The aim of this study is to apply the membrane bioreactor technology in an oxidation ditch in submerged conditions. This new wastewater filtration process will benefit rural areas (<5,000 population equivalent) subject to chronic water shortages by reusing this water for irrigation of green areas. For this purpose, the membranes developed without support are immersed in an aeration well and work in suction mode. The development of the membrane without support and more precisely the performance of spacers are approached by computational fluid dynamics in order to provide the best compromise between pressure drop/flow velocity and permeate flux. The numerical results on the layout and the membrane modules' geometry in the aeration well indicate that the optimal configuration is to install the membranes horizontally on three levels. Membranes should be connected to each other to a manifold providing a total membrane area of 18 m². Loss rate compared to the theoretical throughput is relatively low (less than 3%). Preliminary data obtained by modeling the lagoon provide access to its hydrodynamics, revealing that recirculation zones can be optimized by making changes in the operating conditions. The experimental validation of these results and taking into account the aeration in the numerical models are underway.

INTRODUCTION

Developing solutions for recycling and reusing wastewater has become a part of a growing industry. Thus, many researchers are working on developing new solutions to purify wastewater and fully enable reuse. In particular, rural areas are subject to difficulties in the use of water in the Mediterranean area, in Europe, and in areas with a long dry season around the world. The wastewater treatment in rural areas is mainly constituted by lagoon processes. Nowadays, artificial lagoons are mechanically aerated to (1) supply oxygen necessary for a good purifying efficiency and (2) maintain the circulation of the fluid in the basin. The quality of the treated water leaving the system complies with the standards for release to the environment, but does not meet the standard for different possibilities for reuse. Membrane bioreactors combine a biological degradation and a liquid/solid separation by using porous membranes. Ultrafiltration membranes used allow a complete retention of the total suspended solid (microorganisms and some viruses). This feature permits disinfection of the wastewater without using oxidant chemicals which can lead to the production of carcinogenic molecules (Wisniewski 2007). In addition, the use of a membrane bioreactor can increase and improve the control of residence time of the sludge (Gander et al. 2000; Kim et al. 2001), which reduces processing costs. The LAGUMEM project aims to provide a solution for rural communities who want to reduce their impact on the environment or those who are exposed to chronic water deficit, by reusing the wastewater for irrigation of crops or green areas by coupling the lagoon process and membrane filtration. Since this technology is unfortunately not an option for rural areas from an economic (investment and operating costs) and technical (complex working) point of view, the aim of this study is to adapt membrane technology for the treatment of urban wastewater in an aeration lagoon by a low-cost manufacturing of the immersed module and less intensive energy process compared to classical membrane bioreactors. In this work, the purpose is to immerse membranes into aerated lagoons through the development of self-supported membranes and the implementation of these self-supported membranes in the aeration well of lagoons. The installation of self-supported submerged membranes requires the addition of spacers acting as mechanical stabilizers to strongly reduce the cost of the membrane. Although the addition of these materials appears to be essential, they inevitably produce an increase of pressure losses and thus production costs. These effects were analyzed in various publications (Thomas 1965; Da Costa et al. 1991; Hicks & Mandersloot 1968; Schock & Miquel 1987; Fimbres-Weihs et al. 2006; Fimbres-Weihs & Wiley 2007). Over the past two decades, numerical simulation (computational fluid dynamics (CFD)) has become a widely used tool to solve complex problems related to hydrodynamics (Ghidossi et al. 2006). In this work, CFD has been used to characterize the flow between the membranes at a microscopic scale, to determine the geometry and optimum arrangement of membrane modules in the aeration shaft of the lagoon, and to describe the hydrodynamics expected in the basin. In order to validate the results, an experimental pilot-scale setup is installed in the laboratory with an external module and a submerged module to compare their filtration efficiency.

MATERIAL AND METHOD

The aeration tank

The oxidation ditch used in this project is made by EAU PURE company and designed as Oxylag®. Air is injected into the bottom of a deep well in order to maintain the loop circulation of water in the lagoon, to aerate the biological environment, and to reduce the fouling of the membrane. Oxylag has been widely used as a modified activated sludge biological treatment process due to its reliability, simplicity of operation, and low sludge production. Since the process efficiency depends heavily on the flow field in an Oxylag, a good understanding of hydrodynamics of the ditch is needed for a successful design. An assumption of an ideal flow pattern, ignoring the real hydrodynamic characteristics of the ditch, was often made in design (Luo et al. 2005; Stamou 2008). Today, the improvement of the computational power and the availability of CFD codes have led to more accurate prediction of the flow pattern in oxidation ditches. Therefore, aeration and hydrodynamics are optimized by CFD in order to limit the size of the lagoon. The lagoon is oblong in shape with a length of 30 m, a width of 15 m, and a depth of 1.5 m. The depth in the aeration well is 3 m. Numerical simulations of the flow in the basin are made with StarCCM+ software (CD-Adapco) using the finite volume method. The Navier–Stokes equations are solved using a high-precision computer. The size of the geometry requires a mesh containing a large number of elements (approx. 60 million) and refinement in areas with high gradient and near walls. In addition, the hydrodynamic behavior is very heterogeneous in the lagoon; it is essential that the mesh is also suitable for different operating conditions. A turbulent multiphase flow model with a degassing boundary is used to take into account the aeration in the well. The case uses a simple aeration tank, where air enters the tank through five gas injectors (air) at the bottom of the tank and escapes through the top surface. The degassing surface (the top boundary) is modeled using a frictionless phase-permeable wall. This wall can share a region with a flow-split outlet or pressure outlet boundary.

Membrane processes in the aeration tank

To date, no organic membrane meets the criteria of fouling resistance for application in lagoons. The company ORELIS Environment produces a range of membranes that serve as support for our study. Therefore, the proposed solutions are based on the development of flexible polyvinylidene fluoride and polyacrylonitrile membranes fixed on a grooved fabric support, forming a self-supported membrane with the spacer inside. In the proposed configuration, the membranes are stuck to each side of a spacer and on three borders, thus forming a pocket which is connected to a vacuum pump for the extraction of the permeate (Figure 1).

Figure 1

(a) Front view and (b) schematic cross-section view of the developed membrane.

Figure 1

(a) Front view and (b) schematic cross-section view of the developed membrane.

Spacers

The spacers are used to prevent the flexible membrane from sticking during suction mode. They are located on the permeate side as opposed to their traditional use in spiral modules. To optimize the flow, a lot of spacers are tested. They have an optimum velocity and low pressure drop. Moreover, the flexible membrane is adhered by suction on the spacer; it is therefore essential to use a spacer with optimal shear forces to control the fouling of the outer membrane and limit the contact points (areas of non-permeation). Figure 2 shows the geometry of a sample spacer discretized by a mesh containing about 5 million elements.

Figure 2

Numerical representation of a spacer with the boundary conditions at the left and the mesh at the right.

Figure 2

Numerical representation of a spacer with the boundary conditions at the left and the mesh at the right.

The boundary conditions are established as follows: a mass flow rate is imposed at the inlet, a pressure P0 at the outlet, symmetry planes on the lateral walls, and solid walls on the upper and lower faces representing the membrane without permeation.

Experimental setup

A membrane bioreactor of 300 L has been set up in the laboratory to compare filtration performances of an external module and a submerged module with membranes without support (Figure 3) developed in this study. To control fouling of the membrane, a membrane air diffuser was fixed at the bottom of the plexi-glass tank. The diffuser also mixes the reactor contents (activated sludge) and maintains aerobic conditions within the reactor. The activated sludge feeding is made in batch mode from wastewater treatment plant sludge and the residence time is less than 5 days. The wastewater was from domestic households (pH = 7.98). A recirculation pump covering a wide flow range and a suction pump are installed respectively for the external module and the submerged module. The permeate and the retentate are re-injected into the biological reactor. The residence time of the sludge does not exceed 5 days.

Figure 3

Experimental setup combining an external and a submerged membrane bioreactor with recycling of the permeate and retentate.

Figure 3

Experimental setup combining an external and a submerged membrane bioreactor with recycling of the permeate and retentate.

Experiments are initially performed for a biomass concentration of a typical lagoon with a constant air flow rate and then the concentration is increased slightly.

RESULTS

Assessment of pressure drop in the spacers

To compare the performance between different spacers, the average speed is assessed throughout the discretized field. It is possible to model only a very small portion of the sample as the field is assumed to be very large, thus reducing the calculation costs significantly. The simulations indicate a heterogeneity of the velocity field in the direction of the flow whatever the initial conditions and the spacer tested (Figure 4(a)). Moreover a cross-section view of the flow reveals the instability and the swirling nature of the flow.

Figure 4

(a) Velocity field of a parallel cross-section; (b) pressure gradient as a function of the mean velocity for three spacers.

Figure 4

(a) Velocity field of a parallel cross-section; (b) pressure gradient as a function of the mean velocity for three spacers.

Figure 4(b) shows the variations of the pressure gradient as a function of the average speed. Thus, the spacer 1 presents the best compromise between pressure drop (energy consumption) and velocity of fluid.

Membrane permeability

In the submerged configuration, the membranes are immersed into the biomass, which avoids the use of a recirculation pump. Biomass is not sheared, making filtration conditions softer and reducing energy consumption. The air injected into the bottom of the pilot allows a continuous cleaning of the membrane and a supply of oxygen for the biomass, thus reducing process costs. Figure 5(a) presents the curve of the variation of the permeate flux as a function of the time for a total suspended solids (TSS) concentration of 3.5 g L−1, a transmembrane pressure (TMP) of 0.55 bar, an air flow rate of 7 L min−1 and 25 hour in continuous filtration. At 20 °C the flux is about 21 L hm−2 bar−1, which is in good agreement with typical membrane bioreactors.

Hydrodynamic characterization of the basin

The simulations were carried out initially with clean water without air, without membranes, in a closed system (‘batch’), and assuming a stationary three-dimensional flow. A velocity of 0.3 m s−1 is imposed at the inlet closed to the aeration well. Preliminary results indicate a fluid velocity three times higher near the outer wall. Singularities also appear, disturbing the flow structure, with recirculation zones in the middle of the basin and detachment zones close to the inner wall (Figure 6): eventual settling problems may occur in these locations. The CFD results of the flow field are in good agreement with experimental data found in the literature (Yang et al. 2011), and the variation of the velocity is at the same order.

Figure 5

(a) Permeability of the immersed membrane at the pilot test; (b) experimental setup.

Figure 5

(a) Permeability of the immersed membrane at the pilot test; (b) experimental setup.

Figure 6

Streamlines in the aeration tank [Inlet velocity = 0.3 m s−1].

Figure 6

Streamlines in the aeration tank [Inlet velocity = 0.3 m s−1].

It is therefore possible to change operating conditions to optimize hydrodynamics, and add the effects of aeration in models of computation and positioning of membranes.

Influence of aeration and membranes in the well

For this simulation (Figure 7), results indicate the influence of the aeration on the increase of the water velocity between the membrane patterns is about 40%. No literature data were found regarding the interaction of the air phase with membranes due to the new concept of this work. Khalili-Garakani et al. (2009) have studied the wall shear stress of the air phase and water phase in an airlift membrane bioreactor (AMBR) between two membranes, and this study seems to be the closest technology to ours. Values of the wall shear stress at the surface of the membrane were underestimated compared to our study and overestimated for the water phase. Although we are at the same order of magnitude, the difference is mainly due to the recirculation of the water phase in the AMBR, which led to the confinement of the bubble beam.

Figure 7

Velocity field of the water phase in the well (air inlet velocity = 0.3 m s−1, water inlet velocity = 0.1 m s−1).

Figure 7

Velocity field of the water phase in the well (air inlet velocity = 0.3 m s−1, water inlet velocity = 0.1 m s−1).

CONCLUSION

For this work, the use of CFD allows the prototyping of the developed membranes without support and testing of the aerated lagoon with clean water. The membranes without support have been developed, manufactured, and are now being tested at the laboratory scale. The spacers acting as a spacer between the membranes have interesting results by CFD, indicating that some samples have optimal hydrodynamic qualities with low pressure drop. The optimization of the arrangement of these membranes in the well of the lagoon, for the overall performance of the process in terms of water production and especially for the hydrodynamics of the lagoon (performance of biological degradation of wastewater), is performed by numerical simulation. The consideration of the multiphase mixture in the aeration well for the CFD also brings good results and acts to predict the hydrodynamics in the latter.

REFERENCES

REFERENCES
Da Costa
A. R.
Fane
A. G.
Fell
C. J. D.
Franken
A. C. M.
1991
Optimal channel spacer design for ultrafiltration
.
Journal of Membrane Science
62
(
3
),
275
291
.
Fimbres-Weihs
G. A.
Wiley
D. E.
2007
Numerical study of mass transfer in three-dimensional spacer-filled narrow channels with steady flow
.
Journal of Membrane Science
306
(
1–2
),
228
243
.
Fimbres-Weihs
G. A.
Wiley
D. E.
Fletcher
D. F.
2006
Unsteady flows with mass transfer in narrow zigzag Spacer-filled channels: a numerical study
.
Industrial & Engineering Chemistry Research
45
(
19
),
6594
6603
.
Gander
M.
Jefferson
B.
Judd
S.
2000
Aerobic MBRs for domestic wastewater treatment: a review with cost considerations
.
Separation and Purification Technology
18
(
2
),
119
130
.
Ghidossi
R.
Veyret
D.
Moulin
P.
2006
Computational fluid dynamics applied to membranes: state of the art and opportunities
.
Chemical Engineering and Processing: Process Intensification
45
(
6
),
437
454
.
Khalili-Garakani
A.
Mehrnia
M. R.
Mostoufi
N.
Sarrafzadeh
M. H.
Vali
M.
Sadeghi
F.
2009
Flow characteristics in an airlift membrane bioreactor
.
Process Biochemistry
46
(
5
),
1138
1145
.
Kim
J.-S.
Lee
C.-H.
Chang
I.-S.
2001
Effect of pump shear on the performance of a crossflow membrane bioreactor
.
Water Research
35
(
9
),
2137
2144
.
Luo
L.
Li
W.-M.
Deng
Y.-S.
Wang
T.
2005
Numerical simulation of a combined oxidation ditch flow using 3D k-ε turbulence model
.
Journal of Environmental Sciences
17
(
5
),
808
812
.
Stamou
A. I.
2008
Improving the hydraulic efficiency of water process tanks using CFD models
.
Chemical Engineering and Processing: Process Intensification
47
(
8
),
1179
1189
.
Wisniewski
C.
2007
Membrane bioreactor for water reuse
.
Desalination
203
(
1–3
),
15
19
.