Membrane fouling was studied using three pilot-scale submerged membrane bioreactors operated at a series of permeation and aeration conditions to treat municipal wastewater. The transmembrane pressure increases were used to calculate the fouling ratios to compare the relative fouling rates. The results showed that the trends of fouling resistances differed greatly, depending on the permeate flux and mixed liquor characteristics. A stable fouling resistance can result when the filtration is operated at sustainable permeate flux conditions. At the unsustainable permeate flux conditions, the fouling resistance increased exponentially as the filtration progresses. In all the cases, the fouling ratios increased with permeate flux and decreased with aeration intensity. Furthermore, the effects of aeration intensity on fouling ratio is independent of the permeate flux and vice versa. Finally, the variation of fouling ratios at different operating conditions strongly depends on the sludge characteristics of mixed liquor and it appears that more than one parameter of mixed liquor is needed to define their relationship.
Membrane bioreactors (MBRs), combining conventional wastewater activated sludge processes with membrane filtration, yield excellent solid-liquid separation efficiency and high quality effluent. Additional advantages include a small footprint, robust resistance to influent variations, reduced sludge production and modular design. However, their capital and operating costs have limited widespread application in wastewater treatment and reuse even though these costs have drastically declined in recent years due to the process improvement and technological advance in membrane production (Belfort, et al. (1994); Cui and Taha (2003)). Between 1994 and 2000, treatment costs declined by 80% while energy requirements dropped by 85% (Husain, 2005).
Further advances to reduce process costs are yet desirable.
Sparging aeration for effective fouling control accounts for a large portion of the operating energy costs for submerged MBRs. Fouling, which occurs through the deposition and adsorption of dissolved and suspended solids, increases the hydraulic resistance, thereby, limiting the permeate flux across the membrane surface and necessitating the frequent membrane module cleaning and replacement. It is generally accepted that fouling occurs in the form of adsorption, pore deposition, cake formation, concentration polarization and/or biofouling. Their relative magnitudes depend on membrane material and configuration, operating conditions and mixed liquor characteristics. Particles may be transported to the membrane surface by the permeate flux. Depending on the particle-particle interactions, the particle-membrane interactions and the hydrodynamic condition on the membrane surfaces, accumulated particles can be transported back into bulk liquid via Brownian diffusion, inertial lift and shear induced dispersion (Wiesner, et al. (2005)).
Design modifications and operation improvements have led to substantial increases in process efficiency by improving the dispersion of accumulated solids. In early applications, sidestream recirculation was employed to maintain high crossflow velocities to control cake build-up. This method, however, is uneconomical in large scale applications due to process complexities, high energy cost associated with pumping and increased fouling from the breakup of large floc (Shimizu et al. (1996)). With the introduction of submerged operation in the mid 1990’s, the membrane modules were immersed directly in the bioreactor while the effluent was withdrawn by vacuum. Vigorous sparing aeration coupled with the lower fluxes resulting from this operating mode greatly reduces fouling. Nevertheless, the energy required to provide sufficient sparging air usually exceeds that to run a conventional activated sludge plant.
Numerous researches have shown that the membrane fouling decreases with the airflow rate (Chang and Judd (2002); Engelhardt et al. (1998)). A long-term reduction in aeration can also lead to a rapid accumulation of fouling material on the membrane surface (Chang et al. (2002)). However, it has also been shown that for submerged MBRs there is an optimum aeration rate which maximizes fouling suppression (Bouhabila et al. (1998); Le-Clech et al. (2003)). Chua et al. (2002) suggested that beyond this critical value, the fouling may actually increase because the shear force breaks up larger particles and the biomass film. Further, there may be an optimal aeration duration and sequence. Guinzbourg (2003) found that reducing the aeration on-time to off-time ratio by a factor of 6 did not produce a significant change to the TMP across the membrane surface.
Further studies showed that the rates of membrane fouling are specific to each mixed liquor. The extracellular polymer substances (EPS) have been considered among the most important components affecting the membrane fouling but different extraction and analytical techniques have produced conflicting results (Martin (2005)). Differing results have also been reported for the effects of MLSS concentration (Yamamoto et al. (1989)) but generally it is agreed that it poses little direct influence (Rosenberger et al. (2002)). Rather it is the particle size distribution and the quantity of colloidal particles (Fan (2005); Wisniewski and Grasmick (1998)) that appear to contribute significantly to cake formation. Additionally, particle surface charge and pH have been found to affect the filterability of mixed liquor.
The objectives of this study are to assess the effects of key operating conditions on membrane fouling including permeate flux, aeration intensity and aeration sequence, and to examine the roles of sludge characteristics at different operating conditions.