NTPC
NTPC
NTPC
View Export
Online Citation
Abstract. A solar seawater desalination plant with glass-covered compound parabolic (CPC) collectors, flash evaporation
chamber and multiple-effect distillation (MED) with 8 stages, operating in Vallur, India, is modeled in the Ebsilon
thermodynamic simulation software. Standard equations from literature for solar collectors and for saltwater heat
exchangers are selected and implemented as scripts. Model results are a distillate/steam ratio of 7.0 and average temperature
steps of 2.8 K between effects. With the simulations it has been possible to identify performance-relevant deviations of the
INTRODUCTION
A solar desalination plant uses the energy from solar irradiation for the production of fresh demineralized water
from seawater as feedstock. NTPC operates a new commercial scale solar desalination plant in the vicinity of a thermal
power station on the coast in Vallur near Chennai, India. The plant consists of a solar collector field with glazed CPC
solar collectors, and a multiple-effect distillation (MED) system, top brine temperature of around 70°C and a nominal
production rate of 5 tons of distillate per hour. For periods without sunshine there is an auxiliary heater operated with
4-bar-steam from the adjacent thermal power station.
The solar field supplies hot water as energy source to the MED system. The MED is operating at relatively low
temperature of 70-75°C or lower due to the sub-atmospheric pressure inside the evaporator stages (effects). This low
temperature reduces negative effects of fouling/scaling of the saltwater on the heat exchanger tubes and makes solar
hot water an attractive energy source for this process for water production.
For the purpose of verifying the operation and performance of the solar desalination plant, NETRA and DLR both
use Ebsilon Professional software for simulation of thermodynamic cycles, processes and systems, and the model is
set up in this powerful engineering software environment (www.ebsilon.com). Ebsilon has the typical thermodynamic
process components integrated in the library, including turbines, heat exchangers, pumps, and piping, and also
concentrating solar collectors, piping manifolds etc., as well as saltwater thermodynamic properties. Missing
components in Ebsilon are however: CPC solar collector, steam-driven evaporator for saltwater, flash evaporation for
saltwater, and steam-driven preheater and condenser for saltwater.
140008-1
Solar Energy from CPC collectors
Solar CPC collectors consist of evacuated absorber tubes with linear-curved mirrors behind the tubes in order to
reflect and slightly concentrate solar irradiation onto the absorber tube. The absorber tubes are evacuated heat pipes
with water/steam or organic fluid at low pressure as heat transfer medium to transport the collected heat by evaporation
and condensation to a collecting manifold at the top of the collector, where the heat is transferred from the condenser
bulb of the heat pipe in the dry socket connector to the water of the solar circuit. The collector is specified as Apricus
Solar Co. Ltd model ETC-CPC-24, with 24 absorber tubes of about 1.7 m length, and a gross aperture of 5.12 m² per
collector (Figure 1). The collector has an additional glass cover. Performance data is taken from the Solar Keymark
certificate SC0064-19, reported in 2019 by Intertek Testing Services Shenzhen Ltd. Guangzhou.
The collectors are installed in the field, close to the ground, tilted by 20° to south, at a latitude of 13°14'N. The
solar field is configured in 5 parallel subfields (bays), each subfield consisting of 22 parallel rows of 5 collectors in
series each, resulting in a total collector gross area of 2816 m². For the plant modelling, piping lengths, diameters, and
insulation thickness are extracted from drawings and implemented in the model for heat loss and pressure loss
estimates.
The hot water from the solar field flows to a flash chamber at sub-atmospheric pressure to partly evaporate to low-
pressure steam flowing into the first effect of the MED system (“motive steam”), where it condensates releasing its
energy to the first stage (effect) of the desalination process. The nominal temperature and pressure in the flash chamber
are 72°C and 0.34 bar respectively. A pump feeds the remaining saturated water and the condensate from the first
effect back to the solar circuit, thus requiring a geodesic height difference to avoid cavitation in the pump.
Multiple-Effect Distillation
Distillation at industrial scale is widely used in many processes, and the most common way for desalination of
seawater to produce demineralized or drinking water, beside high-pressure membrane processes. The high amount of
thermal energy needed to evaporate water of around 2.26 kJ/kg (630 kWh/ton) makes this process only economically
efficient, if the condensation energy is recovered in subsequent stages by recovering it for evaporation at decreased
pressure, and if the energy source is sufficiently cheap and available. In the multiple-effect distillation the saltwater
preheated close to boiling temperature at the respective stage pressure level is sprayed onto the horizontal heat
exchanger tubes of the so-called effect (Figure 2 and 3). The evaporation takes place on the heat exchanger surface
under vacuum conditions and with excess liquid. The heat transfer from the condensing distillate inside the tubes to
the boiling saltwater in the shell side is very high under these phase-change conditions (2 kW/m²K) which is
advantageous for low temperature gradients between the stages and allows for good energy recovery. The formed
vapor is fed to the condensing side (inside tube) of the next stage as heat source, and distillate is collected from the
individual stages. Brine at increased salt concentration remaining after the evaporation flows in the same direction
(from higher to lower temperature) through the stages allowing for flash evaporation in the subsequent stages at lower
pressure generating some additional distillate vapor. Part of the distillate vapor generated in the stages is used for
preheating the entering feed of saltwater to boiling temperature in separate preheaters. Behind the last (coolest) stage
a condenser condenses the vapor from the last stage by cooling with excess seawater stream. Part of the brine with
increased salt concentration is recycled together with fresh seawater input into the evaporation stages, which saves
some thermal energy because the brine from the last stage is still at elevated temperature, and reduces the amount of
rejected brine. A higher number of stages reduces the specific thermal energy consumption (kWh/m³), but the number
140008-2
of stages is limited by the temperature difference between motive steam temperature and condenser, and by the cost
associated to the heat exchangers. Pumps and pumping (electric) energy are needed to pump distillate and brine from
the sub-atmospheric pressure level at the condenser pressure out of the system.
100 1
boiling point elevation, in K
90
80 0,8
temperature, in °C
Figure 4 shows the relation between temperature and pressure at the boiling/condensation point of water and the
“boiling point elevation”, the increase of boiling temperature due to the dissolved salt in seawater and brine. The
relevant temperature range for the MED is 75 to 30°C, and the related effect and condenser pressure range is 0.39 to
0.04 bar.
The NTPC installation in Vallur has eight effects and three intermediate seawater preheaters. It uses steam from
the solar field formed in the flash chamber as heat source for the first effect. The cold end in the condenser is cooled
by excess seawater, part of which becomes preheated seawater to the feed of the effects. According to design a fraction
of 2/3 of the total brine flow from the last effect, at elevated temperature, is recirculated together with preheated fresh
seawater as feed into the effects. In addition to the condensation in consecutive effect stages, fractions of the distillate
vapor from effects 2, 4 and 6 are used to preheat the feed water.
140008-3
MODEL IMPLEMENTATION
For practical reasons in development and simulation, the Ebsilon model for the solar desalination is originally
separated into two blocks. The solar field including pump and including backup steam supply is one block, that can
be simulated separately and in time-series mode. The second block comprises the whole MED system including energy
supply in form of hot water to the flash chamber. This block can be used to simulate the operation parameters of the
MED separately. Figure 5 shows the complete model.
applying the interpolated longitudinal and transversal IAM coefficients on direct irradiance, and the diffuse IAM Kd
on diffuse and ground-reflected irradiance for determining the absorbed solar power. The typical ground albedo value
of 0.2 is used. The influence of ground reflected irradiance is very low, could be negligible, under the conditions of
this application and low tilt angle. For completeness the equation has been fully implemented.
Following collector performance equations are implemented:
Eabs η0b χ Edir Kθ Long,coll θlong Kθ Trans,coll θtran Ediff Erefl Kd optical gains (2)
Eloss a1 ΔT a2 ΔT ² thermal loss (3)
Qout ncoll AG Eabs –Eloss field thermal output (4)
140008-4
with ΔT as difference between mean collector temperature and ambient temperature. The collector output power can
be negative in periods without solar irradiance if fluid temperature is higher than ambient temperature. The applicable
parameters and symbols are listed in table 1.
The concept of collector cleanliness χ is implemented in this model. This parameter specifies optical loss on the
collector glass (between incoming solar irradiation and absorber) that is not considered in the collector efficiency
under clean test conditions. Dust is depositing on the collectors every day (can be additional 1-2% per day), also called
soiling, and has to be cleaned off by washing with water spray and brush regularly to maintain performance.
Cleanliness varies strongly depending on the ambient conditions (wind, dust, dew) and the cleaning frequency and
intensity. In field application this parameter can be measured on flat glass with a reflectometer, but this is not a
common technique. In the modelling this parameter is foreseen to be used as overall factor for performance reduction
of any type.
TABLE 1. CPC collector and field parameters (Apricus model ETC-CPC-24)
Block 1 and 4: Solar Keymark Certificate data, block 2: solar field layout data, block 3: cleanliness effect on optical performance
Name variable value unit
collector gross aperture AG 5.12 m²
optical efficiency η0b 0.54 -
1 thermal loss coefficient 1st order a1 1.21 W/(m²K)
thermal loss coefficient 2nd order a2 0.004 W/(m²K²)
IAM for diffuse irradiance Kd 0.98 -
number of collectors per collector row ncoll 5 / 110 loop / field
2 collector field azimuth (to south) Cazim 0 °
collector installation tilt angle to south β 20 °
sun elevation angle α variable °
3 collector cleanliness χ 0.9 -
angle θ 0° 10° 20° 30° 40° 50° 60° 70° 80° 90°
4 iam Kθ Trans,coll 1.00 1.02 1.03 1.04 1.05 1.12 1.18 0.79 0.39 0.00
Table 2 presents the design parameters for the whole solar field as modeled with the mentioned parameters.
Implementing MED
The second Ebsilon model block simulates the MED system fed by hot water on the hot end and cooled by seawater
on the cold end of the process. Information from the specific plant layout and from the literature [3 - 8] has been used
to implement the details of such system. Inlet stream apart from the solar circuit is the seawater, outlet streams are
cooling seawater, brine, distillate product (Figure 5 right end).
To implement the solar desalination plant as stationary simulation model in Ebsilon, there are conventional
components used, such as valves, pipes, manifolds, pumps, control units, sun. Several specific units of the system
however are not available as components in Ebsilon, in particular not for saltwater, and have been programmed as
macro components with EbsScript code. These are:
1. the condensing tubular heat exchangers with specified heat transfer coefficients for preheaters and condenser
2. the tubular evaporators with phase change on both sides of the tubes and saltwater spray evaporation with
specified heat transfer coefficients, and
3. the flash evaporation of saltwater.
140008-5
Each component has the specific balances (equations) on energy, water mass, and salt mass and pressure boundary
conditions at the respective boiling points of the specific saltwater/brine conditions or water(distillate) to be fulfilled.
Design sea water mass flow in each evaporator and condenser is 4 t/h and 90 t/h respectively. Design heat transfer
area of three preheaters varies in the range 4 m² to 16 m². Design heat transfer area of condenser and evaporator are
around 75 m² and 100 m² respectively.
According to the most common reference work by El-Dessouky [4], the overall heat transfer coefficient U in each
effect, preheater and condenser is a function of condensation temperature. The values are determined for a fouled
condenser, including a fouling resistance of 0.08 m²K/kW, with the equations [4]:
condenser/preheater: 𝑈 1617.5 0.1537 𝑇 0.1825 𝑇 0.00008026 𝑇 (5)
evaporator (effect): 𝑈 1939.1 1.40562 𝑇 0.0207525 𝑇 0.0023186 𝑇 (6)
where saturation temperature Tsat is in °C, and U is in W/m²K. Typical results are in the range of 1800 and 2800
W/m²K, for condenser and 1st effect respectively, directly determining the resulting temperature difference between
hot (condensing) and cold (liquid or boiling at lower pressure) side of the heat exchangers.
Pressure loss of the steam side is neglected (set to 0). Pressure loss of the saltwater side is not calculated, as there
is no influence on thermal performance of the system. Vapor loss with the non-condensable gases (NCG) to the
vacuum system is set to 0. Non-condensable gases, which might influence heat transfer coefficients, are not considered
in the model [5]. Non-equilibrium allowance in the effects is neglected [3, 9].
The condenser and the preheater components use the same equations and the identical program code.
Heat transfer is limited to the full condensation of the heating vapor (no subcooling). The effect of possible further
scaling on the evaporator tubes can be accounted for by a fouling factor, that reduces the heat transfer coefficient U.
RESULTS
The plant in its design configuration reaches modelling results according to table 3. The results show that there is
some design margin for the production of the desalination plant, which is specified to 5 t/h of distillate with the
specified solar field net power.
TABLE 3. Modelling results of solar MED plant performance
Net power from solar field 611 kW
Top brine temperature 65°C
Distillate production 6.6 t/h
Specific thermal energy for distillation 93 kWh/t
Recovery ratio distillate/feed 0.21
Performance ratio distillate/steam 7.0
Temperature step between effects 2.8 K
140008-6
110
900
100 1000
800
90
Temperature in °C
Qnet χ=0.80
Qnet solar in kW
600 70
Qnet χ=0.70
500 60 600
Qnet χ=0.44
400 50
40 400
300 SolarToFlash °C
30 SolarReturn °C
200
20 Power FlashChamber kW 200
100 10 Flash steam flow kg/h
0 0 0
0 200 400 600 800 1000 0 200 400 600 800 1000
solar GHI in W/m² solar GHI in W/m²
(a) (b)
FIGURE 6. (a) Solar field net outlet power for cleanliness of 0.90 to 0.44,
simulation results with flow rate 35 t/h and solar field inlet temperature 70°C
(b) Flash chamber inlet and outlet temperature and power depending on solar GHI for design mass flow 35 t/h,
for sun position on 27-June 13:00h, simulation results of combined model with cleanliness χ=0.80, Tamb=30°C, Tseawater=29.4°C
For the combined model of solar field and MED it results that flash chamber and top brine temperatures are directly
related to the amount of energy coming from the solar field. With the maximum specified feed temperature of the
solar part to the flash chamber, the power is limited at 611 kW. Figure 6 shows this relation.
The model component parameterization has been adapted to also reproduce operation data with lower specific
distillate production. The model parameters for tuning the Ebsilon model with the objective of good agreement with
the operation data include in particular demister pressure loss, condenser pressure and preheater steam mass flow.
Parameters such as feed flow rate and brine recirculation ratio are apparently of low influence on the performance.
The simulation with the model and design parameters shows a good agreement with the available design data.
Distillate production rate under ideal design conditions is above the design value, according to expectations. The
comparison of the process engineering and design data and plant data from 7 days of trial operation with solar heating
60
pEff
0,24 600
7
m distillate in t/h
6
50 0,2 500
5
p_Effect
Q_Effect
t_Effect
40 0,16 400
4
30 0,12 300 3
design
20 0,08 200 2
Feed 19 t/h
1 June27 conditions
10 0,04 100
0
0 0 0 0 200 400 600 800 1000
1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 Q in kW
Effect # Effect #
The models can be applied for simulation of the solar field alone and the MED plant alone. The combination of
both model blocks into one single model is demonstrated to simulate the entire plant with specific operating parameters
or as time series, in particular concerning the solar irradiation input data variation.
Simulation results of the combined model show that the flash chamber temperature strongly depends on the amount
of energy from the solar field, determining also the top brine temperature in the first effect of the MED. Other control
parameters of the plant have only low influence on this top brine temperature level.
140008-7
The models are set up and prepared with components and process parameters for a specific commercial project,
but the new components can be flexibly used with other parameters or modified programs for different configuration
of a solar desalination with multiple-effect distillation.
CONCLUSION
The model in Ebsilon introduces new components for the solar collector rows, for saltwater flash evaporation,
saltwater evaporators, preheaters and condenser. Technical data has been implemented as specification values for
components and streams. With the provided variations in parameters such as condenser and flash pressure, preheater
steam flow and demister pressure loss it is possible to reproduce trial operation data of the plant from June 2020. This
serves as validation of the developed model.
The simulation shows for the solar field that at high irradiance the temperature limitation at 85°C in combination
with the mixing valve control leads to collector temperatures above 140°C. To maintain flow through the collectors,
elevated pressure behind the pump would be required, otherwise the collectors (or some of them) go into stagnation
and no energy will be delivered from the affected loops to the solar circuit.
Further observation concerns the saltwater feed preheaters. According to the operation data, the transferred heat in
the preheaters is too low in comparison to the model results. In order to reproduce the lower performance of the
preheaters in the model, the mass flow split to the preheater from the respective effect can be set in the model. The
reason for the low performance of the preheaters must be a too low steam-side mass flow and is probably the steam-
side pressure level which could be influenced by some valve or other pressure drop or by geodesic height difference
to the effects. This reduces the production rate and increases the specific energy consumption.
The Ebsilon model can be used as digital twin for the physical plant to analyze the influence of operation
parameters such as mass flow and temperature of solar circuit, seawater to condenser, feed to MED, and brine
recirculation under different solar energy input levels on the distillate production and the key performance parameters.
The parametric studies are suggested to be performed in particular with the MED part of the model and using the time
series function of Ebsilon to produce tables of calculation results. The performance of the solar field can be analyzed
ACKNOWLEDGMENTS
The work was part of the German-KfW-(development bank)-funded NETRACST project of NETRA and DLR.
REFERENCES
1. E. L. Maxwell “A Quasi-Physical Model for Converting Hourly Global Horizontal to Direct Normal Insolation”,
SERI/TR-215-3087 USA, (1987) https://www.nrel.gov/grid/solar-resource/disc.html
2. K.B.H. Liu, R.C. Jordan “Daily insolation on surface tilted toward the equator”, ASHRAE Journal 10, pp. 53-
56, ASHRAE Transactions 3, 67, 526-541 (1961)
3. K.H. Mistry, M.A. Antar, J.H. Lienhard “An improved model for multiple effect distillation”, Desalination and
water treatment 51(4):1-15. (2012) https://doi.org/10.1080/19443994.2012.703383
4. H. T. El-Dessouky, H. M. Ettouney “Fundamentals of Salt Water Desalination”, (Elsevier 2002)
5. H. T. El-Dessouky, H. M. Ettouney, F.Mandani “Performance of parallel feed multiple effect evaporation system
for seawater desalination”, Applied Thermal Engineering 20, 1679–1706 (2000) https://doi.org/10.1016/S1359-
4311(99)00098-8
6. J. Wellmann “Conceptual design of a concentrating solar power plant for a combined electricity and water supply
of the city El Gouna”, Dissertation Technical University Berlin (2015) https://doi.org/10.14279/depositonce-
4943
7. E.P. Dall “Integrating desalination with concentrating solar power: Large scale cogeneration of water and
electricity”, Master Stellenbosch University (2017) http://scholar.sun.ac.za/handle/10019.1/101238
8. A. Cipollina, M. Agnello, A. Piacentino, A. Tamburini, B. Ortega, P. Palenzuela, D. Alarcon, G. Micale
“A dynamic model for MED-TVC transient operation”, Desalination 413, 234-257 (2017)
http://dx.doi.org/10.1016/j.desal.2017.03.005
9. I.S. Al-Mutaz, I. Wazeer “Development of a steady-state mathematical model for MEE-TVC desalination
plants”, Desalination 351 9–18 (2014) http://dx.doi.org/10.1016/j.desal.2014.07.018
140008-8