Dynamic modeling and simulation of a multi-effect distillation
  plant
 
Lidia Roca1 Luis J. Yebra1 Manuel Berenguel2 Alberto de la Calle1
1 CIEMAT-PSA, Ctra. Senés s/n, 04200 Tabernas, Almería, Spain
 
Centro de Investigaciones Energéticas, MedioAmbientales y Tecnológicas
  Plataforma Solar de Almería
2 Dep. Lenguajes y Computación, Universidad de Almería,
  Ctra. Sacramento s/n, 04120 Almería, Spain
{lidia.roca,luis.yebra,alberto.calle}@psa.es,beren@ual.es
 

 
Abstract tems is gaining more acceptance as a result of their
  lower energy requirements, higher heat transfer coef-
This
  paper describes a model which simulates the dy- ficients, compactness,
  high product water quality and
namics of a multi-effect distillation system in different low pre-treatment [2, 7]. In the literature there is a
operating conditions. It has been designed to improve wide variety of steady-sate models for MED plants
the operation of the process and develop a control [3, 5, 6, 9]. One of the last works is the one developed
strategy which optimizes the distillate production. The in [13], which shows a hybrid system that combines a
physical models are based on conservation equations desalination system with solar and wind energies. In
of mass and energy. They also include experimental that paper, the model includes the distillation unit, the
correlations for heat transfer coefficients. Conserva- flat-plat collectors and the wind system. Regarding dy-
tion laws are applied in different components, such as namic models, the literature about multi-effect distil-
the heater, the effects and the preheaters. The results of lation systems is scarce [4, 8].
the mathematical model simulation of the whole pro- The innovation of the present paper is that the dynamic
cess show promising outcomes. model has been developed with the object-oriented
Keywords: solar desalination, multi-effect distilla- Modelica language using the Dymola tool and the
tion, modeling Modelica.Thermal library. This framework has al-
lowed us to develop new libraries to make simulations
easier and improve the operating procedure.
1 Introduction
One of the challenges today is the production of fresh- 2 Description of the system
water for those population areas with high water stress.
For places close to the sea, the desalination process The AQUASOL system (Figure 1) at CIEMAT-
provides an excellent way to tackle this problem. The Plataforma Solar de Almería (PSA), located in the
use of desalination plants in these regions with plen- South of Spain, proposes a solar distillation technol-
tiful seawater resources is becoming a technological ogy that consists of a compound parabolic collector
way to produce freshwater. Since large-scale desalina- (CPC) solar field, two 12m3 water storage tanks, a
tion typically requires large amounts of energy, a so- multi-effect distillation unit with a 3m3 /h nominal dis-
lution is coupling desalination plants with renewable tillate production, and a double effect (LiBr-H2 O) ab-
energies [10]. This process can be performed in var- sorption heat pump (DEAHP) [1].
ious ways, for instance, using solar energy in which The desalination plant at CIEMAT-PSA is a
the source that provides the heat for the desalination forward-feed multi-effect distillation unit manufac-
process is collected in a solar field. tured and delivered by Weir ENTROPIE (Paris,
Multi-effect distillation plants (MED) raise a great France) in 1987. It has 14 cells, or effects, in a vertical
interest in industry due to its efficiency when they are arrangement. The original first cell that worked with
coupled with a solar thermal system. This kind of sys- low-pressure saturated steam (70 °C, 0.31 bar [11])

DOI Proceedings of the 9th International Modelica Conference 883

10.3384/ecp12076883 September 3-5, 2012, Munich, Germany
Dynamic modeling and simulation of a multi-effect distillation plant

  V1
V2 • the preheater-14 is considered as the final con-
  CPC SOLAR
Secondary
Water
 
Primary
Effect1
Seawater
denser.
Water
FIELD Tank Tank
MED
Each component of the plant (the heater, the effects
and the preheaters) has been modeled in the Model-
ica language using Modelica.Thermal library. Figure
Gas Boiler DEAHP Brine Distillate 2 shows the final model of the MED unit, which in-
cludes the heater, 13 effects and 14 preheaters.
Figure 1: AQUASOL diagram The inputs of the model are the feedwater mass flow
rate to the heater, ṁM , the inlet temperature to the
heater, TiM , the salt concentration of the seawater com-
was replaced in the AQUASOL project by a new one, ing into the heater, CB0 , the pressure in each effect, the
which works with hot water coming directly from a seawater mass flow rate, ṁsw , and the inlet seawater
thermal storage tank. For optimal operation, the inlet temperature to the preheaters 14 and 13. The outputs
feed-water temperature in the first cell must be around of the model are the outlet temperature from the heater,
66.5°C. It is possible to reach this temperature with ToM , and the distillate production, ṁd . The correlations
heat from a solar field as well as with steam generated of the heat transfer coefficients included in the models
by an auxiliary gas boiler coupled to a double effect were obtained from experiments carried out in the real
absorption heat pump that can work at variable steam plant [12]. Nomenclature and subscripts are shown in
loads (from 30% to 100%). Tables 1, 2.
Seawater is preheated on its way toward the first
cell of the plant, which is at the top of the desalina- Table 1: Nomenclature
tion tower. Vapour is produced in this first effect (or
heater) using the hot water from the storage system. Name Description Units
This vapour flows to the preheater-1 and part of the A Surface area m2
latent heat is transferred to the seawater that flows in- BPE Boiling Point Elevation K
side this preheater, increasing the temperature of the C Concentration %
seawater. The steam produced in the first effect goes Cp Specific heat capacity J/kgK
to the effect-2, where it is condensed in a tube bun- dT Temperature difference
dle sprayed with the more concentrated brine which between successive effects K
falls by gravity from the previous effect. The latent h Specific entalphy J/kg
heat released by condensation of the vapour allows ṁ Mass flow rate kg/s
part of the seawater entering the second effect to evap- M Mass kg
orate at a lower temperature/pressure. This condensa- P Pressure Pa
tion/evaporation process is repeated in the successive Q Heat transfer rate W
effects. Finally, the vapour produced in the effect-14 T Temperature K
is condensed in a final condenser cooled by seawater. U Overall heat trasfer coefficient W /m2 K
λ Latent heat of vaporization J/kg

3 The dynamic model Table 2: Subscripts

Name Description
The model of the MED unit is based on the following
B Brine
assumptions:
d Distillate
e Effect
• no heat losses with the atmosphere,
h Heater
• no flash vapour is produced, i Inlet
k Effect identification number
• the final distillate production is the sum of the dis- M MED heating water
tillate produced in each effect, o Outlet
p Preheater
• the temperature drop in each effect is equal to the
temperature difference in the preheaters,

884 Proceedings of the 9th International Modelica Conference DOI

September 3-5, 2012, Munich Germany 10.3384/ecp12076883
 Lidia Roca, Luis J. Yebra, Manuel Berenguel and Alberto de La Calle

Ambi1 mMi heater preheater1

mMo
effect2

ambi2
effect3

effect4

effect5

effect6

effect7

effect8

effect9

effect10

effect11

effect12

V3 T3
effect13
T4

effect14
T2
V2

T1

Figure 2: Modelica model of the MED unit

sw Seawater Distillate mass flow produced in the heater can be

v Saturated vapour estimated using the latent heat of vaporization, λ :
Qh
3.1 The heater ṁdh = (3)
λ
The first effect of the MED plant is the heater. Hot Since the vapour pressure of the aqueous solution is
water coming from a storage system enters the heat ex- lower than that of pure water at the same temperature,
changer and produces the first evaporation of the sea- the boiling point of the solution will be higher than
water. Fig.3 shows the model of the heater. that of the water. Therefore, the temperature of the
The heat transfer rate for the first effect can be cal- brine can be obtained using the boiling point elevation,
culated from the MED heating water mass flow rate, BPE:
ṁM , and the MED heating water temperature differ-
ence in stationary conditions as follows: TB1 = Tv1 + BPE (4)
The BPE is a brine property and depends on the
Qh = ṁM ·C p · (TiM − ToM ) (1)
brine salinity and temperature.
Using the log-mean temperature and the overall heat The mass flow rate and concentration of the brine
transfer coefficient, Uh , the heat transfer rate can be can be obtained applying mass and energy balances.
written as: Mass balance:

(TiM − Tv1 ) − (ToM − Tv1 ) d

Qh = Uh · Ah · (2) (MB1 ) = ṁsw − ṁB1 − ṁd1 (5)
dt
ln TToM
iM −Tv1
−Tv1
Salt mass balance:
where Tv1 is the saturation temperature of the vapour d
generated in the heater. (MB1 ·CB1 ) = ṁsw ·CB0 − ṁB1 ·CB1 (6)
dt

DOI Proceedings of the 9th International Modelica Conference 885

10.3384/ecp12076883 September 3-5, 2012, Munich, Germany
Dynamic modeling and simulation of a multi-effect distillation plant

 
Using the overall heat transfer coefficient:
CB0 (Tp(k−1) − Tvk ) − (Tpk − Tvk )
Tsw1 msw
Q pk = U pk · A p · Tp(k−1) −Tvk (9)
ln Tpk −Tvk
TiM mM

P
3.3 The effects
The vapour that has not been condensed in the pre-
Tv1 md1 heater flows to the following effect, where the seawa-
ToM mM
ter with a higher brine concentration flows by gravity
from the previous effect. Then, the vapour condenses
and transfers its latent heat to the seawater producing a
TB1 mB1 new evaporation. Fig.5 shows the model of one effect.
CB1 CB(k-1) TB(k-1) mB(k-1)

Figure 3: Modelica model of the heater dT

Energy balance: Tvk mdk

d P
(MB1 · hB1 ) = ṁsw · hsw − ṁB1 · hB1 − ṁd1 · hv1 (7)
dt

CB(k) TB(k) mB(k)

3.2 The preheaters
The vapour produced in the heater flows to the
preheater-1 located besides, where it condenses as the Figure 5: Modelica model of the effect
temperature of the seawater that flows inside the pre- The heat transfer rate equation for each k-effect
heater tubes increases. This process is repeated in the evaporator is:
successive effects and preheaters. Figure 4 shows the
model of the preheater. Qek = Uek · Ae · (dTk + BPE) (10)
where dT is the temperature difference between suc-
Tp(k-1) msw cessive effects, which is calculated in the prehetear
component.
dT The distillate mass flow rate in the k-effect is:
P ṁdk =
Qek
(11)
Tvk mdk λ
Tp(k) msw The model of each cell or effect is based on mass
and energy balances taking into account the distillate
produced and the brine mass flow rate from the previ-
ous cell:

Figure 4: Modelica model of the preheater d

(MBk ) = ṁB(k−1) − ṁBk − ṁdk (12)
dt
The heat transfer rate for each k-preheater was cal- d
culated using the measured seawater mass flow rate, (MBk ·CBk ) = ṁB(k−1) ·CB(k−1) − ṁBk ·CBk (13)
dt
ṁsw , and the temperature difference between the out-
let and the inlet: d
(MBk · hBk ) = ṁB(k−1) · hB(k−1) − ṁBk · hBk − ṁdk · hvk
dt
Q pk = ṁsw ·C p (Tp(k−1) − Tpk ) (8) (14)

886 Proceedings of the 9th International Modelica Conference DOI

September 3-5, 2012, Munich Germany 10.3384/ecp12076883
 Lidia Roca, Luis J. Yebra, Manuel Berenguel and Alberto de La Calle

4 Simulation results Acknowledgements

The developed model can be used to improve the oper- The authors would like to thank the CIEMAT Research
ation of the plant, studying the effect of the variation in Centre and the Spanish Ministry of Economy and
the operating conditions on the MED unit performance Competitiveness for funding Project DPI2010-21589-
and production rate. C05-02.
The final distillate production will be the sum of the
amounts of vapour produced in each effect as follows:
k=14 References
ṁd = ṁdh + ∑ ṁdk (15)
k=2 [1] D. Alarcón-Padilla, L. García-Rodríguez, and
Figure 6 shows the results obtained simulating the J. Blanco-Gálvez. Assesment of an absorp-
developed model and using the following inputs: tion heat pump coupled to a multi-effect distil-
lation unit within aquasol project. Desalination,
• MED heating water mass flow rate, ṁM , is 12
212:303–310, 2007.
kg/s,
• MED inlet heating water temperature, TiM , varies [2] M.T. Ali, H.E.S. Fath, and P.R. Armstrong. A
between 338 and 345 K (as shown in Fig. 6) comprehensive techno-economical review of in-
direct solar desalination. Renewable and Sustain-
• seawater mass flow rate inside preheaters 1-13, able Energy Reviews, 15(8):4187–4199, 2011.
ṁsw , is 1.94 kg/s,
[3] N.H. Aly and A.K. El-Figi. Thermal perfor-
• preheater-13 inlet seawater temperature, Tp14 is
mance of seawater desalination systems. Desali-
about 303 K (see Fig .6).
nation, 158(1-3):127–142, 2003.
As it can be observed in Fig.6, the MED outlet heat-
ing water temperature, ToM , is about 3.3 K less than [4] N.H. Aly and MA Marwan. Dynamic re-
TiM . Nevertheless, if the inlet temperature increases, sponse of multi-effect evaporators. Desalination,
this difference also increases slightly. As it was ex- 114(2):189–196, 1997.
pected, higher temperatures cause higher thermal con-
sumption. On the other hand, the higher the inlet tem- [5] H. El-Dessouky, I. Alatiqi, S. Bingulac, and
perature, the more distillate is produced. H. Ettouney. Steady-state analysis of the multiple
Therefore, the model may be an efficient tool to es- effect evaporation desalination process. Chemi-
timate the thermal consumption depending on the de- cal engineering & technology, 21(5):437, 1998.
tillate demand. This means that we can predict if the
solar resource is enough to reach the production goals [6] A.M. El-Nashar. Predicting part load perfor-
or if we should combine it with the use of the heat mance of small med evaporators-a simple simu-
pump. lation program and its experimental verification.
Desalination, 130(3):217–234, 2000.

5 Conclusions [7] M.A. Eltawil, Z. Zhengming, and L. Yuan. A

review of renewable energy technologies inte-
In this paper, a multi-effect distillation unit has been grated with desalination systems. Renewable and
modeled. Physical equations for each one of the main Sustainable Energy Reviews, 13(9):2245–2262,
components (the heater, the effect and the preheater) 2009.
have been developed using the object-oriented Model-
ica language. The whole plant has been defined with [8] A. Husain. Integrated Power and Desalination
multiple instances of the effect and preheater subsys- Plants. EOLSS Publishers Ltd., 2003.
tems properly interconnected between them. First sim-
ulation results are promising and the model may be [9] MH Khademi, MR Rahimpour, and A. Jahan-
used to improve the operation in the real plant. The miri. Simulation and optimization of a six-effect
main purpose of the model is the prediction of the ther- evaporator in a desalination process. Chemical
mal dynamics of the heater as well as the prediction of Engineering and Processing: Process Intensifi-
the distillate production rate. cation, 48(1):339–347, 2009.

DOI Proceedings of the 9th International Modelica Conference 887

10.3384/ecp12076883 September 3-5, 2012, Munich, Germany
Dynamic modeling and simulation of a multi-effect distillation plant

305
Tswi13 [K]
304

303

302
0 1000 2000 3000 4000 5000 6000
345
temperatures [K]
Heater

340
TiM ToM
335
0 1000 2000 3000 4000 5000 6000
2.6
md [m3/h]

2.4
·
2.2
0 1000 2000 3000 4000 5000 6000
Relative time [s]

Figure 6: Simulation results of the MED unit model

[10] E. Mathioulakis, V. Belessiotis, and E. Delyan-

nis. Desalination by using alternative en-
ergy: Review and state-of-the-art. Desalination,
203:346–365, 2006.

[11] B. Milow and E. Zarza. Advanced med solar

desalination plants. configurations, costs, future–
seven years of experience at the plataforma solar
de almeria (spain). Desalination, 108(1-3):51–
58, 1997.

[12] P. Palenzuela, D. Alarcon, J. Blanco, E. Guillen,

M. Ibarra, and G. Zaragoza. Modeling of the heat
transfer of a solar multi-effect distillation plant
at the plataforma solar de almeria. Desalination
and water treatment, 31(1-3):257–268, 2011.

[13] İ.H. Yılmaz and M.S. Söylemez. Design and

computer simulation on multi-effect evaporation
seawater desalination system using hybrid re-
newable energy sources in turkey. Desalination,
2012.

888 Proceedings of the 9th International Modelica Conference DOI

September 3-5, 2012, Munich Germany 10.3384/ecp12076883