Pisa, 22-24 September 2004

TWO-PHASE FLOW MODELLING IN A SOLAR CONCENTRATOR

WITH AMMONIA EVAPORATION
,
Naghelli Ortega* °, Octavio García-Valladares*, Roberto Best*

*Centro de Investigación en Energía-UNAM, Apdo. Postal 34, 62580, Temixco, Morelos, México
°Centro de Investigación en Energía-UNAM, Posgrado en Ingeniería-Energía,
Apdo. Postal 34, 62580, Temixco, Morelos, México, [email protected]

ABSTRACT

A solar concentrator of the compound parabolic concentrator type was studied as a vapour generator in an ammonia-water
absorption solar refrigerator. A methodology was developed in order to analyse the evaporation process, where the
temperature lift associated with the ammonia evaporation and the forced convection two-phase flow were taken into
consideration.
The system was considered in steady state and a one-dimensional analysis was applied. In order to model the ammonia-
water mixture behaviour during the evaporation process, the local fluid temperature, the ammonia concentration in liquid and
vapour phases, and the vapour-liquid relation at the two-phase intermixture must be determined. Local thermodynamic
equilibrium between the liquid and vapour phases was supposed; since the mixture temperature changed with position, the
equilibrium conditions always changed.
A one-dimensional numerical simulation of the thermal and fluid-dynamic behaviour of two-phase flow had been
developed. The governing equations (continuity, momentum and energy) inside the tube, together with the energy equation in
the tube wall and the thermal analysis in the solar concentrator, were solved iteratively in a segregated manner. The discretized
governing equations in fluid flow were coupled using an implicit step-by-step method.

1 INTRODUCTION troughs are used for applications where temperatures are in

the order of 150°C and 300°C, respectively [5].
In great part of the world, the deficiencies in electric power Since 1990, solar parabolic troughs have been used to
supply have implications in refrigeration availability [1]. In evaporate water and to produce steam directly on the receiver
the last few years, there have been advances in research in [10]. The technology developed is known as direct steam
order to develop new refrigeration and cooling systems. The generation (DSG). The vapour produced is mainly applied for
main research line has been focused on thermal systems, power generation and, recently, for desalinisation processes
applying alternative energy sources, as residual, geothermal [11]. DSG presents many advantages compared to the oil
and solar energy. based technology, since DSG eliminates costly synthetic oil,
Solar energy has the advantage that the cooling load is intermediate heat transport piping, special type equipment to
generally required when solar radiation is available [2]. This run the high temperature oil, and oil to steam heat exchanger
feature represents the main reason for sustained research into [12].
solar cooling devices for three decades, which include the Based on the main advantages of DSG, a CPC was
combination of solar energy technologies with some thermal designed in order to directly evaporate ammonia [13]. The
cooling technologies as absorption, adsorption and desiccant ammonia vapour would be utilized in an ammonia-water
[2-4]. absorption solar refrigerator. To our knowledge, there is not
Solar thermal technologies can be divided in flat-plate another study about direct ammonia vapour generation in a
collectors, and concentrators. The main difference between solar concentrator.
them is the proportion among the aperture area and the In the present work, the evaporation process in the CPC
receptor area. In flat-plate collectors this relation is 1.0 and in designed was analysed. A detailed one-dimensional numerical
concentrators is higher than 1.0. This geometrical relation is simulation of the thermal and fluid-dynamic behaviour of
known as the concentration ratio. two-phase flow was developed.
Flat-plate collectors are mainly used for hot water
production in the residential and the commercial sectors, 2 COMPOUND PARABOLIC CONCENTRATOR
where the maximum temperature reached is approximately
80°C. This feature, in addition to its sensitivity to A trough CPC consists of three primary components:
environmental temperature and insolation, limits the reflector, receptor and cover. The reflector is developed by
application of these devices [5]. two segments, an involute section and a parabolic segment,
There are several types of solar concentrators, such as where the AB arc is the involute of the AT arc (Figure 1).
compound parabolic concentrators (CPCs), parabolic troughs,
Fresnel reflectors, and central receivers [6-9]. The main
differences between them are the concentration ratio that each
one can reach, and its applications, e.g. CPC and parabolic
 4.1 Subcooled Liquid Region

The Gnielinski correlation [16] was used in order to

calculate the heat transfer coefficient at the single-phase
region:

α f , l = max (α f ,l ,3.66 ) (2)

Where:

( f 8 )( Re− 1000 ) Pr k
α f ,l = (3)
1 + 12.7 ( f 8 ) ( Pr 2 3 − 1) Di
Figure 1. CPC with tubular receptor

The reflector shape must guarantee that in any point P, the

f = (1.82 log10 Re− 1.64 )
reflector normal NP bisects the angle between the tangent to −2
(4)
the receptor PT and the line QP, which forms a θc acceptance
half-angle with the CPC axis. The CD line represents the
cover. The friction factor was evaluated from the expression
In the study case, the receptor was tubular and was located proposed by Churchill [17]. In the subcooled boiling region
on the focus of the CPC. (if it exists) the heat transfer coefficient was estimated
The CPC designed [13] had an inside and outside diameter according to Kandlikar [18].
tube of 0.0525m and 0.0603m, respectively; an acceptance
half-angle of 15° that corresponds to a concentration ratio of 4.2 Equilibrium Two-Phase Region
3.86, and a height of 1.66m.
If the CPC reflector is fully developed the ends almost do Separated two-phase flow was considered; the void
not contribute to concentrate the solar radiation, for this fraction was estimated from the semi-empirical equation of
reason the CPC was truncated up to a height of 0.76m that Premoli (cited by Rice [19]). For the convective heat transfer
corresponds to a real concentration ratio of 3.5. At this height coefficient, two different models were applied, the correlation
the cover width was 0.66m. proposed by Mishra, Varma, and Sharma with the coefficients
for ammonia-water mixture obtained by Rivera and Best [20];
3 MATHEMATICAL FORMULATION and the Zürcher, Thome and Favrat flow boiling model [21].
The friction factor was calculated from the same equation
Local knowledge of the shear stresses, the two-phase flow as in the case of subcooled liquid flow using a correction
structure and the heat transfer through the solar concentrator, factor (two-phase frictional multiplier Φ) according to Friedel
tube and fluid were necessary in order to achieve the one- [22].
dimensional model.
Since there was thermodynamic equilibrium in the 5 NUMERICAL RESOLUTION
subcooled liquid and two-phase regions, temperature, mass
fraction and all the thermodynamic properties could be
Figure 2 shows the receptor tube cross-section. The
calculated using matrix functions of the pressure and enthalpy, subcooled ammonia-water mixture enters the tube at position
i.e.: 0 with a mass flow m , and an inner temperature Tf,i. The
receptor tube receives a useful energy gain qu. Ammonia
φ = φ ( P, h) where φ = T , xg , ρ ,... (1) starts to evaporate at a certain length Ls, where saturation
temperature Tf,s is reached. Finally, the two-phase mixture is
For the ammonia-water mixture, these matrices were out at position L with an outside temperature Tf,o.
obtained using REFPROP v7.0 [14].
Some transport properties (thermal conductivity, and
viscosity) were calculated with the correlations proposed by
Selim and Elsayed [15], as functions of mass concentration
and temperature.

4 EMPIRICAL COEFFICIENTS EVALUATION

The analysis of the flow inside the receptor tube in the

CPC was divided in two sections: a section where the
ammonia-water mixture was subcooled until it reached the Figure 2. Receptor tube cross-section
saturation temperature, and the two-phase section where
ammonia was evaporated and a two-phase mixture was The numerical analysis was carried out by means of a
obtained. control volume (CV) method on the receptor tube. The
The convective heat transfer coefficient and friction factor discretized equations were coupled using a fully implicit step-
were calculated in different ways for each receptor tube by-step method in the flow direction. From the known values
section: subcooled liquid region and two-phase region. at the inlet section and guessed values of the wall boundary
conditions, the variable values at the outlet of each CV were
iteratively obtained from the discretized governing equations. Where, in terms of the mass flow rate, gas and liquid
This solution (outlet values) was the inlet values for the next velocities were:
CV. The procedure was carried out until the end of the
m (1 − xg )
 
receptor tube is reached.  mxg 
Vg = Vl = (13)
5.1 Discretization of Equations ρ g ε g At ρl (1 − ε g ) At

Fluid flow analysis. For each CV, a set of algebraic The energy governing equation was written as:
equations was obtained by governing equations discretization
 
(continuity, momentum and energy). ∂ P V2   ∂  Vx2 
The transient terms of the governing equations were ρV  h f − + x  + m  f
h +  = qu (14)
discretized using the approximation: ∂t  ρ 2  ∂x  2 

∂φ ∂t ≅ (φ − φ o ) ∆t where φ = h, P, T , ρ ,... (5) The outlet enthalpy was calculated from the discretized of
Eq.(14), and Eq.(8):
 
Where superscript “o” indicates the value of the previous
( 2π Din ∆x ) qwall − ami + bmi −1 + cAt ∆x ∆t
instant. hi =   (15)
The receptor tube was divided in N control volumes: mi + mi −1 + ρtpo At ∆x ∆t

i = 0,1,..., N φ = φi (6) Where:

The variables averages were estimated by the arithmetic qwall = α f (Ta , i − T f , i )

mean between their inlet and outlet values:  
a =  xgVg + (1 − xg )Vl  − hi −1
2

i
φi ≅ φi ≡ (φi + φi +1 ) 2 (7)   (16)
b =  xgVg + (1 − xg )Vl 
2
− hi −1
i −1
 
Based on the numerical approaches indicated above, the
final form of the governing equations is given below.
c = 2 ( pi −1 − p o
i −1 ) − ρ (h o
tp i −1 ) (
− 2hi o−1 − ρVi −21 − ρ oVi o−12 )
The continuity governing equation was:
This formulation of conservation equations was applied to

∂ρ ∂ ρVx
+
(
=0
) (8)
transient two-phase flow, as well as particular cases of steady
flow or single-phase flow.
∂t ∂x
Receptor tube wall. The conduction in the internal tube
The outlet mass flow rate was obtained from the discretized wall was calculated assuming the following hypotheses: one-
of Eq.(8): dimensional transient temperature distribution. A
characteristic CV is shown in Figure 3, where P represents the
At ∆x central node, E and W indicate its neighbours. The CV-faces

mi = mi −1 −


∆t
( ρtp − ρtpo ) (9) are indicated by e, w, n and s.

Where the two-phase density was obtained from:

ρtp = ε g ρ g + (1 − ε g ) ρl (10)

The momentum governing equation was:

  
Figure 3. Discretized receptor tube wall
 ∂V  ∂V  ∂P ∂ 2V
ρ  x + Vx x  = − + µ 2x (11) Integrating the energy equation over this CV, the following
 ∂t ∂x  ∂x ∂x
equation was obtained:

The outlet pressure was obtained from the discretized of  ∂h  
Eq.(11): ( 
)
qwall ps − qu pn ∆x + qw − qe Ata = m
∂t
( ) (17)
  
 f m2 m − mo  
π Din Φ +  Where qwall was evaluated using the convective heat
 4 2 ρtp At2 ∆t  transfer coefficient in the fluid flow, and the conductive heat
∆x   m  


( )
 
Pi = Pi −1 − +  xgVg + (1 − xg )Vl   (12) fluxes were evaluated from the Fourier law, this was:
At   ∆x i 
  m  
  ∂T    ∂T 
( )
 
−  xgVg + (1 − xg )Vl   qe = −α e  a  qw = −α w  a  (18)
  ∆x  i −1   ∂x e  ∂x  w
 The following equation was obtained for each node of the

( )
grid: −1
 α t−, a1 − c α t−, r1− ex + (α ra , r − c + α t , a − r ) 
−1

 1 
aTa ,i = bTa ,i +1 + cTa ,i −1 + d U L ,i = + (23)
(19) α t , c − ex α t−, a1 − c + α t−, r1− ex + (α ra , r − c + α t , a − r )−1 
 
Where the coefficients were:
Where individual heat loss coefficients were:
k A k A A ∆x
a = w ta + e ta + α f ,i ps ∆x + ta ρ Cp α t ,c − ex = α co ,c − en + α ra ,c − sk
∆x ∆x ∆t
k A k A α t , r − ex = α co ,r − en + α ra , r − sk
b = e ta c = w ta (20) (24)
∆x ∆x α t , a − c = α co , a − c + α ra , a − c
Ata ∆x
d = (α f , i psT f , i + qu , i pn ) ∆x + ρ CpTwo,i α t , a − r = α co , a − r + α ra , a − r
∆t
Ac
The coefficients mentioned above were applicable for α co ,c − en = (5.7 + 3.8υ ) (25)
2 ≤ i ≤ N − 1 ; for i = 1 and i = N adequate coefficients were Aa
used to take into account the axial heat conduction or
temperature boundary conditions. The set of heat conduction Ar
discretized equations was solved using the algorithm TDMA α co , r − en = (5.7 + 3.8υ ) (26)
Aa
[23].

5.2 Boundary Conditions Where the reflector area Ar was calculated by:

 π 
The ammonia-water mixture was subcooled at the receptor
 ϕinv
2 φ par + θ C + ϕ − cos (ϕ − θ C ) 
1
tube inlet, then mass flow, pressure, temperature and
composition were known, therefore inner enthalpy could be
Ar = Do ∆x  + ∫ 2
2 φinv 1 + sen (ϕ − θ C )
32
dϕ 
 4 
calculated. The heat flux or useful energy gain qu through the  
receptor tube wall had to be known. The energy balance in (27)
the CPC components had to be work out in order to obtain
this heat flux. This value was calculated taken account the
solar energy transferred through the solar concentrator α co , a − r = 3.25 + 0.0085
(T a ,i − Tr , i )
(28)
components until the receptor tube wall. Next section details 2 Do
the procedure.

α ra , r − sk = ε rσ (Tr2,i + Tsk2 ) (Tr ,i + Tsk )

Ar
5.3 Useful Energy Gain (29)
Aa
The useful energy gain in each CV was calculated by [13]:
α ra , c − sk = ε cσ (Tc2,i + Tsk2 ) (Tc ,i + Tsk )
Ac
(30)
= c S − U L ,i (Ta ,i − Ten )
A Aa
qu ,i (21)
Aa
σ (Tc2,i + Tr2,i ) (Tc ,i + Tr ,i ) Ar
α ra , r − c = ⋅ (31)
1 − ε c (1 − ε r )  Ac  Aa
Cover and receptor area were defined as:
+  
Ac = w∆x Aa = π Do ∆x (22) εc ε r  Ar 

The useful energy gain depends on solar absorbed energy S σ (Ta2,i + Tc2,i ) (Ta ,i + Tc ,i )
that is equal to the incident solar energy reduced by optical α ra , a − c = (32)
losses in the concentrator [24]. Thereby, absorbed energy S is 1  Ac  1 
+   − 1
distributed as useful energy gain and thermal losses in the ε c  Aa  ε a 
cover, the reflector and the receptor tube, represented as the
overall heat loss coefficient UL.
α co , a − c = 3.25 + 0.0085
(T a ,i − Tc , i )
(33)
Solar absorbed energy. Solar absorbed energy S is a 2 Do
function of radiative properties of CPC components
σ (Ta2,i + Tr2,i ) (Ta ,i + Tr ,i )
(reflectance, emittance, absorptance, and transmittance) and
environmental conditions that depends on solar time (solar α ra , a − r = (34)
radiation, solar position, and environmental temperature). 1 − ε r (1 − ε a )  Ar 
Solar absorber energy was calculated with the method showed +  
εr ε a  Aa 
by Duffie and Beckman [24].

Overall heat loss coefficient. This coefficient UL,i depends

on temperatures of CPC components through the individual
heat loss coefficients:
 Component temperatures. Temperatures of cover and established by the previous model [13]: an input temperature
reflector were necessary in order to resolve Eq.(21). Both of 60ºC, an initial ammonia concentration of 38.7% and a
were determined by means of the energy balances in each generator pressure of 1.1MPa. These conditions would allow
CPC component [13]: producing 6kW of cooling power in an average solar day of
operation. Analysis was made for a CPC installed in Temixco,

Tc ,i =
(α ra , a − c + α co, a − c )Ta ,i + α ra , c − sk Tsk
+
Morelos, Mexico (18°50.36’ N, 99°14.07’ W), and for March
15 at solar noon, when solar absorbed energy per unit of area
α ra , a − c + α co , a − c + α ra ,c − sk + α co ,c − en − α ra , r − c -2
is 649.3Wm (the day with the highest radiation level during
(35)
α co, c − enTen − α ra , r − cTr , i the 2000 year).
The materials specified for the CPC designed were: carbon
α ra , a − c + α co, a − c + α ra , c − sk + α co, c − en − α ra , r − c
steel for the receptor tube (thermal conductivity of 16.3W
-1
(mK) , absorptance of 0.85, and emittance of 0.80), glass for
the cover (absorptance of 0.05, emittance of 0.85, and

Tr ,i =
(α ra , a − r + α co , a − r )Ta ,i + α co ,r − enTen
+
reflectance of 0.05) and highly polished aluminium for the
reflectors (absorptance of 0.10, emittance of 0.03, and
α ra , a − r + α co , a − r + α co , r − en + α ra , r − sk − α ra , r − c
(36) reflectance of 0.85).
α ra , r − sk Tsk − α ra , r − cTc ,i The receptor tube length established for the analysis was
α ra , a − r + α co, a − r + α co, r − en + α ra , r − sk − α ra , r − c 45m, except for the cases were the evaporation process did
not start at 45m. The mass flow rate was of 0.048kg/s for all
the calculations, with exception of the length for the
5.4 Numerical Solver saturation analysis presented in Figure 7.
The component temperatures, mixture quality, convective
At each time step the solution process was carried out on heat transfer fluid coefficient, useful energy gain and overall
the basis of a global algorithm that, in each CV, solves in a heat loss coefficient were recorded for certain conditions. The
segregated manner the flow inside the receptor tube, the heat analysis shows that the pressure loss in the receptor tube is
conduction in the receptor tube wall and the heat transfer in not significant.
the solar concentrator, in order to calculate all the variables at Figure 4 shows the temperatures of the reflector, the cover,
the outlet section, with the known values at the inlet section the receptor tube wall and the ammonia-water mixture,
and the boundary conditions. calculated with Rivera and Best model. The temperature lift
The coupling between the three main subroutines was in the subcooled zone was about 40°C, this is because of the
performed iteratively for each time step following the low fluid mass rate, and the low thermal conductivity of the
procedure: receptor tube.
140
Fluid flow inside the internal tube. The equations were
solved considering the receptor tube wall temperature 120
distribution as boundary condition, evaluating the convective Ta
heat transfer in each CV.
TEMPERATURE / (° C)

100

In the internal tube wall. The temperature distribution was 80

recalculated using the fluid flow temperature, the convective Tf

60
heat transfer coefficient evaluated in the preceding step, and
Tr
considering the useful energy gain as boundary condition. 40
Tc

Solar thermal analysis. The useful energy gain was 20

obtained by means of the thermal analysis in the CPC
components, and receptor tube wall temperatures calculated 0
0 5 10 15 20 25 30 35 40 45
in the previous steps.
LENGTH / (m)

Convergence was verified at each CV using the following Tf Ta Tc Tr

condition: Figure 4. Mixture and components temperatures for March 15.

 φi* − φi −1  Figure 5 presents the convective heat transfer coefficient in

1 −  < δ (37) the fluid flow, calculated by the correlations of Rivera and Best
 φi − φi −1  [20], and Zürcher et al. [21]. Figure 6 shows the useful energy
gain obtained from the CPC collector. From both figures it is
Where φ refers to the dependent variables of pressure and inferred that the value of the convective heat transfer
enthalpy; and φ* represents their values at the previous coefficient in the fluid flow produces small changes in the
iteration. overall heat losses coefficient at the CPC.
The global convergence was reached when between two Figure 7 shows the variation of the saturation length
consecutive loops of the three main subroutines Eq.(37) was (beginning of evaporation) as a function of the mass flow.
verified for all the CVs in the domain.

6 RESULTS AND DISCUSSION

The calculations were made for a CPC without tracking,

with a tilt angle of 15°, and with the characteristics
 was developed. The designed compound parabolic
1600 concentrator for the evaporation of the ammonia-water
HEAT TRANSFER COEFFICIENT / (W / K M )
2 mixture was thermally analysed in order to calculate the
1400
RB
useful energy gain that reaches the receptor tube wall. The
1200 conduction in the internal tube wall was solved using the
algorithm TDMA.
1000
The numerical analysis was made with a control volume
800 method on the receptor tube, and the discretized equations
were coupled using a fully implicit step by step method in the
600 flow direction.
KTF
400 In order to resolve the one-dimensional model, the
thermodynamic properties of the ammonia-water mixture
200 were calculated using matrix functions of the pressure and
0
enthalpy, and some transport properties were estimated as
0 5 10 15 20 25 30 35 40 45 functions of mass concentration and temperature.
LENGTH / (m) Two distinct models were applied to calculate the two-
phase flow convective heat transfer coefficient, considering a
Rivera and Best (RB) Kattan, Thome and Favrat (KTF)
separated model.
Figure 5. Fluid heat transfer coefficient calculated by both The numerical algorithm resolves, for each control volume,
models: Rivera and Best [20], and Zürcher et al. [21], in three subroutines: for the flow inside the receptor tube, for the
March 15 at solar noon.
heat conduction in the receptor tube wall, and for the heat
transfer in the solar concentrator. All the variables at the
1600
RB
outlet section were calculated with the known values at the
1400
inlet section and the boundary conditions. The coupling
KTF
between the three main subroutines was performed iteratively
USEFUL ENERGY GAIN / (W / m )
2

1200 until convergence was reached.

The analysis was for a CPC with input temperature of
1000
60ºC, initial ammonia concentration of 38.7%, and pressure
800 of 1.1MPa.
The results show that the useful energy gain, and therefore
600
the overall heat loss coefficient, has small changes with the
400 two-phase convective heat transfer coefficient used.
>From Figure 7 is deduced that there is a direct correlation
200
between the mass flow and the saturation length. This is due
0 to the heat transfer and residence time of the fluid in the CV,
0 5 10 15 20 25 30 35 40 45 if the mass flow is low, Reynolds number is low and
LENGTH / (m) residence time is high and, if the mass flow is higher,
Rivera and Best (RB) Kattan, Thome and Favrat (KTF)
Reynolds number is higher and residence time is lower.
Figure 6. Useful energy gain obtained with both heat transfer
coefficient models, in March 15 at solar noon. 8 NOMENCLATURE
2
Aa receptor tube heat transfer area [m ]
70 2
Ac cover heat transfer area [m ]
2
Ar reflector heat transfer area [m ]
60 2
At cross section area of fluid flow [m ]
SATURATION LENGTH / (m)

2
50
Ata cross section area of receptor tube [m ]
-1 -1
Cp calorific capacity [J kg K ]
40 D diameter [m]
f friction factor
-1
30 h enthalpy [J kg ]
-1 -1
k thermal conductivity [W K m ]
20 L length [m]
-1
m mass flux [kg s ]
10
p perimeter [m]
P pressure [Pa]
0 -2
0,02 0,03 0,04 0,05 0,06 0,07 0,08 0,09 0,10 0,11 0,12 0,13 0,14 0,15 qu useful energy gain per unit of area [W m ]
-2
MASS FLOW / (kg / s) qwall heat flux per unit of area from fluid to wall [W m ]
-2
S solar absorbed energy per unit of area [W m ]
Figure 7. Length of receptor tube necessary to start the t time [s]
ammonia evaporation as a function of mass flow. T temperature [K]
-1 -2
UL overall heat loss coefficient [W K m ]
7 CONCLUSIONS V volume [m ]
3

-1
Vx velocity in the axial direction [m s ]
A detailed one-dimensional numerical simulation of the
thermal and fluid-dynamic behaviour of two-phase w cover width [m]
(evaporation process at a solar concentrator receptor tube) x axial coordinate
xg mass fraction (vapour quality) 3. R. Best and N. Ortega, Solar Refrigeration and Cooling,
Renewable Energy, vol. 16, pp. 685-690, 1999.
Greek letters 4. F. Ziegler, State of the Art in Sorption Heat Pumping and
α -1
heat transfer coefficient [W K m ]
-2
Cooling Technologies, Int. J. Refrigeration, vol. 25, pp.
δ convergence factor 450-459, 2002.
ε emittance 5. A. Rabl, Active Solar Collectors and Their Applications,
εg void fraction Oxford University Press, 1985.
φ generic dependent variable 6. J.M. Gordon and N.K. Choon, High-Efficiency Solar
Cooling, Solar Energy, vol. 68, pp. 23-31, 2000.
ϕ angle of involute generation [°]
7. P. Schild, Solar Thermal Power: European Perspective,
µ -1 -1
viscosity [kg m s ] Proc. Symp. Int. SolarPACES on Concentrated Solar
θC acceptance half-angle [°] Power and Chemical Energy Technologies, pp. 647-652,
ρ density [kg m ]
-3
2002.
σ Stefan-Boltzman constant [5.6697 × 10 W K m ]
-8 -4 -2
8. K. Sumathy, K.H. Yeung and L. Yong, Technology
υ velocity of the wind [m s ]
-1
Development in the Solar Adsorption Refrigeration

x spatial discretization step [m] Systems, Progress in Energy and Combustion Science,

t temporal discretization step [s] vol. 29, pp. 301-327, 2003.
Φ two-phase frictional multiplier 9. C. Grass, W. Schoelkopf, L. Staudacher and Z. Hacker,
Comparison of the Optics of Non-Tracking and Novel
Types of Tracking Solar Thermal Collectors for Process
Dimensionless numbers Heat Applications up to 300°C, Solar Energy, vol.76, pp.
Pr Prandtl number 207-215, 2004.
Re Reynolds number 10. R. Almanza, A. Lentz, and G. Jiménez, Receiver
Behaviour in Direct Steam Generation with Parabolic
Subscripts Troughs, Solar Energy, vol. 61, pp. 275-278, 1997.
a receptor tube 11. L. García-Rodriguez, A.I. Palmero-Marrero and C.
c cover Gómez-Camacho, Application of Direct Steam
co conductive Generation into a Solar Parabolic Trough Collector to
en environment Multieffect Distillation, Desalination, vol. 125, pp. 139-
ex external 145, 1999.
f fluid 12. S.D. Odeh, G.L. Morrison and M. Behnia, Modelling of
g gas Parabolic Trough Direct Steam Generation Solar
i inlet Collectors, Solar Energy, vol. 62, pp. 395–406, 1998.
in inner 13. N. Ortega and R. Best, Modelling of a Compound
inv involute Parabolic Concentrator with Direct Vapour Generation
l liquid phase for Refrigeration Applications, Proc. Symp. Int.
o outlet SolarPACES on Concentrated Solar Power and
par parabola Chemical Energy Technologies, pp. 621-627, 2002.
r reflector 14. REFPROP version7.0, Reference Fluid Thermodynamic
ra radiative and Transport Properties, NIST Standard Reference
s saturation Database 23 Gaithersbug, MD 20800, USA, 2002.
sk sky 15. A.M. Selim and M.M. Elsayed, Interfacial Mass Transfer
t total and Mass Transfer Coefficient in Aqua Ammonia Packed
tp two-phases Bed Absorber, Int. J. Refrigeration, vol. 22, pp. 263-274,
1999.
Superscripts 16. V. Gnielinski, New Equations for Heat and Mass
− arithmetical average over a CV Transfer in Turbulent Pipe and Channel Flow, Int.
∼ integral average over a CV Chemical Engineering, vol. 16, pp. 359-368, 1976.
value of previous instant 17. S.W. Churchill, Frictional Equation Spans all Fluid Flow
o
Regimes, Chemical Engineering, vol. 84, pp. 91-92,
1977.
9 ACKNOWLEDGES
18. S.G. Kandlikar, Heat transfer characteristics in partial
boiling, fully developed boiling and significant void flow
This work had been financed by DGAPA-UNAM through regions of subcooled flow boiling, Journal of Heat
of PAPIIT Project IN105602-3. The authors thank to Adrián Transfer, vol. 120, pp. 395-399, 1998.
Oskam Voorduin for its collaboration in the CPC design. 19. C.K. Rice, The effect of void fraction correlation and
heat flux assumption on refrigerant charge inventory
10 REFERENCES predictions, ASHRAE Transaction, vol. 93, pp. 341-367,
1987.
1. L.F. Mendes, M. Collares-Pereira and F. Ziegler, Supply 20. W. Rivera and R. Best, Boiling Heat Transfer
of Cooling and Heating with Solar Assisted Absorption Coefficients Inside a Vertical Smooth Tube for
Heat Pumps: An Energetic Approach, Int. J. Water/Ammonia and Ammonia/Lithium Nitrate Mixtures,
Refrigeration, vol. 21, pp. 116-125, 1998. Int. J. Heat and Mass Transfer, vol. 42, pp. 905-
2. A. Syed, G.G. Maidment, J.F. Missenden and R.M. 921,1998.
Tozer, An Efficiency Comparison of Solar Cooling 21. O. Zürcher, J.R. Thome and D. Favrat, Evaporation of
Schemes, ASHRAE Trans., part I, pp. 877-886, 2002. Ammonia in a Smooth Horizontal Tube: Heat Transfer
 Measurements and Predictions, J. Heat Transfer, vol. 23. S.V. Patankar, Numerical Heat Transfer and Fluid Flow,
121, pp. 89-101. Taylor & Francis, London, 1980.
22. F. Friedel, Improved Friction Pressure Drop Correlation 24. J.A. Duffie and W.A. Beckman, Solar Engineering of
for Horizontal and Vertical Two-Phase Pipe Flow, Thermal Processes, John Wiley and Sons Inc., New
European Two-Phase Flow Group Meeting, Ispra, Italy, York, 1991.
Paper E2, 1979.