Proceedings of Insert Conference Abbreviation

Fig. 1: CSP central receiver plant [8] 
Figure 1 displays an example of a central receiver power plant. 
The central receiver power plant comprises of mirrors 
(heliostats) that track the sun about two axes, a central receiver 
to heat a heat transfer fluid (HTF) which in turn is used to heat 
the working fluid (steam) for a Rankine cycle.
2. Model description 
2.1 Heliostat field 
The site selected for this case study is Upington, a large town in 
the Northern Cape, on the Orange River. The coordinates and 
elevation for Upington is given in table 1. 
Latitude 
Longitude 
Elevation 
-28.395 
21.2368 
879 
Table 1: Site location
Weather data for a typical meteorological year (TMY) for 
Upington was collected from the Solar Radiation Data [9]. 
Hourly averaged time step data was used to model CSP plant 
and it was assumed that the plant operate through successive 
steady states. Solar-PILOT software from National Renewable 
Energy 
laboratory 
(NREL) 
was 
used 
for 
design, 
characterization and optimization of concentrating solar power 
field geometry [10]. Weather data was provided in the format 
required by Solar-PILOT. The heliostat field layout was set up 
at design-point (spring equinox), and DNI, solar azimuth and 
solar elevation angles were specified. Then the heliostat 
geometry, focus parameters, mirror performance parameters 
and optical error parameters were set to determine heliostat 
efficiency at the design point. To determine the annual 
efficiency of the heliostat field, time integration over a full year 
was performed. Parameters required for the integration were the 
hourly DNI over the year, and the position of sun (azimuth and 
altitude angle). The latter were calculated from Duffie and 
Beckman [11]. 
2.2 Central receiver 
To determine the total energy received from the optical 
heliostat field, it is assumed that an external receiver is used 
with solar salt as heat transfer fluid. Crystallization of the salt 
mixture starts at 240 ºC whilst the salt starts to decompose at 
temperature above 600 ºC. The total heat received is calculated 
as: 
ℚ
𝑜𝑓
= (𝐷𝑁𝐼)(𝜂
𝑜𝑓
)(𝐴
𝑓𝑎
)(𝑆𝑀) (1)
Where the field aperture area 
(𝐴
𝑓𝑎
)
is known as 
𝐴
𝑓𝑎
= (𝐻
𝑡
)(𝐴
ℎ𝑒
) (2)
There is no set up for the circumferential difference of the 
average heat flux on the receiver that will change throughout 
the day and the energy balance for the receivers that produces 
the heat transfers to the salt is determined. 
(𝛼)ℚ
𝑜𝑓
= 𝜎𝜀𝐹𝐴(𝛵̅
𝑚𝑟
4
− 𝛵
𝑎
4
) + 𝑈𝐴
𝑟𝑎
(𝛵
𝑚𝑟𝑠
− 𝛵
𝑎
) + ℚ
𝑠𝑎𝑙𝑡
(3)
It is assumed that the receiver heat flux 
𝑞
𝑚𝑎𝑥
′
is limited to 
700 kW
t
/m
2
as suggested by Sargent and Lundy LLC 
Consulting Group [12]. A receiver aspect ratio, 
(𝐿/𝐷)
of 1.6 
was assumed, which allows one to calculate its height L and 
diameter D respectively. Overall heat transfer 
𝑈
is assumed to 
be approximately equal to the air side convective heat transfer 
coefficient. The receiver absorber area is conservatively taken 
as 
𝐴
𝑟𝑎
= 𝜋𝐿𝐷 (4)
For radiation heat losses, it was assumed that the ground, air 
and surrounding structures are at ambient air temperature. Since 
the mean radiation temperature is much greater than the air 
temperature, the impact of this assumption will be insignificant. 
Also assumed that since the receiver is completely enclosed by 
its environment, the radiation shape factor will be 1 and the 
radiation loss is establish from the integration over the receiver 
surface presented by Hoffmann and Madaly [13]. 
ℚ
𝑟𝑎𝑑
= 𝜎𝜀 ∫ [𝑇
4
(𝜉) − 𝑇
𝑎
4
]𝜋𝐷 𝒹𝜉
𝐿
0

---

= 𝜎𝜀𝜋𝐷 ∫ 𝑇
4
(𝜉)𝒹𝜉 − 𝜎𝜀𝜋𝐷𝐿𝑇
𝑎
4
𝐿
0
= 𝜎𝜀 𝜋𝐷𝐿(𝛵̅
𝑚𝑟
4
− 𝑇
𝑎
4
) (5)
For a linear salt temperature distribution in the receiver, the 
mean radiation temperature at the receiver surface is calculated 
by 
𝛵̅
𝑚𝑟
4
= ∫ 𝑇
4
(𝜉)𝒹𝜉
𝐿
0
𝛵̅
𝑚𝑟
4
=
𝛵
𝑚𝑎𝑥
4
+ 𝛵
𝑚𝑎𝑥
3
𝛵
𝑚𝑖𝑛
+ 𝛵
𝑚𝑎𝑥
2
𝛵
𝑚𝑖𝑛
2
+ 𝛵
𝑚𝑎𝑥
𝛵
𝑚𝑖𝑛
3
+ 𝛵
𝑚𝑖𝑛
4
5
(6)
Where 
𝛵
𝑚𝑎𝑥
= 𝛵
𝑜𝑠
+
2𝑞
𝑚𝑎𝑥
′
𝐷
𝑜𝑟
𝑙𝑜𝑔(𝐷
𝑜𝑟
𝐷
𝑖𝑟
⁄
)
𝑘
𝑐
𝛵
𝑚𝑖𝑛
= 𝛵
𝑖𝑠
+
2𝑞
𝑚𝑎𝑥
′
𝐷
𝑜𝑟
log(𝐷
𝑜𝑟
𝐷
𝑖𝑟
⁄
)
𝑘
𝑐
The receiver also experiences heat losses due to convection. To 
determine the convective loss which contains of natural and 
forced (wind focused) convection, a convective heat transfer 
coefficient is determined from 
ℎ = √ℎ
𝑛𝑐
2
+ ℎ
𝑓𝑐
2
(7)
The convective heat transfer coefficient is reliant on the extent 
of the mixed velocity through the receiver which will recover 
both the limiting cases for forced convection when 
ℎ
𝑛𝑐
2
= 0
and 
natural convection for which
ℎ
𝑓𝑐
2
= 0
. The Nusselt number 
presented in Çengel and Ghajar [14] for natural convection 
(𝑁𝑢
𝑛𝑐
)
is determined by Rayleigh number 
(𝑅𝑎)
as 
𝑁𝑢
𝑛𝑐
=
ℎ
𝑛𝑐
𝐷
𝑘
𝑎
= 0.1𝑅𝑎
1 3
⁄
(8)
Where 
𝑘
𝑎
,
the thermal conductivity of air, and the air properties 
are calculated at the mean film temperature. The mean receiver 
surface temperature are also determined by
𝛵
𝑚𝑟𝑠
≈
𝛵
𝑜𝑠
+ 𝛵
𝑖𝑠
2
+
2𝑞
𝑚𝑎𝑥
′
𝐷
𝑜𝑟
𝑙𝑜𝑔(𝐷
𝑜𝑟
𝐷
𝑖𝑟
⁄
)
𝑘
𝑡𝑚
(9)
The receiver is approximated as a cylinder. The forced 
convection heat transfer coefficient for flow across a circular 
cylinder presented by Zukauskas [14] is calculated as 
𝑁𝑢
𝑓𝑐
=
ℎ
𝑓𝑐
𝐷
𝑘
𝑎
= 0.027𝑅𝑒
0.805
𝑃𝑟
0.333
(10)
The air properties are also calculated again at the mean film 
temperature due to the Reynolds 
𝑅𝑒
and Prandtl numbers
𝑃𝑟
. 
Where Reynolds number is defined as 
𝑅𝑒 =
𝜌𝑉(𝑦)𝐷
𝜇
(11)
The exact shape of the wind profile depends on the atmospheric 
stability, but for accessibility the wind speed at the receiver 
height is calculated from the 1/7
th
law for a neutral (adiabatic) 
atmosphere which is 
𝑉(𝑦) = 𝑉
10
(
𝑦
10
)
1 7
⁄
(12)
Standard wind speed measurements are taken at 10 m above the 
ground level. Hence, convectional loss is determined by 
ℚ
𝑐𝑜𝑛𝑣
= 𝜋 𝐷 𝐿 ℎ
𝑓𝑐
(𝛵
𝑚𝑟𝑠
− 𝛵
𝑎
) (13)
2.3 Thermal energy storage 
Figure 2 assumes that thermal energy storage loss is limited to 
the tank side wall, and only up to the salt level inside the tank. 
Since there is a direct connection between the salt inventories 
stored inside the tank and heat loss from the tank, assumed that 
the overall heat transfer coefficient is constant. That is U= 6 
W/ºC which bring about 1.5 % loss to the energy stored per day 
from a fully changed tank [15]. The thermal energy loss is 
given as 
ℚ
𝑡𝑒𝑙
= 𝜋𝐷𝐻
𝑠
𝑈(𝑇̅
𝑚𝑠
− 𝑇
𝑎
) (14)