51736006).

Appendix A

Here we present the governing equations and the near-wall treatment of the model used in the simulations. The instantaneous values of velocity
(uk) and temperature (T) fields are given by the sum of their time averaged mean value (uk , T ) and corresponding turbulent fluctuating component
(uk′, T′).
Governing equations:
∂ui ∂u
ρ +ρ i =0
∂t ∂x i (A1)

∂ui ∂u ∂p ∂ ⎛ ∂ui
ρ + ρuj i = − + ⎜μ −ρui′ u′j ⎞⎟ + Fj
∂t ∂x j ∂x j ∂x j ⎝ ∂x j ⎠ (A2)

∂T ∂T ∂ ⎛ ∂T
ρcp + ρcp ui = λ ⎜ −ρui′ T ⎞ + Si ⎟

∂t ∂x i ∂x i ⎝ ∂x i ⎠ (A3)

∂ (ρk ) ∂ (ρkui ) ∂ ⎡⎛ μ ∂k ⎤
+ = μ + t⎞
⎜ + Gk−ρε
⎟

∂t ∂x i ∂x j ⎢
⎣⎝ σ ⎥
k ⎠ ∂x j ⎦ (A4)

∂ (ρε ) ∂ (ρεui ) ∂ ⎡⎛ μ ∂ε ⎤ G ε2
+ = μ + t⎞
⎜ + 1ε Gk−G2ε ρ
⎟

∂t ∂x i ⎢
∂x j ⎣ ⎝ ⎥
σε ⎠ ∂x j ⎦ k k (A5)
In addition, the form of Fi in Eq. (A2) can be rewritten as ρα (T −Tref ) gi δij according to the Boussinesq approximation for simulating buoyancy
where α, g and Tref represent coefficient of thermal expansion, acceleration due to gravity, and the reference temperature, respectively.
Enhanced Wall Treatment combines a two-layer model with the enhanced wall functions.
Two-layer model:
In the viscosity-affected near-wall region, μt,layer is computed from:

μt ,layer = ρCμ lμ k (A6)

lμ = yCl∗ (1−e−Rey / Aμ ) (A7)

ρy k
where lμ is the length scale, Re y is turbulent Reynolds number (< 200), defined as . Cμ, Cl∗ and Aμ are constants then the two-layer formulation
μ
for the enhanced wall treatment is proposed as:
μt ,ewt = λ ε μt + (1−λ ε ) μt ,layer (A8)
here a blending function, λ ε , is determined by a set of semi-empirical equations varying from 0 to 1.
In the viscosity-affected near-wall region, ε is computed from:
k 3/2
ε=
lε (A9)

lε = yCl∗ (1−e−Rey / Aε ) (A10)

where lε is the length scale, Aε is a constant.
Enhanced Wall Functions:
The law-of-the wall for the entire wall region uses the function below:
+ + + +
u+ = e Γ1ulam + e1/Γ1uturb ;T + = e Γ2 Tlam + e1/Γ2 Tturb (A11)
where subscripts lam and turb represent the laminar and turbulent conditions, and corresponding variables are derivable. The blending functions are
given by:
0.01(y+ ) 4 0.01(Pry+) 4
Γ1 = − +
;Γ2 = −
1 + 5y 1 + 5 Pr 3y+ (A12)
uτ is the friction velocity and dimensionless variables are defined as:
uP + TW −TP
y+ = ρu τ y / μ;u+ = ;T = qW
uτ
ρcp u τ (A13)

References enclosures for engineering applications. The particular case of the parallelogrammic
diode cavity. Appl. Therm. Eng. 63 (1), 304–322.
Blanco, M., Santigosa, L.R., 2016. Advances in Concentrating Solar Thermal Research and
Baïri, A., Zarco-Pernia, E., De María, J.-M.G., 2014. A review on natural convection in Technology. Woodhead Publishing.

57
S. Yang et al.
Clausing, A.M., 1981. An analysis of convective losses from cavity solar central receivers.
Sol. Energy 27 (4), 295–300.
Cui, F., He, Y., Cheng, Z., Li, Y., 2013. Study on combined heat loss of a dish receiver with
quartz glass cover. Appl. Energy 112, 690–696.
Flesch, R., Grobbel, J., Stadler, H., Uhlig, R., Hoffschmidt, B., 2016. Reducing the con-
vective losses of cavity receivers. In: AIP Conference Proceedings. AIP Publishing, pp.
030014.
Foster, A., Swain, M., Barrett, R., D’Agaro, P., Ketteringham, L., James, S., 2007. Three-
dimensional effects of an air curtain used to restrict cold room infiltration. Appl.
Math. Model. 31 (6), 1109–1123.
Fox, J., 1965. Heat transfer and air flow in a transverse rectangular notch. Int. J. Heat
Mass Transf. 8 (2), 269–279.
Green, M.A., Emery, K., 1993. Solar cell efficiency tables. Prog. Photovoltaics Res. Appl. 1
(1), 25–29.
Hogan, R.E., 1994. AEETES—a solar reflux receiver thermal performance numerical
model. Sol. Energy 52 (2), 167–178.
Hughes, G., Pye, J., Kaufer, M., Abbasi-Shavazi, E., Zhang, J., McIntosh, A., Lindley, T.,
2016. Reduction of convective losses in solar cavity receivers. In: AIP Conference
Proceedings. AIP Publishing, pp. 030023.
Kolb, G.J., 2012. Suction-recirculation device for stabilizing particle flows within a solar
powered solid particle receiver. Google Patents.
Leibfried, U., Ortjohann, J., 1995. Convective heat loss from upward and downward-
facing cavity solar receivers: measurements and calculations. J. Sol. Energy Eng.
117, 75.
Mavriplis, D., 1995. Three-dimensional multigrid Reynolds-averaged Navier-Stokes solver
for unstructured meshes. AIAA J. 33 (3), 445–453.
McDonald, C.G., 1995. Heat loss from an open cavity. Sandia National Labs.,
Albuquerque, NM (United States); California State Polytechnic Univ., Pomona, CA
(United States). Coll. of Engineering.
Ngo, L., Bello-Ochende, T., Meyer, J., 2015. Numerical modelling and optimisation of
58
Solar Energy 166 (2018) 50–58
natural convection heat loss suppression in a solar cavity receiver with plate fins.
Renew. Energy 74, 95–105.
Paitoonsurikarn, S., Lovegrove, K., 2002. Numerical investigation of natural convection
loss in cavity-type solar receivers. Proc. Solar 1–6.
Paitoonsurikarn, S., Lovegrove, K., 2006. A new correlation for predicting the free con-
vection loss from solar dish concentrating receivers, Solar. p. 44th.
Reddy, K., Kumar, N.S., 2009. Convection and surface radiation heat losses from modified
cavity receiver of solar parabolic dish collector with two-stage concentration. Heat
Mass Transf. 45 (3), 363–373.
Reddy, K., Veershetty, G., Vikram, T.S., 2016. Effect of wind speed and direction on
convective heat losses from solar parabolic dish modified cavity receiver. Sol. Energy
131, 183–198.
Tan, T., Chen, Y., Chen, Z., Siegel, N., Kolb, G.J., 2009. Wind effect on the performance of
solid particle solar receivers with and without the protection of an aerowindow. Sol.
Energy 83 (10), 1815–1827.
Taumoefolau, T., Paitoonsurikarn, S., Hughes, G., Lovegrove, K., 2004. Experimental
investigation of natural convection heat loss from a model solar concentrator cavity
receiver. J. Sol. Energy Eng. 126 (2), 801–807.
Wang, J., Yang, S., Jiang, C., Yan, Q., Lund, P.D., 2017. A novel 2-stage dish concentrator
with improved optical performance for concentrating solar power plants. Renew.
Energy 108, 92–97.
Wu, S.-Y., Xiao, L., Cao, Y., Li, Y.-R., 2010. Convection heat loss from cavity receiver in
parabolic dish solar thermal power system: a review. Sol. Energy 84 (8), 1342–1355.
Yang, S., Wang, J., Lund, P.D., Jiang, C., Liu, D., 2018. Assessing the impact of optical
errors in a novel 2-stage dish concentrator using Monte-Carlo ray-tracing simulation.
Renew. Energy 123, 603–615.
Zhang, J.J., Pye, J.D., Hughes, G.O., 2015. Active air flow control to reduce cavity re-
ceiver heat loss. In: Proc. Ninth International Conference on Energy Sustainability.
ASME Power & Energy, San Diego.