COULD A VERTICAL TUBE MED BE AN ECONOMIC OPTION FOR THE FUTURE THERMAL

DESALINATION MARKET?

2. November 2018 – Preliminary Conceptual Design - Revision A – by F. Alt

CONTENT

1 INTRODUCTION
2 VT – MED DESIGN CONCEPT
2.1 Basic Criteria
2.2 Selection of Tubes
2.3 Module Design
2.4 Evaporator Configuration and Plant Layout

3 COMPARISON
3.1 Energy Consumption and other Key Data
3.2 Design Details

4 SUMMARY
5 REFERENCES

PAGE 1 of 19
1 INTRODUCTION

The MSF and MED-TVC systems currently in operation in the Middle East region have a specific allocated energy consumption (SAEC = convertible steam
energy + auxiliary power consumption) in the order of 2 to 3 times the power consumption of SWRO systems. The question is, how far the energy
consumption can be reduced and can thermal systems with low energy consumption compete with the SWRO technology in CAPEX and OPEX.
While some companies are concentrating in recent years on the development of horizontal tube MED systems (HT-MED) operating at elevated top
temperatures, vertical tube MED systems (VT-MED) may be a competitive option.
One example of a VT-MED concept described in this paper may be considered as a base for discussion. Since there are many variables regarding system
design and optimization, different engineers may like to apply different process parameters, different effect configuration and structural concepts, or changes
considering optimization of manufacturing procedures etc, to achieve lowest possible cost.

2 VT - MED DESIGN CONCEPT

2.1 BASIC CRITERIA

While it may be required to consider new system configurations, changes in process parameters, new materials etc to achieve significant improvements in
energy consumption and capital cost, it may be on the other side preferred to stay as much as possible with proven process design parameters and proven
materials, which could simplify the development of a new concept and reduce any potential risks of failure. For those reasons, the design concept described
here is kept as simple as possible. Considering that it can be proven, that the design would be feasible and economical, individual components could be
optimized or components could be added like nano-filtration (NF) and/or adsorption cycle (AD).
Information about vertical tube falling film evaporators in the field of seawater desalination as well as for other applications may be obtained from the
reference documents [1-7] or others.

For the example of a large capacity VT-MED the following basic design criteria have been considered:

 a unit capacity of 33 migd (150,000 m3/day)

 a SAEC (specific allocated energy consumption) below 7 kWh/m3
 minimized heat transfer area and evaporator volume to minimize OPEX
 a modular concept, which would allow to adjust unit capacities without a minimum effort of system design for other unit capacities
 keeping the aspect ratio of the system structure (height/width ratio) as low as possible to minimize challenges in the structural design concept

PAGE 2 of 19
2.2 SELECTION OF TUBES

In regards to tube materials, the possibility is given to select traditional materials such as titanium, copper nickel alloys or aluminum brass wherein certain
alloys may be limited to certain ranges of operation temperatures. Other possible materials so far not commonly used in the field of seawater desalination
may be stainless steel, aluminum or plastic materials.
In regards to tube dimensions, tubes with a outside diameter in the order of 40-50 mm have been used in conventional MSF-cross tube evaporators while
tubes with up to 3 inches (76.2 mm) in diameter have been considered in the past [7] for VT-MED design concept. In most existing HT-MED systems (MED-
TVC) a 1 inch (25.4 mm) tube diameter may have been applied for the effect tubing, while in those systems for feed heaters tube diameters of 5/8 inch
(15.875 mm) and ¾ inch (19.05 mm) are common, like for MSF-long tube evaporators. While in existing HT-MED evaporators the length of effect tubes may
be up to 8.0 meters long, in most of the VT-MED concepts the tube length may have been limited to about 3.00 meters, primarily to limit the evaporator
height. In regards to tube shapes, fluted tubes have been considered in a number of design concepts [1, 5, 6, 7] with the consideration that advantages in
heat transfer coefficient would have a positive impact on heat transfer and resulting system cost. Two examples of fluted tubes are shown in FIG 1a. The tube
on the left with an outside diameter of 2 inch (50.8 mm) is extruded out of an aluminum material, while it would be technically also possible to manufacture
fluted tubes with an outside diameter as small as ¼ inch (6.35 mm) made of a copper nickel alloy as shown on the right.

FIG 1a examples of profiled tubes

For the design concept described in this paper, conventional smooth tubes made of titanium or a high grade stainless steel with an outside diameter is 5/8
inches (15.875 mm), a wall thickness of 0.4-0.5 mm and a tube length of about 1.60 meters are considered. Those criteria may be preferred in regards to
market availability, material life cycle and minimization of evaporator dimensions.

PAGE 3 of 19
2.3 MODULE DESIGN

In the design concept a single module as shown in FIG 1b comprises a tube bundle with a feedwater reservoir located on top of the tube bundle to receive the
feedwater discharging from the tube bundle of the effect located above and to re-distribute the feedwater evenly to the individual falling film tubes. The tubes
are fixed in the upper and lower tube sheet with grommets, designed such, that tubes can be inserted from the side of the tube bundle, not requiring a
tubing/re-tubing space above or below the tube bundle. Such tube installation is possible since no intermediate tube support plates would be required in
between the top and bottom tube sheet. Each tube bundle comprises approximately 2880 tubes.

MIST ELIMINATOR

FEEDWATER RESERVOIR
HEIGHT 2.40 m

TUBE BUNDLE

STEEL STRUCTURE EFFECT SEPARATION FLOOR

WIDTH 1.80 m

SINGLE MODULE MULTIPLE MODULE CONFIGURATION

FIG 1b single module and example of multiple modules arranged in vertically stacked effects

Examples of tube grommets and inserts are shown in FIG 1c. The grommets located on the upper and lower tube end would typically have the function to
seal and to hold the tubes in place. The example of the shown insert comprises an upper part with an orifice which would have for each effect a specific
orifice size through which the feed water would pass from the reservoir into the individual tubes. At the discharge of the insert orifice the feed water would
flash down and the two phase vapor liquid flow would be guided by the lower insert part toward the tube wall, where the liquid would form a falling film.

PAGE 4 of 19
 TUBE

INSERT ORIFICE
LOWER TUBE SHEET

LOWER INSERT PART
UPPER GROMMET
LOWER GROMMET

TUBE

UPPER TUBE END LOWER TUBE END

FIG 1c examples of tube grommets and inserts

Vapor generated inside the tubes of a module located in one effect, passes in the module located below in the next effect through mist eliminators before
entering the tube bundle from the outside where it is condensing. The mist eliminators are installed vertically above the feedwater reservoir. The base plate of
the module forms an effect floor, separating the vapor spaces of effects stacked on top of each other.

The steel structure is carrying the module weight, liquid weight and the operation load resulting from differential pressure between effects, in vertical direction
toward the bottom structure as well as in horizontal directions to the evaporator shell. The module has a base plate size of 1.80 x 1.80 meters and a height of
2.40 meters. Multiple modules can be arranged in an effect in one row while multiple rows of such modules may be arranged side by side as required to
arrange the necessary quantity of modules in an effect, respectively to achieve the desired evaporator unit capacity. The space between the tube bundles
serves as vapor flow space as well as access space for maintenance.

2.4 EVAPORATOR CONFIGURATION AND PLANT LAYOUT

The system is designed with two parallel operating evaporators connected with platforms on each effect level where feed heaters are installed as shown in
FIG 2. Each of the evaporators comprises 24 effects and 24 modules per effect (3 rows of modules with 8 modules per row) with a footprint of 5.4 x 14.4
meters and a height of 60 meters (shell dimensions per evaporator excluding stiffeners).

PAGE 5 of 19
 STEAM DISTRIBUTION 1st EFFECT

FEED DISTRIBUTION 1st EFFECT

FEED HEATERS

EVAPORATOR

STEAM SUPPLY
BRINE DISCHARGE

FINAL CONDENSER
DISTLLATE DISCHARGE
COOLING WATER SUPPLY

LAYOUT ‐ VAPOR SUPPLY SIDE LAYOUT ‐ COOLING WATER SUPPLY SIDE PIPING (excluding evaporators and steel structure)

FIG 2 plant layout of a 33 migd unit with SAEC <7 kWh/m3

As shown in FIG 3, below the last effect, modules similar to the effect modules are installed, in which brine is collected and discharged by gravity, distillate is
collected and pumped out by distillate pumps and the vapor generated in the last effect is guided to the final condensers and the parallel operating first feed
heaters. The base steel structure has a height of about 10 meters, a width of about 15.0 meters and a length of about 18.0 meters. As such the entire plant
has a height of about 70 meters an aspect ratio (height/width) of <5 (for comparison, the tower concept of a 62.5 migd (283,875 m3/day) VT-MED [5] with a tower
diameter of about 24 meters and a height of 158 meters has an aspect ratio of about 6.5. Using the same tower concept for a unit capacity of 32 migd, the tower diameter
would change to about 17.5 meters and the aspect ratio would exeed a value of 9).

While the first feed heater is feed with cold seawater in parallel to the final condensers, the last feed heater located on the level of effect 1, is heated directly
by a part of the heating steam. In between, feed heaters are installed at each effect, which are heated by a part of the vapor generated in the effects.

PAGE 6 of 19
 MODULES IN BOTTOM EFFECT

FINAL CONDENSER

BOTTOM FEED HEATER

FEEDWATER PUMPS
BRINE COLLECTION

BRINE DISCHARGE

FIG 3 plant layout – bottom part

The system is designed in a forward feed flow concept with a split feed distribution, where 100% of the feedwater is passing through the feed heaters located
on the bottom 6 effects, 16.7% are split off at each of the effects 18, 12 and 6 and the remaining 50% of the feedwater is distributed over the tube bundles in
the first effect.

3 COMPARISON

3.1 ENERGY CONSUMPTION AND OTHER KEY DATA

For a preliminary comparison with other systems, graphics presented at the IDA congress 2017 in Sao Paulo [9] are used as shown in FIG 4, 5 and 6.
The specific allocated energy consumption (SAEC) for the VTE-MED as described, shows in FIG 4 with 6.7 kWh/m3 a slight advantage over the curves
directly above, which have been calculated for HT-MED and VT-MED systems. Primary reason for the reduced SAEC for this system are assumed lower
fouling rates. The design fouling factor for the effect tubing has been reduced from 0.00020 m2 K/W to 0.00015 m2 K/W while the design fouling factor for the
feed heater tubing has been reduced from 0.00015 m2 K/W to 0.00010 m2 K/W.
Those assumptions may be realistic for the VT-MED since liquid loadings are significantly higher than on HT-MED tube bundles, and as such the degree of
scaling may be reduced. The lower fouling factor for feed heater tubes may be in general realistic for all MED - systems if a ball cleaning system is used like
in a MSF plant.

PAGE 7 of 19
 24

kWh / m3
22

20
specific allocated energy consumption

existing MSF
MED-TVC @ 15 bar(a)
18

16 existing MED-TVC
MSF - 1 mgd - Operation Report 1971 [8]
MSF concepts
14

12

MED concepts VT - MED study 1995 [5]

10

8
MED @ 0.5 bar(a)

6
VT-MED-112C @ 3 bar
VT-MED-112C @ 1.8 bar(a)
4
4 6 8 10 12 14 16 18 20 22 24
performance ratio kg / 2326 kJ

FIG 4 specific allocated energy consumption (SAEC) over performance ratio

Other factors contributing to an improvement of the performance ratio respectively reduced SAEC relatively to the curves shown for other VT-MED or HT-
MED systems is a change of the steam supply condition from 3.0 bar @ 133 C [9] to 1.80 bar @ 117 C which reduces the content of convertible steam
energy.
Since there are many variables for the system design, individual systems currently under development may be designed with different parameters than
assumed for the calculation of the shown curves [9]. Characteristic values of individual concepts may be added into the graphics for comparison.
The VT-MED curve for the systems designed for 112C shows in FIG 5 in regards to the specific heat transfer area due to the higher top temperature an
advantage over the HT-MED concepts designed with a lower top temperature of 105C. The example of a 33 migd VT-MED described in this paper is based
on further changes like a minimized steam supply pressure and reduced design fouling factors. As a result, the 33 migd VT-MED has a ratio of SAEC/specific
heat transfer area somewhere in between a system based on more conservative design parameters [9] and a VT-MED concept study from 1995 [5] (red dot
shown on the bottom left of the graphic FIG 5). Comparing to conventional/existing MSF or MED systems with SAEC in the range of 12-16 kWh/m3, the 33
migd VT-MED concept would require a similar specific heat transfer area while the SAEC would be in the order of 50%. The reduced SAEC is a result of the
evaporator design concept, reduced steam supply condition and reduced auxiliary power consumption.

PAGE 8 of 19
 17

kWh / m3
16
existing MSF MED-TVC-85C @ 15 bar(a)
existing MED-TVC
15

14
specific allocated energy consumption

MED @ 0.5 bar(a) MED-TVC-125C @ 15 bar(a)

13 MSF concepts

12

11

10

9
VT - MED study 1995 [5]
8

VT-MED-112C @ 3 bar
7 HT-MED concepts-85C @ 3 bar

6 HT-MED concepts-105C @ 3 bar

VT-MED-112C @ 1.8 bar(a)
5
2.0 2.5 3.0 3.5 4.0 4.5

specific heat transfer area m2 per m3 per day of distillate production

FIG 5 specific allocated energy consumption (SAEC) over specific heat transfer area

17
MED-TVC-85C @ 15 bar(a)
16 existing MSF
kWh / m3

existing MED-TVC
15

14 MED-TVC-125C @ 15 bar(a)
MSF-B long-tube evaporators
specific allocated energy consumption

13

12
MSF concepts MED @ 0.5 bar(a)
11

10

9
HT-MED concepts-85C @ 3 bar
8 VT-MED-112C @ 1.8 bar(a)
VT-MED-112C @ 3 bar
7

6
(Rev A) VT - MED study 1995 [5] HT-MED concepts-105C @ 3 bar
5
5,000 10,000 15,000 20,000 25,000 30,000

evaporator volume for 20 migd capacity m3

FIG 6 specific allocated energy consumption (SAEC) over evaporator shell volume (calculated for 20 migd unit capacity) – (Rev. A)

PAGE 9 of 19
To compare the evaporator volumes of different systems, a uniform 20 migd unit capacity has been assumed [9]. For the same reasons which are resulting
in a low heat transfer area as described above, the evaporator volume for the described example of the VT-MED is relatively low compared to other
evaporators. In addition, the tube diameter of 16 mm (5/8 inch), the tube length of 1.60 meters and the fact that no re-tubing space is required in longitudinal
direction between two tube bundles, contributes to the reduction of the evaporator volume, considering in comparison that for the design of the HT-MED
systems 25.4 mm tube diameter (1 inch) and a tube length of 4-6 meters has been assumed [9]. As the evaporator volume of the described VT-MED example
is similar to the volumes of the MSF-B evaporators [10] shown in the graphic above, it may be mentioned that those long-tube MSF-evaporators are also
designed with 16 mm (5/8 inch) tube diameter, while the flash stages are also based on a modular multiple tube bundle concept with vertically stacked flash
stages, which provides a more compact design compared to the latest built MSF cross tube evaporators which may have tube diameters in the range of 40
mm. Comparing to the VT-MED study from 1995 [5], it should be mentioned that in the shown volume of this plant, feed heaters are an integral part of the
evaporator tower, while externally arranged feed heaters of HT-MED and other VT-MED plants are not included in the shown evaporator volumes. Also in the
VT-MED study from 1995 tubes with an outside diameter of 50.8 mm and a length of approximately 3.0 meters have been applied.

Another factor which may be important for future thermal desalination plants, is the space requirement. To get an approximate idea of the space requirement,
the VT-MED is compared in FIG 7 to some MSF plant concepts with unit capacities between 20 and 36 migd and performance ratios between 12 and 15
kg/2326 kJ [11].

VT‐MED 33 migd / PR 18.7

MSF‐C2 36 migd / PR 15

MSF‐C1 27 migd / PR 12

MSF‐A 20 migd / PR 12

FIG 7 VT-MED in comparison to MSF- plant concepts [11]

PAGE 10 of 19
While the specific space requirement for a conventional MSF with 20 migd capacity and PR 10 may be in the range of 500 m2/migd, the shown MSF plant
concepts may require about 400 m2/migd for the 20 migd unit capacity with PR12 and about 200 m2/migd for a 36 migd unit with PR 15, while the described
VT-MED concept would require <50 m2/migd.

3.2 DESIGN DETAILS

3.2.1 IN-TUBE VAPOR VELOCITIES

While in HT-MED evaporators vapor is condensing inside the tubes, in the VT-MED evaporators vapor is generated inside the tubes. As such, the vapor exit
velocities on the tubes of a VT- MED evaporator may be a design criteria comparable to the vapor inlet velocities on the tubes of a HT- MED evaporator.
An example of vapor velocities of a HT-MED is shown in FIG 8 together with the tube exit velocities of the vapor in the described VT-MED in comparison to
typical vapor design velocities for mist eliminators.

50

45

40

35
VELOCITIES m/s

30 VT-MED tube exit

MED-TVC tube inlet
25 mist eliminator

20

15

10

0
35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115

VAPOR TEMPERATURE C

FIG 8 vapor velocities on tube inlet of HT- MED and tube outlet on VT- MED compared to mist eliminator design velocities

PAGE 11 of 19
The vapor velocities shown for the mist eliminator are design velocities calculated with a common equation (1)

v = 1.07 * [(DensL – DensV)/DensV] ^0.5 (1)

wherein DensL = density of liquid kg/m3

DensV = density of vapor kg/m3

The vapor velocities shown as an example for a HT - MED are calculated for a MED-TVC with a top temperature of about 65 C which may be similar to the
configuration of the 15 migd evaporator at Yanbu – II, while the vapor velocities shown for a VT-MED are calculated for the evaporator configuration as
described in this paper. The VT-MED shows in the range below 70C operation temperature, similar vapor velocities as the MED-TVC, while the in-tube vapor
velocities in this range are by a factor 2-3 higher than the design vapor velocities for the mist eliminators shown as reference. The in-tube vapor velocities of
the VT-MED are approaching toward the top temperature, similar vapor velocities as the mist eliminator design velocities. While the shown vapor velocities
are only two examples, it may be noted that in-tube vapor design velocities may vary in different evaporators in a wide range, as engineers may make
compromises between vapor velocities, related irreversible temperature losses and evaporator size and related capital cost.

Comparing the influence of in-tube vapor velocities on heat transfer coefficients in a horizontal tube MED versus a vertical tube MED, there may be some
difference. In the horizontal tube MED the vapor with the highest velocity in the tube entrance area is facing an almost dry heat transfer surface, where an
increase of vapor velocity would contribute little to an increase of heat transfer, while on the tube exit side, the amount of accumulated condensate is the
highest, which may partly reduce the heat transfer, while the vapor velocity is toward the tube end the lowest, so that it has in this area also no significant
influence on the heat transfer coefficient. In a vertical tube evaporator a first portion of vapor is generated in the top part of the tube when the feedwater
enters and flashes down, while the amount of vapor and vapor velocity is increasing toward the bottom end of the tubes, so that the vapor may have an
increasing influence on the heat transfer coefficient toward the tube bottom end due to the influence on the annular feedwater film flow, such as shearing
force, reduction of film thickness or a ripple effect in the film flow pattern.

3.2.2 FALLING FILM WETTING RATES AND RELATED FILM CONCENTRATION

Wetting rates () may be defined in some literature as a liquid mass flow per meter width of a falling film on a single heat exchange surface, respectively on
one side of a horizontal tube [12] or a liquid mass flow per meter tube length, corresponding to the falling film mass flow running down on two sides of a
horizontal tube [13]. To compare the wetting rates on horizontal and vertical tubes, wetting rates used in this paper are defined as falling film liquid mass
flowing on a single heat exchange surface. The wetting rate at a VT-MED would be the liquid mass flow per meter inner circumference of the vertical falling
film tubes.
While wetting rates are in the design of HT-MED evaporators a critical factor in regards to scaling potential [13], wetting rates in VT-MED evaporators may be
in general significantly higher. Examples of falling film flows on horizontal and vertical tubes are shown in FIG 9a. The flow pattern shown on the horizontal
tubes, may be in the range between a droplet flow and a column flow as defined in [12] which would correspond approximately with a liquid loading range of 
= 0.030 to 0.035 kg/m/s, which may be assumed as the lower range of wetting rates which would ensure a complete wetting of the entire tube surface.

PAGE 12 of 19
However, laboratory tests [13] show certain differences in the wetting characteristics of different tube materials, wherein a 100% wetting of the tube surfaces
may be ensured in this range of wetting rates only with the application of a wetting agent, wherein a wetting agent may allow also wetting rates below  =
0.030 kg/m/s. In comparison, the discharge of a 2 inch diameter vertical tube with double fluted profile (see also FIG 1a) tested at a wetting rate in the range
of  = 0.300 to 0.400 kg/m/s shows a sheet flow on the discharge, wherein such sheet flow may be achieved already somewhere in the range of a wetting
rate  = 0.100 kg/m/s [12].

FIG 9a Typical falling film flow pattern on horizontal and vertical tubes

In the design of any falling film evaporator a number of parameters may determine the wetting rates. A first factor is the design concentration of the
discharging brine, which determines with the seawater salinity and the design distillate production the available feedwater mass flow. A second parameter,
which can influence the wetting rate in a wide range, is the feed distribution concept. Common concepts may be:

 parallel feed – where the feedwater is distributed over all effect tube bundles in parallel, meaning all tube bundles are exposed to a falling film with
seawater concentration on the top tube rows and maximum film concentration in the bottom tube rows, typically applied to HT-MED configurations
operating at top temperatures <67C like currently operating MED-TVC systems

PAGE 13 of 19
  forward feed – where all the feedwater is distributed only over the tube bundles in the first effect, while the discharging brine from one effect is
passing as feedwater to the tube bundles of the next effect, which may be primarily applied to VT-MED evaporators
 mixed forward/parallel feed – wherein the total number of effects are divided in groups of typically 3 to 4 effects and all the feedwater is distributed
over the tube bundles located in the first effect group while the brine discharging from one effect group is passing as feedwater to the tube bundles of
the next effect group, which may be primarily applied for HT-MED configurations with more than 4-6 effects.
 forward-split feed – wherein only a fraction of the total feed is distributed over the tube bundles in the first effect and the discharging brine from one
effect is passing as feed water to the tube bundles of the next effect like in the forward feed configuration, while the remaining feed water is
distributed by mixing it with the brine discharging from individual effects. Such mixing may be made after each effect or at each first effect of several
groups of effects as applied for the VT-MED described in this paper.

Further, the design wetting rates in HT-MED evaporators may depend on factors like tube diameter, tube length, number of tubes, number of tube columns
and tube rows, while similar influences are applicable in the design of VT-MED evaporators. However, wetting rates may stay in general within certain limits.
For example, three systems are compared. Those are a HT-MED designed as MED-TVC with parallel feed similar to the Yanbu-II system, a HT-MED with top
temperature 112C and a mixed forward-parallel feed configuration comprising 4 effect groups and the VT-MED as described in this paper which is based on
the forward-split feed configuration, wherein 50% of the feed is entering into the tube bundles of the top effect and 16.7% are mixed into the brine
downstream of the 6th, 12th and 18th effect. For all systems a recovery ratio of 30-33%, a seawater salinity of 45,000 ppm and a resulting brine concentration
of 65,000 – 67,000 ppm are considered, which results with assumed tube bundle geometries in the following approximate design values:

for the HT-MED – parallel feed: wetting rate at top tube rows in all effects  = 0.060 kg/m/s @ 45,000 ppm
wetting rate at bottom tube rows in all effects  = 0,042 kg/m/s @ 65,000 ppm
for the HT-MED – mixed forward-parallel feed: wetting rate at top tube rows – first effect group  = 0.060 kg/m/s @ 45,000 ppm
wetting rate at bottom tube rows – first effect group  = 0,055 kg/m/s @ 47,500 ppm
wetting rate at top tube rows – last effect group  = 0.046 kg/m/s @ 58,500 ppm
wetting rate at bottom tube rows – last effect group  = 0,042 kg/m/s @ 65,000 ppm
for the VT-MED – forward-split feed: wetting rate at tube inlet of 1st effects  = 0.400 kg/m/s @ 45,000 ppm
wetting rate at tube exit of 1st effect  = >0,380 kg/m/s @ <47,500 ppm
wetting rate at tube inlet of last effects  = >0.400 kg/m/s @ <65,000 ppm
wetting rate at tube exit of last effect  = >0,400 kg/m/s @ 65,000 ppm

In all 3 examples the design film concentration would increase from 45,000 ppm to about 65,000 ppm. The wetting rates would be in the VT-MED about 6-8
times higher compared to the wetting rates in the HT-MED systems. To compare the possible variations of local falling film concentrations in operation, the
theoretical average concentrations on the bottom of the tube bundles are shown in FIG 9b which are based on an even distribution of all feedwater in
individual effects or group of effects. While the bumps in the characteristic of the HT-MED with mixed forward/parallel feed are a result of the grouping of
effects, the bumps in the characteristic of the VT-MED are resulting from the feedwater mixing with the brine at the effects 6, 12 and 18.

PAGE 14 of 19
 130,000

120,000

110,000

100,000
FALLING FILM CONCENTRATION ppm

ANHYDRATE
90,000
HEMI HYDRATE
GYPSUM
80,000 HT-MED 67C - parallel feed
HT-MED 112C - forward parallel feed
VT-MED
70,000

60,000

50,000

40,000

30,000
30 40 50 60 70 80 90 100 110 120

OPERATION TEMPERATURE C

FIG 9b theoretical average brine film discharge concentrations in different systems

Since in an actual evaporator design respectively in actual operation a perfectly even distribution of the feedwater is not possible, the film concentrations may
deviate locally from the average design film concentrations shown in FIG 9b. Another factor influencing the local wetting rates are variations of heat transfer
or heat flux in different regions of the tube bundles. This influence however has been neglected in the following comparison.
Assuming a distribution factor in the range of about 0.50 to 1.00, wherein the distribution factor 1.0 would mean that the local wetting rate on top of the tube
bundle would be equal to 100% of the design value, while a factor 0.50 would mean that the actual local wetting rate is only 50% of the wetting rate at ideal
even distribution. Assuming a uniform rate of evaporation from the top to the bottom and over the length of the tubes in a tube bundle, the resulting wetting
rates and falling film concentrations on the bottom of the tube bundles would vary for individual systems approximately as shown in FIG 9c. Towards lower
distribution factors in the range of 0.50, the design wetting rates for the HT-MED systems are dropping towards values below  = 0.03 kg/m/s (solid lines on
the bottom), while the wetting rates for the VT-MED would stay with a distribution factor of 0.50 still at about 0.18 kg/m/s (green solid line). As a result, the
brine film concentration in the VT-MED would reach with a distribution factor of 0.50 a concentration of about 50,000 ppm in the top effect (lower green dotted
curve) and <70,000 ppm in the bottom effect (upper green dotted curve). In the HT-MED with parallel feed concept, a distribution factor of 0.50 could result in
a film concentration between 110,000 and 120,000 ppm (red dotted curve), which would be close to the gypsum curve (upper solubility limit shown in FIG 9b
on the top). In the HT-MED with mixed forward-parallel feed a distribution factor of 0.5 would result in the top effect group in a film concentration of > 100,000

PAGE 15 of 19
ppm (lower blue dotted curve) and above 150,000 ppm in the bottom effect group (upper blue dotted curve), which means that in the top effects as well as in
the bottom effects, solubility limits as shown in FIG 9b would be exceeded.

140,000 0.50

130,000 0.45

120,000 0.40

film conc ‐ at HT‐MED ‐ parallel feed

BRINE FILM CONCENTRATION     ppm

110,000 0.35
film conc ‐ at VT‐MED ‐ first effect

wetting rate   kg / m s
100,000 0.30 film conc ‐ at VT‐MED ‐ last effect

film conc ‐ at HT‐MED ‐ mixed feed ‐ top effects

90,000 0.25
film conc ‐ at HT‐MED ‐ mixed feed ‐ bottom effects

80,000 0.20 WR ‐ at HT‐MED ‐ parallel feed ‐ bottom tubes 

WR ‐ at VT‐MED ‐ first & last effect bottom

70,000 0.15
WR ‐ at HT‐MED ‐ mixed feed ‐ top effects
60,000 0.10
WR ‐ at HT‐MED ‐ mixed feed ‐ bottom effects

50,000 0.05

40,000 0.00
0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00
DISTRIBUTION FACTOR

FIG 9c change of wetting rates and brine film concentration relative to distribution factors

Tentatively this comparison may show, that allowable deviations from a design wetting rate are very limited in HT-MED systems, while VT-MED systems may
have tentatively higher design wetting rates and may therefore be less sensitive to deviations from the design wetting rates.

PAGE 16 of 19
3.2.3 DESIGN FOULING FACTORS AND THE INFLUENCE ON PRODUCTION AND PERFORMANCE RATIO

Significant experience is available from MSF-plant operation in regards to controlling the fouling rates of heat exchange surface areas with suitable chemicals
together with ball cleaning systems. Even so, the performance ratios of MSF plants can be increased with periodic acid cleaning, the fouling rates are
increasing after a cleaning cycle relatively fast back toward the design fouling rates. A more permanent reduction of the scaling of the heat transfer areas
would be theoretically possible with a pre-treatment of the make-up water by nano-filtration, which would be equally applicable to HT-MED and VT-MED
evaporators. Theoretically such pre-treatment would also allow to increase the top brine temperature significantly above 112C. However, since the increase of
the top brine temperature requires in the design of evaporators special attention in regards to increased operation pressure and a higher potential of material
corrosion, which can have a significant impact on the capital cost, the comparison described here is limited to a first step, wherein a MSF system and a VT-
MED are compared in regards to change in performance ratio and distillate production as a result of a NF-pre-treatment, while leaving on both systems the
top temperature constant at 112 C.

Both systems would have in the basic design without NF pretreatment a unit capacity of about 33 migd (approximately 150,000 m3/day). The MSF unit
applied in the comparison is a type MSF-C unit with a performance ratio of about 15.7 kg/2326 kJ [11] when operating without any pre-treatment, while the
VT-MED as described in this paper would have a performance ratio of about 18.7 kg/2326 kJ without pre-treatment. For the recovery section of the type MSF-
C evaporator a design fouling factor of 0.000100 K m2/W is assumed, while for the feed heaters of VT-MED the same fouling factor is assumed as for the
MSF-C, while for the falling film tube bundles a design fouling factor of 0.000150 K m2/W is assumed.

In a first theoretical calculation step a comparison is made, assuming that the TDS of the feedwater of the VT-MED respectively the make-up water of the
MSF system would be reduced by the NF from 45,000 ppm to 34,650 ppm or about 23%, while the fouling factors for the heat exchange surfaces would
remain unchanged. The resulting change of performance ratio and distillate production are illustrated in FIG 10 with the changes from point 1 to point 2. In a
second theoretical calculation step it is assumed that with the NF pretreatment the design fouling factors can be reduced toward 66.7 % of the above
mentioned design values. The changes of distillate production and performance ratio resulting from the reduced fouling rates are illustrated with the curves
from point 2 to 3.

The first theoretical step, which considers only a reduction of the TDS, reduces the boiling point elevation, which results at both systems in a small increase of
the performance ratio and the distillate production. The assumed reduction of fouling rates has similar influences in the performance ratio of both systems,
while it has substantially no influence in the distillate production of the MSF system, but increases the distillate production of the VT-MED by more than 30%.

Reasons for the differences in the characteristic changes between the MSF and the VT-MED are, that the reduction of the fouling factor in the MSF is
resulting primarily in a reduction of the approach temperatures on the tube bundles and a resulting reduction of the temperature difference between brine
heater inlet and outlet, which reduces the consumption of heating steam and increases as such the performance ratio. Since the total flash range in the MSF
is remaining substantially unchanged, there is no increase in the distillate production.

The feed heaters of the VT-MED have substantially the same characteristic as the MSF system. The amount of vapor condensing on the individual feed
heaters is substantially equal to the vapor released by the flash down of the feedwater (brine) and distillate in each effect, like in an MSF. The reduction of the
fouling factor in the feed heaters results like in the MSF in a reduction of the heating steam required in the top feed heater or in the top effect to bring the

PAGE 17 of 19
feedwater to the top temperature, which increases the performance ratio. The reduction of fouling factor results on the heat exchange surface of the effect
tube bundles in a higher heat flux rate, so that more vapor can condense on the outside of the tubes while more vapor is released from the falling film on the
inside of the tubes.

240,000 21.0

230,000 20.5
3
220,000 20.0

210,000 3 19.5

performance ratio     kg/2326 kJ
distillate production      m3/day

200,000 2 19.0

1
190,000 18.5

180,000 18.0 VT‐MED Prod

MSF‐C Prod
170,000 17.5 VT‐MED PR

3 MSF‐C PR
160,000 2 17.0

3 2
150,000 1 16.5
2
140,000 16.0

1
130,000 15.5

120,000 15.0
65.0 70.0 75.0 80.0 85.0 90.0 95.0 100.0 105.0

% of desing  fouling rate

FIG 10 change of performance ratios and production over fouling factors

This results in the shown increase of the distillate production. Looking into options in the system design to benefit from this advantage, one option when
considering to maintain the design output of 150,000 m3/day would be, to reduce the tube length per effect by about 0.40 meters which would reduce the
required heat exchanger surface area by about 25% and the evaporator by about 15%, compared to values shown for the basic design concept of the VT-
MED in FIG 5 and FIG 6.

PAGE 18 of 19
4 SUMMARY

Considering that VT-MED evaporators with top temperatures in the range of 112 C have been designed and tested already more than 50 years ago [1] and
for large unit capacities, the vertically stacked configuration has been considered for a long time as the most economical configuration [5, 6, 7], the described
preliminary design concept of a VT-MED may be a competitive alternative for the future desalination market. While key parameters like SAEC, specific heat
transfer areas and specific evaporator volumes described in this paper are indicating certain advantages over other thermal desalination systems, a more
detailed analysis may be required to compare tentative CAPEX and OPEX for such VT-MED systems to other thermal desalination concepts currently under
development.
Further it may be evaluated how far CAPEX and OPEX of such systems could be reduced if components like nano-filtration (NF) and/or adsorption cycle (AD)
would be added.
Last not least, it may be evaluated, how such system would fit into a future power & water production system in regards to operation parameters, flexibility in
operation cycles and plant layout.

5 REFERENCES

[1] OSW Report No. 559 – Freeport Test Facility and Vertical-Tube Evaporator Test Bed
[2] BP Engineering – Industrial Falling Film Evaporators http://www.engineering-bp.com/en/equipment-and-machines-for-production-of-juices-and-
concentrates/concentration-section/falling-film-evaporators
[3] GEA – Falling Film Evaporators https://www.gea.com/en/products/falling-film-evaporator.jsp
[4] SUEZ – RCC - Zero Liquid Discharge – Technical Paper TP1044EN.pdf
[5] Metropolitan Water District – Preliminary Design Report No. 1084 – VTE Evaporator 75 mgd – GOR 23
[6] Institute of Nuclear Energy Technology (INET) Tsinghhua University, Beijing, PR China, International Symposium on Peaceful Applications of Nuclear
Technology in the GCC Countries, Jeddah 2008
[7] Conceptual Design of a 250 mgd Vertical Tube Evaporator for Seawater Desalination – September 1967 - Oakridge National Laboratory – Doc. No.
ORNL-TM-1901
[8] Operation of the Multi-Effect Multi-Stage Flash Desalination Plant – San Diego / California – April 1971 – 0SW Report No. 668
[9] Specific Allocated Energy Consumption of MSF and MED Plants – IDA Congress – Sao Paulo 2017 – by F. Alt – Doc No. IDA17WC-57912_Alt
[10] The Potential of Traditional MSF-Technology – IDA Congress – Dubai 2009 – by F. Alt – Doc No. DB09-044
[11] DTRI 171208 - Future Thermal Desalination Systems – Revison A.pdf, provided to SWCC in December 2017
[12] Engineering Handbook III – Chapter 14, Wolverine Tube Inc.
[13] Improvement of Tube Wetting by Wetting Agents in Mulitiple-Effect Distillers – IDA World Congress 2015 – by Heike Glade – Ref.
IDA15WC_Glade_51526

PAGE 19 of 19