Membrane Distillation: Hollow Fiber Module Design and CFD Modeling

Tuesday, June 12, 2012: 10:00 AM
Poydras (InterContinental New Orleans)
Xing Yang, singapore membrane techonology centre (SMTC), Nanyang Technological University , Singapore, Singapore

Membrane distillation: hollow fiber module design and CFD modeling

Xing Yang, Hui Yu, Rong Wang, Anthony G. Fane Singapore Membrane Technology Centre, Nanyang Technological University, Singapore 639798

Summary

Membrane distillation (MD), which involves mass- and heat-transfer phenomena, is a promising technique for desalination. There are two main challenges for its practical applications, namely the availability of highly permeable membranes and suitable modules with improved hydrodynamics. Hence, this study focuses on: (1) designing novel configurations and accessing module performance for an enhanced MD process in terms of flow distribution, temperature-polarization (TP) effect and mass flux; and (2) computational fluid dynamics (CFD) modeling to analyze the heat-transfer phenomenon in a hollow fiber module. Overall, performance metrics (e.g., heat-transfer coefficients, temperature profiles, TP coefficients, mass flux and thermal efficiency) are examined. The aim of this combined study is to provide a deep insight into the heat- and mass-transfer phenomena in the direct contact MD (DCMD) process and guide further optimization of MD operation for enhancement.

Keywords

Membrane distillation, heat transfer, module design, hydrodynamics, numerical simulation <>1. Background

Membrane distillation (MD) is a thermally driven process that produces high-quality pure water by utilizing a hydrophobic microporous membrane. In spite of many potential benefits, such as low sensitivity to salt concentration and high salt rejection, mild operating temperature and no applied pressure as well as the feasibility to utilize low-grade heat, the MD mass flux tends to be low compared with other membrane processes [1]. Severe TP, which is attributed to poor hydrodynamics in underperforming modules, reduces the actual transmembrane temperature difference and leads to a low water flux. Relatively low thermal efficiency due to conduction through the membrane matrix has also hindered MD commercialization [2]. In recent years, attention has been paid to the MD membrane development and energy analysis [3]. However, there has not been a comprehensive investigation of MD module design. Also, no prior studies have addressed the alleviation of TP and enhancing flux by altering fiber geometries or introducing turbulence aids in a hollow fiber module. On the other hand, to reveal the attributes of a well-designed module and enhanced heat transfer, temperature distributions on membrane surfaces should be fully described. The CFD modelling has proved to be a valuable tool in analyzing fluid dynamic behaviour in membrane modules [4]. It can be used to assist the process design, determine the controlling step in the heat-transfer process and provide guidance for scale-up. However, to date there is very limited CFD modelling of hollow fiber modules in the literature.

Therefore, in a practical manner, this study attempts to develop novel hollow fiber module designs to improve module performance; while on a theoretical level, a series of CFD numerical simulations are conducted to analyze the effects of process parameters in improving module performance. <>2. Material and methods

2.1. Membrane materials

In this study a hydrophobic polyvinylidene fluoride (PVDF) hollow fiber membrane developed for MD applications was characterized experimentally. Synthetic seawater (i.e. 3.5 wt% sodium chloride solution) was used as feed solution and pure water as permeate.

2.2. DCMD experiments

In the MD experimental setup, both the feed and permeate solutions were cycled through the hollow fiber module in a countercurrent mode. On the shell side, the feed solution (313-360K) was circulated (0-12 L°¤min-1). On the lumen side, the permeate (294-327K) was circulated (0-4 L°¤min-1). In the experiments, five novel configurations with modified geometries were fabricated and compared with a randomly-packed module; for CFD model verification, single fiber modules were tested.

2.3. CFD computational domain and algorithms

A series of 2D axial-symmetric single-fiber domains were built. The simulations were carried out using the software Fluent 6.3, with SIMPLE (Semi-Implicit Method for Pressure Linked Equations) algorithm for pressure-velocity coupling and QUICK (Quadratic Upstream Interpolation for Convective Kinetics) algorithm for discretization of the conservation equations.

3. Results and discussion

3.1. Performance comparison of novel hollow fiber configurations

Experiments were performed to study the effects of the Reynolds number of the feed, Ref, on the fluxes in different membrane modules. Fig. 1 shows the permeation flux as a function of Ref for six modules. Compared to the randomly packed module, a one to three-fold flux enhancement could be obtained even at an extremely low flow velocity in four modified configurations. The highest flux enhancement (>300 %) was achieved by the modules with undulating membrane surface (spacer-knitted and curly fibers). This might be due to a more even flow distribution and the occurrence of secondary flows in the modified module that improves the fluid dynamics and heat mass transfer.

 

3.2. Heat transfer analysis by numerical simulations

Fig. 2 shows the local and overall heat-transfer coefficients for membranes with increasing MD coefficient C as functions of fiber length. For different membranes operated under the same hydrodynamic conditions, hm increases significantly with increasing C. For a membrane with a low C of 2.0E-7 kg°¤m-2°¤s-1°¤Pa-1, the heat-transfer resistance of the membrane 1/hm is much higherthan those on the feed/permeate sides. Hence, 1/hm is considered to play a dominant role in the heat transfer process. In this situation, the process enhancement focus should be on the membrane instead of the fluids. As C increases to 1.0E-6 kg°¤m-2°¤s-1°¤Pa-1, local heat-transfer resistances of the membrane and feed flow play equally important roles. In this case, the introduction of turbulence aids is expected to show great effect on the heat transfer.

Fig. 1 Effect of feed velocity on permeation

flux for novel designs

Fig. 2 Local heat-transfer coefficients

distributions for membranes with different C

4. Conclusions

This work focuses on the development of novel designs and CFD studies in hollow fiber modules for DCMD applications. Experimentally, it was found that the improved module performance was attributed to the altered flow geometries, which caused effective boundary layer surface renewal and more uniform flow distribution. Numerically, a new heat-transfer model was established and applied to analyze the dominant effect in the MD process. Strategies for an enhanced system were proposed.

Acknowledgment

Support from Siemens Water Technology is gratefully acknowledged.

References

[1] K. W. Lawson and D. R. Lloyd, Membrane distillation, J. Mem. Sci., 124 (1) (1997) 1-25

[2] R. W. Schofield, A. G. Fane and C. J. D. Fell, Heat and mass transfer in membrane

distillation, J. Mem. Sci., 33 (3) (1987) 299-313

[3] H. Susanto, Towards practical implementations of membrane distillation, Chem. Eng. Process., in press (2011)

[4] H. Yu, X. Yang, R. Wang and A. G. Fane, Numerical simulation of heat and mass transfer in

direct membrane distillation in a hollow fiber module with laminar flow, J. Mem. Sci., 384 (1-2)

(2011) 107-116


Extended Abstract: File Not Uploaded
See more of this Session: Membrane Distillation
See more of this Group/Topical: General Submissions