-
A NANO FILTRATION (NF) MEMBRANE PRETREATMENT OF SWRO FEED AND
MSF MAKE-UP
(PART I)
A. M. Hassan, M. AK. Al-Sofi, A. M. Farooque, A. G. I. Dalvi, A.
T. M. Jamaluddin, N. M. Kither, A. S. Al-Amoudi, and I. A. R.
Al-Tisan
Research and Development Center,
Saline Water Conversion Corporation P.O. Box #8328, Al-Jubail
31951, Kingdom of Saudi Arabia
For the first time ever, a new approach to seawater desalination
processes was
developed at Saline Water Conversion Corporation (SWCC), R&D
Center, by
integrating the nanofiltration (NF) membrane pretreatment
process with one of the
conventional desalination processes to form, for example, an
NF-SWRO or an NF-MSF
or a combination thereof, such as an NF-SWROreject-MSF. The
process was successfully
applied to those cases on a pilot plant scale with remarkable
results. The seawater
treatment first with the NF membrane removed from it turbidity
and microorganism,
caused significant rejection of the scale forming hardness ions,
e.g., SO4= by up to 98%,
reduced TDS in Gulf Seawater by up to 65 %, and produced a new,
partially
desalinated seawater product, considerably different and
superior to seawater in
qualities and without the problems normally associated with
seawater of high
concentration of scale forming ions, high TDS, high turbidity
and high bacteria count.
The said desalination arrangements led to a significant
improvement in the seawater
desalination processes, for example by doubling the SWRO product
water output and
recovery ratio and the production of high purity permeate (TDS
< 200 ppm) from one
single stage SWRO. It also allowed for the successful operation
at high recovery ratio
of the MSF as part of an NF-MSF unit or as part of a trihybrid
NF-SWROreject-MSF
desalination system, where the reject from SWRO constituted the
make-up to the MSF
unit, at a top brine temperature (TBT) of 120 oC, without the
addition to the MSF make-
up of acid or antiscalant or antifoam and its operation under
those conditions without
scale formation. Moreover, by this trihybrid process, up to 90%
of the NF permeate was
converted to potable water. In addition to the above findings,
the report describes the
* Issued as Technical report #3807/98008-I Part of the work
titled A New Approach to Membrane and Thermal Seawater Desalination
Processes Using Nanofiltration Membranes Part 2 was presented at
4th WSTA Conference held at Bahrain, Feb. 13-17, 1999.
-
effect of long term operation on the performance of those
seawater hybrid desalination
processes.
As for the execution of the project, to be referred to hereafter
as Part I, it was started on
22/3/1997 and after seven weeks of system modification and dual
media cleaning to
remove the residual Fe3+ (from its media), the nanofiltration
(NF) operation, using one
module with two NF membrane elements, was started on 10/5/1997
followed by the
installation and operation of 5 NF modules (10 elements) on
2/6/1997. The NF-SWRO,
NF-MSF and NF-SWROreject-MSF pilot plants were first operated on
3/6/1997,
21/9/1997 and 10/2/1998, respectively. So far the NF pilot plant
unit has been in
operation for more than 9700 hours, while the SWRO and MSF pilot
plants were
operated on NF product as feed to the former and make-up for the
latter for over 5000
and 2300 hours, respectively. The hybrid system
NF-SWROreject-MSF was operated for
a total of 170 hours (further operation shall be conducted at a
later date). Continuous
non-interrupted operation of the NF unit without shutdown was
maintained from the
start of the project until now. The execution of the project
proceeded at a fast pace
ahead of schedule and all tasks were completed on time. Results
were presented in a
paper to IDA World Congress, Madrid 97 [1]. A series of lectures
on the same topic
were presented at SWCC. A two parts paper was published in the
International IDA
D&WR Quarterly Magazine, May & September issues, 1998,
on request of the editor
[2]. Two additional papers were presented at EDS Conference,
Amsterdam, 1998, and
both papers were published in Desalination 118 (1998) 35-61 and
123-129 (3, 4). One
paper was accepted for presentation at the WSTA 4th Gulf Water
Conference, Bahrain,
Feb. 1999. Three paper abstracts were submitted to organizers of
IDA conferences on
desalination, San Diego, 1999, and the process is now in a
patent pending status in
USA, Saudi Arabia and the International Patent Cooperation
Treaty (PCT) Countries.
In view of the positive results obtained in this part of the
R&D investigation, it is
recommended to continue work in this new field of seawater
desalination. Summary of
an eight part proposed R&D work is covered under
Recommendation, Section 9, Page
(26) of this report.
-
INTRODUCTION
With an increase in world population and rise in their living
standard, there is an
increase in demand for good quality water. To meet this rise in
demand, water treatment
in all its forms, is also on the rise. This tends to be the case
also for water desalination
which has raised within the last two decades by over twenty
folds and by the end of
1995 it stood at 20,300,000 m3/d [6]. Most of this world
capacity is made of seawater
desalination by the MSF process (48.15%) and to a lesser extent
by the RO desalination
process (35.9%). This latter number represents all forms of RO
desalination: seawater
RO (SWRO), Brackish water RO (BWRO), industrial and membrane
softening
processes.
By comparison to other forms of water desalination, seawater
desalination is by far the
most complicated and complex process. It has the lowest water
recovery ratio (30% to
35%). Operation is restricted to certain operation conditions.
Moreover, it tends to
require extensive pretreatment, especially if the feed is taken
from an open seawater
intake. The process is an energy intensive process, and for all
the above factors the
seawater desalination is the most expensive among all
desalination processes. The
major cause for the high expense and process complexity is the
seawater itself which is
characterized by having: (1) high degree of hardness, (2)
varying degrees of turbidity,
and (3) high TDS at pH of about 8.2. These seawater properties
give rise to three major
problems in seawater desalination which exert severe limitation
and have pronounced
effects on the performance and productivity of seawater
desalination plants.
As the world leader and the major producer of desalinated water
from the sea by both
the multi stage flash (MSF) and seawater reverse osmosis (SWRO)
processes, SWCC
has a major interest in both the thermal and membrane
technologies. SWCC has a keen
interest in solving the above three problems. For the last ten
years, SWCC R&D Center
was engaged in finding solutions to those problems. First work
dealt with studying the
existing SWCC SWRO plants by establishing their performance and
identification of
major problems encountered in their operation [7]. This was
followed by a meticulous,
systematic study of factors affecting the seawater feed
pretreatment and investigation
of the various parameters affecting the conventional coagulation
filtration process [8].
The
-
study led to the use of membranes in the pretreatment of
seawater feed to SWRO plants
[9], first by using the ultrafiltration (UF) membrane process as
a secondary pretreatment
to SWRO feed along with the evaluation of other microfiltration
(MF) and
nanofiltration (NF) membranes for pretreatment of feed to SWRO
plants. The UF
experiment proved the effectivity of this membrane process as a
secondary
pretreatment. Like the MF membrane pretreatment, the UF membrane
pretreatment
succeed only in the removal of turbidity and bacteria from the
feed, but kept the ionic
composition of seawater the same. The NF membrane pretreatment,
however, was
found to be successful in removal of turbidity, in significant
removal of hardness and in
lowering of the seawater TDS [1-5]. With this pretreatment SWCC
which introduced
during the late 1980s the new hybrid concept by the addition of
the two 15 mgd Jeddah
SWRO plants to the existing MSF - Power facilities, introduced a
new concept to
seawater desalination by combining the NF membrane process with
one or more of the
conventional seawater desalination processes in one fully
integrated process system. In this process, the NF product
constitutes the feed to SWRO plants or the make-up to thermal,
e.g., MSF, VC, ME, desalination plants. Alternatively, in a
trihybrid system,
the reject from SWRO of the NF-SWRO unit is used as the make-up
to the MSF unit. This
concept was evaluated on an NF-SWRO, an NF-MSF and an
NF-SWROreject -MSF pilot
plant units using Gulf seawater. From the results obtained so
far together with
preliminary technoeconomic and cost analysis performed on the
system, the process
was found to merit further consideration and evaluation on a
demonstration plant, which
is currently ongoing, and on commercial SWRO plants as well as
its application to
thermal seawater desalination processes. The report describes
the long-term operation
and results obtained so far on pilot plants scale (capacity 20
m3/d), using 4" x 40"
commercial NF membranes with this new concept of NF-seawater
desalination
processes, which is now in a patent pending status. 2-SOME MAJOR
PROBLEMS IN SEAWATERDESALINATION &THERREMEDIES
As mentioned under Introduction, by comparison to other forms of
water desalination,
seawater desalination is the most complicated, complex and
expensive water
desalination process. The major cause for the high expense and
process complexity is
the seawater itself which is characterized by having high degree
of hardness due to
-
presence of Ca++, Mg++, SO4= and HCO3 ions at relatively high
concentrations, high
TDS, varying degrees of turbidity, presence of micro particles,
macro and
microorganisms at pH of about 8.2. For illustration,
compositions of Gulf and normal
(ocean) seawater are given in Table 1. These extreme seawater
characteristics give
rise to three major problems in seawater desalination, which are
discussed along with
their remedies in the following sections.
Problems in Seawater Desalination
2.1.1 Problem 1: Seawater Hardness and its Effect on Seawater
Desalination
The high degree of hardness constitutes a problem inherent to
all forms of desalination,
be it thermal (MSF, ME, VC) or membrane type (SWRO, seawater
ED). The seawater
desalination processes are separation processes in which fresh
water is extracted from
saline water. This way the salts and hardness ions are left
behind in the brine with the
effect that both the brine TDS and hardness concentrations are
increased. Because
hardness ions are sparingly soluble in seawater, the increase in
their concentration and
under certain operation conditions could lead to their
precipitation on the desalination
equipment, e.g., tubes and membranes causing them to scale.
Depending on the desalination operating conditions, two types of
scale could form: an
alkaline soft scale made of CaCO3 and Mg(OH)2 and a non-alkaline
hard scale
consisting of CaSO4, or CaSO4.H2O or CaSO4.2H2O. The formation
of non-alkaline
scale becomes exaggerated at temperature above 120 oC, since the
CaSO4 solubility
decreases as the solution temperature is increased. To prevent
and avoid alkaline scale
formation certain antiscalant additives are added to the feed.
For example, in thermal
desalination processes such as MSF a scale control chemical is
added to the make up
seawater to allow for MSF operation without scaling. There are
three classes of such
scale control chemicals. These chemicals are: (i) polyphosphate
(ii) polymeric type
chemicals, e.g., polyphosphonates or polycarboxylic acid or
(iii) mineral acids (H2SO4
or HCl). In the same order as above they are dosed at the rate
of 3-5, 1-4 or (110-120)
parts per million (ppm) for operation at top brine temperature
(TBT) of 90 oC, 115 oC
and 120 oC, respectively. In spite of this, the product water
recovery as a fraction of
product to make-up remains low, 30% to 35%. Operation at higher
TBT requires
removal from seawater of sulfate or Ca++, which was done in the
past by the ion
-
exchange process [10]. This way higher water recovery and more
importantly at a
reduced unit production cost could be achieved. In SWRO
operation, acid and in some
cases SHMP or other antiscalants are normally added to prevent
membrane or SWRO
plant scaling. Again, water recovery tends to be limited, e.g.,
for Gulf seawater to about
35% or less.
2.1.2 Problem 2: Impurities and marine organisms and their
Effect on Seawater Desalination
Another problem in seawater desalination is the impurities in
seawater feed to the
desalination plants. Their presence in seawater feed could and
many cases cause the
desalination system to foul. Thus, the presence of
macroparticles and macroorganisms
(mussels, barnacles, algae) requires their removal from feed to
both SWRO and thermal
desalination plants. In presence of bacteria in the feed,
disinfection of the feed tends to
be a requirement also for both the SWRO and the thermal
processes. Removal of
turbidity and fine particulates, measured as total suspended
solids (TSS), from feed
destined to SWRO plants is an essential process but is not
required for the thermal
processes. Although, as illustrated in this study, removal of
particulates from the make-
up to MSF plant reduces foaming and thereby eliminates the
addition of antifoam
chemicals. Removal of Cl2 from feed to chlorine sensitive SWRO
membrane is also a
must.
2.1.3 Problem 3: Seawater TDS and its Effect on Osmotic Pressure
and SWRO Desalination
A third problem in seawater desalination is the seawater feed
high TDS. This
constitutes a major problem to the SWRO process but not as much
to the thermal
processes. However, an increase in seawater TDS is always
accompanied by an increase
in hardness which as mentioned above in Section 2.1 constitutes
a problem to thermal
desalination processes i.e., by limiting distillate recovery.
Per se, the TDS without or
with very low concentration of hardness ions has a minimum
effect on distillate
recovery in the thermal desalination process. This is
demonstrated later in the text,
section 5.3 where high degree of distillate recovery was
achieved from high TDS feed
from which a significant portion of the hardness ions up to 90%
were removed.
Nevertheless, increased TDS hence elevated brine recycle
salinity would influence
-
distillation processes. The increased brine recycle salinity
would give rise to solution
specific gravity (SP) and boiling point elevation (BPE). These
in turn could marginally
give rise and in the same order as they appeared above to
pumping power plus heat
input and distillate production due to increases in required
differential temperatures and
pressures between flash zones and condensing zones.
The feed osmotic pressure increases as the ionic molar
concentration and, therefore,
TDS in seawater are increased. The feed osmotic pressure f in
bar is calculated from
the following equation:
f = 8.308 x 0.9 (Tf + 273.15) Mi / 100.0195 (1)
where Tf is the feed temperature and Mi is the sum of ionic
concentration in moles
while the factor 100.0195 is to convert kilo pascal to bar [11].
For example f (of feed)
is 31.08 compared to 27.03, 24.94 and 14.19 bars for Gulf,
Mediterranean, Open Ocean
(normal) and Baltic seawater with TDS of 43800, 38,088, 35145
and 20,000 ppm,
respectively. The average decline in f is 0.70968 bar per 1000
ppm drop in seawater
TDS. The ratios of f value for seawater from Baltic sea to that
from normal sea and to
f of Gulf seawater are: 0.4566: 0.8024: 1.0, respectively.
From the principles of SWRO, the applied pressure Pappl is
necessarily used to overcome
the osmotic pressure and the remaining pressure is the net water
driving pressure
through the membrane (Pnet). Thus, the net driving pressure can
be represented simply
as:
Pnet = Pappl - (2)
Where Pappl and are the differential applied and osmotic
pressures across the
membrane, respectively. Both Pappl and are calculated from the
following equation
[].
Pappl = Pf - (Pfb / 2) - Pp (3)
= fb - p (4)
-
Thus, Eq. 2 can be rewritten as:
Pnet = [Pf - (Pfb / 2) - Pp] - [fb - P] (5)
where the subscripts f, b and p denote the feed, brine and
product, respectively.
The fb in bar is given by the following equation:
fb = [0.2654 Cfb [T + 273.15] / [1000 - Cfb / 1000] [100.0195]
(6)
The concentration Cfb is the average concentration of feed and
brine:
Cfb = (Cf + Cb) / 2 (7)
While Cb in terms of the recovery ratio is given as:
Cb = Cf / [1 - ] (8)
The product water quantity (Qp) or the rate of permeate passage
through the membrane
is given by:
Qp = Kw (P - ) (A / ) Tc Mf = Kw Pnet (A / )Tc Mf (9)
where
- Kw is the membrane permeability coefficient
- A is the membrane surface area
- is the membrane thickness
- Tc and Mf are the correction factor effects for temperature
and membrane
flux, respectively.
The salt quantity in permeate (Qs) is proportional to membrane
salt permeation
coefficient (Ks) and the salt concentration differential across
the membrane (C) and is
defined by the following equation:
-
Qs = Ks (Cfb) (A/) (10)
From Eqs. (2) And (5), the less is the osmotic pressure , the
greater is the Pnet and the
greater is the amount of pressure available to drive the
permeate water through the
membrane and the greater is the quantity of product. This is
apparent in Eq. 9 as Qp and
Pnet are directly proportional. Likewise, the process energy is
directly related to
pressure as shown later by Eq. 11. Unlike permeate flow, the
salt passage through the
membrane (Qw) is independent of pressure and depends on the salt
concentration
differential across the membrane. Therefore, raising the RO feed
pressure increases
water flow and normally without increasing salt passage which
leads to a decrease in
permeate TDS, i.e., it leads to salt dilution in permeate.
The effect of varying feed TDS on fb and Pnet in the SWRO
process at an applied
pressure of 60 bars and final brine TDS of 66,615 ppm is shown
in Figure 1. The
available useful pressure to drive the water though the membrane
(Pnet) marked by the
shaded area increases as the feed TDS and therefore fb is
decreased and vise-versa.
The fraction of the Pappl which equals to fb is considered to be
a wasted energy
although it is necessary in the SWRO process. Since the permeate
flow through the
membrane is directly proportional to the Pnet, therefore, any
process that lowers the feed
TDS not only reduces the wasted energy but it increases the
fresh water permeation (Qp)
through the membrane (Eq. 9) and decrease the permeate TDS. As
shown later, this
case of increasing Pnet by lowering TDS of feed by NF
pretreatment of feed is being
exploited with advantage in the combined NF-SWRO process by
increasing at an
applied low pressure the permeate flow and reducing their TDS
(see Section 5.2).
Remedies of Problems in Seawater Desalination
2.2.1 feed pretreatment in conventional seawater desalination
plants as means of improving feed quality
The above problems in seawater desalination caused by the
seawater properties and
measures used to alleviate them are summarized in Table 2 along
with the quality
requirements of feed to SWRO plant and make-up feed to MSF
plants when the feed is
taken from an open sea (surface) intake. In the past it was
demonstrated that use of
beachwell can effectively remove turbidity and improve SWRO
plant performance. The
-
same effect can be achieved by using ultrafiltration (UF)
membranes as a secondary
pretreatment to the SWRO coagulation filtration process [9]. Use
of antiscalants proved
effective in preventing scale formation but failed to increase
significantly water
recovery in both the membrane and the thermal processes. Simple
water filtration,
which is employed in some brackish or drinking water treatments
or coagulation -
filtration processes are employed in the removal of fine
particles. This process,
however, does not remove the very fine particles with sizes of
less than 1 to 2
micrometer.
2.2.2 Application Of Membrane Filtration In Treatment Of Feed To
Brackish And Low Salinity Water (Drinking & Waste Water)
For the removal from the feed of fine particles with sizes less
than 1 micrometer,
microfiltration (MF), UF, NF and hyperfiltration/reverse osmosis
membrane filtration
are employed. The MF is used for particle separation having
sizes in the range of 2.0
down to about 0.08. The UF membrane process is suited for the
separation of finer
particles having sizes in the range of 0.1 down to 0.02 and of
molecular weight
(MW) in the range of 10,000 g/mole and above. On the other hand,
the RO process,
which is based on the well-known RO desalination principle,
deals with separation of
ionic size particles in the range of 0.001 or less, molecular
weight 200 g/mole or less.
The NF Membrane falls in-between the RO and UF separation range,
and is suited for
the separation of particle sizes in the range of 0.01 to 0.001,
MW of 200 and above.
The rejection by NF, however, is based on three principles. The
rejection of neutral
particles is done according to their sizes, as is the case with
the MF and UF membranes.
Rejection of part of the ions and the very fine particles is
done by the reverse osmosis
process. Like the RO membranes, the permeate flow through the NF
membranes
increases, as illustrated in Section 5.1.2, as the Pnet is
increased. Also, by a third
process, the ionic rejection of inorganic matter is achieved by
their electrostatic
interaction with the negatively charged membrane [12]. The
negatively charged
membrane repulses the anion increasing their rejection by the
membrane. To maintain
electro-neutrality in the solution the cations are also rejected
to the same degree as the
anion. For example, in certain membranes the rejection of
sulfate ions is in the order of
80-90% or better with a similar rejection for the cations Mg++
and Ca++. This compares
-
to 10-60% rejection for the Na+ and Cl- ions. Thus, the degree
of rejection by the NF
membrane is lesser for mono valent ions such as Cl- and Na+ than
that for the divalent
SO4= and Ca++. The above ion selectivity allowed for the use of
NF in the removal of
hardness from low salinity water. For this reason, the process
is gaining acceptance as a
major water softening treatment replacing the conventional lime
softening treatment.
This is most evident in the State of Florida where many of the
ground waters in coastal
areas are classified as hard water. Presently the total NF
plants capacity in Florida alone
is over 60 mgd and is on the rise [13]. In some of the plants
both NF & RO (hybrid) are
included to allow for greater removal of dissolved salts.
Moreover, use of NF serves in
the removal of hardness allowing for higher water recovery by
the RO plant. Not only
the NF process removes hardness but also the product water
quality was found to meet
or to exceed the drinking water standards.
Recently there is an increased interest in the application of NF
membrane filtration in
brackish water and drinking water softening, removal of color,
turbidy, removal of
dissolved organic which are precursors to disinfection
by-products (THM) [14&15].
The water utilities justify the slightly higher cost of membrane
softening over the lime
process by the superior water quality it produces especially
when the feed has high color
and other impurities. To meet the various Drinking Water Acts in
USA, UF and MF
membrane processes have been used also in the treatment of
drinking water [16,17&18].
The NF has been used in other applications to treat salt
solution and landfill Leachate
[19], demineralization of whey, removal of sulfate from seawater
to be injected in off-
shore oil well reservoirs [20,21 &22] oil water separation
[23], removal of natural
organic matter and precursors of disinfection by-products from a
highly colored ground
water. Cross-flow microfiltration have been used in the
treatment of industrial waste
water for the removal of toxic heavy metals and other suspended
particles [24&25]. It
was also used to achieve 95% water recovery, which is essential
in the treatment of
waste and industrial waters [26]. Use of NF filtration in
treatment of various low
salinity waters is described in the literature.
-
2.2.3 Membrane Pretreatment Of Feed To Conventional Seawater
Desalination Plants
Use of UF and MF membrane in the pretreatment of SWRO feed was
reported in earlier
work [9,27&28]. The MF pretreatment, as the only primary
treatment to seawater feed,
however, required membrane washing at a frequency of 30 to 40
minutes. It was
concluded that the membrane filtration pretreatment approach
could be suitable as an
alternative to the conventional coagulation-filtration method,
but recommended further
investigation [27& 28]. Use of UF as secondary pretreatment
at SWCC, however, gave
good results reducing the SDI of feed from the primary
pretreatment from 2.5 0.2 to
less than SDI of 1 for the feed receiving UF secondary
pretreatment. Moreover, the
differential pressure across the SWRO membranes remained steady
at less than 4 psi
during the 12,000 hours of SWRO pilot plant operation [26].
However, Use of NF
membrane as part of a combined NF-seawater desalination plant to
pretreat feed to
seawater desalination plants, e.g., SWRO, MSF, etc., or as a
secondary pretreatment to
remove hardness, to lower TDS, to remove turbidity, has not been
reported prior to this
work. This has been tried here for the first time in the
pretreatment of feed to SWRO
pilot plant and make-up to MSF pilot plant unit in the
hybridization process of NF-
SWRO, NF-MSF and NF-SWROreject -MSF [1-5].
A combination of simple dual media filtration (without
coagulation) with NF membrane
was found adequate not only in removal of turbidity and
improvement of seawater feed
quality but also performed the important role of reducing
seawater hardness and
lowering of its TDS.
Investigation of a new process utilizing NF in the pretreatment
of feed to SWRO
plant and make up to MSF.
Study the effect of NF pretreatment on removal of turbidity and
reduction of
seawater hardness.
Study the effect of NF pretreatment in improving the performance
of : (1) SWRO
and (2) MSF (in particular reduction of hardness and antiscalant
consumption)
Technoeconomical evaluation of the NF-SWRO Desalination
Process.
-
All experimental work were done on a pilot plant scale. A
schematic flow diagram of
the NF-SWRO pilot plant is given in Figure 2, while Figure 3
shows the integration
of the NF-SWRO with an MSF pilot plant distiller comprising 2
and 4 stages of heat
rejection and recovery, respectively. Moreover, this arrangement
allows for utilization,
especially in winter season, of the seawater from MSF heat
rejection section as feed to
the NF unit. The NF-SWRO pilot plant consists of seawater supply
system, dual media
filter followed by fine sand filter, 5-micron cartridge filter,
feed tank, the NF unit and
the SWRO unit. The SWRO and MSF pilot plant set-up excluding the
NF unit were
described in earlier work [8&29]. The NF unit consists of
the high-pressure pump and
NF modules each containing two membrane elements (size 4"x40")
which were bought
about four years back. The arrangement of the modules is as
shown in Figure 2 where
the feed is supplied to the first two modules arranged in
parallel and the reject of each is
fed to its following module which is connected to it in series.
Reject from the latter two
modules constitutes the feed for the final fifth module. The
SWRO unit is made of a
high pressure pump followed by six SWRO modules, each contains
one spiral wound
membrane element (size 2.5"x40"), all arranged in series as
shown in the figure.
After its filtration without coagulation the filtrate was passed
to the NF membrane under
pressure. This was followed by passing the product from the NF
unit to the SWRO unit
where it is separated under pressure of 40 to 60 bar into
product (permeate) and reject.
Alternatively, the NF permeate or SWRO reject can be fed to the
MSF unit as shown in
the Figure 3.
During the entire experiment, no coagulant was added to the NF
feed, which was
nonchlorinated seawater except in a very few occasions when the
seawater, due to
intake maintenance, was received from intake to Al-Jubail MSF
Phase-I plant. Chlorine
when present in the feed was removed prior to feed entry into
the NF membranes.
During the first 2200 hours of NF pilot plant operation no acid
was dosed in the feed.
Acid, however, was added to the feed, at the rate of about 25
ppm, to bring the feed pH
to 6.7 when the NF permeate recovery was raised above 50%. No
antiscal or acid,
however, were added to the feed to SWRO or make-up to MSF pilot
plants when they
were operated in the dual desalination system of NF-SWRO or
NF-MSF or the trihybrid
-
desalination system of NF-SWROreject-MSF. For comparison
purposes, however, acid
was added in some trials to MSF make-up.
Chemical and biological analyses were done according to latest
standard analysis
methods, which are already fully established for them at RDC,
chemical and biological
laboratories. These analysis were performed on a routine basis
and as required, for the
seawater feed, permeate, reject from NF and SWRO units as well
as for feed, make-up,
brine recycling and blow down in the MSF pilot plant.
The work was done in five steps: first by passing the feed to
the NF unit and only after
establishing its performance the product from the NF unit was
passed to the SWRO unit
(second step). The NF unit was operated continuously without
adding acid except
occasionally and for short time to reduce the feed pH to 7.0.
During the winter months,
to offset the effect of decline in NF feed temperature on NF
product flow, the NF feed
pressure was raised to reach 31 bars. The SWRO unit was operated
initially at pressure
from 56 bars reduced thereafter to 40 bars when the NF filtrate
TDS was reduced to
about 16,000 ppm by increasing the NF feed pressure. After
completing this phase, the
NF permeate from the NF desalination unit was fed to the MSF
unit (third step). The
MSF pilot plant was operated at 120oC, which is the temperature
operation limit of the
present MSF pilot plant, without addition of antiscalant or acid
or antifoam, which are
generally used to overcome the scaling problems and to prevent
product quality
degradation due to excessive foaming. The MSF pilot plant was
also operated on
SWRO reject from the NF-SWRO unit in an integrated trihybrid
NF-SWROreject -MSF
system (step 4). The effect of varying feed temperature &
pressure on NF performance
was investigated in step 5.
5.1 NF- Trials
Figure 4 shows the reduction in hardness ions in seawater
receiving NF pretreatment
utilizing FilmTec NF-70 membranes, while Figure 5 shows the
percentage reduction
for the same plus other ions (Cl-, Na+) TDS and conductivity in
NF permeate. Table 3
lists the concentration of the various seawater ions in Gulf
seawater before and after the
NF treatment along with their percent salt rejection. When the
seawater feed is passed
-
at a pressure of 18 bar through 5 NF modules, the average ion
concentration of Ca++,
Mg++, SO4= and HCO3 was 93, 193, 206 and 46 ppm, respectively,
while their average
salt rejection was to 80.7, 87.7 93.3 and. 63.3% and again in
the same order. Total
hardness was reduced by 86.5%. The SO4= ion concentration,
however, decreased with
operation time to about 65 ppm (rejection of 98%). By comparison
at the same applied
NF feed pressure, the concentration of the hardness ions of
Ca++, Mg++, SO4= and HCO3
in NF permeate when using one NF module was 63,105,55 and 37
ppm, respectively, as
compared to their concentration in seawater of: 481, 1608, 3200
and 128 ppm in the
same order. The rejections of those ions Ca++, Mg++, SO4= and
HCO3 from the feed
was: 87, 92, 98 and 71%, respectively. The M-Alkalinity (of NF
permeate) as CaCO3
was reduced from 45 ppm at a pressure of 18 bars without acid
dosing to the feed to less
than 25 ppm with acid dosing to pH of about 6.7.
In addition to the reduction of hardness ions by the NF
pretreatment, the Cl ions are
also reduced from 22,780 ppm in seawater feed to an average of
abut 16,692 ppm in NF
permeate or a reduction of about 26.7%. Similar reduction is
expected for the Na+ and
K+ ions. The net effect of this reduction by the NF treatment in
Cl , Na+ and K+ ions
together with the reduction in hardness ions caused a reduction
in TDS from 44,046
ppm in seawater to an average of 27,720 ppm for the NF
pretreated feed, i.e., a
reduction of 37.3%. Due to hardness ions reduction, the pH of
the feed of 8.2 is also
reduced to an average of 7.85 in the NF permeate.
Raising the NF feed pressure to 22 bar reduced the Ca++, Mg++,
SO4= and HCO3-
concentration in the filtrate from the 5 modules to 50, 96, 72
and 30 ppm, respectively,
for a remarkable rejection of 89.6, 94, 97.8 and 76.6%. Total
hardness was reduced by
93.3%. The Cl-, Na+ and TDS were also reduced, the latter from
44046 to 20230 for a
remarkable rejection of 54%, (Table 3). Moreover, reduction in
seawater hardness and
TDS the latter value by up to 63% was achieved by raising the NF
feed pressure to 31
bar.
Typical bacteria count, expressed as colony forming units/ml
(CFU), was as shown in
Figure 6. The CFU number in the NF permeate of 1.3 x 103 which
was observed on
26.7.97 is, unusually high and is not expected, since the
average size of bacteria is in the
order of 1 m or over and is much larger than the NF membrane
pore size of less than
-
0.01m. So, how can this relatively large size bacteria (i.e. 1m)
pass through such tiny
NF membrane of pore size less than 0.01m ? Moreover, the
differential pressure
across the SWRO membranes (P) remained steady and low in value
at less than 1 bar,
during the entire NF-SWRO experiment indicating no biofouling.
Also the NF
membrane autopsy and analysis, after 12000 hours of operation,
showed no biofouling.
The bacteria count on membrane surface was within the acceptable
limit, i.e., CFU
-
Figure 9 illustrates the NF permeate flow versus operation time,
for about 9700 hours
of operation. Figure 9a shows the operation conditions, while
Figure 9b, c and d
shows the actual and normalized permeate flow for feed flow,
temperature and pressure.
Data in Figure 9c are normalized only for temperature effect.
Figure 9a shows that
during the cold seasons, operation hours 6000 to 9000 hours,
adjustment of feed
pressure allowed for raising the product flow by off-setting the
decline in permeate flow
due to the seasonal lowering of feed temperature. Full
normalization of the data,
however, shows that a decline in flow occurred after the
cleaning of the NF membranes
by a commercial phosphate - based type detergent, DMCA-14/BIZ,
made by Cheyma
Inc., Monterial, Canada. Use of various cleaning and flushing
procedures failed to
improve the permeate flow to its level prior to cleaning. Use of
this detergent in NF
membrane cleaning should be avoided. However, after cleaning the
NF permeate flow
remained nearly steady at its level before cleaning (Figure 9b).
Prior to this cleaning the
membrane retention coefficient (MRC) calculated from Qt/Qi was
about 80%, where Qt
and Qi are the quantity of permeate flow at time t = 5000 hour
and initial time over the
first 100 hours.
5.1.2 Effect Of Operating Conditions On NF Membrane
Performance
The effect of feed applied pressure, flow and temperature at
different operating
conditions on NF permeate flow, recovery and conductivity are
shown by a family of
curves, one for each operating case, in Figures 10-13 Both
permeate flow and recovery
are noticed to increase as either or both the feed applied
pressure or temperature are
increased. On the average, this increase was about 6% and 3.4%
for a rise in applied
pressure by 1 bar and for a rise in temperature by 1C,
respectively. Increasing feed
flow had lesser effect on increasing permeate flow than that
occurred when increasing
either feed applied pressure or temperature, but it is the
recovery, which is markedly
increased with decrease in feed flow. For example, under same
operating conditions of
applied pressure of 37.5 bars, the permeate recovery at the
reduced feed flow rate of 20
l/min is 60% compared to only 43% when the feed flow rate was
increased to 33 l/min
(Figure 10). On the other hand, permeate conductivity tends to
decrease as the feed flow
rate or applied pressure or both are increased, while it tends
to increase as the feed
temperature is increased (Figures 10, 12 and 13).
-
In the NF membrane process, permeation of the NF permeate
through the membrane
occurs by two means, either by passage through the fine pores or
by the RO permeation
process. In both cases the permeate flow is dependent on the
applied feed pressure,
where up to the membrane flux limitation, the flow increases as
the pressure increases.
This is illustrated in Figure 11, where both the permeate flow
and recovery increased
while permeate TDS decreased as the applied pressure is
increased. As shown in
Figure 7, more specifically, the observed increase in permeates
flow and recovery is
due to the net driving pressure (Pnet). As illustrated by Eq. 9
shown also in Figure 11,
a good portion of the applied pressure is lost in overcoming the
osmotic pressure ().
For example, at the applied pressure of 37.9 bars, the Pnet, the
net driving pressure, is
only 12.6 bars for a loss in the applied pressure of 25.3 bars
(Figure 11).
From the above trials it can be concluded that the three
operating variables: feed
pressure, feed temperature and feed flow exert different
influence on NF recovery and
product quality. Increasing feed pressure increases both
permeate flow as well as
recovery and improves its quality. Improvement in permeates flow
and recovery can be
achieved also by increasing feed temperature which leads to a
moderate decline in
permeate quality. Increasing feed flow improves both permeate
flow and quality but it
has a marked influence on lowering permeate recovery. For proper
plant operation, by
optimizing NF permeate yield and quality, a balancing act of
operating the NF plant at
best values of feed: flow, temperature and pressure are to be
identified and selected.
Those operation criteria are being further investigated
thoroughly for the proper
operation of large NF plants by this NF-seawater desalination
process.
5.2 NF-SWRO Trials
During the early operation, from the start up to 2200 hours at
the applied pressure of 60
kg/cm2 and feed temperature of 33 1 oC, the conductivity of the
NF product, which
constituted the feed to the SWRO unit in the NF-SWRO hybrid
system, was 41000
s/cm. The SWRO permeate flow and recovery ratio were maintained
steady at 5 l/min
and about 50%, respectively. This is illustrated in Figure 14,
which also shows the
SWRO permeate flow, product recovery and conductivity plotted
versus operation time.
During the final phase of SWRO operation, because of a drop in
NF permeate
-
conductivity to about 25000 s/cm, the same above SWRO permeate
value and
recovery ratio were maintained at an applied pressure of only 40
bars instead of 60 bars
as shown in Figure 14. Operation at the normally applied
pressure of 60 bars increased
both the permeate flow and permeate recovery each by about 40
and 25%, respectively,
Figure 14. Most important, the SWRO membranes maintained, a
steady high
performance, which did not decline with operation time when
operation was done at
same operating conditions.
The effect of applied pressure on SWRO permeate flow, recovery
and conductivity are
illustrated in Figure 15, which for reasons of comparison shows
also the same for the
conventional operation of same SWRO membranes under identical
conditions but
without the NF pretreatment. Passage of the NF permeate to SWRO
unit under pressure
gave satisfying results with P remaining steady and constant at
2 bars during the entire
operation. As shown in Figure 15, because of the low hardness of
SWRO and TDS of
feed (see Table 3 for ions conc. at 31 bar) it is possible to
obtain a recovery of up to
80% when the pressure is raised to about 70 bars. This high
product recovery was
achieved with NF permeate compared to one half this value or
less for normally
pretreated seawater feed. The product flow (Qp) and recovery
ratio are also increased
directly with the applied pressure Figure 15. The SWRO permeate
from the combined
NF-SWRO desalination system were much greater than those for the
SWRO alone
when the two systems are operated at the same pressure and
temperature. For example,
at 40 bars the permeate flow and recovery from the conventional
SWRO are 1 l/min and
16.7%, respectively, as compared to a much higher flow and
recovery ratio of 4.8 l/min
and 48% from the new process of NF-SWRO, i.e., for an increase
of 480% in flow and
by 3 folds for permeate recovery. Even at the frequently
employed pressure of 56 bars
(800 psi) the SWRO product flow and recovery for NF-SWRO: SWRO
alone are in the
ratio of 2.43:1. Moreover, the quality of the permeate product
from SWRO process is
4500 s/cm at an applied pressure 40 kg/cm2 and drops to 2300
s/cm at 60 kg/cm2 as
compared to less than 500 s/cm for SWRO permeate from the
NF-SWRO process
obtained from the same membrane at the same pressure range. The
SWRO membranes
used in this test, Figure 15, are old membranes and the NF
pretreatment revives their
low performance. This process (NF-SWRO) is expected to extend
the life of an
otherwise ready to be replaced membranes.
-
5.3 NF-MSF Trials
In this trial NF permeate was used as make-up to MSF at a flow
rate of 1.5 m3/hour
replacing normal seawater at low concentration of scale forming
ions of alkaline and
non-alkaline types in the NF permeate to the MSF pilot unit as
shown in Table 3,
while Table 4 shows for comparison these concentrations in the
brine recycle stream
of the MSF operated with NF or SWRO reject from an NF-SWRO unit
as make-up or
seawater as make-up (conventional MSF). Also, listed in Table 4
are pH and
conductivity values of the brine recycle streams. From this
table, it can be seen that
scaling potential in the MSF system have been significantly
reduced and it was safe to
operate the MSF plant with NF make-up for over 2320 hours, or
with reject of SWRO
from the NF-SWRO unit for 270 hours at high temperature of 120
oC without addition
of antiscalant or antifoam chemicals. At the same operating
conditions, the
concentration of the scale forming ions of Ca++, and SO4= of 168
and 410 ppm in the
NF-MSF case, Table 4, and 232 and 1020 in the SWRO reject
make-up case, Table
4, are low when compared to 882, and 5830 ppm in the brine
recycle stream of
conventional MSF. These observation, especially the drastic
reduction in SO4=, Ca++
and Mg++, are encouraging to project MSF operation at higher TBT
in the range of 120 oC to 160 oC without inducing scaling [30] thus
improving plant production and hence
water cost. Operation of MSF plants at higher temperature should
increase the gain
output ratio (GOR) in Kgproduct/Kgsteam and the performance
ratio (PR) in Kgproduct/1000
Kj, while decreasing the energy consumption in Kj/Kg product
(Figure 16). Finally, the
MSF plant operation on NF make-up remained steady with operation
time.
The result obtained so far from the NF-SWRO pilot plant were
encouraging to consider
the application of the NF membrane process first in a
demonstration NF-SWRO plant,
which is being done now, to be followed by its actual
application to an existing
conventionally operated SWRO plant, which is now under
consideration. For the
present report the performance of SWCC Jeddah SWRO plant with
and without NF
-
unit was conceptually evaluated using the results obtained from
the pilot plant studies
described above [1-3].
When NF feed pressure was set to 31 bar the hardness ion
concentration in the NF
permeate, which constitutes the feed to SWRO plant was as shown
in Table 3. At an
applied pressure of 40 bar and SWRO product recovery of 50%, the
reject from the
SWRO contains low concentration of hardness ions of 96, 253, 410
and 42 ppm for
Ca++, Mg++, SO4= and HCO3-, respectively. The TDS of reject
brine of 30,640 ppm in
the same Table 3 is also low when compared for example to the
reject from Jeddah
SWRO plant of about 66,615 ppm at the applied pressure of about
60 to 65 bar
[31&32]. This suggests that higher recovery of more than 35%
can be achieved from
the Jeddah SWRO plant if it is modified to operate with an NF
pretreatment in a
combined NF-SWRO system. This is illustrated in Figure 17 which
is a schematic
flow diagram of the desalination part of Jeddah SWRO plant with
and without NF unit.
Figure 17a represents the actual Jeddah SWRO plant feed, product
and reject flows
along with the product water recovery percentage (%), the brine
flow/modules and the
energy required for the desalination part alone. Energy was
calculated from the
equation [33]:
Energy (KWH/m3) = [Qf . Hf / 366 Qpe] (11)
where:
- Qf and Qp are the quantity of feed and product in m3/hr,
respectively,
- H is the pressure head in (m),
- density of seawater (1.03), and
- e pump efficiency ( 0.85).
Figure 17b, and c simulates the results of operation of Jeddah
SWRO in a combined
NF-SWRO system utilizing the present SWRO desalination set-up as
it is now (case b)
and with reject staging (case c), respectively. The Jeddah SWRO
plant receives feed,
with TDS of 43,300 ppm from a conventional coagulation
filtration unit at the rate of
6760 m3/hr and produces from 1480 modules at an applied pressure
of about 60-65 bar
2370 m3/hr of fresh water for a product recovery of 35%. The
total quantity of reject is
-
4390 m3/hr with TDS of about 66,678 ppm. The product and reject
flow per
module/hour is 1.6 and 2.97 m3/hr, respectively. The energy
requirement for the SWRO
desalination part alone is 6.14 kwh/m3 of product, assuming
plant operation at 60 bars
and rises to 6.65 kwh/m3 at an operational pressure of 65
bar.
Each of the hollow fine fiber membrane modules used at Jeddah
contains two SWRO
membrane elements arranged in series with brine staging where
the feed is first passed
to the first set of elements (1st desalination step) and the
remaining feed after extraction
of a fraction of it as product is passed as the SWRO reject to
the second set of elements
constituting the 2nd desalination step which in turn extracts a
second fraction of product
(Figure 17). The ratio of product quantities extracted by the
first set of elements and the
second set of elements can be computed through Eq. 9 and the
following ratio equation:
Q setQ set
K P A T MK P A T M
pst
pnd
w appl fb c f
w appl fb c f
=
12
1
2
( ) ( / )( ) /
=
( )( )
PP
PP
appl fb
appl fb
net
net
=
1
2
1
2 (12)
where Pnet-1 and Pnet-2 are the net driving pressure force for
all the first set and second
set of elements (1 and 2), respectively. Also as demonstrated in
Figure 15, the product
flow (Qp) should increase as Pnet is increased.
To establish the value of Pnet-1 and Pnet-2 the recovery ratio
for the first set of elements is
to be established. For first approximation the recovery of first
set of elements is
estimated at 20% making the second set recovery equals 15% of
the 6760 m3/hr feed for
a total system recovery of 35%. With this first approximation of
recovery for the first
set of elements the water product recovery ratio of 1st set: 2nd
set of elements is 0.5705:
0.4295 as compared to different ratio of Pnet-1 : Pnet-2 of
0.6175 : 0.3825. The Pnet values
are calculated from Eqs. 5 and 6. Regression analysis lead to
the values shown in
Figure 13, where the product water ratio of 1st set: 2nd set of
elements is 0.62: 0.38 or
in the ratio of 1:0.61 which is not only identical to the same
ratio of Pnet-1 : Pnet-2 but also
the same, with some deviation, as the observed ratio values
which were established
experimentally at our pilot plant by using the same SWRO
membrane as used at Jeddah
-
SWRO desalination plant. The product recovery ratio of 1st set:
2nd set of elements
when compared to their feed is 21.75%: 17.01% and with product
ratio of 1470: 900
m3/hr, respectively. But the recovery (%) of each of 1st set:
2nd set of elements
compared to feed of 6760 m3/hr is 21.75%: 13.31% for a total
product recovery of
35.059%.
The same data treatment used in case "a" for Jeddah SWRO was
employed in
establishing the potential performance of Jeddah SWRO when
combined with NF (NF-
SWRO), cases "b" and "c" in Figure 17. Again, the product flow,
recovery and energy
were as shown in Figure 17b, and c. As in the previous case "a"
the SWRO
desalination is assumed to occur in two steps. In case "b" of
Figure 17, the first set of
elements, step 1, is assumed to treat the NF-product 27,300 ppm
to yield reject with
comparable salinity (TDS) as that for the actual feed to Jeddah
SWRO plant, which
constitutes the feed to the second stage elements. The second
stage set of elements
allows for extraction of product from this feed to yield brine
with TDS of 68258 ppm.
This arrangement yields 2724 m3/hr of product at a recovery of
about 41% compared to
a product flow of 1282 m3/hr and recovery of 32% of the feed to
the second step. The
overall recovery is 60%. But in this latter case the brine flow
per module of 1.83 m3/hr
will be less than the minimum brine flow requirement of 2.0
m3/hr per module. This
could be tolerated since the hardness content of the SWRO reject
is very low (Table 3).
However, SWRO feed has an SDI
-
turn is expected to improve the yield, recovery and product
water quality of Jeddah
SWRO plant when combined with an NF pretreatment.
The above results obtained with the NF pretreatment of seawater
feed in the removal of
hardness, lowering of TDS, pH when added to the improvement of
NF permeate as feed
to SWRO and the observed gain in SWRO product water recovery
ratio are remarkable
ones and should allow for the overcoming of the three problems
in seawater
desalination described earlier, Section 2, which are due to the
composition of seawater
and what it contains such macro and micro particles and
organisms. Moreover, because
of the high purity of product from the SWRO unit (TDS 200 ppm)
the NF
pretreatment should make it quite easy to produce fresh water
from the sea by the
SWRO process in one single stage SWRO, thus eliminating the
second brackish RO
stage with savings in each of capital investment and O&M
cost by over 10%. This is in
addition to increasing the plant output at least by 15% since
the elimination of the
second stage allows for the recovery of all products from the
first stage SWRO unit.
The above treatment of adding the NF to SWRO plant can be
extended to NF addition
to MSF plants. Earlier work showed that operation of MSF plant
at 135 oC and 150 oC
without scale formation was possible when the sulfate in the
feed (Mediterranean sea)
was reduced, using ion exchange, from 2900 to 1200 ppm [10].
With the NF
pretreatment the sulfate ions in Gulf seawater is reduced from
3200 ppm to less than
206 ppm when using pressure to the NF feed of 18 bars and to
only 72 ppmwhen using
a pressure of 21 bars. Further reduction in level of sulfate to
less than above values is
expected from the NF treatment of seawater in other seas, e.g.,
Ocean, Mediterranean.
This should allow for the use of the NF permeate as make-up to
MSF plant in a
combined NF product - MSF unit permitting the operation at TBT
of 120 oC to 150 oC or
higher with an increased gain in distillate output. Furthermore,
the sulfate content in
SWRO reject from the NF-SWRO pilot plant is less than 420 ppm
and Ca++ less than
100 ppm when their concentration in the NF product is 230 and 52
ppm, respectively.
Lower values are expected when theCa++ and SO4= levels in the NF
product are less
than the above values. Again, this should allow for its use as
make-up to MSFplant in
a hybrid NF-SWROreject/MSF system. This last desalination system
arrangement should
allow, as already demonstrated in this study, for the recovery
of up to 90% of the NF
product as potable fresh water, where 60 to 70% of the NF
product is converted to
-
potable water by the SWRO unit and with the remaining 20 to 24%
derived from the
conversion of 80% of the SWRO reject also into potable
(distillate) water by the MSF
Unit. Work is in progress on further evaluation of the two
concepts of NFproduct-MSF
and NF-SWROreject-MSF system by operation at TBT of 120 oC.
Future work will
explore TBT elevation to as high as 160 oC in the MSF pilot
plant at SWCC RDC, Al-
Jubail. Results of work in progress as well as future work are
to be described in
separate reports.
(Cost of Water from Conventional SWRO and NF-SWRO Processes)
The cost of water production in SR/m3 is calculated for Jeddah
1, Jeddah 2 and Yanbu
SWRO plants with and without NF pretreatment unit. In all cases,
the production for
each SWRO plant when operated alone without NF pretreatment is
set to the actual
Jeddah 1 and 2 SWRO plants capacities at 56880 m3/d, while the
installed SWRO plant
cost is adjusted to reflect this limitation in plant size for
the Yanbu SWRO plant. The
power and chemical consumption, the costs of spare parts,
membrane replacement,
micron cartridge filters, other consumables and O&M
including labor as well as plant
availability of 90% are set to the actual values established for
Jeddah -1 SWRO plant
[34]. However, no coagulant is used in the NF-SWRO process and
the H2SO4 is added
at the reduced rate of about 30 ppm to lower feed pH to about 7
in order to allow for
product recovery 60% from the NF unit without alkaline scaling.
For the combined
NF-SWRO case 'b' in Figure 13 the energy is calculated from Eq.
11 based on 60%
recovery of both the NF and SWRO units. The total energy
E(NF-SWRO) kwh/m3 was
computed from the equation:
ET(NF-SWRO) = Ef + Eb1 + ENF + Eb2 + ESWRO + Ep+ Eothers
(13)
Where E is the electrical energy (kwh/m3) delivered by the
various pumps. The letters f,
b, and p denote the feed, booster, and product, respectively,
ENF and ESWRO are the
energy delivered by the high-pressure pump(s) to the
desalination part of the NF and
SWRO units, respectively. The figure 8.35 kwh/m3 is the reported
actual energy for the
total SWRO process including intake, pretreatment and post
treatment steps in Jeddah
SWRO plant by [34], i.e., equals the sum of energies as given by
Eq. 13. The energy
-
required for the desalination parts of the NF & SWRO
process, i.e., ENF and ESWRO, in
Eq. 13 computed from Eq. 11 at recovery ratio of 60% for each
process is 1.655 and
3.31 kwh/m3, respectively, for a total of 4.97 kwh/m3. This
value compares to 6.14
kwh/m3 for the conventional SWRO process or in the ratio of the
latter to the former
case of 1.0: 0.81. As established from Eq. 13 the total energy
for the NF-SWRO
process is 6.28 kwh/m3 and the total energy ratio of ESWRO:
ENF-SWRO is 8.35: 6.28
kwh/m3 or in the ratio of 1.0: 0.75. Thus the non-desalination
part of the total process
energy in Eq. 13, i.e., Ef + Eb1 + Eb2 + Ep + Eothers, for the
conventional ESWRO : ENF-
SWRO is 2.21: 1.32 kwh/m3 or 1.0: 0.597 and product ratio for
ESWRO: ESWRO is /.32 :
2.21 kwh /m3 or 0.584 : 1.0 which explains the reduction in
energy of the NF-SWRO to
that of the conventional SWRO process (see Eq. 11). The
reduction in energy for the
NF-SWRO as compared to conventional SWRO is due to the expected
increase in plant
productivity.
Operation of the combined NF-SWRO requires the following
modifications: (1) the
addition of NF unit to the existing SWRO plant and (2) an
increase in feed quantity by
introducing additional feed line with its necessary electrical,
mechanical and civil
works. The latter works were computed from actual prices at
about 18% of total
contract value for the Jeddah-2 SWRO plant and Yanbu plant and
at about 25% for the
Jeddah-1 SWRO plant. As for the cost of the NF unit it was
estimated from a quotation
for its desalination part alone (high pressure pump, NF modules,
pipe and control)
without other auxiliary equipment.
Table 5 lists the cost in SR/m3 product for six cases: Jeddah-1,
Jeddah-2 and Yanbu
SWRO plants with and without NF pretreatment unit for two energy
prices. In case "a"
the cost of power was set at SR 0.05/ kwh as was done in
reference [34] and in case "b"
the power cost was increased to SR 0.375 ($0.1)/ kwh. The annual
fixed charge rate
was assumed at 10% per year of the installed cost. The interest
on investment is zero.
This case is assumed to equal plant depreciation over 20 years
with 7% annual interest.
In all cases, the cost of product from the NF-SWRO plants was
lower than that when the
conventional SWRO plant is operated alone without the NF
pretreatment. A drop in
cost of product by about 28% can be realized when the plant is
operated with NF
pretreatment. The product water ratio of SWRO: NF-SWRO is
18,685,080: 31,977,504
m3/yr or 1: 1.71. The product cost SR/m3 is in the ratio of 1:
0.72 for SWRO: NF-
-
SWRO for Jeddah-2 and Yanbu SWRO plants. More details of the
results are shown in
Table 5. The techno-economic results are in conformity with
results of the simulated
model in that the operation of the SWRO plants with NF unit is
superior to SWRO plant
operation alone without the NF pretreatment. Moreover,
improvement in plant
productivity accompanied with reduction in water cost as well as
reduction in energy
consumption are expected when NF feed pressure and NF and SWRO
recovery ratios
are raised.
1. The NF membrane treatment of noncoagulated dual media
filtered seawater feed
to desalination plants removes from it (1) very fine turbidity,
(2) residual
bacteria, (3) scale forming hardness ions, in some case by up to
98% and (4)
lowers its TDS, depending on operation conditions by 35 to
60%.
2. With this NF feed treatment the otherwise complex
conventional seawater
desalination process, are simplified since the effects on
seawater desalination by
the above four factors, which constitute the major problems in
seawater
desalination by the conventional processes, are eliminated.
3. Feeding this low turbidity new NF product to SWRO membrane
resulted in
remarkable increase in membrane product (permeate) quantity and
improved its
quality. Recovery ratio increased by over 100%.
4. With this new NF SWRO process addition of a second stage
(brackish RO) is
not required.
5. Use of this new NF product as MSF make-up increased
distillate output up to
80%. Operation at the same condition of MSF with make-up made of
seawater
was not possible due to increased potential of non-alkaline
scaling.
6. With this new NF source of make-up to the MSF process it
becomes feasible to
operate at TBT above 120 oC due to its low potential of
non-alkaline scaling.
-
Table 1. Composition of Gulf Seawater, Al-Jubail and Normal
Seawater
Constituents Gulf Seawater, Al-Jubail
Normal Seawater
Cataions (ppm) Sodium, Na+ Potassium, K+
13440 483
10780 388
Calcium, Ca2+ * Magnesium, Mg2+ *
508 1618
408 1297
Copper, Cu2+ Iron, Fe3+ Stronsium, Sr2+ Boron, B3+
0.004 0.008
1 3
-- -- 1 --
Anions (ppm) Chloride, Cl-
24090
19360 Sulfate, SO4= * Bircarbonate, HCO3- *
3384 176
2702 143
Carbonate, CO3= Bromide, Br- Flurried, F- Silica, SiO2
--- 83 1
0.09
-- 66 1.3 --
Other Parameters Conductivity (S) H
Dissloved Oxygen (ppm) CO2 (ppm) Total Suspended Solids
(ppm)
62800 8.1 7
2.1 20
-- 8.1 6.6 2 --
Total Dissolved Solids (ppm)
43800 35146
* Hardness scale forming ions
-
Table 2. Pretreatment and Quality Requirements of Feed Taken
from an Open Sea (Surface) Intake
Problems in Seawater Desalination Due to
Seawater Characteristics
Pretreatment and Quality Requirement of Feed to
SWRO Thermal High Degree of Hardness (Ca++, Mg++, SO42-,
HCO3-)
Requires Removal or Inhibition of precipitation
by addition of antiscalant, and by
Operation at correct conditions
Requires: Removal or Inhibition of precipitation by
adding antiscalant Operation at correct conditions
High TDS Requires lowering of TDS which in turn Lowers Waste due
to Increases Recovery Ratio Lowers Energy /m3 Lowers Cost /m3
Lowering of TDS beneficial by reducing concentration of hardness
ions
High Turbidity (TSS, Bacteria, etc.).
Requires complete removal Requires Partial Removal Complete
removal of
turbidity, however, reduces foaming and, therefore, eliminates
need for addition to make-up of antifoam
-
Table 3. Chemical Composition and Physical Properties of
Seawater, NF Filtrate, and NF and SWRO rejects at Different NF Feed
pressure
Element / Parameter Seawater NF Filtrate (5 modules)*
SWRO Reject
NF Reject
Ion. Conc. Ion Conc. Rejection (%) Ion Con. Rejection (%) Ion.
Con. Rejected % Ion Con. Ion Con. NF Feed Pressure (Bar) 18 18 22
22 31 31 31 31 A. Hardness Ca++ (ppm)
481
92
80.9
50
89.6
52
89.2
96
701
Mg++ (ppm) 1608 192 88% 96 94.0 143 91.1 253 2200 Total Hardness
(ppm) 7800 1014 87 520 93.3 720 90.8 1280 10800 SO4-- (ppm) 3200
206 93.3 72 97.8 230 92.8 414 4950 HCO3- (ppm) 128 46 63.3 30 76.6
24 81.3 42 133 B. Other Ions Cl- (ppm)
22780
16,692
26.7
12320
46.3
9640
57.7
19570
29350
Na+ (ppm) (12860) (9426) 26.7 (6904) 46.3 (5442) 57.7 C. Total
Dissolved Solids TDS (ppm)
44046
27,720
37.3
20230
54.1
16400
62.8
30,640
63640
pH 8.2 7.85 7.92 6.38 7.08 7.46 Conductivity (s/cm) 60,000
40,470 31100 24600 43300 68600 Module arrangement 2:2:1 , each
module contains 2 NF elements, arrangement equals two parallel
lines each having 5 elements in series.
Table 4. Operation and Performance Parameters of NF-MSF and
NF-SWROreject-MSF Desalination Hybrid Systems Vs Conventional
Seawater MSF Desalination System (Operation of MSF unit with NF
product or SWRO Reject from NF- SWRO unit was done without addition
to make-up of antiscalant or antifoam)
Trials Brine Recycle Data* Performance Parameters TBT
(oC) Ca++
(ppm) SO4=
(ppm) M- alkalinity
(ppm) Conductivity
(s/cm) pH Make-up
(m3/hr) Product (m3/hr)
Recovery ratio (%)
A. NF-MSF With acid 120 160 390 26 59000 8.19 1.5 0.97 65
Without acid** 120 168 410 65 62000 8.63 1.5 0.97 65 B. NF-SWRO
reject-MSF ***
Without acid 120 232 1020 72 87400 8.50 1.5 0.97 65 C. Seawater
MSF
With acid 120 882 5830 30 92000 7.99 1.5 0.97 65 With acid 90
661 4460 20 74000 7.57 1.5 0.59 39 With acid 120 561 3330 14 61000
7.50 4.0 0.92 23 With acid 90 581 4000 18 66000 7.51 2.1 0.94
45
* Brine recycle flow rate was maintained between 6.5 to 6.8
m3/h. ** Make-up is NF product having Ca++ 52, Mg++ 143, SO4--230,
M-Alkalinity 24 and TDS 16400 ppm *** Make-up is SWRO reject from
NF-SWRO unit with composition : Ca++96, Mg++ 253, SO4-- 414,
M-Alkalinity 42 and TDS 30640 ppm
-
Table 5. Cost (SR) of Product Water from SWRO Plants with and
Without NF Pretreatment (a). Cost of Power SR. 0.05 KWH
Cost Component Jeddah -1 SWRO
Jeddah -2 SWRO
Yanbu SWRO
Jeddah -1 NF-SWRO
Jeddah-2 NF-SWRO
Yanbu NF-SWRO
Annual fixed charge rate 16,545,566 36,293,785 37,938,113
16,545,566 36,293,785 37,938,113 Cost of NF unit - - - 5,158,188
5,158,188 5,158,188 Cost of additional Feed - - - 4,376,353
4,376,353 4,376,353 Chemicals 9,32,940 9,32,940 9,32,940 574,500
574,500 574,500 Electrical 7,798,599 7,798,599 7,798,599 9,278,700
9,278,700 9,278,700 Spare parts 2,249,100 2,249,100 2,249,100
2,500,000 2,500,000 2,500,000 Membrane replacement 2,330,835
2,330,835 2,330,835 3,000,000 3,000,000 3,000,000 Micro cartridge
filter 316,800 316,800 316,800 400,000 400,000 400,000 O&M -
labor 3,500,000 3,500,000 3,500,000 3,600,000 3,600,000
3,600,000
Total Cost (SR) 33,673,840 53,422,659 55,066,387 45,433,307
65,181,526 66,825,854
Product cost (SQ/m3) 1.786 2.833 2.92 1.421 2.038 2.089 Ratio
SWRO : NF-SWRO 1 1 1 0.791 0.719 0.716
(b). Cost of Power SR 0.3754 / KWH ($ 0.1 KWH)
Cost Component Jeddah -1 SWRO
Jeddah -2 SWRO
Yanbu SWRO
Jeddah -1 NF-SWRO
Jeddah-2 NF-SWRO
Yanbu NF-SWRO
Total cost without electric 25,875,241 45,623,460 47,267,788
36,154,607 55,902,826 57,547,154 Energy cost 58,551,881 58,551,881
58,551,881 69,664,480 69,664,480 69,664,480 Total cost 84,427,122
104,175,341 105,819,669 105,819,087 125,567,306 127,211,634
Cost (SR/m3) 4.478 5.525 5.612 3.309 3.928 3.978
Ratio 1 1 1 0.739 0.711 0.709
Total Product (m3/d) 18,685,080 18,685,080 18,685,080 31,977,504
31,977,504 31,977,504
-
0
10
20
30
40
50
60
70
80
20000 25000 30000 35000 40000 45000TDS (PPM)
Fee
d-b
rin
e O
smo
tic
Pre
ssu
re (
bar
)
Seawater
M S FTO MSF
HPP
SWRO SECTIONSWRO REJECT TO MSF
PERMEATE
NF REJECT
ROFEED
NF PRODUCT TO RO FEED TANK
CF FEEDTANK
NF SECTION
M S F
Dualmedia Fine Sand
Pump
Fig2Schematic Flow Diagram of NF-SWRO Desalination Pilot
Plant
Seawater
CF FEEDTANK
Duel Media Fine Sand Media
HIGH PRESSUREPUMP
R e j e c t t o M S F
PERMEATE
SWRO UNIT
A/A
MSF UNIT
4 H . R . C S t a g e s
B r i n e h e a t e r
BOOSTER PUMPFEED TANK
H. RJ
SW
D
B.B
B.R
N F U n i t
PRODUCT
N F R E J E C T
Pump
S e a w a t e r f r o m M S F H e a t R e j e c t i o n S e c t
i o nS W R O U n i t
Fig3Schematic Flow Diagram of NF, SWRO and MSF Pilot Plant
Fig 1 Effect of Seawater Feed TDS on Osmotic Pressure of
Feed-brine (fb) Keeping SWRO Brine Concentration at 66615 ppm
Pnet= Pappl-
Pappl
-
Ca++ Mg++ SO4-- HCO3-- Total Hardness
48192
1608
192
3200
206 128 128
7800
1014
0
1000
2000
3000
4000
5000
6000
7000
8000
Con
cent
ratio
n (P
PM)
Ca++ Mg++ SO4-- HCO3-- Total Hardness
SeawaterNF Filtrate (AV )
Fig4 Effect of New Process on Removal of Hardness Ions (Ca++,
Mg++, SO4--, HCO3--) From Gulf Seawater (Using 5 Modules, 10 NF
Elements)
-
Ca+
+
Mg+
+
SO4-
-
HC
O3-
-
Tot
al H
ardn
ess
Cl-
Na+
TD
S
Con
duct
ivity
8188
93
63
86
27 27
3732
0
10
20
30
40
50
60
70
80
90
100
Rej
ectio
n %
Ca+
+
Mg+
+
SO4-
-
HC
O3-
-
Tot
al H
ardn
ess
Cl-
Na+
TD
S
Con
duct
ivity
Fig 5 NF Percent Rejection of Ions, TDS and Conductivity
-
26 / 07/ 97 12 / 10 / 97
0
1
2
3
4
5
thou
sand
BDMFADMFAMCFNFP
Fig 6 Bacterial Count Colony Forming Unit /ml (CFU) from
Different Sampling Points in NF Pilot Plant at 0 hr.
CFU
-
BD
MF
AD
MF
AM
CF
NF
B
NF
P
MAY
JUN.
JUL.
AUG.
SEPT.
OCT.
Nov
DEC
1.00E+01
5.01E+03
1.00E+04
1.50E+04
2.00E+04
Mon BDMF ADMF AMC F NFB NFPMAY 1.0E+3 2 .8E+3 2.8E+3 6.2E+ 3
6.2E+1JUN 2.6E+3 2.4E+3 2.1E+3 4 .3E+3 2.1E+2JUL 2.9E+3 8.5E+3
1.4E+3 2.4E+3 2.1E+2 AUG 1.2E+4 7.9E+3 7.9E+3 1 .06E+4 3 .1E+2S EP
1.7E+4 3.2E+3 7.2E+3 1.9E+3 1.5E+2O C T 1.1E+2 1.9E+2 6.0E+1 6.0E+2
3 .0E+1 NO V 1.2E+3 9 .5E+1 1.0E+2 8.5E+2 7.1E+1DEC 2.7E+3 9.5E+1
7.5E+1 1.2E+3 5.7E+1
CFU
Fig 7 M onthly AVG Bacteria Count in Colony Forming Unit /ml
(CFU) from Different Sampling Points in NF Pilot Plant at 0 hrs
(BDM F& ADM F are Before and After Dual M edia Filter
respectively, AM CF is After M icron Cartridge Filter, , NFB&
NFP are Nanofiltration Brine and Permeate)
-
0
10
20
30
40
0 200 400 600 800 1000 1200
Feed
Flow(l/mi
Temp (
0C)
feed flow temperature(oC) pressure (bar)
y = 19.383x-
05
1015202530
0 200 400 600 800 1000 1200
Flow (l/
iCleaning
05
1015202530
0 200 400 600 800 1000 1200
Flow (l/
iCleaning
and Feed Flow
y = 39.456x-
05
1015202530
0 200 400 600 800 1000 1200
Flow (l/
i
-
20.7 24.1 27.6 31.0 34.5 37.9
y = 3.5199xR2 = 0.9987
y = 1.0519xR2 = 0.9989
0
2
4
6
8
10
12
14
16
18
20
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0Pnet (bar)
Flo
w (
l/min
.)
0
5
10
15
20
25
30
35
40
45
50
Rec
ove
ry (%
)
Pf = Pappl. (bar)
Pnet = [Pf - (Pfb/2) - Pp] - [fb - p]
f = Feed; b = Brine; p = Product
Recovery
Flow
-
80 100 120 140 160
-
Q F = 11266 m3/H
Q F = 6760 m3/h
2774 + 1282 = 4056m 3/h
Q f = 6760 m3/H
1470 + 900 = 2370 m 3/h
Q f = 11266 m3/H
3057 + 933 = 3990m 3/h
=C f 43300ppm
C f 43300= ppm
C f 43300= ppm
66600 ppm
2771 m 3/h
68250 ppm
2704 m 3/h
66678 ppm
4390 m 3/h
a. Jeddah SW R O (A ctual)
b. N F-Jeddah SW R O
C F = 27300 ppm 3986 m 3/h
46300 m 3/hSWRO SWRO SWRO
3.33
SW R O SW R O SW R ON F
Q F = 6760 m3/h
C F = 27300 ppmC f = 67300 ppm
C f = 67300 ppm
c. N F-Jeddah SW R O W ith Brine Staging
SW RO unit
SW RO unit
N F
SW RO unit
1st Stage2nd Stage
-
IDA
Desalination and Water Reuse Quarterly
EDS Conference
Desalination
Water Science and Technology Association
IDA World Congress
IDA World Congress
IDA World
Congress
Desalination
ASTM Desalination, Desalination, Desalination
-
Desalination Desalination
Desalination Water Supply Aqua, Desalination Desalination
Filtration and Separation The 1995 Thirteen Membrane Technology
Conference
Journal of
Membrane Science, Journal of Membrane Science Desalination, IDA
World
Congress World Congress
IDA World Congress
IDA
World Congress
Inst. Engg. Research report
Desalination
IDA World Congress
-
NWSIA National Desalination and Water Reuse Conference,