You are on page 1of 20

Desalination 435 (2018) 3–22

Contents lists available at ScienceDirect

Desalination
journal homepage: www.elsevier.com/locate/desal

Transient analysis of steam accumulator integrated with solar based T


MED-TVC system

V. Samson Packiaraj Raphaela, , R. Velrajb, Purnima Jalihala
a
Energy and Fresh water Group, National Institute of Ocean Technology, Chennai, India
b
Institute for Energy Studies, Anna University, Chennai, India

A B S T R A C T

In a solar field, the steam production rate and pressure is not constant, but varies and becomes maximum at solar
noon. However solar thermal applications such as multi effect desalination with thermo vapour compressor
(MED-TVC) system are usually designed for a particular pressure and flow rate. Hence a buffer storage system is
needed in solar MED-TVC system to adjust the differences between the steam production and the consumption
rates. In the present work steam accumulators were used as buffer storage in direct steam generation (DSG) for
correcting the mismatch between production and consumption and also for energy storage. A parametric ana-
lysis was performed to investigate the impact of steam accumulator volume and also the irradiance disturbance
in solar field on MED-TVC system's performance. This paper presents the transient simulation results of the steam
accumulator's output and the corresponding output of the MED-TVC yield from 9:00 h to 20:00 h in a day.

1. Introduction (LFR) are examples for line concentrators. Different types of heat
transfer fluids (HTF) are used for transferring this thermal energy from
In most of the tropical coastal regions, there is a scarcity of fresh the concentrators to the solar power plant or process heat applications.
water due to increase in population, industrialization, urbanization and One option is the utilization of water/steam in the concentrators as
increased life standards of people. Naturally available fresh water HTF, the so-called direct steam generation (DSG). Michael Berger [5]
sources are not capable of meeting fresh water demands. Desalination is et al. have presented the first year's operational experience of a LFR
one of the solutions to meet the water demand. The most common based DSG solar thermal plant which supplied steam to a pharmaceu-
desalination plants in operation follow the principle of reverse osmosis tical company in Jordan. A non-linear simulation model was developed
(RO), multi stage flash evaporation (MSF) and multi effect desalination in the modelica language by Markus Eck et al. [6] for a study of the
with thermal vapour compression (MED-TVC). Conventional desalina- interaction between the collector loop and the control system in a
tion plants are operated by thermal energy and electricity generated by parabolic trough based DSG plant. The simulation result showed a
fossil fuel. Hence desalination has shown increased threats of CO2 significant reduction in the temperature deviations in a recirculation
emissions and a severe impact on environment. Desalination can be mode type DSG plant compared to the simple proportional–integral
made sustainable if it is coupled with renewable energy or utilization of controller type DSG plant caused by the feed forward control. In some
waste heat available in coastal regions. The use of thermocline from the solar thermal power plants synthetic oil is used as HTF in the con-
sea in a single stage flash evaporative desalination process called LTTD centrators and molten salt is used as thermal storage system (TES) when
[1–2], in MED process called ST-MED [3], coastal power plant's con- excess energy is available. This solar field is connected to the steam
denser reject heat in single stage flash evaporative desalination process turbine circuit via a heat exchanger. Synthetic oil based plant is simple
[4] and the use of solar energy in MED process show great potential in from the control point of view and does not involve two phase flow
the coastal regions. complications in solar field unlike in DSG. Jan Fabian Feldhoff et al. [7]
In the solar field, solar energy is converted into thermal energy in compared the efficiency and levelized electricity costs (LEC) of two
solar thermal concentrating systems. These systems are classified by promising solar power plant concepts with different HTF and TES. In
their focus geometry as point concentrators or line concentrators. A both concepts, 100 MWel gross turbine with a 9 h storage capacity was
Central receiver and a parabolic dish are examples for point con- considered as a bench mark. In one concept, DSG in parabolic trough
centrators. Parabolic trough collector (PTC) and Linear Fresnel reflector collector was considered with two parts TES, a sensible part of three


Corresponding author.
E-mail address: samson@niot.res.in (V. Samson Packiaraj Raphael).

https://doi.org/10.1016/j.desal.2017.12.045
Received 28 August 2017; Received in revised form 21 December 2017; Accepted 23 December 2017
Available online 17 January 2018
0011-9164/ © 2017 Elsevier B.V. All rights reserved.
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Nomenclature τ Phase change relaxation time, s


Γ Mass flow rate of seawater on one side of the tube,
A Heat transfer area kg.m− 1.s− 1
a Steam-water interface concentration, m2.m− 3 μ Dynamic viscosity, N.s.m− 2
H Total enthalpy, kJ α Thermal diffusivity, m2.s− 1
h Specific enthalpy, kJ.kg− 1
ha Heat transfer coefficient, W.m− 2.K− 1 Subscripts and superscripts
M Mass, kg
ṁ Mass flow rate, kg.s− 1 B Boundary parameter
P Pressure, bar c Condensation
Q̇ Heat transfer rate, W e Evaporation
r Latent heat of evaporation/condensation, kJ.kg− 1 in Inlet
T Temperature, °C out Outlet
t Time, s 1 Water
V Volume, m3 2, S Steam
v Specific volume, m3.kg− 1 21 Interfacial transfer from steam to water
n Total number of effects ‘ Saturated water
ΔT Temperature difference, °C "
Saturated steam
F Feed seawater flow rate, kg.s− 1 1 First effect
B Brine flow rate, kg.s− 1 n last effect
D Distillate(vapour) flow rate, kg.s− 1 1a Condensation component in liquid form inside the tube
L Latent heat, kJ.kg− 1 1b Condensation component in vapour form inside the tube
X Salinity, PPM 2a Evaporation component in liquid form outside the tube
R Resistance 2b Evaporation component in vapour form outside the tube
hin Inside heat transfer coefficient, W.m− 2.K− 1 ou, o outer
hou Outside heat transfer coefficient, W.m− 2.K− 1 in, i inner
g Acceleration due to gravity, m.s− 2 i ith effect
k Thermal conductivity, W.m− 1.k− 1 n nth effect
ν Kinematic viscosity, m− 2.s− 1 1 1st effect
q Heat transfer rate, W tot Total
d Tube diameter, m r entrained vapour
h fg Latent heat of vaporization, kJ.kg− 1 b brine
C Specific heat, kJ.kg− 1.k− 1 d Distillate, discharged steam from TVC
CR Compression ratio fw feed water
ER Expansion ratio fi Inside fouling
fo Outside fouling
Greek symbols wall Tube wall

ρ Density, kg/m3

molten salt tanks and a latent part with phase change material (PCM). He has also presented the climatic and sea water data for the plant site.
The latent PCM storage used sodium nitrate (NaNO3) as material, In this plant 1064 evacuated tube collector panels assembled in 76 rows
showing a melting temperature of 306 °C. In the other concept, the of 14 panels in each row were used. The absorber plate area of each
synthetic oil was heated in a parabolic trough collector from 295 °C to panel was 1.75 m2, giving a total collector area of 1862 m2. In this
about 393 °C and two tank molten salt storage systems with one hot and plant, water was used as the heat transfer fluid in the solar field and
one cold tank were considered as TES. The efficiency of the DSG plant also heat storage medium. Three vertical cylindrical tanks having total
was reported as about 8% better than synthetic oil plant, whereas its capacity of 300 m3 were used for hot water storage for the supply of
project investment was about 10% higher. This caused higher levelized heat about 16 h after sunset. The sensible heat of this hot water was
electricity costs of the DSG concept solar power plant by about 6%. used in the first effect of the MED evaporator. Variations were reported
When storage capacity was lower than 3 h, DSG with a steam accu- in the performance ratio of the plant. These variations were between
mulator storage concept's LEC was less compared to synthetic oil HTF 12.5 and 14.0 all through for most of the days during the year 1985
with two-tank molten salt storage concept and molten salt HTF with the except during the month of August. The overall heat transfer coefficient
same molten salt storage concept [8]. Hence DSG with steam accu- for condenser and pre heater was reported to be in the range of 1600 to
mulator as buffer storage concept is found to be the better option for 1700 kcal/m2h °C (1860 to 1977 W/m2 K) and for the evaporator, it
solar MED-TVC process, as it avoids the potential risk associated with was 2600 to 3000 kcal/m2 h °C (3024 to 3489 W/m2 K). The conclusion
synthetic oil usage in drinking water plant. Also oil water heat ex- was that only a fraction of the solar radiation was converted into useful
changer is not necessary in the DSG plant and expensive oil can be energy for the production of distillate while another part was lost to the
replaced by simple water. environment by heat convection and radiation. Heat losses in summer
Solar energy based MED process has been studied extensively in months were seen as lower than in winter months due to higher am-
medium size plants. Nashar [9] has presented the performance para- bient temperatures in the summer than in the winter.
meters observed at the 18 effect 120 m3/day capacity solar MED plant Bernhard Milow et al. [10] has presented his seven years opera-
located in Abu Dhabi, U.A.E during the first year (1985) of operation. tional experience of the 72 m3/day capacity 14 effect solar MED plant

4
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

installed in 1988 at the Plataforma Solar de Almeria, Spain. The plant simultaneous decline in the energy quality for both stages that resulted
consisted of 2672 m2 single axis tracking parabolic trough collectors in cascade utilization of thermal energy. Buschle et al. [18] have ush-
(PTC), which delivered 6.5 mWht thermal energy daily. In this plant, ered in the concept of Phase Change Material (PCM) enhanced varying
synthetic oil was used as a heat transfer fluid in the solar field and also pressure steam accumulator. In this concept, a pump ensures the cir-
heat storage medium. A 115 m3 thermocline vessel was used as the culation of water between a tube register with externally arranged PCM
thermal storage system, which stored 5 MWht at charge/discharge to the steam accumulator. The same concept was analyzed using
temperatures of 300 °C/225 °C. Vapour at 70 °C produced in a low Modelica and the results were presented. Baldini et al. [19] have pro-
pressure boiler was supplied to the first effect. The project was im- posed a thermodynamic model for solar parabolic trough based steam
plemented in two phases. In first phase reliability and technical feasi- generator with steam accumulator for sizing the solar field and accu-
bility of coupling solar thermal energy to MED process were demon- mulator. Exergy efficiency of the system was analyzed using this model.
strated. In the second phase, a double-effect absorption heat pump and The conclusion was that the largest exergy destruction took place in the
a steam ejector based vacuum system were introduced in the system. solar collector due to temperature rise of water in the solar field, while
Implementation of absorption heat pump and steam jet ejector based the contribution of the steam accumulator was relatively small. Baozhi
vacuum system was reported as reducing the thermal energy con- Sun et al. [20] validated the non-equilibrium thermodynamic model
sumption by 44% and electricity consumption by 12%. The thermo- proposed by Stevanovic et al. [16] through testing with a lab-based
economic analysis of this plant was presented by Gracia-Rodriguez et al. steam accumulator system for accurate prediction of the dynamic
[11]. characteristics of steam catapult's steam accumulator and proposed a
A 6 m3/h capacity 6 effect solar MED-TVC plant was commissioned revised relaxation time for evaporation and condensation.
at Ramanathapuram, India at 2013 [12–13]. Water was used as HTF in The reliability of the MED plant performance depends on the
solar field in this plant. Linear Fresnel reflector of area 1404 m2 con- mathematical models considered for arriving at the MED process
centrating the insolation 50 times was used for converting the water parameters and correlations used for arriving at the Heat Transfer
into steam. A 12 m3 volume steam accumulator was used as buffer Coefficient (HTC) while designing the plant. Hence consideration of
storage to store the steam. This steam accumulator corrected the dif- suitable mathematical model and use of appropriate HTC are essential
ferences between steam production and consumption rates and deliv- when designing a new plant. El-Dessouky et al. [21] have proposed
ered steam at constant pressure of 5 bar and a constant flow rate of extensive mathematical models for multi effect desalination with
635 kg/h to the process, out of which 35 kg/h was used in steam jet thermo vapour compressor process in steady state conditions to de-
ejector based vacuum system and 600kg/h was fed to the TVC through termine the effect of important design and operating variables. These
control valves. The plant produced fresh water at quality < 5 ppm. mathematical models were solved by a modified fixed point iteration
Direct storage of high pressure steam in a pressure vessel is not algorithm. The algorithm consists of 10 calculation blocks and 6 logical
economical due to low volumetric energy density whereas, saturated or blocks was implemented using the L-A-S computer aided language. The
superheated steam is stored in the steam accumulator as sensible heat conclusion was that the plant thermal performance ratio was in-
in pressurized saturated liquid water (up to 1.2 kWh/m3) due to its high dependent of the top brine temperature and strongly depended on the
heat capacity also act as a buffer storage for meeting the difference number of effects, whereas the specific heat transfer area increased by
between production and consumption [14]. In this accumulator, char- raising the number of effects and reducing the top brine temperature.
ging takes place when saturated or superheated steam which is gener- Bin Amer [22] has developed a steady state mathematical model for
ated in the solar field enters the accumulator, which contains relatively parallel feed MED-TVC process using an Engineering Equation Solver.
low pressure saturated steam and water initially. There is an increase in The model was validated with three commercial MED-TVC units. He
the temperature and pressure of the water in the accumulator when the has also proposed two approaches to find the optimum operating and
accumulator is getting charged with superheated steam. This is due to design conditions for MED-TVC unit: (1) Smart Exhaustive Search
the condensation of the superheated steam. But the variation in the Method (SESM) and (2) sequential Quadratic Programming (SQP). The
liquid mass is very little. During the discharging process pressure conclusion was that the results obtained by SQP method were better
sliding takes place in the accumulator. This results in flash evaporation than those obtained by SESM and the corresponding total execution
of high pressure water and generation of saturated steam in the accu- time was also less. The results led to the inference of variations between
mulator. The volume specific thermal energy density depends on the 8.5 and 18.5 for 4 and 12 effects for the maximum gain ratio with an
saturation temperature during pressure sliding in discharging process optimal top brain temperature ranging from 55.8 to 67. 5 °C. Mabrouk
and its values are in the range of 20–30 kWh/m3 [15]. Since water is et al. [23] have proposed a mathematical model for predicting the
used both as working fluid and storage in direct steam generation based CaCO3 formation and CO2 release in a triangular pitch tube bundle
solar-MED process, steam accumulator is found to be well suited for arrangement. The conclusion was that increasing the feed seawater to
meeting the storage requirement in this process. Steinmann et al. [15] twice that of the original flow caused reduction in the scale thickness by
have proposed a physical model of a sliding pressure steam accumulator 15% whereas the scale deposite thickness at a tube pitch of 2 ∗ d was
for compensation of transients in insolation for a solar thermal system 30% lower than that at tube pitch of 1.3 ∗ d.
using steam as working medium. Stevanovic et al. [16] have developed Kamali et al. [24] have developed mathematical models for predicting
a non-equilibrium numerical model for steam accumulator based on the design data for plate type evaporator and shell and tube type eva-
mass and energy balance equations for predicting steam accumulator porator of 1200 m3/day capacity MED-TVC unit and compared the eva-
capacity. Using this model the pressure transients of the steam accu- porator results. They have also presented the influence of design and
mulator in the coal drying plant during charging and discharging of the operating parameters like Number of effects, first effect steam tempera-
steam were predicted and the results were compared with results seen ture, salt concentration ratio on performance ratio of the system. Ibrahim
in the equilibrium model. A numerical model for the two stage thermal et al. [25] have proposed a steady state mathematical model and analyzed
storage system was developed by Fengwu Bai et al. [17]. In this model the influence of important design and operating variables on the perfor-
concrete was considered in the high temperature heat storage stage and mance of the plant. The conclusion was a variation in the gain output
steam accumulator was considered in the low temperature stage. The ratio between 7.34 and 15.04 for 4 effects and 12 effects. A dynamic
exiting temperature of steam in both the steam accumulator and the mathematical model for MED-TVC process was proposed by Mohammad
concrete storage was seen decreasing with discharging time for such a et al. [26]. The transient and steady state behavior of this model was
two stage thermal storage system. This was the result of this investigated by applying disturbance and also changing the operating

5
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

point which happened in a real condition. Muhammad et al. [27] have pressure ranged from 0.93 to 3.60 kPa and sea water ranged from 15,000
proposed a mathematical model for hybrid system consisted of multi ef- to 90,000 mg/l based on the experimental results [28]. In the proposed
fect desalination with Adsorption Desalination (AD). In this system the correlation, a new term was introduced to capture bubble agitation effects
last effect was connected to the adsorbent bed where silica gel was packed as compared to the conventional falling film evaporation correlations.
in the form of cakes. This combination of MED and AD brought down the The conclusion was that at low saturation temperatures, the bubble agi-
temperature of last effect below ambient due to the high affinity of va- tation phenomena boosted the heat transfer coefficients by as much as
pour and helped insertion of more number of effects and hence more two to three fold as compared to the conventional Han & Fletcher cor-
recoveries and also reduce the corrosion chances. This hybridization of relation. The effect of feed flow rate in the evaporator, tube diameter,
MED with AD was also reported to increase the production more than bundle effect, wall temperature etc. were studied experimentally by Raju
threefold with the same top brine temperature as compared with tradi- Abraham et al. [29] and this results were validated with Computational
tional MED plants. The same author has proposed a correlation for hor- Fluid Dynamics (CFD) results and also compared with published data
izontal tube falling film evaporation for sub-atmospheric conditions, available in the literature. Heat transfer coefficient was reported as the

Fig. 1. Schematic of solar MED-TVC process.

6
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

highest for the top tube in the bundle and decreased proportionately to evaporated vapour of the first effect is supplied to the next effect and
the position of the tube in the bundle and it stabilized within the first 10 serves as the heating steam for the second effect and the same process is
rows. The heat transfer performance increased with feed rate, but de- repeated in successive effects due to slightly lower pressure maintained in
creased with tube diameter. The experimental and CFD results suggested the effects than the previous effect. The remaining part of the vapour from
that In-line tube configuration as found to be better compared to stag- the last effect, after a portion is entrained in the thermo vapour com-
gered tube configuration. pressor, condenses in the condenser. A small fraction of the generated
Many researchers have published mathematical models for multi ef- vapour is used for preheating the feed seawater in pre-heaters. A total of
fect desalination processes for simulation and economic evaluation pur- six effects are considered in the evaporator and except in the sixth effect,
poses. Operational data of several plants are available in literature for the in each effect one pre heater is employed.
MED-TVC process. More experimental results are available for predicting
heat transfer in horizontal tube falling film evaporation. However there is 3. Discretization and simulation analysis
no literature which focusing the influence of steam accumulator in solar
MED-TVC process. In the present study, the effects of two different types Three sets of governing equations have been considered in this
of steam accumulators, namely varying pressure steam accumulator and study. One set of governing eqs. (A1 to A6 in Appendix A) proposed by
constant pressure steam accumulator on solar MED-TVC process were Stevanovic et al. [16] have been utilised for steam accumulator ana-
investigated. Computational codes have been written in MATLAB for this lysis. A second set of governing eqs. (B1 to B23 in Appendix B) derived
purpose and the simulation results discussed. by basic heat and mass balance have been considered for calculating the
fresh water generation rate in MED-TVC process. A third set of gov-
2. Process description erning eqs. (C1 to C7 in Appendix C) have been utilised for predicting
heat transfer area in the MED evaporator. The discretization method
A schematic diagram of a solar MED-TVC desalination process is used was the finite difference method, and the most common finite
presented in Fig. 1. The major components involved in the process are difference representation of derivatives is based on Taylor's series of
solar field, steam separator, steam accumulator, thermo vapour com- expansion. In this study, partial derivatives in the governing equations
pressor, horizontal falling film evaporator (effects), pumps, condenser and proposed by Stevanovic et al. [16] for steam accumulator have been
feed pre heaters. In this study, the solar field parameters in Ramanatha- replaced by algebraic difference quotient; thus the original partial dif-
puram plant have been considered [12]. LFR type solar thermal con- ferential equations were discretized. The previous time step values were
centrating system is available in the solar field. The LFR is in two modules used in this simulation study for the calculation of the current time step
are connected in parallel. Each module having a collector area of 702 m2 as per finite backward difference method. 39,600 time steps have been
and hence total collector area in the solar field is 1404 m2. In each used to simulate the actual duration of 11 h between 9:00 h and
module, an array of eight linear mirrors of each 1.8 m width and 48.75 m 20:00 h. Computational codes were written in the MATLAB for this si-
long has been arranged in the tracking system. These mirrors focus sun mulation. It is the language of technical computing which is widely
light on a stationary linear receiver. In a linear receiver, eight absorber used by scientists and engineers. In this simulation study an open
tubes which are connected in series, made of SS 304 having dimensions of source code known as Xsteam was used for obtaining water and steam
outer diameter 33.4 mm and inner diameter 26.7 mm are housed in a thermodynamic properties at different temperatures, pressures and
trapezoidal cavity. This trapezoidal cavity has glass cover at the bottom enthalpies. In this open source code, steam and water properties are
and insulation at the other three sides to reduce heat loss. The reflected based on the International Association for Properties of Water and
sun rays pass through the glass cover and are incident on the absorber Steam Industrial Formulation 1997(IAPWS IF-97). It provides accurate
tubes. Water, while passing through the absorber tubes, gets heated and data for water and steam properties from 0 to 1000 bar and from 0 to
subsequently becomes steam. As the insolation varies in the solar field, 2000 °C. In this simulation, the temperature, enthalpy, and mass of
the dryness fraction of the steam also varies in the solar field. Hence solar water and steam in the accumulator in each time step and pressure of
field is supplied with surplus of feed water by Variable Frequency Drive the accumulator in each time step have been found and the distillate
(VFD) in the feed water pump for avoiding dry-out and overheating of the output was calculated on the basis of the governing equations men-
absorber tubes and to enable achievement of around 0.5–0.6 dryness tioned for MED-TVC process. When pressure and its corresponding
fraction of steam in the absorber tubes. This two phase mix generated in enthalpy of steam accumulator were used to calculate the transient
the solar field is fed into the steam separator from side, where the baffle value of the water specific volume, the results will not be correct [20].
plates reduce the momentum of the water particles available in the two Hence only the pressure of the steam accumulator was used to calculate
phase mix forcing them to settle down at the bottom of the steam se- the water specific volume in this simulation. The flow chart for the
parator thereby removing the water particles. This saturated water col- MATLAB code is presented in the Appendix D.
lected at the steam separator is re-circulated along with the treated make- In this study, three different groups of simulations have been per-
up water from MED output to solar field. Dry steam from separator is fed formed. In the first group of simulations, transient behavior of steam
into the steam accumulator. Steam accumulator supplies the constant accumulator and the corresponding MED-TVC system yield have been
pressure and constant mass flow rate motive steam to the thermo vapour analyzed. In the second group, the influence of steam accumulator
compressor through pressure control cum flow control valve. A part of volume on the varying pressure type and on the constant pressure type
vapour from the last effect is also entrained into the thermo vapour steam accumulator has been analyzed. In the varying pressure type
compressor. Hence both motive steam and entrained vapour are fed accumulator, no water was fed or removed from the accumulator
through de-super heater into the tubes of the first effect at discharge (ṁ 1in = ṁ 1out = 0), hence there was very little variation in this liquid
pressure, where they condense and supply its latent heat to the feed sea mass. Charging and discharging of steam took place; hence there were
water which is sprayed over the tubes as falling film. This latent heat of change in pressure and temperature in this type accumulator. In con-
condensation increases the temperature of the falling film water to top stant pressure type steam accumulator, charging and discharging of
brine temperature. Evaporation takes place in the shell side of the effect steam took place and also water was fed into the accumulator when
and a fraction of the feed sea water is converted into vapour and the pressure of the accumulator exceeded the set pressure of the accumu-
remaining liquid becomes brain. This is the result of low pressure which is lator, hence pressure was almost constant during operation. In this
less than the saturation pressure of falling film water maintained in the study, the maximum set pressure of the constant pressure type accu-
shell side of the effect by the steam jet ejector based vacuum system. The mulator was considered as 8 bar and when the pressure of the

7
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

accumulator exceeded 8 bar, 3 kg water at 28 °C was considered as fed initial pressure in the accumulator which was 2.182 bar in the simula-
into the accumulator. In the third group of simulation, the influence of tions. The initial water and steam mass in the accumulator were calcu-
cloud cover above the solar field (irradiance disturbance) on both type lated by the initial liquid water level, which specified in each simulation
of steam accumulator was analyzed. The results of the simulations have in terms of percentage of total volume of the accumulator. In all the si-
been discussed in three different sections. mulations performed, pressure of the steam (Psf) and mass of the steam
All simulations were performed with the same values as the empirical (Msf) generated in solar field of RAMDAS plant [12] were considered as
parameters in the model considered by Stevanovic et al. [16]: the con- input parameters. Three conditions were considered in all simulations.
densation and also evaporation relaxation time was (τ) = 85 s and the The first condition is, if solar field pressure is greater than the accumu-
product of the heat transfer coefficient from steam to liquid water and the lator pressure, steam is fed into the steam accumulator from the solar field
steam-water interface area concentration was (ha)21 = 50,000 W/m3-K. or no steam is fed into the steam accumulator, In practice this first con-
The assumption in all performed simulations was that in the initial state, dition appears valid. The second condition is, if pressure in the steam
liquid water and steam were in thermal equilibrium determined by the accumulator is > 5 bar, steam discharged from the accumulator is

Fig. 2. Transient simulation results of 12 m3 volume steam ac-


cumulator.
Solar Insolaon Steam generaon rate in solar field a
Solar Insolaon W/m2

Steam generaon rate kg/s


1000 1
800 0.8
600 0.6
400 0.4
200 0.2
0 0
8 10 12 14 16 18 20
Time, h

Steam accumulator pressure Solar Field Pressure b


61
51
41
Pressure, bar

31
21
11
1
8 10 12 14 16 18 20
Time, h

Steam charged into the steam accumulator


Steam discharged from steam accumulator to MED c
0.25
Rate of steam mass, kg/s

0.2
0.15
0.1
0.05
0
8 10 12 14 16 18 20
Time, h

Mass of water in accumlator Mass of steam in accumlator d


7000 70
Mass of water, kg

Mass of steam, kg

6000 60
5000 50
4000 40
3000 30
2000 20
1000 10
0 0
8 10 12 14 16 18 20
Time, h

8
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

0.176 kg/s (635 kg/h), else 0 kg/s, since 5 bar pressure is the process to decrease in solar insolation. As the accumulator was getting charged
requirement as per Shinu M. Varghese et al. [12]. The third condition is, if with this super heated steam from solar field, the temperature and
the volume of water in the accumulator is > 10% of the total volume of pressure of the water and steam increased in the accumulator due to
the accumulator, steam discharged from the accumulator is 0.176 kg, else condensation of superheated steam. Initially the pressure in the accu-
0 kg, as it is always safe to maintain water level above 10% of accumu- mulator increased quickly up to 5 bar as the discharge valve was in a
lator total volume in the accumulator to ensure submergence of steam closed condition and set pressure to open the discharge valve was only
charging nozzle in water. 5 bar. After attaining 5 bar, both charging of steam from solar field to
accumulator and discharging of steam from accumulator to process at
the rate of 0.176 kg/s took place simultaneously, hence pressure raise
4. Result and discussion was gradual and the maximum pressure attained by the accumulator
was 9.65 bar. After that, as the solar field pressure was less than the
4.1. Transient behavior of steam accumulator on solar MED-TVC process accumulator pressure, charging stopped and only discharging of steam
took place. During discharging process pressure sliding took place in
In this study, the mass and pressure of steam generated in the solar the accumulator. This resulted in flash evaporation of high pressure
field at experimental plant [12] were considered as input parameters. water and generation of saturated steam in the accumulator and re-
The mass of steam generated in the solar field and the solar insolation at duction of water mass. Hence again a slight increase in pressure took
that duration is presented in Fig. 2a. In this simulation, the volume of place in the accumulator. When the volume of water was < 10% of the
the accumulator was considered as 12 m3 and 50% of the accumulator volume of the accumulator, discharge of steam also stopped as it was
volume was considered as filled with water initially. No water was fed the set volume to ensure submergence of steam charging nozzle in
into or removed from the accumulator, hence ṁ 1in = ṁ 1out = 0. The water. The total duration of discharging in the present simulation was
simulations results are discussed in this section. 27,046 s. The rate of steam mass charged into the accumulator from
A comparison of the solar field pressure and its influence on steam solar field and rate of steam mass discharged from the accumulator for
accumulator's pressure is shown in Fig. 2b. The pressure increased process are presented in Fig. 2c.
steadily in solar field from 2.18 bar to 49.31 bar, due to increase in The mass of water and mass of steam in the steam accumulator are
solar insolation. Then the solar field pressure decreased gradually due

Fig. 3. Temperature and Thermal energy in 12 m3 volume steam


Temperature of water in steam accumulator

200
Temperature of steam in steam accumulator a accumulator.
Temperature, OC

180
160
140
120
100
8 10 12 14 16 18 20
Time, h
Instantaneous energy charging into the accumulator
Instantaneous energy discharging from the accumulator b
700
600
500
400
Energy, kJ

300
200
100
0
8 10 12 14 16 18 20
Time, h

Termal energy stored in terms of hot water


c
Thermal energy stored in terms of Steam
8.00E+06
Total thermal energy stored in accumulator
6.00E+06
Energy, kJ

4.00E+06

2.00E+06

0.00E+00
8 10 12 14 16 18 20
Time, h

9
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Table 1 charging and discharging were less. There was no rapid charge or dis-
Input data. charge of steam from the accumulator. Hence temperature difference
between water and steam was only negligible, though non-equilibrium
Parameter Unit Value
model was considered in this analysis. The instantaneous energy
Number of effects 6 charged into the steam accumulator and discharged from the steam
Pressure of motive steam bar 5 accumulator is presented in Fig. 3b. The energy stored in the form of
Flow rate of motive steam kg/s 0.167
o
pressurized hot water, super heated steam and the total energy avail-
Temperature of feed seawater C 28
Flow rate of feed seawater (‘Ftot’) kg/s 4.7267 able in the steam accumulator are presented in Fig. 3c.
Salinity of feed seawater ppm 35,000 During this simulation, the duration of the discharge of steam from
Temperature raise considered in effects °C 5 the accumulator was 27,046 s and the amount of steam discharged from
Saturation temperature of first effect °C 63 the accumulator was 0.176 kg/s. Out of this, 0.167 kg/s was supplied
Saturation temperature of last effect °C 50
for the operation of TVC and remaining was considered for steam jet
Length of the tubes m 1.6
Outside diameter of the tubes mm 25.4 ejector based vacuum system operation. The input parameters con-
Liquid load (Γ) Kg/m-s 0.01 sidered for obtaining the fresh water yield of the MED-TVC process are
presented in Table 1.The process simulations of MED was carried out
and the results are shown in Tables 2 and 3. In this study the first effect
saturation temperature have been considered as 63 °C and the sixth
Table 2 effect saturation temperature as 50 °C. Hence the temperature differ-
TVC parameters. ence between the effects was 2.6 °C. As parallel feed arrangement was
considered in the evaporator, in each effect 0.79 kg/s feed sea water at
Parameter Unit Value
35000 PPM salinity was considered to be sprayed over the tubes. The
Pressure of motive steam bar 5 corresponding Reynolds number of the feed sea water was 21.96, which
Flow rate of motive steam Kg/s 0.167 was well above the minimum Reynolds number(Remin = q/48) re-
Vapour pressure in 6th effect bar 0.124
commended by Fujita et al. [34] to be maintained for avoiding dry-out.
Entrainment ratio 0.975
Rate entrainment from 6th effect Kg/s 0.171 The temperature raise of feed seawater in each pre heaters was also
Discharge pressure of TVC bar 0.245 considered as 2.6 °C. The total temperature raise of feed sea water in
Discharge flow rate of TVC Kg/s 0.338 final condenser and steam jet ejector's condenser was considered as
17 °C. Hence the feed seawater temperature in the effects were 58 °C,
55.4 °C, 52.8 °C, 50.2 °C, 47.6 °C and 45 °C from first to sixth effects
respectively. As the entrainment ratio of the TVC was 0.975 [12], the
presented in Fig. 2d. During this process, despite the condensation of amount of steam supplied to the first effect was 0.338 kg/s as shown in
superheated steam taking place, variation in liquid mass in the accu- Table 2. The governing Eqs. (B1) to (B23) in Appendix B have been
mulator was less due to discharging also took place simultaneously. considered for arriving at the fresh water yield and the corresponding
When charging of steam stopped and only discharge took place, mass of fresh water generation rate from first effect to sixth effect obtained in
the water decreased gradually and mass of steam increased due to flash this study were 0.33 kg/s, 0.32 kg/s, 0.30 kg/s, 0.28 kg/s, 0.26 kg/s
evaporation of high pressure water. When both charge and discharge of and 0.24 kg/s respectively as shown in Table 3. The average recovery
steam stopped, a small fraction of water was flash evaporating up to the ratio of the first effect was 42% and it was only 29% at the sixth effect.
end of simulation due to slight non-equilibrium between water and This decrease in recovery ratio was due to a reduction in the heat flux
steam in accumulator. This non-equilibrium process occurred mainly induced in the tubes from 1st effect to 6th effect while the overall re-
because the water was still in a superheated state and also due to covery ratio was 36.6%.
thermal inertia of water. The governing Eqs. (C1) to (C7) in Appendix C have been con-
The temperatures of water and of steam in the steam accumulator sidered for arriving at the overall heat transfer coefficient of the MED
are presented in Fig. 3a. During this process, the variation in rate of evaporator and the corresponding overall heat transfer coefficient from

Table 3
MED-TVC process parameters.

Parameters Unit Effect 1 Effect 2 Effect 3 Effect 4 Effect 5 Effect 6

Saturation pressure of effects bar 0.229 0.203 0.180 0.159 0.140 0.124
Saturation temperature of effects °C 63.0 60.4 57.8 55.2 42.6 50.0
Temperature of feed sea water (Sprayed over the tubes) °C 58.0 55.0 53.0 50.0 48.0 45.0
Flow rate of feed sea water kg/s 0.79 0.79 0.79 0.79 0.79 0.79
Reynolds number of feed sea water 21.96 21.96 21.96 21.96 21.96 21.96
Temperature of vapour (inside the evaporator tubes) °C 65.2 62.6 60.0 57.4 54.8 52.2
Velocity of vapour inside the tubes m/s 0.0012 0.0013 0.0015 0.0016 0.0018 0.002
Heat load of evaporator kW 793 765 730 686 634 574
Heat load of Pre Heaters (PH) and Final condenser (FC) kW 8.56 (PH1) 17.12 (PH2) 25.68 (PH3) 34.24 (PH4) 42.80 (PH5) 150 (FC)
Vapour generation rate in effects kg/s 0.33 0.32 0.30 0.28 0.26 0.23
Brine salinity (leaving the tube bundle) ppm 60,237 58,666 56,760 54,595 52,244 49,778
Overall HTC W/m2.K 1665 1649 1633 1616 1599 1581
Total heat transfer area m2 220 214 206 195 182 166
Specific heat transfer area m2/kg/s 667 669 687 696 700 722
Number of horizontal tubes Nos 38 38 38 38 38 38
Number of vertical tubes Nos 46 45 43 41 38 35
Fresh water yield of the MED-TVC system kg/s 1.73
Gain output ratio (GOR) 10.36
Universal Performance Ratio (UPR) 87.29
Thermodynamic Limit (TL) 10.54%

10
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Fig. 4. Analysis results of varying pressure type accumulator.


4cu.m 8cu.m 12cu.m 16cu.m
a
Steam charging rate into

0.25
the accumulator, kg/s

0.2
0.15
0.1
0.05
0
8 10 12 14 16 18 20
Time, h
4cu.m 8cu.m 12cu.m 16cu.m
25
b
accumulator, bar
Pressure of the

20
15
10
5
0
8 10 12 14 16 18 20
Time, h

4cu.m 8cu.m 12cu.m 16cu.m c


9000
Mass of water in the

8000
accumulator, kg

7000
6000
5000
4000
3000
2000
1000
0
8 10 12 14 16 18 20
TIme, h

4cu.m 8cu.m 12cu.m 16cu.m d


Steam discharging rate from

0.2
the accumulator, kg/s

0.15

0.1

0.05

0
8 10 12 14 16 18 20
Time, h

the first effect to the sixth effect was 1665 W/m2-K, 1649 W/m2-K, entire MED-TVC process, the total yield obtained from the steam dis-
1633 W/m2-K, 1616 W/m2-K, 1599 W/m2-K and 1581 W/m2-K re- charging duration of 27,046 s was 46,695 kg and the Gain Output Ratio
spectively. The heat transfer areas required for the MED evaporator (GOR) was 10.36 in this study. As 2.6 kWh electrical energy and
were 220 m2, 214 m2, 206 m2, 195 m2, 182 m2 and 166 m2 from first 77.66 kWh thermal energy were consumed per cubic meter of fresh
effect to sixth effect respectively. As shown in Table 3 the maximum water generation, the Universal Performance Ratio (UPR) [35] of this
vapour velocity through the tubes was below the technical limit of process was 87.29 and Thermodynamic Limit (TL) of this process was
60 m/s and velocity through demister was lower than 8 m/s [23]. 10.54%.
Considering the total fresh water generation rate of 1.73 kg/s in the

11
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Fig. 5. Analysis results of constant pressure type accumulator.


4cu.m 8cu.m 12cu.m 16cu.m
a
Steam charging rate into

0.250
the accumulator, kg/s

0.200
0.150
0.100
0.050
0.000
8 10 12 14 16 18 20
Time, h
4cu.m 8cu.m 12cu.m 16cu.m
10 b
8
Pressure of steam
accumulator, bar

0
8 10 12 14 16 18 20
Time, h
4cu.m 8cu.m 12cu.m 16cu.m
10000 c
8000
Mass of water in the
accumulator, kg

6000
4000
2000
0
8 10 12 14 16 18 20
TIme, h

4cu.m
0.200
8cu.m 12cu.m 16cu.m
d
from the accumulator, kg/s
Steam discharging rate

0.150

0.100

0.050

0.000
8 10 12 14 16 18 20
Time, h

4.2. Influence of volume of the steam accumulator on solar MED-TVC discharging rate in the varying pressure type accumulator are presented
process in the Fig. 4a to d respectively. Fig. 4a shows increase in the charging
duration with increases in the accumulator volume. This was due to
The influence of accumulator volume has been analyzed on both increase in mass of water, which was the result of an increase in volume
varying pressure type as well as constant pressure type steam accu- of accumulator, which stored the energy of the charged steam as a
mulator. There was variation in the accumulator volume from 4 m3 to sensible heat. At the same time, an increase in accumulator volume
16 m3 in increment of 4 m3 and the initial volume of water was con- decreased the maximum pressure attained by the accumulator as shown
sidered as 50% of the total accumulator volume. The mass and pressure in Fig. 4b. The maximum pressure attained by accumulators were
of steam generated in the solar field at experimental plant [12] was 19.78 bar, 12.14 bar, 9.56 bar and 8.41 bar in 4 m3, 8 m3, 12 m3 and
considered as input parameters as presented in the previous Section 4.1. 16 m3 capacity accumulators respectively. Fig. 4c shows the small
The steam charging rate, pressure, mass of water and steam variation in water mass compared to the initial water mass in the

12
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Fig. 6. Duration and mass of production in the varying pressure


30000 55000 type accumulator coupled MED-TVC process.

Amount of producon, kg
Duration of production, s

28000 51000 Duraon of


producon in
26000 47000 seconds

24000 43000

22000 39000 Amount of


producon in kgs
20000 35000
0 4 8 12 16 20

Volume of the steam accumulator in m3

accumulator during charging process. When the solar field pressure was Fig. 5c shows increase in water mass due to charging of sub cooled feed
less than the accumulator pressure the charging stopped and only dis- water in the accumulator. As discussed earlier when the solar field
charge took place. Pressure sliding took place during this period and pressure was less than the pressure of the accumulator, only discharge
there was a reduction in the mass of water due to flash evaporation in of steam and also reduction in mass of water took place in the accu-
the accumulator. This is shown in Fig. 4b and c. mulator as in Fig. 5c.
The pressure dropped quickly during the pressure sliding process The steam discharging rate from the varying pressure and constant
then increased slightly before finally stabilizing. The occurrence of this pressure type accumulator to MED-TVC are presented in Fig. 4d and
phenomenon was due to the entry of a fraction of the steam into the Fig. 5d respectively. The 4 m3 volume accumulators (both constant as
steam space directly before an ample heat exchange and equilibrium well as varying type) discharged the steam quickly for MED operation
with water during charging time, this effect leading to a quick raise in also stopped the steam at the earliest, whereas 16 m3 steam accumu-
pressure during charging process. When charge stopped and discharge lators took a longer duration for discharge of steam for MED operation
took place, the steam in the steam accumulator condensed and trans- and also it took a longer duration to stop the discharge of steam. As the
formed into water because of the energy potential difference between mass of water which stored the energy as sensible heat was less in 4 m3
the steam and water phase, which caused a quick drop in the pressure. accumulator as compared to 16 m3, it quickly attained above 5 bar due
Then due to non-equilibrium effect between steam and water and also to condensation of charged super heated steam. Whereas in the 16 m3
due to thermal inertia of water, flash evaporation happened and the accumulator the water mass was more, hence it took time to attain
pressure of accumulator was increasing slightly until it reached a new 5 bar which was the set pressure to open discharge valve. Following the
equilibrium state. stopping of charge in the accumulator, as the water mass was less in
The Fig. 5a to d show the steam charging rate, pressure, mass of 4 m3 accumulator as compared to 16 m3, it quickly attained 10% of the
water, and steam discharging rate in the constant pressure type steam total volume of the accumulator during discharge. Whereas in 16 m3
accumulator respectively. In the constant pressure type accumulator, accumulator the water mass was more, it took a longer duration to
the maximum pressure attained by the accumulator was 8 bar as it was attain 10% of the accumulator volume. The duration of production and
the set pressure. Fig. 5b shows increase in the duration for attaining this the corresponding yield in both varying pressure and constant pressure
maximum pressure with increases in the accumulator volume. This was type accumulator coupled MED-TVC systems are presented in Figs. 6
due to increase in mass of water, which was the result of increase in and 7 respectively. The figures show increase in accumulator volume
volume of accumulator, which stored the energy of the charged steam causing increase the amount of fresh water production. It is also re-
as a sensible heat. Following the attainment of 8 bar pressure, the vealed that for the same volume of the accumulator, constant pressure
variation in the pressure was seen as less despite charge and discharge type accumulator coupled MED-TVC system gives more fresh water
of steam taking place simultaneously. This was due to 3 kg water at yield. This is due to fact that the charging of sub cooled feed water at
28 °C fed into the accumulator when pressure was exceeding 8 bar. The intervals, decrease the pressure and increase the amount of steam

Fig. 7. Duration and mass of production in the constant pressure


32000 55000 type accumulator coupled MED-TVC process.
Duration of production, s

Amount of producon, kg

31000 53000

Duraon of
30000 51000 producon in
seconds
29000 49000

28000 47000 Amount of


producon in
27000 45000 kgs
0 4 8 12 16 20

Volume of the steam accumulator in m3

13
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Fig. 8. Pressure and mass of water in the varying pressure type


Distrubance No disturbance a accumulator.
14
Pressure of steam accumulator,

12
10
8
6
bar

4
2
0
8 10 12 14 16 18 20
Time, h
Disturbance No disturbance
b
7000
6000
5000
Mass of water, kg

4000
3000
2000
1000
0
8 10 12 14 16 18 20
TIme, h

Fig. 9. Pressure and mass of water in the constant pressure type


Distubance No disturbance a accumulator.
Pressure of steam accumulator,

9
8
7
6
bar

5
4
3
2
1
0
8 10 12 14 16 18 20
Time, h

8000
Disturbance No Disturbance b
7000
Mass of water, kg

6000
5000
4000
3000
2000
1000
0
8 10 12 14 16 18 20
TIme, h

14
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

charged into the constant pressure type accumulator compared to production was 19.32% in varying pressure type accumulator coupled
varying pressure type accumulator. MED-TVC system. This was due to cloud cover of 30 min. Similarly the
reduction in fresh water production was 10.77% in the constant pres-
4.3. Influence of irradiance disturbances on the steam accumulator on solar sure type accumulator coupled MED-TVC system. Hence as compared to
MED-TVC process the varying pressure type accumulator, the constant pressure type ac-
cumulator stores and also discharges more steam for the specified solar
The influence of cloud cover of a 30 min duration has been analyzed field conditions. However both the type of accumulators of 12 m3 ca-
on both varying pressure type as well as constant pressure type steam pacity is capable of meeting 30 min steam demand.
accumulator. In this analysis 39,600 time steps, which is 09:00 h to
20:00 h has been simulated. Out of which, the period of 13:00 to 5. Conclusion
13:30 h was considered as cloud cover above the solar field and no
steam generation during this period has been considered. The simula- Steam accumulator is an attractive buffer storage system in solar
tion results have been compared with the results obtained without MED-TVC process for adjusting the differences between steam pro-
cloud cover case. In this analysis, 50% water level is considered at duction and consumption rates. But the challenging task is the predic-
12 m3 capacity accumulators. tion of the transient pressure changes in the accumulator under dy-
A comparison of the pressure and mass of water at varying pressure namic operating conditions due to variations in solar field pressure and
type accumulator with cloud cover and without is shown in Fig. 8a and steam generation rate and also the constant discharge of steam from
b. At 13:00 h, the accumulator attained 8.18 bar and mass of water was accumulator to MED-TVC process. Another challenging task in steam
6118 kg in both cases. During cloud cover time, only discharge of steam accumulator design is the prediction of accumulator volume that can
occured in the accumulator and pressure sliding took place in the ac- provide the required storage capacity between the specified maximum
cumulator and it attained 7.37 bar at 13:30 h. There was a decrease in and minimum pressure limits. Hence transient simulation and also
the mass of water from 6118 kg to 4745 kg in the cloud cover case as a parametric analysis were performed for the investigation of the impact
result of flash evaporation whereas there was attainment of 8.81 bar in of different parameters of steam accumulator on solar MED-TVC pro-
a situation without cloud cover with the mass of water increasing from cess's performance. The main conclusions from the simulation results
6118 kg to 6153 kg. This is shown in Fig. 8a and b. There was a steep are as follows:
increase in pressure in the accumulator after cloud cover time whereas
in the case without cloud cover the pressure increase was gradual. This 1) The steam accumulator stores energy in water and so water mass
phenomenon occurred because lesser water mass (due to reduction in decides the energy storage capacity of the steam accumulator.
water mass during cloud cover period) stored the energy in the case Hence, as the accumulator volume increases, the mass of water
with cloud cover. The corresponding duration of production was which stores energy as sensible heat is also increasing. This results in
27,046 s and 21,821 s in the case without and with cloud cover re- increase in the duration of charge time and duration of steam dis-
spectively. The production of fresh water was 46,695 kg and 37,674 kg charge time from the accumulator. At the same time any increase in
for the case without and with cloud cover respectively. Similarly, there accumulator volume cause a reduction in the maximum pressure
was a reduction in the mass of water from 6175 kg to 4823 kg and attained by the accumulator for the same solar field capacity.
reduction in the pressure from 7.99 bar to 7.17 bar in the constant 2) For the same accumulator volume, the constant pressure type ac-
pressure type accumulator during cloud cover period. This was due to cumulator stores and discharges more steam compared to the
flash evaporation of high pressure water as a result decrease in pressure varying pressure type accumulator. Also constant pressure type ac-
of accumulator from 7.99 bar to 7.17 bar. In the case without cloud cumulator operates within the specified maximum and minimum
cover, there was an increase in the mass of water from 6175 kg to pressure limits eliminating the need for higher wall thickness in
6374 kg, with pressure maintained constant at 8.00 bar. This is shown accumulator shell.
in Fig. 9a and b. The corresponding duration of production was 3) 12 m3 capacity constant pressure type accumulator is capable of
29,819 Seconds and 26,608 s and amount of production of fresh water meeting the 30 min steam demand during the cloud cover for the
were 51,482 kg and 45,939 kg for the case without and with cloud specified solar field condition.
cover respectively. The results show the overall reduction in fresh water

Appendix A. Governing equations for steam accumulator

Liquid mass balance in the accumulator is given as


dM1
= ṁ 1B + ṁ PT 1
dt (A1)
where, ṁ 1B is the net mass balance of liquid water due to inlet and outlet of water in the accumulator. ṁ PT1 is the net liquid mass balance due to
evaporation and condensation of water in the accumulator.
Similarly the steam mass balance in the accumulator is given as
dM2
= ṁ 2B + ṁ PT 2
dt (A2)
where, ṁ 2B is the net mass balance of steam due to inlet and outlet of steam in the accumulator. ṁ PT2 is the net steam mass balance due to
evaporation and condensation of steam in the accumulator.
As the volume of the accumulator is constant and there is no work applied on the steam accumulator, the enthalpy of both liquid and gaseous
phase of the accumulator is given as follows, when neglecting the potential and kinetic energies.
Enthalpy change with respect to time of liquid phase is given as,

15
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

dh1 1 ⎡ dp dM1 ⎤
= ̇ )1B + ṁ PT 1 h" + Q̇21 + V1
(mh − h1
dt M1 ⎣ dt dt ⎦ (A3)
where, (ṁ h)1B is the net energy balance of liquid water due to inlet and outlet flows at the accumulator. Q̇ 21 is the heat transfer rate from the
superheated steam to the liquid.
Similarly steam phase enthalpy is given as
dh2 1 ⎡ dp dM2 ⎤
= ̇ )2B + ṁ PT 2 h" − Q̇21 + V2
(mh − h2
dt M2 ⎣ dt dt ⎦ (A4)
Further, the specific volumes of liquid water and steam are written as functions of pressure and corresponding enthalpies. The volume balance is
differentiated by time and the following equation is obtained

dM1 dM2 ⎛ ∂v dp ∂v dh1 ⎞ ⎛ ∂v2 dp ∂v dh2 ⎞


v1 + v2 + M1 ⎜ 1 + 1 ⎟ + M2 ⎜ + 2 ⎟ = 0
dt dt ⎝ ∂p h
dt ∂h p dt ⎠ ⎝ ∂p h
dt ∂h p dt ⎠ (A5)
Substituting Eqs. (A3) and (A4) in eq. (A5) leads to the following

⎛ ⎛h ∂v1 − v1 ⎞
dM1 ∂v
+ ⎛h2 ∂h2 − v2 ⎞
dM2 ⎞
1
⎜ ⎝ ∂h p ⎠ dt
⎝ p ⎠ dt

⎜ − ∂v1 + ṁ PT 2 h" − Q21 ]⎟⎟
dp
⎜ ∂h ( ) p
̇ )1B
[(mh + ṁ PT 1 h" + Q̇21] − ( )
∂v 2
∂h p
̇ )2B
[(mh ̇
= ⎝ ⎠
dt ⎛ ⎛ ∂v1 ∂v
⎜ + v1 ∂h1 ⎞ M1 + ⎛ ∂v2 + v2
∂v 2
⎞ M2 ⎟⎞
∂p
⎝⎝ h p⎠ ⎝ ∂p h ∂h p
⎠ ⎠ (A6)
Eq. (A1), (A2), (A3), (A4) and (A6) constitute a set of five first-order ordinary differential equations for the prediction of liquid water and steam
masses, liquid water and steam enthalpies and steam accumulator pressure respectively.

Appendix B. Governing equations for MED-TVC process

Applying mass and energy conservation laws to the evaporator and thermo vapour compressor, the following mathematical models are proposed.
The following assumptions are considered in the models:

• The evaporator and the thermo vapour compressor are perfectly insulated, hence heat loss to the surrounding is negligible.
• Equal temperature difference between feed heaters and effects.
• Distillate is salt free (as compared to feed sea water, the distillate salinity is negligible).
• Non condensable gas effects are negligible.
The energy balance of the TVC and its compression (CR), expansion (ER) and entrainment ratio are calculated as presented by Shinu M. Varghese
et al. [12]:
ṁ 2, out h2, out + ṁ r hr = (ṁ 2, out + ṁ r ) hd (B1)
ṁ 2, out
ṁ r
. h2, out + hr
hd = ṁ 2, out
1+ ṁ r (B2)
Pd
CR = = 1.975806
Pn (B3)
Ps
ER = = 40.32258
Pn (B4)

ṁ 2, out
= 0.97561
ṁ r (B5)
The temperature difference between the effects (Δ T) is assumed to be the same, Hence temperature difference between the effects is calculated as
follows:
T1 − Tn
∆T =
n−1 (B6)
where T1 the first is effect temperature and Tn is the last effect temperature. If the temperature in the effect is Ti, then the temperature in the next
effect Ti + 1 is calculated as follows:
Ti + 1 = Ti − ∆T (B7)
Feed seawater Ftot is distributed equally to all effects, hence feed flow at each effect and brine from each effect are calculated as follows:

16
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Ftot
Fi =
n (B8)
Bi = Fi − Di (B9)
Total brine is calculated as follows:
n
B= ∑ Bi
i=1 (B10)
Motive steam (ṁ 2, out ) from steam accumulator and entrained vapour(mr) from last effect are fed into the tubes of the first effect at discharge
pressure(Pd), where it condenses and supplies its latent heat(Ld) to the feed sea water which is sprayed over the tubes as falling film. This latent heat
of condensation increases the temperature of the feed water(Tfw(1)) to top brine temperature(T1). Evaporation takes place in the shell side of the
effect and a fraction of the water is converted into vapour and remaining becomes brine. This is due to low pressure which is less than the saturation
pressure of water maintained in the shell side of the effect. The vapour generated in the first effect is calculated as follows:
(ṁ 2, out + mr ) (Ld ) T1 − Tfw (1) ⎞
D1 = − F1 C ⎛ ⎜ ⎟

L1 ⎝ L1 ⎠ (B11)
The generated vapour temperature is less than the boiling temperature (Ti) by Boiling point elevation (BPEi) and it is calculated as follows from
the correlations proposed by Dessouky and Ettouney [30]:
BPEi = AX + BX 2 + CX 3 (B12)
where,
A = 8.325 × 10−2 + 1.883 × 10−4 Ti + 4.02 × 10−6 Ti 2

B = −7.625 × 10−4 + 9.02 × 10−5 Ti − 5.2 × 10−7 Ti 2

c = 1.522 × 10−4 − 3 × 10−6 Ti − 3 × 10−8 Ti 2


where Ti is the temperature in °C and X is the salt weight percentage. The above equation is valid over the following ranges: 1 ≤ X ≤ 16%,
10 ≤ Ti ≤ 180 °C.
Pressure loss occurs when the vapour flows through mist eliminator in the evaporator, Temperature drop occurs in the vapour as a result of this
pressure drop. The correlation for pressure drop in the demister(Δ Pme) has been developed by El-Dessouky et al. [30] for industrial type wire pads.
This correlation is given by
∆Pme = 3.88178 (ρp )0.375798 (V )0.81317 (δ w )−1.56114147 (B13)
where Δ Pme is the demister pressure drop in Pa/m, δw is the wire diameter in mm, V is the vapour velocity in the demister in m/s, and ρp is the
demister density in kg/m3.
The pressure drop in the lines connecting the vapour space in effect i and the evaporator tubes of the next effect are calculated using Unwin
formula [30] as follows:

∆Pcl =
0.0001306 m2 L 1 + ( 3.6
δi )
ρv δi5 (B14)
where m is the mass flow rate of the vapour stream in kg/s, L is the tube length in m, δi is the inner diameter in m, ρv is the vapour density in kg/m3
and Δ Pcl is the pressure drop in Pa/m.
Hence the condensation temperature of vapour inside the tube bundle of the next effect is less than the boiling temperature(Ti) by boiling point
elevation(BPEi) and the temperature depression(Δ Tme(i)) associated with the pressure loss due to mist eliminator(Δ Pme) and the temperature de-
pression(Δ Tcl(i)) associated with the pressure loss due to friction in connecting lines(Δ Pcl(i)). Vapour condensing temperature in the next effect is
calculated as follows:
TV (i + 1) = Ti − (BPEi − ∆Tme (i) − ∆Tcl (i) ) (B15)
Pre heater load in each effect is calculated as follows:
PHi = i Fi C (Tfw (i) − Tfw (i + 1) ) (B16)
th
Vapour generated in second to n effect is calculated as follows:

((Di − 1) Li − 1) − PHi − 1 Ti − Tfw (i) ⎞


Di = − Fi C ⎛ ⎜ ⎟

Li ⎝ Li ⎠ (B17)
As the feed arrangement is a parallel feed arrangement, the feed seawater salinity in all effects is the same(Xfw). The salt mass conservation law is
applied on the assumption of the distillate being free of salt and the brine salinity which leaves the tube bundle is calculated as follows:
Fi Xfw = Bi Xb (i) + Di Xd (B18)
where Xd = 0

17
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Fi
Xb (i) = Xfw
Bi (B19)
The process performance of the MEE-TVC is evaluated in terms of Gain Output Ratio (GOR) as follows:
n
D= ∑ Di
i=1 (B20)

D
GOR =
ṁ 2, out (B21)
The Universal performance ratio [35] and Thermodynamic Limit of this MEE-TVC is evaluated as follows:
hfg
UPR =
3.6
2
∑i = 1 {φ ( 1
kWhele
m3 )+φ ( TBT
2
kWhther
m3 )} (B22)

UPR
TL = X 100
828 (B23)

Appendix C. Governing equations for calculating heat transfer area of the MED evaporator

Condensation heat transfer co-efficient is calculated using modified Nusselt correlation [31] as follows
1
3 4
hin = 0.728α3/4 ⎛ K1a . ρ1a . (ρ1a − ρ1b ). g . hfg − 1b . ⎟⎞

⎝ μ1a . din. (T1b − Twall ) ⎠ (C1)


When the axial vapour velocity inside the tube is high and turbulent, Nusselt correlation is no longer valid [28]. In this study the dimensionless
mass velocity (jg*) of vapour inside the tube is < 0.5 and Lockhart-Martinelli parameter (Xtt) is < 1.0, hence the condensate profile is wavy and
stratified. The fraction the heat transfer area where stratified layer exists is accounted for by the term 0.728α3/4 in the above Nusselt correlation
[28].
Evaporation heat transfer co-efficient is calculated using Sernas correlation [32] as follows
1 0.24 0.66
g ⎞ 3 ⎛ 4Γ ⎞ ⎛ ν2a ⎞
hout = 0.01925 k2a ⎛ 2 ⎜
⎜ ⎟ ⎟ ⎜ ⎟

⎝ 2a ⎠ ⎝ μ 2a ⎠
ν ⎝ α2a ⎠ (C2)
Sernas derived the above empirical equations of the average heat-transfer coefficient for a tube of 25.4 mm diameter from experimental data. It
fits best to the tube diameter considered in this work. The mean quadratic deviation of the above correlation is 3.1% and only one point out of 33
exceeded such a deviation [33].
In this analysis, for calculating the overall heat transfer coefficient and heat transfer area, the value of T1b − Twall in equation: (C1) is unknown.
Hence the following iteration procedure (six steps) is adopted.
Step 1: Rt is calculated using equation: (C3)

r
Rt = o ∗Rfi +
ro ∗ln ( ) +R
ro
ri
fo +
1
ri k hout (C3)
Step 2: A value is assumed for T1b − Twall.
Step 3: hin is calculated from equation: (C1)
Step 4: U is calculated from equation: (C4)
1
U=
r
ro ∗ ln ⎛ o ⎞
ro 1 ro ⎝ ri ⎠ 1

ri hin
+ ri
∗Rfi + k
+ Rfo + hout (C4)
Step 5: T1b − Twall is recalculated from equation: (C5)
T1b − Twall = ΔT (1 − Rt ∗U ) (C5)
where ΔT is the temperature difference between condensing vapour and feed seawater
ΔT = T1b − T2a (C6)
Step 6: The calculations from step 3 repeated and the iteration is continued until T1b − Twall converges.
Q
A=
U ∗ ( T1b − Twall ) (C7)

18
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

Appendix D. Flow chart

19
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

20
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

References equilibrium temperature difference and flash evaporation in a single stage-


evaporator and an investigation of effect of process parameters on the liquid
flashing in a LTTD desalination process, Desalin. Water Treat. 57 (2016)
[1] Phanikumar V.S. Sistla, G. Venkatesan, Purnima Jalihal, S. Kathiroli, Low tem- 27152–27168.
perature thermal desalination plants, Proceedings of Eighth ISOPE Ocean Mining [3] Muhammad Wakil Shahzad, Muhammad Burhan, Noreddine Ghaffour, Kim
Symposium, 2009, 20–24 September Chennai, India. (ISBN 978-1-880653-75-3; Choon Ng, A multi evaporator desalination system operated with thermocline en-
ISSN 1946-0066). ergy for future sustainability, Desalination (2017), http://dx.doi.org/10.1016/j.
[2] Dilli Balaji, Experimental study on the effect of feed water nozzles on non- desal.2017.04.013.

21
V. Samson Packiaraj Raphael et al. Desalination 435 (2018) 3–22

[4] G. Venkatesan, S. Iniyan, Purnima Jalihal, A theoretical and experimental study of a industrial thermal applications, Int. J. Thermodyn. 12 (2) (2009, June) 83–88.
small-scale barometric sealed flash evaporative desalination system using low grade [20] Baozhi Sun, Jiamin Guo, Lei Yu, Longbin Yang, Yanjun Li, Guolei Zhang, Simulation
thermal energy, Appl. Therm. Eng. 73 (2014) 627–638. and verification of a non-equilibrium thermodynamic model for a steam catapult's
[5] Michael Berger, Mirko Meyer-Grünefeldt, Dirk Krüger, Klaus Hennecke, steam accumulator, Int. J. Heat Mass Transf. 85 (2015) 88–97.
Marwan Mokhtar, Christian Zahler, First year of operational experience with a solar [21] H. El-Dessoukey, I. Alatiqi, S. Bingulac, H.M. Ettouney, Steady-state analysis of the
process steam system for a pharmaceutical company in Jordan, Energy Procedia 91 multiple effect evaporation desalination process, Chem. Eng. Technol. 21 (1998) 5.
(2016) 591–600. [22] A.O. Bin Amer, Development and optimization of ME-TVC desalination system,
[6] Markus Eck, Tobias Hirsch, Dynamics and control of parabolic trough collector Desalination 249 (2009) 1315–1331.
loops with direct steam generation, Sol. Energy 81 (2007) 268–279. [23] A.A. Mabrouk, K. Bourouni, H.K. Abdulrahim, M. Darwish, A.O. Sharif, Impacts of
[7] Jan Fabian Feldhoff, Kai Schmitz, Markus Eck, Lars Schnatbaum-Laumann, tube bundle arrangement and feed flow pattern on the scale formation in large
Doerte Laing, Francisco Ortiz-Vives, Jan Schulte-Fischedick, Comparative system capacity MED desalination plants, Desalination 357 (2015) 275–285.
analysis of direct generation and synthetic oil parabolic trough power plants with [24] R.K. Kamali, A. Abbassi, S.A. Sadough Vanini, M. Saffar Avval, Thermodynamic
integrated thermal storage, Solar Energy 86 (2012) 520–530. design and parametric study of MED-TVC, Desalination 222 (2008) 596–604.
[8] Edouard González-Roubaud, David Pérez-Osorio, Cristina Prieto, Review of com- [25] Ibrahim Halil Yilmaz, Mehmet Sait Soylemez, Design and computer simulation on
mercial thermal energy storage in concentrated solar power plants: steam vs. multi-effect evaporation seawater desalination system using hybrid renewable en-
molten salts, Renew. Sust. Energ. Rev. 80 (2017) 133–148. ergy sources in turkey, Desalination 291 (2012) 23–40.
[9] Ali M. El-Nashar, Performance of the solar desalination plant at Abu Dhabi, [26] Mohammad Taghi Mazini, Alireza Yazdizadeh, Mohammad Hossein Ramezani,
Desalination 72 (1989) 405–424. Dynamic modeling of multi-effect desalination with thermal vapour compressor
[10] Bernhard Milow, Eduardo Zarza, Advanced MED solar desalination plants config- plant, Desalination 353 (2014) 98–108.
urations, costs, future - seven years of experience at the plataforma solar de Almeria [27] Muhammad Wakil Shahzad, Kim Choon Ng, Kyaw Thu, Bidyut Baran Saha, Won
(Spain), Desalination 108 (1996) 51–58. Gee Chun, Multi effect desalination and adsorption desalination (MEDAD): a hybrid
[11] L. Garcia-Rodriguez, C. Gomez-Camacho, Thermo-economic analysis of a solar desalination method, Appl. Therm. Eng. 72 (2014) 289–297.
multi-effect distillation plant installed at the plataforma solar de Almeria (Spain), [28] Muhammad Wakil Shahzad, Aung Myat, Won Gee Chun, Kim Choon Ng, Bubble-
Desalination 122 (1999) 205–214. assisted film evaporation correlation for saline water at sub-atmospheric pressures
[12] Shinu M. Varghese, Raju Abraham, E.P. Hariprasad, C. Sureshkumar, Design and in horizontal-tube evaporator, Appl. Therm. Eng. 50 (2013) 670–676.
Performance Analysis of Solar Assisted Multi Effect Desalination System for Coastal [29] Raju Abraham, A. Mani, Heat transfer characteristics in horizontal tube bundles for
Regions, InDACON 2012, Feb 8–9, Mumbai, (2012). falling film evaporation in multi effect desalination system, Desalination 375 (2015)
[13] S.P. Viswanathan, Raju Abraham, Vineet Saini, Sanjay Bajpai, Public private part- 129–137.
nership on water - lessons from RAMDAS project on solar desalination, [30] H.T. El-Dessouky, H.M. Ettouney, Fundamentals of sea water desalination, (2002)
International Conference on Innovative Technologies and Management for Water (ISBN:0-444-50810-4).
Security, 12–14 February 2014 (Chennai, India). [31] Sadik Kakac, Hongtan Liu, Heat exchangers selection, rating and thermal design,
[14] Marc Medrano, Antoni Gil, Ingrid Martorell, Xavi Potau, Luisa F. Cabeza, State of (2002) (ISBN:0-8493-0902-6).
the art on high-temperature thermal energy storage for power generation part 2 - [32] Lixi Zhang, Shuai Zhou, Hefei Zhang, The process Researches for the desalination
case studies, Renew. Sust. Energ. Rev. 14 (2010) 56–72. using residual heat of flue gas, Proceedings of the International Multi Conference of
[15] Wolf-Dieter Steinmann, Markus Eck, Buffer storage for direct steam generation, Sol. Engineers and Computer Scientists 2009, Vol-II IMECS 2009, March18–20 (Hong
Energy 80 (2006) 1277–1282. Kong).
[16] Vladimir D. Stevanovic, Blazenka Maslovaric, Sanja Prica, Dynamics of steam ac- [33] Petr Kracík, Ladislav Snajdárek, Martin Lisý, Marek Balás, Jirí Pospísil, Correlation
cumulation, Appl. Therm. Eng. 37 (2012) 73–79. of heat transfer coefficient at sprinkled tube bundle, Mater. Technol. 50 (4) (2016)
[17] Fengwu Bai, Chao Xu, Performance analysis of a two-stage thermal energy storage 479–483.
system using concrete and steam accumulator, Appl. Therm. Eng. 31 (2011) [34] Y. Fujita, M. Tsutusi, Experimental investigation of falling film evaporation on
2764–2771. horizontal tube, Heat Transfer Jpn. Res. 27 (1998) 609–618.
[18] J. Buschle, W.D. Steinmann, R. Tamme, Analysis of steam storage systems using [35] Kim Choon Ng, Muhammad Wakil Shahzad, Hyuk Soo Son, A. Osman, Hamed, An
Modelica, Modelica, 2006, September 4th–5th. exergy approach to efficiency evaluation of desalination, Appl. Phys. Lett. 110
[19] A. Baldini, G. Manfrida, D. Tempesti, Model of a solar collector/storage system for (2017) 184101.

22

You might also like