This study is to evaluate the utility of a strategy of stepwise increased selection pressure for cultivating aerobic granules. Three sequencing batch reactors (SBR; R1-R3) were used. Substrate ammonium-nitrogen (NH4+-N) concentration was a microbial selection pressure for selecting nitrifying bacteria. It was stepwise increased from 50 to 200mg l−1 in R2 and fixed at 50mg l−1 in R1 and 200mg l−1 in R3. Results showed that aerobic granules failed to be developed in R3 due to free ammonia (FA) inhibition, while they formed quickly in both R1 and R2. Filamentous dominated granules in R1 exhibited poor performance and started to disintegrate from day 131. Stepwise increased selection pressure in R2 selected slow-growing nitrifying bacteria gradually, and this suppressed filamentous growth and improved the stability of aerobic granules. During the 283 days operation, R2 produced aerobic granules with incrementally higher biomass concentration, better settleability, and larger granule size. Larger granules further favored denitrification and the optimal granule size is suggested to be controlled within 2-3mm. This study demonstrates the successful application of the strategy of stepwise increased selection pressure for avoiding the failure of aerobic granulation and improving the stability and performance of aerobic granules.

This paper examines the influence of activated sludge properties such as the mixed liquid of suspended solids (MLSS) concentration, sludge particle size distribution (PSD), extracellular polymeric substances (EPS), soluble microbial products (SMP), suspended solids in supernatant (SSs), dynamic viscosity (µ), relative hydrophobicity (RH), and zeta potential on membrane fouling. ctivated sludge samples from different full-scale and lab-scale membrane bioreactor processes were used to study their impacts on membrane fouling. The influence of activated sludge properties on membrane permeation was identified with statistical methods. The results showed that MLSS concentration had an exponential relationship with membrane fouling resistance. The sludge particle size (rp=−0.730) was correlated inversely to the membrane fouling resistance significantly. The total EPS (rp=0.898) and protein (rp=0.810) had strong positive effect on membrane fouling resistance, but carbohydrate (rp=0.626) had a moderate correlation with membrane fouling resistance due to its low amounts. SMP (rp=0.757), SSs (rp=0.810), dynamic viscosity (rp=0.691), RH (rp=0.837), and zeta potential (rp=−0.881) also had significant influence on membrane permeability. However, protein (rp=0.936), SMP (rp=0.725), SSs (rp=0.783), dynamic viscosity (rp=0.633), RH (rp=0.877), and zeta potential (rp=−0.953) mainly resulted from the change of EPS concentration. These results confirm that MLSS concentration, PSD and EPS were the predominant factors affecting membrane fouling during membrane filtration of sludge suspension. The membrane fouling resistance can be predicted using a simple model based on MLSS concentration, PSD and EPS.

Cake layer formed on membrane surface presents a major challenge to membrane permeation, and it can be considered as a porous media. Determination of the cake layer permeability is critical for an accurate analysis and design of membrane filtration. A permeation model, based on fractal theory and Darcy's law, for evaluating cake layer permeability in microfiltration of activated sludge wastewater was developed. The cake layer permeability was derived and found to be a function of the pore-area fractal dimension and microstructural parameters. The validity of the model was studied systematically in this paper. The permeation model was applied to study the effect of mixed liquid suspended solids (MLSS) concentration, particle size distribution (PSD) and extracelluler polymeric substance (EPS) on cake layer permeability in a submerged membrane bioreactor. Results showed that the permeation model was a useful tool to study the micromechanism of membrane fouling. There was a close correlation between MLSS concentration and cake layer permeability. There were a slight and a distinct decrease of the cake layer permeation as MLSS increased less and larger than 10,000mg/L, respectively. PSD and EPS are two significant factors affecting cake layer permeability in membrane bioreactor. Effect of activated sludge on cake layer permeability was mainly caused by EPS and small particles attached on membrane surface.

大连香格里拉大饭店采用膜生物反应器处理洗浴污水，自2001年10月投产以来运行良好，出水水质达到《生活杂用水水质标准》（CJPT 48-1999）。另外，还将蒸汽冷凝水经与自来水混合降温后用于员工浴池及洗衣房的热水供应，经济效益明显，值得推广应用。

The membrane bioreactor (MBR) is an important process widely used in wastewater treatment. However, membrane flux declines rapidly due to membrane fouling, which limits the application of MBR. In order to alleviate membrane fouling, a precoated dynamic membrane (PDM) was studied systematically. The membrane module consisted of of 56-µm terylene filter cloth as the support membrane and powder-activated carbon (PAC) as the precoating reagent. The precoated dynamic membrane bioreactor (PDMBR) was used in municipal wastewater treatment practice. Meanwhile, from the viewpoint of retention capacity, flux recovery and alleviating membrane fouling, the suitable thickness of the PAC dynamic membrane was also investigated. The experimental results showed that the formation of PDM from PAC on a support membrane belonged to the cake filtering model, and the relationship of PDM resistance (Ra) and its thickness (D) was in accord with log-function; the suitable thickness was controlled around 0.3 mm. It was found that while the flux of PDM was kept at a high value, the operating pressure of PDM rose to 42 kPa after 43 days of stable operation of the PDMBR, and the removal efficiency of effluent COD and NH4+-N were as good as traditional hollow membrane bioreactors. Furthermore, the flux of the fouled membrane could be totally recovered just after being cleaned by brushing, and it did not consume any chemical reagents. Moreover, it was also found that the PDM could prevent biomass contaminations diffusing from the surface inwards by scanning electron microscopy (SEM).