INVESTIGATIONS INTO THE EFFECTS OF FEED SPACER FILAMENT SPACINGS
ON FOULING PROPENSITIES OF REVERSE OSMOSIS (RO) MEMBRANE SURFACES
USING COMPUTATIONAL FLUID DYNAMIC TECHNQUES
Asim Saeed, Rupa Vuthaluru and Hari B Vuthaluru
International Conference on Desalination, Environment And Marine Outfall Systems
13-16 April 2014, Sultan Qaboos University. Muscat
Presenter: A/Prof Hari Vuthaluru (H.Vuthaluru@exchange.curtin.edu.au)
Presentation Outline
 Introduction and motivation
 Details of CFD Model
Flow conditions
Geometric parameters
Assumptions
Modelling approach
 Results and Discussion
 Conclusions
Introduction
 Available fresh water resources are depleting at an
alarming rate, due to
expansion of industrial and agricultural activities
rise in population rate
enhanced living standards
 Available fresh water is only 0.5% of the total earth’s water.
 Oceans, regarded as the world’s major water reservoir,
comprise of 97% of the total Earth’s water
 Remaining 2% is locked in icecaps and glaciers (Khawaji
et al., 2007)
Desalination
 Different techniques have evolved which separate salts and
unwanted materials from sea water and make it suitable for human
consumption and irrigation (via desalination).
 Desalination can be classified widely into two groups on the basis of
the separation mechanism adopted i.e.

Membrane based desalination:

Thermal desalination:
Water is allowed to diffuse through a
membrane whereas salts are retained on the other side of the membrane
condensation.
Salts are removed from water by evaporation and
Desalination technologies
Thermal desalination
technologies
Membrane based desalination
technologies
Multi-stage flash distillation (MSF)
Reverse osmosis (RO)
Multi-effect distillation (MED)
Nanofiltration (NF)
Vapour compression distillation (VCD)
Electrodialysis (ED)
Out of above mentioned techniques, Reverse Osmosis (RO) and Multi-Stage
Flash distillation (MSF) are widely used commercially
Motivation
 Material build up in RO membranes
Operational issues
Maintenance issues
 Remedial options include
Pre-treatment
Periodic cleaning
Modifications to membrane
 Modifications to membrane configuration
Computational Fluid Dynamics (CFD)
predict flow behaviour and concentration patterns quite accurately when
applied to cross-flow membrane operations
Role of Feed Spacer
SWM in partly unwound state
[adapted from Li & Tung, 2008]
 Promotes mixing and keep the membranes clean
by enhancing mass transfer and disrupting the solute concentration boundary layer.
at the same time they are responsible for the pressure drop and creation of limited
dead zones.
 Efficiency of a membrane module depends therefore heavily
on the efficacy of the spacers to increase mass transport away from the membrane
surface into the fluid bulk by increasing shear rate at the membrane surface at
moderate pressure drop
Present Work
 This study is an extension of our previous work
which considered the impact of spacer filament orientation on
hydrodynamics at fixed spacer mesh length.
 Systematic variation of mesh length of the feed side
spacer
to study the resulting impact on
wall shear stress and mass transfer coefficient of a mono-valent solute (such
as NaCl)
using ANSYS FLUENT for flow visualization to explain the variations
in wall shear stress and similarities in mass transfer coefficients for the two
membrane walls
Flow Conditions Simulated
 In the present study, steady state and laminar flow
conditions are employed
to investigate the impact of filament dimensionless mesh spacing on
mass transfer coefficient and shear stress exerted on the membrane walls
which dictate fouling propensities during membrane operations.
 In most of the real life cases flow through spacer filled
modules do fall
in the Reynolds number category where the flow is steady and laminar
(Fimbres-Weihs and Wiley, 2007)
justifies our choice of steady-state and laminar flow regime.
Geometric Parameters
 Channel height (hch) used as the
characteristic length as well as to
non-dimensionalize spacer
geometric parameters (sum of the
top and bottom filament diameters )
 Channel height is kept as 1mm
for all the simulations in this work
for the sake of convenience.
 Non-dimensionalized filament
spacing for both the top (L1) and
bottom (L2) filaments are taken as:
Boundary Conditions
 The two opposite vertical faces
perpendicular to the flow direction (xdirection) are defined as mass flow inlet
and pressure outlet.
 Mass flow rate is specified in flow
direction (x-direction) and varied to get
the desired hydraulic Reh
 Solute mass fraction kept zero at the
inlet.
 Translational periodic boundary
conditions are defined for the two
vertical surfaces parallel to top
filaments.
Computational Domain
 Filament surfaces are defined as walls.
Cases Simulated
16 cases in total
Configuration
L1
L2
Configuration
L1
L2
SP22
2
2
SP42
4
2
SP23
2
3
SP43
4
3
SP24
2
4
SP44
4
4
SP26
2
6
SP46
4
6
SP32
3
2
SP62
6
2
SP33
3
3
SP63
6
3
SP34
3
4
SP64
6
4
SP36
3
6
SP66
6
6
(Top)
(Bottom)
Modelling Approach
 Membrane surfaces are assumed to be impermeable walls
with no slip conditions
 Membrane walls have a constant higher value of solute
mass fraction than that defined for the inlet condition.
 In all the simulations
the solute mass fraction at the membrane walls are taken as 1
the solute mass fraction is defined as zero at the inlet.
Modelling Procedure
 Working fluid is assumed to be
a binary mixture of water and a monovalent salt, such as NaCl having a mass
diffusivity of 1.54e-09 m2/s (Capobianchi et al., 1998).
Working fluid is further assumed to be isothermal and incompressible
with constant density (998.2 Kg/m3), viscosity (0.001 Kg/(m s)) and solute diffusivity.
 Re defined on the basis of hydraulic diameter and effective fluid
velocity in the domain is used to calculate
mass flow at the inlet of the computational domain.
 Considering the degree of accuracy of the results needed,
computational time required and available computational
capabilities
a grid size of 716,880 was chosen as an adequate grid size for SP22.
Similarly adequate grid sizes for different spacer arrangements were determined to
ensure the solution is grid independent.
Governing Equations & Solution Methods
 Continuity, three x, y and z momentum equations and
concentration equations are the five governing equations
(Navier-Stokes equations) solved by
 ANSYS FLUENT using pressure based segregated solver.
 QUICK (Quadratic Upstream Interpolation for convective
Kinetics) scheme is used
 for discretising momentum equations,
 SIMPLEC (Semi-Implicit Method for Pressure linked
Equations, Consistent) algorithm is used for
 pressure velocity
FLUENT, 2009).
coupling
(Versteeg
and
Malalasekera,
2007,
Incorporation of Mass Transfer into Model
 For spacer filled narrow channels having impermeable
membrane walls, the local and average mass transfer
coefficients can be defined respectively by the following
equations (Fletcher et al., 1985, Kang and Chang, 1982, Shakaib et al., 2009):
Where:
kl and kav are the local and average mtc’s
Yw, Yb and δy represents mass fraction of solute at the
membrane wall, mass fraction of solute in the bulk and
gradient of mass fraction at the membrane wall
A and Ai refer to the membrane surface area and face
area of any computational cell
 Mass transport equation was incorporated in the model by
a user defined function (UDF).
Spacer Configuration Efficacy (SCE)
 The concept of SCE has been used in this study to grade
the spacers.
SCE= Sherwood number / Power number
𝜌2 h4ch
Pn = SPC
µ3
 Spacers yielding higher mass transport effect with
moderate energy requirement will be having higher SCE
values and vice versa.
Velocity Vectors Near Two Membrane Walls
(a) close to top membrane (Z=0.95mm)
(b) close to bottom membrane
(Z=0.05mm) for SP44 at Reh=100
Near Top Membrane Wall
 Bulk of the fluid, in the vicinity of the
top membrane wall, follows the
main flow direction.
 Wall shear stress and mass transfer
coefficient local values follows the
same trend
i.e. they increase or decrease
simultaneously at different locations with the
exception of very small regions where the
flow separates and reattaches from and to
the top filament.
(a) close to top membrane (Z=0.95mm)
Bottom Membrane Wall
 There are strong three
dimensional effects seen in the
vicinity of the bottom membrane
wall due to
flow reattachment
separation phenomena
covering a larger portion of the bottom
membrane
(b) close to bottom membrane
(Z=0.05mm) for SP44 at Reh=100
 Due to this local wall shear stress
and mass transfer coefficient do
not follow the same trend as in
case of top membrane wall
Velocity Vectors for SP44 (Reh=100)
Contours of velocity magnitude overlayed by the velocity
vectors (fixed length) at vertical plane (y=0 mm)
 There is a reduction in cross sectional flow
area due to the presence of bottom
(transverse) filament and the fluid tends to
accelerate when crossing over the bottom
filament.
This phenomenon induces a nozzle like effect which results
in higher local wall shear stress and mass transfer coefficient
values at the top membrane wall directly above the bottom
filaments.
 In the vicinity of the bottom membrane,
fluid tends to
reattach with the bottom membrane surface in the middle of
the two consecutive bottom filaments and further undergoes
flow reversal and recirculation.
Bottom wall shear stress and mtc distribution along flow direction at y=0mm
This recirculation induces a scouring action on major
portion of the bottom membrane and hence results in higher
values of mass transfer coefficient for major part of the
bottom membrane.
Velocity Vectors for SP44 (Reh=100)
Contours of velocity magnitude overlayed by the velocity
vectors (fixed length) at vertical plane (y=0 mm)

There are three regions where the wall shear stress
shows almost zero values but the mass transfer
coefficients are higher. Out of the three regions
two regions represent the area just after and just before the transverse
upstream and downstream filament represented as separation region in
Figure (b).
The third region where the mass transfer coefficient curve shows local
peak despite minimum value of local wall shear stress resides somewhere
in the middle of the two bottom transverse filaments and presented as
reattachment region in Figure (b).

they increase or decrease simultaneously at different locations with the
exception of very small regions where the flow separates and reattaches
from and to the top filament.
Bottom wall shear stress and mass transfer coefficient
distribution along flow direction at y=0mm

Top wall shear stress and mass transfer coefficient
distribution along flow direction at y=0mm
For top wall, local shear stress and mtc values follows
the same trend
Trends in figures indicate that lower local value of wall
shear stress does not necessarily mean lower local
value for mass transfer coefficient
Results at Reh=100.
Top
Bottom
filament
filament
dimensionless dimensionless
Spacer
spacing
spacing
configuration
(L1)
(L2)
Bottom wall
shear stress
2
SP22
SP23
SP24
SP26
SP32
SP33
SP34
SP36
SP42
SP43
SP44
SP46
SP62
Sp63
Sp64
Sp66
2
2
2
2
3
2
2
2
4
4
4
4
6
6
6
6
2
3
4
6
2
3
4
6
2
3
4
6
2
3
4
6
N/m
0.22
0.23
0.19
0.24
0.19
0.19
0.15
0.22
0.17
0.17
0.14
0.21
0.16
0.14
0.14
0.21
Top wall
shear stress
Wall
Bottom wall
shear stress
mass transfer
Pressure drop coefficient
ratio
Top wall
mass transfer
coefficient
Mass
transfer
coefficientr
ratio
Pa/m
9344.74
7509.98
6250.16
4517.87
6775.23
5285.31
4232.24
3019.78
5859.32
4479.46
3536.12
2526.83
5109.48
3816.73
2975.04
2131.75
m/s
3.94E-05
4.04E-05
4.34E-05
3.65E-05
3.78E-05
3.78E-05
3.74E-05
2.90E-05
3.70E-05
3.64E-05
3.59E-05
2.71E-05
3.61E-05
3.48E-05
3.22E-05
2.52E-05
0.88
1.01
1.11
0.92
0.92
1.02
1.00
0.79
0.94
1.05
0.97
0.75
0.95
1.09
0.89
0.70
2
N/m
1.77
1.39
1.13
0.84
1.52
1.18
0.94
0.72
1.41
1.08
0.86
0.66
1.30
0.99
0.78
0.60
8.13
6.05
6.06
3.46
8.10
6.26
6.30
3.25
8.07
6.53
6.06
3.07
8.09
6.98
5.71
2.86
m/s
4.46E-05
4.01E-05
3.90E-05
3.98E-05
4.10E-05
3.69E-05
3.75E-05
3.69E-05
3.95E-05
3.47E-05
3.69E-05
3.60E-05
3.81E-05
3.20E-05
3.62E-05
3.58E-05

Configurations having L2 = 6 yields lower values for the top wall mass transfer coefficient compared to the bottom wall and
hence would lead to relatively quick fouling of the top membrane wall than the bottom. For this reason they are not at all
suitable at all for any efficient membrane separation process.

In addition to that, SP22 and SP64 also have a lower ratio of the top to bottom mass transfer coefficient and would increase
the fouling propensity of top membrane surface compared to the bottom surface and are not suitable to be used in efficient
membrane separation processes.
Comparison of short listed configurations at Reh=100

The average mass transfer coefficient in this table is taken as arithmetic average of the two mass transfer
coefficients for the two membrane walls.

Spacer configurations that have the two average mass transfer coefficients values quite close to each other and
would result in almost the same fouling tendency of the two membrane surfaces and could be suitable for
membrane operations in real life.
Comparison of different spacer configurations at Reh = 100
Further Shortlisting (based on Pn)
SP44
SP63
SP34
SP43
SP33
Sherwood numner (Sh)
45.00
41.00
37.00
33.00
 Different spacer arrangement are
compared at the same Power
number in terms of Sh.
29.00
25.00
1.00E+05
 At the same Reh different spacer
arrangements tend to have
different energy loses. It therefore
appears to be more reasonable to
compare Sherwood number (Sh)
for different spacer arrangement
at the same Power number (Pn).
4.00E+05
7.00E+05
1.00E+06
Power number (Pn)
1.30E+06
SP44 spacer arrangement tends to have
higher values of Sh for the Range of
Power number considered in this work
and tends to assure greater mass
transport of solute away from the
membrane surface compared to the rest
of the arrangements considered.
Validation of model results

Comparison of Sh for different spacer
arrangement with previous experimental and
numerical studies (Da Costa et al. 1994; Li et
al. 2004; Shakaib et al. 2009) at Sc=1350
showed good agreement. Grober equation,
however, presents a relatively higher value
for SP22 (approximately 30% higher) due to
the fact that Grober equation predicts the
mass transfer rate with + 30% error (as
mentioned in their study)

Comparison of wall shear stress, pressure
drop and average mass transfer coefficients
for different spacer arrangements with
literature data at Reh=100 showed good
agreement
Conclusions
 Mass transfer coefficient values for the two walls are not significantly different for the
spacer arrangement having low to moderate bottom filament spacing (L2 = 2 to 4)
although the wall shear stress at the top membrane surface is always higher than that for
the bottom wall.
 When the bottom filament spacing is further increased (L2 = 6), there is a sharp decline in
the pressure drop but the area weighted mass transfer coefficient for the top membrane
wall showed a sharp reduction compared to the bottom membrane wall suggesting high
fouling propensity of the top membrane wall (not a desirable feature in RO operations).
 SP44 is found to be the best spacer arrangement (for the range Reh=75 to 200) having
higher SCE values (throughout the Re range) among the spacer arrangements
considered
 yielding moderate pressure drop
 leads to nearly equal and higher values for mass transfer coefficient for the two walls
Practical Implications
Results emanated out of the current study are
considered to be of significant value and could
potentially lead to the development of efficient
membrane modules with optimum spacer
arrangements for RO operations.