235
Desahatim, 23 (1977) 235-243 8 Ekvier Scientific Pubbsbiag Company, Amsterdam -Printed
DESALINATION
OF SEA
WATER
in The Netbedsnds
BY REVERSE
OSMOSIS
USING
TUBULAR
MODULE
HIDE0
TSUGE+
, CHOTA
YANAGI
Mechanical Engineering Kobe (Japan)
AND
Research
KENJS
MORI
Laboratory.
E&be Steel,
Ltd.,
SUMMARY
In the process usually
pretreated
prevent
fouling
of desalination
to keep
by reverse
membrane
of membrane,
osmosis,
perfomance.
removing
raw
For
of suspended
sea water
example,
solid
to
is usually
recommended. In this first
stage
using
NRO
study,
to simplify
membrane tubular
module
membranes
was made
membranes
of di-
with similar
ones
removing
performance end three
from
types
cellulose
of suspended
system
was
of membrane.
diacetate And also
and trf -acetate. taken
process
in two stige
by
One of these
and others these
solid,
sutdied
were
data were
blended
compcred
with NaC1 solution.
INTRODUCTION
The advantages for desalination cross
section
primarily
ease
9
of the tubular
fouling
of physicai
present
result
and its uniformity
characteristics of ninimal
of tubular
ad&e~s:
tendency, in situ
reverse from
the large
module
module
cleaning
yield,
with proper
fractional
recovery
of the membrsne Ltd.,
Fukiai-Iw
configurations
characteristics
in the flow direction.
higher
Kobe Steel
osmosis
These
design. operation,
surface
Kobe.
fluid
flow
inherent advantage and
in comparison
Japan-
is
with other reverse
configurations
osmosis This
stage
feed
study
treatment
tubular
EXPERIMENTAL
brane And
were
with those
operation
wus carried
this
chemicals.
basic
engineering
data for
at minimized
first
pre-
pressure
with NaCI
aqueous
aqueous salt
filter
collected
were
solution
analysed
was
and sea water.
long
coefficient
was
from
used
term
and change
surface.
the membrane
sample
was
and residue
dried
absorption
were
as pretreatment.
filtered
by atomic
on mem-
of sea water
Continuous
the flux decline
solution , collected Each
solution
rejection)
solution.
cartridge
matter
and concentration
and dissolved was
deter -
spectrophotometer
and chromium.
Through
the experiment
(Cellulose
the hydrolysis
NRO-A
di and triacetate
The pH of sea water
3.
completely.
coefficient
(especially
fouling
the cleaning
NRO-B
studied
100 micron
operation
Pihrltes
rate.
out to obtain
and 50 -
To select
with iron
flow
of NaCl
to various mind.
to obtain
system,
module.
characteristics
compared
After
out
and flux decline
of feed
performance
of rejection
to be clarified
in this
PROCEDURES
effects
desalting
not need
carried
system
using
The
dose
was
of two-stage
Consequently,
of membrane.
was
controlled 1) of membrane.
(Cellulose
blended
diacetate
membrane)
by hydrochloric
membrane)
modules
acid
were
to about
and used.
6 to avoid
EQUIPMENT
Modules The NRO Ltd.
modules
and Nitto Electric Eighteen
length,
connected
in series length,
surface
is 1.75
were
to form
meter.
developed
by Kobe
Steel
Co.
inch in diameter module.
11 centimeter
square
which
fiber-glass-reinforced
one-half
2.6
used,
Industrial
of bored
nominally
meter
were
epoxy
Modules
outdiameter,
tubes,
with turbulence generally and its
are
effective
2.5
meter
promoter
in
are
fabricated membrane
in
za7
Salt 5000
ppm.
Reverse
rejection &cl
of membranes
feed
solution
was
about
and 42 Kg/cm2
98% under applied
the condition
of
pressure.
osmosis Schematic
the concentration
flow diagram of feed
water.
of apparatus
are
shown
in Fig.
concentrate
was
returned
1.
to feed
To control tank.
T Fig. 1 RESULT
Effect
AND
of feed
!WamaticdCqramoftestaqoipmant
DlsCUSSION
flow
rate,
concentration
and anplied
pressure
on membrane
nerforrnance Effects formance
are
flux and salt remarkable. higher high-salt usudly
of flow shown rejection Hence
concentration rejection lowR.
at this
rate
in Fig.
and concentration As
2.
decrease, the higher permeate.
membranes condition
of NaCl
increasing especially
recovery And should
more
feed
concentration,
the effect ratio at first
to lower
stage
of modules
As
perproduct
on the later induce
the concentration
be required.
number
on membrane
is the
of permeate.
flux of these should
is
be required.
N&I
-a4n
Feed flow Fig. 2
rate
04 fed (Iimin):
mfn
A : 1.4.
fXf e : 1.0,
o
: 0.6
ion and feed flow Effect of feed canon podilct flux and ralt r.$ection
Feed @3ncnCdOIX Feed flow rate Fig. 3
5
4
1
Effect of operating and salt rejection
4%
:
h2.
rati
tia)
2Sllmin
pressure an
product
flux
239
In Fig. of feed
flow
le/min.
increase curve tr2tion
of feed 3,
stage
drop
performance 2re
on salt from
pressure
performances
2nd NRO-B
triacetzte.
of pressure
2), 0.7-c
rate.
however, So,
with the increase
2
shouId
is shown
improved
rejection
point
of view
be over
linearly inclination
in
with of
of the concen-
46 kg/cm2.
material
Membrane NRO-A
the effect
on membrane
46 kg(cm2, operating
flow
is improved
flux and rejection
pressure,
over
of permeate.
Membrane
as feed
the product
easy
to avoid
concentration
of operating
turns
performance
But in order
is recommended
In Fig.
3.
the membrane
rate.
Effect Fig.
2,
are
respectively.
membrane
because
0.04
made
from
NRO-B
cellulose was
of its higher
more
and NRO-8 diacetate suited
are
Camcarbon
shown
and cellulose
than NRO-A
rejection.
i
ai
Fig. 4
of NRO-A
of diacetaae and Mend membranes
in Fig.4. di-
for first
240 Comparison desalting
between
To apply desalination sea
water
The
of sea water. was
data
and NaCl
taken
aqueous
solution
on
the relation
results
is obtained
sea
These
NaCl
chamctmirlia
water
relations
are NaCl
than
are
expressed
= 1.154
F (NaCl)
- 0.011
R (sea
water)
= 0.842
R (NaCl)
+ 15.7
F (sea
water);
R (NaCl)
stage
Flux ; Flux
water);
of sea water of
Rejection : Rejection
performance
to between
NaCl
of a33 rratar end f&Cl Oohrtion
solution
water)
R (sea
solution
characteristics
summA rized
F (sea
F (NaCl)
aqueous
of separation
of such experiment with
from
investigated.
Relationship of demhtti~
in conductivity.
First
of sea water
the standard
and NaCl
Fig. 5
result
effect
characteristics
of sea water of NcCl
(1)
____-_____-
(2)
day)
solution
(m3/m2
(‘%) solution
Better
equations.
----------
aqueous
5.
concentration
ty following
(m3/m2
aqueous
in Fig.
at same
(%)
day)
z
¶5-
d
I’
2” =5
.
.
The
0~0
O
of
results
concentration
micron
filter
was
and flux decline
1000
0
0
0
0
.
of modules used
0
0
0
0
0
hours operation
of sea
Concentration
Fig. 6.
93%
0
=
;QlS 5 o” 3 EQlO-
mean
.o”
water
was adjusted
in practical
m was
to about
0.076.
The
flux salt
module
was
are
4% which In this
equipment. initial
for pretreatment,
coefficient
with NRO-A
shawn
is expectedas
operation.
0. X6 m3/rn2
rejection
1.
Performance k eed Sea Water Turbidity
0.4
(ppm) 43.780
Conductivity
Product 1st Stage to.2
2nd Stage co.
2
220
5.940
(iW M alkalinity
(ppm
CL-
(ppm)
S042-
(ppm)
NO;
(ppm)
T -+“e
(PPm)
NP K TDS
415 0.3
’
0.05 0. a8 10,200
F,, m (ppm)
375 34,200
0 (2.0 128
(3.0
<3.0
;=0.2
to.2
2.2
2s SK+)
Mn
1,820
16.300
(ppm) (ppm
< 3.0 26.8
5,350
T hardness
SiO2
98.6
as CaCO3)
0.4
0.05
<0.03
03
co_02
to.
02
860 72.5 3.040
18.6 2.4 -=50
50
day
decreased
to 91%.
TABLE
in
from
242 The series
analysis
of NRO-B
operation,
of product
modules
100 micron
shown
in Fig.6. observed.
tion using
fitter
As
tubular
was
a result
in Table
50% recovery The
used.
100 micron
of use
module,
with pH adjustment
shown
at 40 -
In spite
hardly
is
of Fig.6.
using
was carried
results filter,
of sea
50 -
with
out.
of this
operation
1. in sea water
100 micron
3)
three
En this
the flux decline
and Table
the pretreatment
and filtration
Operation
1.
water
are
wzs desalina-
can be simplified
microstrainer
or
filter.
Selection
of cleaning The
hours
brown
was
25 -
Major
solved
analysis
cleaning
quantity
deposit
metallic were
Na.
solutions and Cr
of Fe
by treatment
and citric
TABLE
observed
was
on the membrane
compornents
of the deposit
.4) , K, Sl
and Cr.
Fe5)
after
500
by emission
and its ignition
loss
acid
for
into cleaning
summa
are were
the deposit
rized
effective
were
solutions
in Table2.
The
investigated. and the weight
Sodium
loss
dichionate.
disof
oxalic
chemicals.
2.
Solubility
of membrane
foulant
to chemical
solution
Solution No. 1
Water
2
HCI
3
Citric
4
”
acid *
5
Ammonium
6
Oxalic
7
EDTA
+
Each
500mg
for
30 min.
(mol/k’)
Fe
6.70
Cr
1.11
2.28
% 11.6 13.2
2.00
2.20
2.10
2.20
114.
17.5
26.0
2
0.0952
4.00
4.00
189
34. 3
38.7
0.0952
5.00
5.05
155
64.0
35.3
0.0952
4.00
4.30
256
39.2
45.3
4.00
4.85
278
22.0
31.5
added
and fzltered
5.85
0.28
0.0952
2
was
Filtrate
0.01
*
scale
Solution 5.40
2
citrate
acid
ton Content in Scale filtrate mp/& 106s
PH (wt%)
mixed
surface
31%.
The
acid
deposit
operation.
spectroscopic
solution
to 500 ml prepared
under
reduced
solutions
pressure.
respectively,
245
Filter
;
*
pH of prepared
The
0.2
micron
Sartorius solution
membrane was
filter
controlled
with HCi or NaOH
REFERENCES
1. 2.
I. NUSBBAUM. A.P.HATCHER K.D.VOS, Desalination. 5 (1968) 157. H. OHYA, T. MORIYAMA, S. SUmEQ AND
AND
F. 0. BURRIS
S. ISHIZAKA,
Jr.,
Desalination,
4.
16 (1975) 235. K. C. CAA?ABASAPPA AND J. J.STROBEL. 5th International Symposium on Fresh Water from the Sea, 4 (1976) 267. 63 J. W. MCCUTCHAN AND J.S. JOHNSON, Chem. Eng. Progress,
5.
J(*967) . M. JACKSON 9o -
3.
AND
D. LANDOLT,
Desalination,
12 ( 1973) 3 61.