Elsevier

Desalination

Volume 285, 31 January 2012, Pages 285-294
Desalination

Dynamic in-series resistance modeling and analysis of a submerged membrane bioreactor using a novel filtration mode

https://doi.org/10.1016/j.desal.2011.10.015Get rights and content

Abstract

This study is focused on the physical filtration characteristics of a flatsheet membrane bioreactor (MBR) operated under a novel filtration mode. The objective of this research was to demonstrate the possibility of running an MBR with high MLSS concentration for prolonged periods without frequent blocking of the membranes. Current MBR designs, mostly dictated by the manufacturers, have restrictions on the level of MLSS due to fouling. It has been observed that this restraint can be eliminated by applying high shear rates for better removal of cake layer from membrane surface. A pilot scale MBR was setup at the inlet works of a domestic sewage treatment plant. The system was dynamically modeled and calibrated for flux, hydraulic permeability, transmembrane pressure using the in-series resistance model. Resistance components were experimentally determined and compared against the results of dynamic simulations. Intrinsic membrane resistance (Rm) and fouling resistance (Rf) were the major components contributing to total resistance with fractions of 69% (Rm/Rt) and 30% (Rf/Rt) respectively. It was found that cake resistance did not have major impact on the total resistance which was linked to the high aeration intensity. Proposed model was validated by experiments which indicated its potential use on other MBR systems.

Highlights

► In-series resistance model dynamically calibrated to match the experimental data. ► High aeration intensities resulted in low cake resistance throughout the study. ► High shear stress changed the longterm filtration and fouling characteristics. ► MBR plant was operable with less relaxation periods under high aeration intensity. ► Novel filtration mode enabled longterm operation at high MLSS with low viscosity.

Introduction

Membrane bioreactor (MBR) systems proved to be reliable in providing biological treatment and efficient solids–liquid separation in a series of tanks; they attract interest as a viable alternative to conventional activated sludge systems. Many small to large-scale MBR applications have already been constructed and are in operation globally [1]. Since the launch of the commercial membranes housed in membrane modules, manufacturers have been in constant process of improving the filtration characteristics of the membranes to mitigate the negative effects of fouling and loss of permeability. Membrane fouling has become the most important phenomenon in MBR systems; fouling can be defined as the state of operation where the membranes are unable to sustain the operating flux due to an increasing trend in the transmembrane pressure. Existence of extracellular polymeric substances (EPS), soluble microbial products (SMP), biomass concentration, activated sludge viscosity, permeate flux, particle size distribution (PSD), air scouring are known to be the major contributors to the membrane fouling rate.

Meng et al. [2] concluded that biomass (MLSS) concentration, PSD and EPS had considerable impact on the fouling of the membrane. An empirical equation was derived to calculate the fouling resistance with respect to these factors. Wu et al. [3] adopted a novel filtration mode involving high instantaneous flux for 120 s followed by a lower flux for 290 s and a backwash in each filtration cycle. It was concluded that this mode prevented the cake layer from getting compressed and strongly attached to the membrane surface. Arabi and Nakhla [4] compared the fouling behavior of conventional and simultaneous nitrification and denitrification MBRs; 20% higher EPS and 23% increase in the fouling resistance were observed in the MBR performing simultaneous nitrification and denitrification (SNdN) along with larger floc sizes compared to the conventional MBR. Germain et al. [5] concluded that only the permeate flux, solids concentration, carbohydrates in the EPS and the aeration velocity were found to influence the fouling rate on hollow fiber membranes. Ji and Zhou [6] studied the effect of aeration intensity and its impact on membrane fouling. It was found that increasing aeration rates resulted in the decrease of EPS within the flocs attached to the membrane surface which was linked to reduced membrane fouling with high scouring rates.

The fouling behavior and the filtration characteristics of membrane bioreactors encompass numerous variables and many models have been proposed to predict the permeate flux or the increase of TMP with time. The resistance model is the oldest and widely used which is based on the cake filtration theory [7]. A number of approaches exist to estimate the cake and fouling resistance without direct measurements which have been validated using experimental procedures [8].

The phenomenon of cross flow velocity which is affected by the aeration intensity in submerged systems has been defined to be the most critical parameter that has an impact on fouling [5], [6], [8], [9]. The aeration intensity and hence cross-flow filtration sweeps away the particles from the surface of the membrane, thereby maintaining a stable cake layer thickness. Defrance et al. [9] stated that the flux could be increased linearly with increasing crossflow velocity. Germain et al. [5] concluded that high membrane aeration velocities were required to maintain low fouling rates at high solids concentration and that an increase in membrane air flow velocity reduced the membrane fouling rate. Fouling mitigation via aeration in hollow fiber systems were mainly due to the agitation of the membrane fibers and any decrease in air flow resulted in an increasing fouling in parallel to the ascending trend for TMP measurements [10].

A novel filtration mode was adapted to a membrane bioreactor equipped with flatsheet membranes to demonstrate the effect of high shear rate and high aeration intensity on the removal of the cake layer and its subsequent effect on fouling. Long term dynamic modeling and resistance analysis were conducted in this study using the in-series resistance model which was supported by direct (experimental) measurements of the intrinsic membrane resistance, cake resistance and fouling resistance. The filtration model was setup and successfully calibrated to explain the filtration and fouling behavior of the system. Verification of the modeling results with experimental tests showed that the proposed model could be used in predicting the filtration and fouling characteristics of other MBR systems. In this context the objective of this study was to understand and evaluate the filtration and fouling behavior of a MBR system operated under high aeration intensity using modeling and experimental analysis. This was important to assess the viability of operating MBRs at MLSS levels higher than supplier recommendations. It was shown that this MLSS restraint could be eliminated by applying higher shear rates and hence aeration intensity for the effective removal of cake layer.

Section snippets

Materials and methods

The MBR pilot plant used in the study was setup at a municipal sewage treatment plant located in the city of Gebze, Turkey, 50 km east of Istanbul and it was fed with actual raw sewage throughout the study. The crude sewage taken from the inlet channel of the main plant was fed to the MBR system after passing two stages of screening with 1 mm static screen and 0.75 mm bidirectional mesh basket screen. Process scheme was based on predenitrification where both the return activated sludge and the

Experimental

A novel filtration mode based on prolonged continuous filtration for 12 h followed by a long relaxation period for 10 min with increased air scouring was used in the study (Fig. 2). Permeate production was stopped during the relaxation period. Specific continuous scouring rate was 1.0 m3/m2-h and this was increased to1.5 m3/m2-h during the relaxation mode which was much higher than the recommended rates for commercial membranes in the market. The adopted novel filtration mode is quite different

Critical flux and aeration intensity

The concept of critical flux as reported by Field et al. [21] can simply be defined as a flux that does not result in a decline with time when the operating flux is below this value. Above it, there is noticeable flux decline that leads to fouling. The fouling incidents for plants operated with fluxes below the critical flux value is referred to as subcritical flux fouling. Majority of the full scale MBR plants are operated with fluxes lower than the critical flux. Wisniewski et al. [22]

Critical flux evaluation

Short term experiments were conducted during the resistance experiments just after the removal of cake layer with sponge balls. During this time the system was at steady state and the MLSS concentration was approximately 15,000 mg/L. The viscosity of the sludge was steady at 1.0 mPa-s. Cake layer was removed prior to start of the experiment for each aeration intensity investigated.

The applied step height for the filtration experiments was chosen as 3 L/m2-h and the step length was 15 minutes. Six

Conclusions

The following conclusions can be drawn from this study:

  • 1.

    In-series resistance model was successfully calibrated to mimic the measured long term dynamic flux, TMP and hydraulic permeability of the pilot MBR operated with a novel filtration mode which was very much different than the current commercial full scale MBRs.

  • 2.

    Experimental verification indicated that the proposed filtration and fouling model can be applied and used for other MBR systems as the filtration and solids removal mechanisms are

Acknowledgments

This study was conducted as part of the Research & Development Project jointly sponsored by The Scientific & Technological Research Council of Turkey (Project No:TIDEB 3030146) and MASS Treatment Systems Inc., with technical contributions of Istanbul Technical University, Environmental Engineering Department.

References (36)

Cited by (30)

  • Explainable multisensor fusion-based automatic reconciliation and imputation of faulty and missing data in membrane bioreactor plants for fouling alleviation and energy saving

    2023, Chemical Engineering Journal
    Citation Excerpt :

    The detailed explanation of RIS is presented in Section 3.2 and Table S4 in the SI. Transmembrane pressure (TMP) was calculated based on the total resistance (Rt), which includes four resistances: clean membrane (Rm), pore fouling (Rp), dynamic sludge cake (Rdc), and stable sludge cake resistance (Rsc) [53]. During the MBR operation, several interaction mechanisms occur between the ASM-SMP and RIS models.

  • Effect of different diafiltration process on the protein fractionation of skim milk by cross flow microfiltration

    2021, LWT
    Citation Excerpt :

    Therefore, if some decontamination methods were available such as been treated by ultrasonic wave or back wash, the ultimate whey protein removal rate under VVD mode might be the highest among these four modes. CP in skim milk UF was brought about by a build-up in concentration of retained milk protein near the membrane surface due to the rejection of these species (Agashichev, 2006), it exerted a resistance towards mass transfer of solute from the bulk solution to the membrane surface, which caused the gradient of the flux drop(Sarioglu et al., 2012). The observed differences in RCP can therefore be attributed to differences in mass transfer coefficient, which was a function of cross-flow velocity, diffusion coefficient of the solute, density of the feed solution and filtration configuration and dimensions.

View all citing articles on Scopus
1

Tel.: + 90 212 285 33 01; fax: + 90 212 285 63 86.

View full text