Modelling bioprocesses and membrane fouling in membrane bioreactor (MBR): A review towards finding an integrated model framework

Bioresource Technology 122 (2012) 119–129

Contents lists available at SciVerse ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Modelling bioprocesses and membrane fouling in membrane bioreactor (MBR): A review towards finding an integrated model framework M.F.R. Zuthi, H.H. Ngo ⇑, W.S. Guo Centre for Technology in Water and Wastewater, School of Civil and Environmental Engineering, University of Technology Sydney, Sydney, NSW 2007, Australia

h i g h l i g h t s " State-of-the art in mathematical modelling studies of MBRs were summarized. " Integrated mathematical model for MBRs and membrane fouling is yet to be developed. " Interaction between MBR processes and membrane limit the application of ASMs to MBR. " Empirical expressions of model parameters for calibration are essential. " Switching functions between model variables could exclude less important parameters.

a r t i c l e

i n f o

Article history: Available online 2 May 2012 Keywords: Membrane bioreactor Membrane fouling Activated sludge model Integrated model

a b s t r a c t The bioprocesses taking place in activated sludge wastewater treatment system itself are characterized by great complexity and yet incomplete understanding of some of the phenomena involved. The MBR technology inherent deficiencies for its simulation due to additional intrinsic complexities resulting from the interaction between concurrently occurring and dynamic biological processes with membrane filtration and the straightforward adoption of the activated sludge models’ (ASM) frameworks or their modified variations. In this backdrop, this paper compiles a brief overview of the previous developments to the current state-of-the-art mathematical modelling approaches of the MBR system. With extended discussions on particular topics such as applications of modified ASMs to MBR modelling, ASM extensions incorporating soluble microbial products (SMP)/extracellular polymeric substances (EPS) concepts, this paper also provides a guide for different end-users of mathematical models of MBR systems. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Wastewater treatment has been a challenge throughout the past few decades due to varying influent characteristics and stringent effluent regulations. Standalone biological wastewater treatment systems have been able to handle most of these difficulties, but at the expense of huge economical cost to achieve the desired effluent water quality particularly at medium to large wastewater treatment facilities. In this backdrop, Membrane bioreactors (MBRs) are being increasingly implemented to treat and reuse wastewater due to their many advantages over conventional activated sludge (CAS) processes. Besides membrane fouling and the high cost of membranes as identified main obstacles for wider application of MBRs, recent studies have reported several crucial specificities of the treatment system. Although several innovative lab-scale MBRs could show convincing achievements to reduce few of the problems, those innovations are yet to be reflected into implemented treatment ⇑ Corresponding author. Tel.: +61 2 95141693; fax: +61 2 95142633. E-mail address: [email protected] (H.H. Ngo). 0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.04.090

facilities of MBRs. There exists, therefore, a pressing demand to develop appropriate integrated models which can translate experimental results of various lab-scale MBRs and also can validate those against true observations at implemented MBR systems. This obviously will require intelligent modelling of the biokinetics, membrane fouling and hydrodynamics of the MBR system resulting desired outcomes against treatability and fouling control targets. The efforts for modelling of wastewater treatment systems, so far, have always targeted either the treatment quality targets or various aspects of system management. In an attempt to develop mathematical models for the MBR system focusing on the biological processes only, the activated sludge models (ASMs) (Henze et al., 1987, 1995, 1999; Gujer et al., 1999) formerly developed for CAS processes, have been applied with or without modifications to simulate biomass kinetics of the MBR systems. However, buildup of SMPs and/or extracellular polymeric substances (EPSs) can cause reduction in membrane permeability (Rosenberger et al., 2006; Ahn et al., 2006). For this reason, formation and degradation kinetics of SMPs (Lu et al., 2001; Wintgens et al., 2003; Jiang et al., 2008) and EPSs (Ahn et al., 2006) have been introduced in

120

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

the ASMs derived hybrid models assuming their useful role for understanding and controlling fouling phenomena. Different approaches have also been developed for modelling the physical and biological aspects of membrane fouling, i.e., fractal permeation model (Meng et al., 2005), empirical hydrodynamic model (Liu et al., 2003) and sectional resistance model (Li and Wang, 2006). On the other hand, to improve the knowledge about the correlation between the biological processes and various membrane fouling phenomena, integrated (hybrid) MBR models have been proposed which combines biomass kinetic models with membrane fouling models. However, most of these integrated MBR mathematical models (Lu et al., 2001; Lee et al., 2002; di Bella et al., 2008; Zarragoitia et al., 2008; Janus and Ulanicki, 2010; Mannina et al., 2011) have focused only on few targeted aspects of steady-state biological processes linked to membrane fouling models leaving many unresolved issues in models’ calibration. Several reviews of modelling studies on MBR have highlighted some issues in this regard, i.e., mini-review of MBR modelling studies (Ng and Kim, 2007), ASM based modelling of MBR (Fenu et al., 2010), bioprocess modelling of MBR system and fouling predictions (Patsios and Karabelas, 2010). This review aims to synthesize previous studies and current state-of-the-art in MBR modelling studies with special regard to some specificities of the treatment system. Fundamentals of correlation between MBR bioprocess and fouling phenomena are discussed first in order to guide the reader to understand missing links in unmodified and modified ASM based modelling of MBR systems. Particular emphasis is placed on the discussion about the MBR bioprocesses and SMP/EPS modelling which have influences on membrane fouling. Significant efforts have been paid to identify key variables of mathematical MBR models which may establish an integrated framework for coupling MBR bioprocesses with fouling. 2. General overview of MBR process kinetics and stoichiometry 2.1. Crucial specificities of MBR biological processes Though the CAS and MBR systems are similar from a biochemical engineering viewpoint, recent studies have reported several crucial specificities of the MBR system, i.e., medium to very high sludge retention times (SRT), high mixed liquor suspended solids (MLSS) concentration, accumulation of SMPs rejected by membrane filtration, and high aeration rates for scouring and good nitrification performance (Ng and Kim, 2007; Fenu et al., 2010). High MLSS concentrations in MBR system causes its operation at high viscosities effecting the energy requirement for pumping, air scour of the membranes and oxygen supply of the micro-organisms. The difference between MBR and CAS systems in terms of sludge characteristics and performance is especially pronounced at high SRTs when deterioration of sludge settleability and effluent quality of the CAS is observed at relatively higher SRTs. Above all, the major process problem associated with MBRs is the membrane fouling which often contributes to huge system operating cost. The rejected constituents in the retentate tend to accumulate at the membrane surface, and as a consequence a reduction in the permeate flux at a constant transmembrane pressure (TMP) or conversely an increase in the TMP at a constant flux is observed. A balance between flux, TMP, energy demand and cleaning frequency is, therefore, very crucial in case of MBR process design. There are further operational issues such as greater foaming propensity, a less readily dewaterable sludge product and generally greater sensitivity to shock loads of MBR compared to CAS. 2.2. Process kinetics of membrane fouling in MBR Membrane fouling in an MBR can be classified into different categories at different stages of operation of an MBR. Reversible

fouling refers to fouling that can be removed by physical means such as backflushing, while irreversible fouling refers to fouling which can only be removed by chemical cleaning (Judd, 2006). The fouling that occurs over long periods cannot be removed by any cleaning and termed as irrecoverable fouling. Since deposits are brought to the membrane mainly by convective transport, the rate of fouling dominantly depends on the velocity orthogonal to the surface – the permeate flux (Drews, 2010). Traditionally, three factors thought to affect fouling are membrane, sludge characteristics and operation (Chang et al., 2002; Le-Clech et al., 2006). Aeration ports and module dimensions have been added to the original three factors and make up the group of relevant design parameters for the MBR system (Judd, 2006). In fact, the rate of fouling depends on various other interrelated parameters making the correlation between flux and fouling rate a dynamic variable. As there still remains a lack of fundamental understanding of the kinetics involved, the term fouling is often used to lump all phenomena that lead to a loss in permeability. Numerous researches have been conducted to identify factors that reduce fouling and thus to increase permeate flux i.e., Ngo et al. (2008) have identified that increasing the attached growth biomass by introducing polyesterurethane sponges in the bioreactor MLSS increases sustainable flux with consequences to reduced membrane fouling. Membrane fouling in a typical MBR system occurs due to the following general mechanisms (Meng et al., 2009): (i) adsorption of solutes or colloids within/on membranes; (ii) deposition of sludge flocs onto the membrane surface; (iii) formation of a cake layer on the membrane surface; (iv) detachment of foulants attributed mainly to shear forces; and (v) the spatial and temporal changes of the foulant during the long-term operation. While there are more or less established theories to explain fouling mechanisms by adsorptiondeposition, cake layer formation and detachment of foulants from the cake layers during short-term operation of an MBR system, researchers are still faced with the difficulties of linking the spatial and temporal changes of foulants during its long term operation. Since the long-term operation of an MBR is typically conducted at a flux lower than the critical flux, the rate of particle convection towards the membrane surface is usually balanced by the rate of back transport and hence the particulate fouling is not a dominant problem to deal with. Experimental evidences suggest that the critical thing to deal here is the change of bacterial community and biopolymer components in the cake layer with time. The rate of membrane fouling aggravates with time principally due to the high rate of cell lysis of the biopolymer components of the cake layer. This gradually results in more of the membrane pore closure due to the higher rate of trace foulants’ adsorption as compared to the rate of desorption and back-transport of such trace foulants. These ultimately increase the specific cake resistance, cake compressibility and irreversibility during the long-term operation of an MBR system. 2.3. Fractions of MLSS responsible for membrane fouling Unlike the simple mathematical expressions proposed (Ishiguro et al., 1994; Fan et al., 2006; Liang et al., 2006; Busch et al., 2007; Guglielmi et al., 2007) to describe the relationship between the concentration of dissolved organic matter (DOM) and membrane flux decline or TMP change, the MLSS or the mixed liquor volatile suspended solids (MLVSS) concentration has a rather complex relation to MBR fouling. Kornboonraksa and Lee (2009) found that MLSS and sludge floc size are the dominant factors that control the membrane filterability in an MBR treating wastewater. Different research groups have reported various empirical mathematical expressions describing membrane flux or fouling rate which included MLSS/ MLVSS concentration, although contradictory findings about the effect of these parameters on membrane filtration have also been

121

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

reported (Le-Clech et al., 2006; Meng et al., 2009). MLSS/MLVSS concentration, as a lump parameter, represents different kinds of suspended organic matter with possibly different fouling propensity. The major focus with regard to fouling has recently turned to sticky substances which can be bound to the flocs or freely suspended. These groups of compounds are mostly termed EPS when they are bound to the flocs or SMP when freely suspended in the supernatant (Drews, 2010). However, there still remains disagreement among the researchers regarding the definition of EPS and SMP. Patsios and Karabelas (2010) have defined soluble EPS (sEPS) or SMP as a biodegradable fraction of DOM. Recently, the terms biopolymers or biopolymeric clusters (BPC) (Wang et al., 2007; Lin et al., 2009; Sun et al., 2011), neither microbial nor EPS are nonfilterable organics, and the transparent exopolymer particles (TEP) (De la Torre et al., 2008, 2010), specially sticky fraction of EPS, have also come into use. Table 1 gives few empirical relationships between these fractions of MLSS and membrane fouling based on findings reported in literature. By definition, all these groups of compounds are produced and excreted by microorganisms. However, what is analysed as EPS, SMP, BPC or TEP by commonly agreed on methods is not only necessarily of microbial origin but can also be terrestrial or man-made (Judd, 2006). Although, the location of the fouling relevant fraction of these fractions and also the conditions that shift it to different locations are still unknown. 2.4. Correlation between biological process variables and fouling in MBR The biodegradable and non-biodegradable organic matter present in the mixed liquor supernatant (usually referred to as DOM) appear to play a significant role during membrane filtration process of the MBR system. A large portion of DOM consisting of soluble and biodegradable organics of microbial origin is frequently referred to the literature as sEPS or SMP. The DOM are adsorbed onto and into the membrane during membrane filtration, block membrane pores and form a partly irreversible gel structure on membrane surface and into the membrane pore causing fouling. Reviewing the relationship between membrane fouling and DOM concentration Le-Clech et al. (2006)have identified proportional relationship between the loss of hydraulic performance and DOM. Regarding relevant influence of DOM composition, a direct relationship between the carbohydrate levels in DOM with parameters of fouling propensity has been observed. Rosenberger et al. (2006)have observed linear relationship between filtration resistance over time and polysaccharide (PS) concentration in sludge supernatant. The observed correlations of DOM and filtration resistance have also been accounted in various mathematical modelling efforts.

Ishiguro et al. (1994) proposed a simple mathematical expression to describe membrane flux as proportional to the difference of DOM concentration between the mixed liquor and permeate. Liang et al. (2006) developed mathematical model for both reversible and irreversible fouling rate where permeate flux decline or TMP rise can be determined with model inputs of DOM concentration. Fan et al. (2006) recommended an empirical mathematical expression based on the observed relation between the critical flux and colloidal total organic carbon (TOC). Busch et al. (2007) presented a detailed model of fouling mechanisms for a submerged hollow fiber filtration in which different model variables for different fouling mechanisms literally represents the DOM concentration. Guglielmi et al. (2007) established a subcritical flux fouling model that could predict the time at which a sharp change in the TMPtime profile occurs considering it inversely proportional to the concentration of the DOM. The concentration of EPS seems to play an important role in the regulation of DOM concentration assuming these two organic fractions are closely interrelated. Patsios and Karabelas (2010) recently conducted an analysis by scanning electron micrographs (SEM) of suspended biomass aggregates which found bacteria are embedded and mostly immobilized within the slime matrix of EPS. Xuan et al. (2010), after comparing the contribution of granular and flocculent sludge to fouling, concluded that membrane filtration decreases with increasing EPS content in flocculated sludge. However, EPS are found to influence considerably activated sludge structural characteristics as well as physico-chemical properties (Le-Clech et al., 2006; Meng et al., 2009) which make their impact on the fouling process rather complicated. Based on the observation of the same nature of DOM and EPS in LC-OCD chromatograms, Rosenberger et al. (2006) hypothesized that their relative concentrations are under a dynamic equilibrium which can easily be shifted by changing conditions in the mixed liquor environment. Various processes and/or conditions that result either in the biosorption of DOM by the bioflocs, or the hydrolysis of EPS and the release of DOM in the bulk liquid (Nielsen et al., 1997) affect the EPS–DOM equilibrium. 3. Applications of biomass kinetic models to MBR 3.1. ASM model families The first product of ASM model families, called activated sludge model No1 (ASM1), is the outcome of the work of a task group formed in 1983 by the International Association on Water Pollution Research and Control (now known as the International Water Association – IWA). The model presented in 1987 (Henze et al. 1987) was

Table 1 Fractions of MLSS and their relationship with membrane fouling (modified after Meng et al., 2009). Fractions of MLSS

Relation with membrane fouling

Reference

EPS

Carbohydrates of EPS",? (tends to influence) clog membrane EPS, hydrocarbon components and inorganic matters govern membrane fouling layer " extracellular polymeric substances (EPS), ? high membrane resistance indicating severe membrane fouling " EPS, fouling rate" Bound EPS affects cake specific resistance

Dvorák et al. (2011) Pendashteh et al. (2011) Chae et al. (2006) Drews et al. (2006) Cho et al. (2005)

SMP

Hydrophilic fraction (in carbohydrate): major cause for membrane fouling "SMP, "fouling rate and membrane fouling index (MFI) Hydrophilic neutrals (carbohydrates) responsible for high fouling potential at short SRTs SMP influence fouling only under certain conditions such as low sludge age and large pore size " MLSS concentrations, ;normalize permeability SMP and polysaccharides influence fouling more than MLSS.

Pan et al. (2010) Arabi and Nakhla (2009) Liang et al. (2007) Drews et al. (2008) Trussell et al. (2007) Zhang et al. (2006)

BPC

Concentration ", fouling" BPC along with SMP and EPS governs membrane fouling "BPC concentration, "filtration resistance

Sun et al. (2011) Lin et al. (2009) Wang et al. (2007)

TEP

TEP more important for fouling than CH, proteins or total EPS " TEP, ? reach the critical flux sooner, ;the mixed liquor filterability

De la Torre et al. (2010) De la Torre et al. (2008)

122

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

basically intended to model biological wastewater treatment process for organic carbon removal, nitrification and denitrification but without phosphorus removal. IWA later introduced other models of the ASM families which were expanded and improved upon the reference ASM1 but additionally included the features of Phosphorus removal from the wastewater and other relevant issues. Table 2 presents some important parameters of ASM model families which are relevant to the following discussions. Two main concepts have been incorporated into the formulation of ASM1 using total chemical oxygen demand (COD) as the suitable parameter for defining the organic matter in the wastewater. The first concept is that readily biodegradable COD (RBCOD) can immediately be used by organisms for synthesis, whereas slowly biodegradable COD (SBCOD) must be broken down to be metabolized. The second concept in the model is death-regeneration. Many of the model’s concepts like Monod type kinetics, the use of COD as organic matter are still considered state-of-the-art and are widely employed by most of the CAS process models. Nevertheless, a number of simplifications and assumptions made in the model have limited its application, especially when industrial wastewaters dominate the wastewater characteristics. ASM2 (Henze et al., 1995) extended the capabilities of ASM1 by including biological phosphorus removal. The model incorporates a new group of organisms to the biomass, called the phosphorus-accumulating organisms (PAOs), which originally consist of heterotrophs and autotrophs. PAOs are assumed to be incapable of denitrifying activity and can only grow on stored cell internal organic material XPHA under aerobic condition. Separate process rates are provided in the model to capture all losses of biomass due to respiration

or death. ASM2d has additionally incorporated that PAOs could use internal cell organic storage for denitrification and grow under anoxic conditions leading to the addition of two corresponding rate processes in the model. In 1999, ASM3 (Gujer et al., 1999) was presented by the IWA task group in which the major changes compared to ASM1 is the inclusion of internal cell storage compounds in heterotrophs, shifting the focus from hydrolysis to the storage of organic substrates. The death-regeneration concept of ASM1 is replaced in ASM3 by the growth-endogenous respiration model. The model inherent deficiencies to model industrial wastewater treatment with changed wastewater characteristics. 3.1.1. ASMs to MBR modelling: requisite features and deficiencies The application of ASMs are presumably meant for CAS process operation in the ranges of operating parameters, e.g., SRT range 3–15 d, HRT range of 3–5 h and MLSS range 1.5–4 g/L for completely mixed systems (Tchobanoglous et al., 2003). A recent study on design and operating experience with municipal MBRs in Europe has reported the ranges of various parameters (Itokawa et al., 2008). Some key differences in operating conditions between CAS and MBR process parameters along with the aforementioned crucial specificities of MBR process, therefore, obviously raise the question as to what extent the ASM framework is applicable to MBR processes. Studies so far suggests that the applicability of ASMs for modelling MBR, in their original form, needs to be carefully verified especially to understand the effects of higher SRTs and mixed liquor suspended solids (MLSS) concentrations on biomass. A bioprocess mathematical sub-model for MBR systems should necessarily

Table 2 Comparison of ASM models with regard to the simulation of MBR bioprocess. Model Major components

ASM1

Total parameters Total processes Bioprocess simulation – COD removal – Nitrification – Denitrification – Hydrolysis Phosphorus removal Key variable for MBR fouling predictions – SMP – EPS

Yes Yes Yes Yes No

Yes Yes Yes Yes Yes

Yes Yes Yes Yes No

No No

No No

No No

Active heterotrophic biomass (XB,H) Active autotrophic biomass (XB,A) Readily biodegradable substrate (SS) Slowly biodegradable substrate (XS) Oxygen (negative COD) (SO) Ammonia/ammonium nitrogen (SNH) Nitrate and nitrite nitrogen (SNO) Particulates from biomass decay (XP)

Aerobic growth of XB,H Anoxic growth of XB,H Aerobic growth of XB,A Decay of XB,H Decay of XB,A

– – – – – – – –

ASM3

Different reaction process rates depend on parameters such as maximum growth rates lH, lA; saturation coefficients KS, KN,H, Ko,H, KO,A, heterotrophic yield coefficient YH, decay coefficients bA, bH 13 8

Major rate processes

Major parameters/ coefficients in rate processes

ASM2/ASM2d

– Dissolved oxygen (SO) – Inert particulate organic matter (XI) – Readily biodegradable organic substrate (SS) – Active heterotrophic organisms (XH) – Nitrate and nitrite nitrogen (SNOx) – Cell internal storage product of XH (XSTO) – Ammonium plus ammonia nitrogen (SNH4 ) – Suspended solids (XSS) Aerobic/anaerobic/anoxic hydrolysis Hydrolysis aerobic/anoxic storage of SS Aerobic growth of XH on SF/SA Aerobic/anoxic growth of XH Anoxic growth of XH on SF/SA Aerobic/anoxic endogenous respiration of XH Storage of XPHA, XPP, Lysis of XH, XPAO, XPP, XPHA Aerobic endogenous respiration of XSTO Aerobic growth of XAUT Aerobic growth and endogenous respiration of XA iN,Xs and iP,Xs for N and P contents of XS, saturation Y STO;O2 and YSTO,NO for aerobic and anoxic coefficients K for oxygen and ammonium, storage of XSTO, fXI for aerobic and maximum growth rate of autotrophs lAUT, Yield of endogenous respiration of XI, yield heterotrophs and PAO, qPHA , qPP rate constants for coefficient sYH for XH and XPAO, etc. storage etc. 19/20 13 19/21 12

– – – – – – – –

Fermentable substrate/products (SF/SA) Nitrate and nitrite nitrogen (SÞNO3 SNO3) Dissolved oxygen (SO2 ) Ammonium and ammonia nitrogen (SNH4 ) Inorganic soluble phosphorus (SPO4 ) Total suspended solids (XTSS) Phosphate accumulating organisms (XPAO) Cell internal storage products of PAO (XPHA)

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

describe the elevated concentration of DOMin MBR and the higher MLSS concentration. As the physicochemical characteristics of the mixed liquor suspension and supernatant inevitably affect the filtration performance, models of biomass activity in MBR should have a dual perspective of adequately describing the complex biological processes as well as accounting for the biomass characteristics that affect membrane filtration performance (Patsios and Karabelas, 2010). In particular, a basic model of biomass kinetics in MBR should at least be capable of providing estimates of the EPS concentration in the activated sludge flocs and the concentration of DOM in the bioreactor supernatant taking into account the existence of SMP. 3.2. Applications of unmodified ASMs to MBR modelling The application of ASMs for mathematical modelling of MBR system started soon after the ASM1 had gained popularity for modelling CAS processes. An early investigation (Chaize and Huyard, 1991) using the default parameter values of ASM1 has found that a non-calibrated ASM1 is able to give a reasonable estimate of the MBR effluent COD and TKN while it fails to predict fairly accurate solids concentration particularly at very low HRT and very high SRT in MBR systems. However, recent efforts have presented various aspects of systematic calibration of ASMs’ key sensitive parameters along with emphasis on the influence of influent wastewater characterization in terms of various ASM based fractions (Delrue et al., 2010; Spèrandio and Espinosa, 2008). The wastewater characterization, in this regard, has conventionally been done using two different approaches: (i) integration of chemical, physical and biological methods; and (ii) application of ‘‘trialand-error’’ procedures. Jiang et al. (2005) compared two methods of wastewater characterization in order to calibrate ASM1 in a side-stream membrane bioreactor. Significant differences between the two approaches were observed, especially the inert particulate organics (XI) fraction was found significantly higher when determined with the ‘‘physical–chemical’’ method causing overestimation of the MLSS. In a simulation of a bench-scale MBR, Spèrandio and Espinosa (2008) observed the MLVSS trend indicating ASM1 is able to predict sludge production at low SRTs, while an overestimation was observed at high SRT = 110 days. ASM3, in contrast, showed underestimation of MLVSS with low SRT values but slightly better prediction of the same at SRT of 110 days. The application of modelling tools for wastewater characterization in a full-scale MBR, on the other hand, usually favours ‘‘trial and-error’’ procedures for influent fractionation. Certain factors of the MBR system may have dominant effects on process kinetic parameters such as the sludge suspended solids (XTSS) impacting the excess sludge production and the oxygen transfer rate (a-factor), the removed and residual nitrogen species (SNH, SNO), the residual phosphorus concentration (SPO4 ) and oxygen consumption rate (OUR, and SO) (Fenu et al., 2010). Different researchers have therefore tried adjusting ASMs’ kinetic and stoichiometric parameters for matching model prediction with experimental data. Jiang et al. (2005) calibrated ASM1 for modelling a side-stream MBR through defining Relative Sensitivity Function (RSF) for several model variables towards all model parameters. The stoichiometric parameter (YH) and kinetic parameters (bH, bA, lmaxH and lmaxA) were found very to moderately influential on the MLSS concentration and effluent quality. However, changes of local sensitivity based on operating conditions indicate also the necessities of sensitivity analysis for each calibration exercise (Fenu et al., 2010). Regarding the associated impact of MBR on the nitrification kinetics, Jiang et al. (2005) observed that ammonium consumption and residual concentration is sensitive to nitrification parameters in decreasing order of influence: lA, bA, KNH, KOA and YA. These findings might however also depend on a different mass balance. Hocaoglu et al. (2011), in a MBR modelling study with modified

123

ASM1found that both nitrification and denitrification kinetics vary as a function of the sludge age and the calibrated values of half saturation coefficients need to be reduced when the sludge age is lowered. In contrast to the nitrification process, denitrification is apparently less modified by the membrane configuration (Parco et al., 2007). Parco et al. (2007) observed that kinetic parameters for biological P-removal are comparable in MBR and CAS. Using ASM2d default parameters, Jiang et al. (2008) found overestimated nitrate concentration but an underestimated phosphorus concentration. The above mentioned parameters deserve special attention regarding MBR modelling studies and need to be measured with proper methods. However, the database of experimental findings of such parameter values is still insufficient and some have dependencies on particular experimental technique. Because of this lack of coherence, it is not appropriate to suggest best set of model parameters for MBR modelling studies. It may rather be useful to suggest empirical correlation in order to calculate/assume parameter values and then apply trial–error procedure for model calibration. 3.3. Modified ASMs’ application to MBR modelling Compared to the CASP, large fractions of flocs, bacteria, biopolymers such as polysaccharides, proteins and organic colloids are mostly retained MBR bioreactor which may significantly change the biodegradation kinetics within the bioreactor. This change is more significant for polysaccharides than for proteins. Current ASM models do not distinguish between protein and polysaccharide fractions although in MBR systems with typically low organic loads, the retained molecules may have a significant impact on the metabolic path, allowing further use of carbon based metabolites (Furumai and Rittmann, 1992). Ignoring SMP and EPS formation could thus lead to a general overestimation of true cellular growth rates and this would severely under predict the COD effluent (Jiang et al., 2008). Mathematical modelling of MBR systems to incorporate the above mentioned specific features have been usually attempted through modifying ASMs with SMP/EPS concepts. When introducing the SMP concentration as a state variable in the model, one of the key issues for COD fractionation is the determination of inert components (Si, Xi) with proper methods. In this connection, an early modification of ASM1 has been proposed by Lu et al. (2001) by modelling a bench-scale MBR fed with synthetic sewage. Baek et al. (2009), most recently, applied modified ASM1 to an aerobic membrane bioreactor. Negligible XBA and Xp were assumed in the model. The model discarded SALK because pH in the bioreactor is well maintained at neutral pH. In general, some models have been developed as stand-alone descriptions of the concepts of production and degradation of both SMP and EPS while others have focused on integrating only the SMP concepts into the ASM type of models. 3.4. Standalone SMP/EPS models 3.4.1. Stand-alone SMP models SMPs are defined as soluble cellular components or debris that are released during cell lysis, diffuse through the cell membrane, are lost during synthesis, or are excreted for some other purpose (Fenu et al., 2010). Substrate utilization, biomass decay, and EPS hydrolysis are the major processes responsible for the SMPs’ formation. A simplified definition of SMPs, however, excludes intermediate products of non-microbiological origin and defines it as any soluble material which leaves as effluent from a biological system that was not present in influent (Barker and Stuckey, 1999). Despite there still remains disagreement about exact kinetics of SMP, it is crucial to include SMPs in the modelling of CASPs/MBRs.

124

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

To model biomass accurately without calibration using only experimental data is one of the specific advantages of SMP model over ASMs. However, researchers have so far encountered great difficulties in the determination of individual fractions of SMP. The correlation of SMP fractions with other parameter makes the MBR modelling a quite complex task resulting in large number of variables in the model formulation. The SMP model also does not incorporate biological phosphorus removal. Namkung and Rittmann (1986) were the first to incorporate SMP formation kinetics in an extended steady state biofilm reactor model. The model was based on SMP sub-division into utilization associated products (UAP) and biomass associated products (BAP). UAP is controlled by the specific substrate utilization rate and comprised of the direct by-products of substrate utilization and cell growth whereas BAP is controlled by the cell concentration and is independent of cell growth rate. BAP can be considered a byproduct of endogenous respiration of cell mass. A recent modelling study (Jiang et al., 2008) has further hypothesized about two types of UAPs (lower and higher molecular weight) based on the utilization of storage associated products. Nevertheless, there is no general consensus regarding the BAP production and degradation mechanisms. BAP is produced solely by EPS hydrolysis (Laspidou and Rittmann, 2002a, b) while Aquino and Stuckey (2008) have recently demonstrated that both sEPS and cell lysis products are the sources of BAP. In addition to the direct degradation mechanisms of BAP, Jiang et al. (2008) have reported that BAP could also be hydrolyzed yielding fermentable COD. Table 3 presents some of the empirical expressions that have been proposed by different researcher from their respective findings about the UAP and BAP formation and degradation kinetics. However, the findings in general suggest that the SMP equilibrium concentration is not a steady state parameter, but a dynamic variable which not only results from production/degradation mechanisms but also SMP retention by the membrane (Fenu et al., 2010). 3.4.2. Stand-alone EPS models Although the importance of EPS to cell aggregation has long been recognized, there has been limited modelling study on EPS formation. Luedeking and Piret (1959) proposed the first model (Eq. (1)) to characterize microbial products formation from the fermentation of lactic acid including only the consistent EPS formation as a growth and a non-growth associated product.

r EPS ¼ k1 lX þ k2 X

ð1Þ

The first term of the equation accounts for EPS formation associated with a first-order growth and the second term represents EPS formation associated with a non-growth term. This very simple model did not include the mechanism of EPS dissolution. An alternative kinetic expression (Eq. (2)) for EPS formation was introduced by Hsieh et al. (1994). They have developed a mechanistic model to express the production of EPS which includes an EPS loss term as indicated by the third term of Eq.

(2). The dissolution was modeled with Monod kinetics but some terms involved in expression containing K3 make the rate changes complicated.

rEPS ¼ k1 lX a þ fK 2 X a  ldiss K 3 EPS

Laspidou and Rittmann (2002a,b) studied the EPS mass balance (Eq. (3)) in a continuous flow reactor which was latter applied in a submerged membrane bioreactor (Jang et al., 2006). The second term in Eq. (3) quantifies the rate of EPS loss due to hydrolysis, using a first order relationship

rEPS ¼ K eps r s X a  khyd EPS

ð3Þ

The modified model has been based on the hypothesis that the formation of bound EPS is only growth associated, and is in direct proportion to substrate utilization. Aquino and Stucky (2008) proposed to model the formation of EPS as a non-growth associated product in anaerobic condition. It is considered in the model that both soluble EPS and cell lysis products are the sources of BAP whereas in Laspidou and Rittman model hydrolysis of EPS is considered the only source of BAP. 3.5. ASM extensions incorporating SMP/EPS concepts Ohron et al. (1989) first attempted to integrate of the formation and degradation kinetics of SMP into the ASM1. The model was calibrated and verified with data from a sequencing batch reactor. The simple mathematical model developed by them includes the so called soluble residual products SRP (equivalent to non-degradable BAP) assuming that only contributed significantly to the soluble COD of the mixed liquor. Later Artan et al. (1990) further developed the model to include UAP. However, strong parameter correlations due to the model’s combined concepts of formation and degradation of UAP and BAP affect their correct determination. Lu and coworkers, on the contrary, incorporated a very complex SMP model into ASM1 (Lu et al., 2001) (Fig. 1(a)) and ASM3 (Lu et al., 2002) in MBR studies. Most original parameters of ASM1 were used, but the denitrification correction factor was enhanced to account for higher sludge concentrations in the system. The model contains eight SMP related parameters which can be determined by trial and error or can be approximated from references in literature. Besides significant underestimation of MLVSS concentration, the ASM1-SMP model predictions have been found close to the experimental observations for an intermittent aerobic MBR system. OliveiraEsquerre et al. (2006) proposed modification of ASM3 by introducing five new SMP kinetic parameters (cMP,H, cMP,A, kMP, fb, Ymp with values adopted from Lu et al. (2001), and two new processes. UAP and BAP are lumped into a general term MP in the modified ASM3 (Oliveira-Esquerre et al., 2006). Evaluation of the both model predictions for a submerged MBR system has showed that the carbonaceous materials were more accurately estimated by the modified ASM3, while the model of Lu et al. (2001) performed slightly better in the estimation of nitrate.

Table 3 Formation and degradation kinetics of SMP fractions (modified after Fenu et al. (2010)). Parameters

Equations

References

UAP formation rate

r UAP ¼ lhet X het or r UAP ¼ laut X aut or rUAP ¼ ðk1 ) quap S=K s þ SÞX bm rUAP ¼ ðcuap =YÞ  l  X

Lu et al. (2001) Laspidou and Rittmann (2002a,b) Janus and Ulanicki (2010)

UAP degradation rate

r UAP ¼ lSMP ðSSMP =K SMP þ SSMP ÞX het rUAP ¼ ð ) quap UAP=K uap þ UAPÞX bm rBAP ¼ K hyd EPS rBAP ¼ K 2 X þ K hyd EPS r BAP ¼ fbap  b  X þ ð1  f3 Þ  K h;EPS  EPS  Y BAP rBAP ¼ K h;bap Sbap X H rBAP ¼ ðquap BAP=K bap þ BAPÞX bm

Lu et al. (2001) Laspidou and Rittmann (2002a,b) Laspidou and Rittmann (2002a,b) Aquino and Stuckey (2008) Janus and Ulanicki (2010) Jiang et al. (2008) Laspidou and Rittmann (2002a,b)

BAP formation Rate

BAP degration rate

ð2Þ

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

125

Fig. 1. Schematic of the (a) ASM1-SMP hybrid model (adapted from Ng and Kim, 2007) (b) ASM1-SMP–EPS hybrid model (modified from Ahn et al. (2006)).

Jiang et al. (2008) has extended the existing ASM2d to ASM2dSMP by introducing kinetics for formation and degradation of SMP by hydrolysis steps, creating three new processes and imposing variations in thirteen other processes. The study has revealed the SRT as the key parameter controlling the SMP concentration. Incorporating both EPS and SMP in a model may offer a rational approach to the biofouling characteristics of the membrane bioreactor process. In this connection, Ahn et al. (2006) have included EPS (Fig. 1(b)) in the ASM1-SMP model (Lu et al., 2001) by including an explicit EPS loss term and assuming BAP is produced uniquely by EPS hydrolysis. Five new processes have been added into the model with eight new parameters for their description. Simulation results have indicated the effect of SRT on EPS rather than SMP. EPS concentration is sensitive to KEPS, KBAP and SRT, while KUAP sensitivity is low. However, kinetic parameters are not reported, SMP behavior is not described, and the model also lacks an appropriate calibration (Jiang et al., 2008). Recently a new model, CES-ASM3 model, has been proposed by Janus and Ulanicki (2010) in order to predict SMP formation and

EPS production in an activated sludge system based on LuedekingPiret (1959) hypothesis. Janus and Ulanicki (2010) reformulated non-growth associated term an additional reaction for EPS hydrolysis;

rEPS ¼ fEPS lX þ fEPS;d bX  kh;EPS EPS

ð4Þ

where production of UAP is biomass-growth and substrate utilization associated while BAP formation is associated with biomass decay and hydrolysis/dissolution of EPS. 4. Standalone membrane fouling model An empirical hydrodynamic model proposed by Liu et al. (2003) investigated the influence of hydrodynamic conditions on the mixed liquor crossflow velocity and the membrane fouling rate in an MBR. Liu et al. (2003), for example, observed the changing course of TMP over different experiments using a submerged MBR system and they concluded that the cake layer on membrane produced ca. 85% of the total filtration resistance while other factors produced ca. 15% of the filtration resistance. The results

126

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

of different experimental results of Liu et al. (2003) is generally represented by the following empirical expression

5. Integrated and hybrid MBR models

K ¼ ð8:933  107 ÞX 0:532 J 0:376 U 3:047 Lr

Various by products of the metabolic activity of bacterial cultures, specially the SMPs and EPSs, have been found to be correlated with floc strength and resistance to shear and to influence various properties of activated sludge (Janus and Ulanicki, 2010). However, early efforts of integrated MBR model development ignored the incorporation of formation and degradation kinetics of SMPs/EPSs. There also exist differences among different empirical expressions that were independently developed for a particular MBR system having system dependent relations between the hydrodynamic variables and membrane fouling. As such, there is no general consensus about the correlation of hydrodynamic variables with fouling independent of the MBR systems. Different researchers adopted different ASM models for bioprocess modelling and then coupled those with membrane fouling or filtrations. The choices of ASM models for such purposes also varied depending on the type and purposes of different MBR systems. Lee et al. (2002) presented an MBR model based on the model Lu et al., 2001 coupled with a resistance-in-series filtration model for simulating fouling phenomena. The concentration of SMP in the model was assumed negligible compared to the TSS. Wintgens et al. (2003) introduced an integrated model employing ASM3 with resistance-in-series model to describe the filtration performance of submerged capillary hollow fiber module in an MBR. The simulation results for permeability evolution over time matched well the data from the pilot plant except for a major deviation at the end of the considered period. di Bella et al. (2008) also adopted the modified version of ASM1, proposed by Lu et al. (2001), combined with deep-bed-theory of the secondary membrane filtration. Zarragoitia et al. (2008) developed an integrated model that couples biomass transformation processes, membrane fouling and the effects of filtration cycles with intermittent coarse bubble aeration. Mannina et al. (2011) have very recently proposed an integrated model which modifies the model proposed by di Bella et al. (2008) with improved analysis of fouling deposition coupled with the deep-bed filtration. The biological phenomena in the modified model have been evaluated extending the ASM1 model with the BAP and UAP degradation by hydrolysis process (Jiang et al., 2008). This simpler model was able to balance COD balance of SMP more accurately and the conversion between COD and TSS is evaluated considering the following equation

ð5Þ

where K is the increasing rate of filtration resistance (m1 h1), X is the suspended solids concentration in the mixed liquour (gl1), J is the filtration flux (lm2 h1) and ULr is observed cross-flow velocity in the tap water (ms1). Since the cross-flow velocity of the activated sludge along the membrane surface interacts with other hydrodynamic variables such as aeration rate and sludge concentration, Liu et al. (2003) at first correlated those with empirical expressions such as the one given below

U Sr ¼ 1:311U 1:226 e0:0105X Lr

ð6Þ

1

where ULr (ms ) is the cross-flow velocity in the activated sludge in the riser. Although the model is easy to use due to some simplified use of empirical hydrodynamic equation, it cannot capture the actual observed phenomena reversible and irreversible membrane fouling. Latter, a model by Liang et al., 2006, quantified reversible fouling and irreversible fouling while it could not explain the change in fouling properties with changing MBR operating conditions. Meng et al. (2005) developed a fractal permeation model based on Darcy’s law. They hypothesized that the microfiltration of activated sludge formed disordered and complicated sludge cake layer on the membrane which could not be modeled by the application of traditional geometry. In order to evaluate the permeability of such irregular cake layer they applied the fractal geometric theory to determine the pore area fractal dimension, Ds, of a cake layer in terms of its average, self-similar properties. The model, however, does not illustrate how transient operational parameters and conditions affect cake layer formation on the membrane. In a submerged MBR system, coarse bubbles from aeration provide a cleaning mechanism for the immersed membrane modules by the application of shear force and thus scouring the surface of the foulant layer. Li and Wang (2006) proposed their sectional resistance model in order to model such uneven cake formation and degradation due to varying shear distribution over the length of the membrane. They divided the membrane surface into equal fractional areas, De, and calculated separate total resistances, R, for each section;

R ¼ Rm þ Rp þ Rsf þ Rsc

ð7Þ

where Rm, Rp, Rsf, Rsc denote inherent membrane resistance, pour fouling resistance, dynamic sludge film resistance and the resistance due to stable sludge film respectively. The advantages of this transient model are that it accounts for cleaning cycles and characterizes fouling development over time with varying sludge concentrations, filtration fluxes, and aeration intensities. However, the model is only able of capturing general trends of membrane fouling and might not be suitable for applications requiring accurate modelling of membrane fouling phenomena. Using a dynamic model, Giraldo and LeChevallier (2006) successfully simulated some of the complex but commonly observed effects such as: (i) exponential increase in TMP due to high MLSS; (ii) reduced fouling rates at increased aeration intensities; and (iii) subcritical operation fouling, and effect of increased particle size on the filterability. The results from the membrane and cake resistance were combined into Eq. (6) to obtain the change of TMP and cake pressure differential as a function of time:

J ¼ Dp=lðRm þ Rbiofilm þ Rcake þ   Þ

ð8Þ

Their modelling approach, however, does not allow modelling the variation in colloidal particles that might take place during dynamic influent flow conditions to the bioreactor. It also does not incorporate, currently, a change in the backward transport rate of cake-forming particles due to differences in particle size.

MLSS ¼ iSS;XI X I þ iSS;XS X S þ iss;BH X B;H þ iss;BA X B;A

ð9Þ

where iSS,XI, iSS,XS, iSS,BH, iSS,BA are the stoichiometric conversion parameters whose values are chosen from literature. The sectional resistance-in-series approach (Li and Wang, 2006) has been used to determine the total membrane resistance (Rts) taking into account the irregular formation of cake layer due to different effects of aeration.

Rts ¼ Rm þ Rp;i þ Rc;i ¼ Rm þ Rp;i þ ðRdc;i þ Rsc;i Þ

ð10Þ

where Rm is the intrinsic resistance of the membrane; Rp,i is the porefouling resistance caused by solute deposition inside the membrane pores; Rc,i is the resistance of cake layer that is considered as sum of the dynamic sludge film (Rdc,i) and of stable sludge cake resistance attached onto the membrane surface (Rsc,i). Rdc,i and Rsc,i are the product of the specific resistance of the biomass in each specific cake fraction, relating to the mass of the sludge (dynamically or stably) attached in each film (Rdc,i = rsc,i Msc,i; Rdc,i = rdc,i Mdc,i) and having empirical correlation with other parameters;




d  rdc ¼ rsc ¼ TMPp =l2 a þ b 1  exp c SSMP=0:8:MLSS

ð11Þ

where a, b, c and d are constants, SSMP and TMPp are the SMP and TMP coefficients. Realizing the inter-relationship and their impact

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

ASM (Key Processes) -

Aerobic/Anoxic growth of Heterotrophs (XB,H) Aerobic growth of autotrophs (XB,A) Decay of heterotrophs/autotrophs Aerobic/Anoxic growth of heterotrophs (XH) Aerobic/Anoxic storage of organic substrate (XS/SS) Aerobic and Anoxic Hydrolysis Fate of DO (-ve COD) concentration

Physical Sub-model - Cake layer formation - Removal by biological membrane - Removal by physical membrane

Biological Sub-model/s Formation and Degradation Kinetics (SMP/EPS/TEP/BPC) (Switching Function) SMP (UAP and/or BAP) -

127

Membrane resistance

Fate of EPS (DOM-EPS equilibrium)

Engineering Control Biological- HRT/ SRT/ loading rate/ Feed type/ O2 / Substrate Operational/maintenance- Backflushin / relaxation/ aeration rate/ intermittency Fig. 2. Conceptual framework of integrated mathematical model for MBR system.

on membrane fouling, Janus and Ulanicki (2010) implemented the theory of production and degradation of SMP and EPS as proposed by Laspidou and Rittmann (2002a, b) in an ASM3-SMP/EPS hybrid model. The model simulations indicated an increased production of SMP and EPS at higher MLSS, lower temperatures and lower SRT. The model also predicted a slight increase in SMP and EPS with increased DO. The above review has pointed out the fact that a complete integrated model especially including phosphorus removal is yet to be developed. Moreover, most of the ASM-SMP/EPS hybrid integrated models developed so far are generally too complicated and over-parameterized. This associated with the lack of coherence among different methods of influent wastewater characterization makes the task of calibration of ASMs for MBR bioprocess modelling a very complicated one. Some of the process variables introduced into the hybrid models are impossible to be determined experimentally in full-scale MBR systems (e.g., UAP and BAP) and, thus, serious identifiability issues are raised (Patsios and Karabelas, 2010). Most of the integrated models developed have been evaluated for a bio-system with rather simple substrate inputs that were considered merely soluble and readily biodegradable. In contrast, the organic matter in influent wastewater of a real treatment plant is very complex and consists of both soluble and particulate fractions with different biodegradability rates. In this backdrop, it is very difficult to prescribe any comprehensive but easily implementable integrated mathematical model framework for MBR system. Some simplification ideas (Fig. 2) can be integrated into the conventional integrated MBR model framework, i.e., incorporation of empirical expressions in order to derive fractions of SMP/EPS from other measurable parameters, shortcut sub-models with variables determined using black box model. These, along with other simplifications, may lead to reformulation of bioprocess modelling using ASM-SMP/EPS hybrid models.

6. Conclusions The application of ASMs or ASM-SMP/EPS hybrid models is usually too complex to be justified for mathematical modelling of fullscale MBR systems treating complex wastewater. An integrated mathematical MBR model with provisions of switching functions between variables could help skip less influential model parameters under certain operating condition and thus accelerate model simulation and calibration. The DOM–SMP equilibrium criteria can be used to derive empirical expressions to reasonably approximate fractions of SMP/EPS responsible for membrane fouling. Such simplifications suggested in the review could be some steps towards formulation of an integrated model implementable for mathematical modelling of full-scale MBR systems.

Acknowledgements This critical review based research was supported by the Centre for Technology in Water and Wastewater (CTWW) – Theme of Wastewater Treatment and Reuse Technologies, School of Civil and Environmental Engineering, University of Technology, Sydney (UTS).

References Ahn, Y.T., Choi, Y.K., Jeong, H.S., Shin, S.R., 2006. Modeling of extracellular polymeric substances and soluble microbial products production in a submerged MBR at various SRTs. Water Sci. Technol. 53 (7), 209–216. Aquino, S.F., Stuckey, D.C., 2008. Integrated model of the production of soluble microbial products (SMP) and extracellular polymeric substances (EPS) in anaerobic chemostats during transient conditions. Biochem. Eng. J. 38, 138– 146. Arabi, S., Nakhla, G., 2009. Characterization of foulants in conventional and simultaneous nitrification and denitrification membrane bioreactors. Sep. Purif. Technol. 69, 153–160.

128

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129

Artan, N., Orhon, D., Baykal, B.B., 1990. Implications of the task group model I. the effect of the initial substrate concentration. Water Res. 24 (10), 1259–1268. Baek, S.H., Jeon, S.K., Pagilla, K., 2009. Mathematical modelling of aerobic membrane bioreactor (MBR) using activated sludge model no. 1(ASM1). J. Ind. Eng. Chem. 15, 835–840. Barker, D.J., Stuckey, D.C., 1999. A review of soluble microbial products (SMP) in wastewater treatment. Water Res. 33 (14), 3062–3082. Busch, J., Cruse, A., Marquardt, W., 2007. Modeling submerged hollow-fiber membrane filtration for wastewater treatment. J. Membr. Sci. 288, 94–111. Chae, S., Ahn, Y., Kang, S., Shin, H., 2006. Mitigated membrane fouling in a vertical submerged membrane bioreactor (VSMBR). J. Membr. Sci. 280, 572–581. Chaize, S., Huyard, A., 1991. Membrane bioreactors on domestic wastewater treatment sludge production and modelling approach. Water Sci. Technol. 23, 1591–1600. Chang, I.-S., Le-Clech, P., Jefferson, B., Judd, S., 2002. Membrane fouling in membrane bioreactors for wastewater treatment. Environ. Eng. Sci. 128 (11), 1018–1029. Cho, J., Song, K.-G., Ahn, K.-H., 2005. The activated sludge and microbial substances influences on membrane fouling in submerged membrane bioreactor: unstirred batch cell test. Desalination 183, 425–429. De la Torre, Lesjean, B., Drews, A., Kraume, M., 2008. Monitoring of transparent exopolymer particles (TEP) in a membrane bioreactor (MBR) and correlation with other fouling indicators. Water Sci. Technol. 58 (10), 1903– 1909. De la Torre, T., Iversen, V., Meng, F., Stuber, J., Drews, A., Lesjean, B., Kraume, M., 2010. Searching for a universal fouling indicator for membrane bioreators. Desalin. Water Treat. 18, 264–269. Delrue, F., Racault, Y., Choubert, J.M., Sperandio, M., 2010. Modelling a full scale membrane bioreactor using activated sludge model no.1: challenges and solutions. Water Sci. Technol. 62 (10), 2205–2217. Di Bella, G., Mannina, G., Viviani, G., 2008. An integrated model for physicalbiological wastewater organic removal in a submerged membrane bioreactor: model development and parameter estimation. J. Membr. Sci. 322, 1–12. Drews, A., 2010. Membrane fouling in membrane bioreactors - characterization, contradiction, causes and cures. J. Membr. Sci. 363, 1–28. Drews, A., Vocks, M., Bracklow, U., Iversen, V., Kraume, M., 2008. Does fouling in MBRs depend on SMP? Desalination 231 (1–3), 141–149. Drews, A., Vocks, M., Iversen, V., Lesjean, B., Kraume, M., 2006. Influence of unsteady membrane bioreactor operation on EPS formation and filtration resistance. Desalination 192, 1–9. Dvorˇák, Lukas., Gomez, M., Dvorakova, M., Ruzickova, I., Wanner, J., 2011. The impact of different operating conditions on membrane fouling and EPS production. Bioresour. Technol. 101, 3876–3883. Fan, F., Zhou, H., Husain, H., 2006. Identification of wastewater sludge characteristics to predict critical flux for membrane bioreactor processes. Water Res. 40, 205–212. Fenu, A., Guglielmi, G., Jimenez, J., Sperandio, M., Saroj, D., Lesjean, B., Brepols, C., Thoeye, C., Nopens, I., 2010. Activated sludge model (ASM) based modelling of membrane bioreactor (MBR) processes: a critical review with special regard to MBR specificities. Water Res. 44, 4272–4294. Furumai, H., Rittmann, B.E., 1992. Advanced modelling of mixed population of heterotrophs and nitrifiers. Water Sci. Technol. 26 (3–4), 493–502. Giraldo, E., LeChevallier, M., 2006. Dynamic mathematical modeling of membrane fouling in submerged membrane bioreactors. Proceedings of Water Environment Foundation, WEFTEC, pp. 4895–4913. Guglielmi, G., Saroj, D.P., Chiarani, D., Andreottola, G., 2007. Sub-critical fouling in a membrane bioreactor for municipal wastewater treatment: experimental investigation and mathematical modelling. Water Res. 41, 3903–3914. Gujer, W., Henze, M., Mino, T., van Loosdrecht, M., 1999. Activated sludge model No. 3. Water Sci. Technol. 39 (1), 183–193. Henze, M., Grady, C.P.L.J., Gujer, W., Marais, G.v.R., Matsuo, T., 1987. Activated Sludge Model No. 1. IAWPRC Scientific and Technical Report No. 1, IAWPRC, London. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C.M., Marais G.V.R., 1995. Activated sludge model No. 2. IWA Scientific and Technical Report No. 3, London, England. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C., Marais, G.V.R., Van Loosdrecht, M.C.M., 1999. Activated sludge model No. 2d, ASM2d. Water Sci. Technol. 39 (1), 165–182. Hocaoglu, S.M., Insel, G., Cokgor, E.U., Orhon, D., 2011. Effect of sludge age on simultaneous nitrification and denitrification in membrane bioreactor. Bioresour. Technol. 102, 6665–6672. Hsieh, K.M., Murgel, G.A., Lion, L.W., Shuler, M.L., 1994. Interactions of microbial biofilms with toxic trace metals: 1. Observation and modelling of cell growth, attachment, and production of extracellular polymer. Biotechnol. Bioeng. 44, 219–231. Ishiguro, K., Imai, K., Sawada, S., 1994. Effects of biological treatment conditions on permeate flux of UF membrane in a membrane/activated-sludge wastewater treatment system. Desalination 98, 119–126. Itokawa, H., Thiemig, C., Pinnekamp, J., 2008. Design and operating experiences of municipal MBRs in Europe. Water Sci. Technol. 58 (12), 2319–2327. Jang, N., Ren, X., Cho, J., Kim, I.S., 2006. Steady-state modeling of bio-fouling potentials with respect to the biological kinetics in the sMBR. J. Membr. Sci. 284, 352–360. Janus, T., Ulanicki, B., 2010. Modelling SMP and EPS formation and degradation kinetics with an extended ASM3 model. Desalination 261 (1–2), 117–125.

Jiang, T., Liu, X., Kennedy, M.D., Schippers, J.C., Vanrolleghem, P.A., 2005. Calibrating a side-stream membrane bioreactor using activated sludge model no. 1. Water Sci. Technol. 52 (10), 359–367. Jiang, T., Myngheer, S., De Pauw, D.J.W., Spanjers, H., Nopens, I., Kennedy, M.D., Amy, G., Vanrolleghem, P.A., 2008. Modeling the production and degradation of soluble microbial products (SMP) in membrane bioreactors (MBR). Water Res. 42 (20), 4955–4964. Judd, S., 2006. The MBR Book: Principles and Applications of Membrane Bioreactors. Elsevier, Amsterdam. Kornboonraksa, Thipsuree., Lee, S.H., 2009. Factors affecting the performance of membrane bioreactor for piggery wastewater treatment. Bioresour. Technol. 100, 2926–2932. Laspidou, C.S., Rittmann, B.E., 2002a. Non-steady state modelling of EPS, SMPs, and active and inert biomass. Water Res. 36 (8), 1983–1992. Laspidou, C.S., Rittmann, B.E., 2002b. A unified theory for EPS, SMPs, and active and inert biomass. Water Res. 36 (11), 2711–2720. Le-Clech, P., Chen, V., Fane, A.G., 2006. Fouling in membrane bioreactors used in wastewater treatment. J. Membr. Sci. 284, 17–53. Lee, Y., Cho, J., Seo, Y., Lee, J.W., Ahn, K.-H., 2002. Modeling of submerged membrane bioreactor process for wastewater treatment. Desalination 146, 451–457. Li, X.Y., Wang, X.M., 2006. Modelling of membrane fouling in a submerged membrane bioreactor. J. Membr. Sci. 278, 151–161. Liang, S., Liu, C., Song, L., 2007. Soluble microbial products in membrane bioreactor operation: behaviors, characteristics, and fouling potential. Water Res. 41 (1), 95–101. Liang, S., Song, L., Tao, G., Kekre, K.A., Seah, H., 2006. A modeling study of fouling development in membrane bioreactors for wastewater treatment. Water Environ. Res. 78 (8), 857–863. Lin, H.J., Xie, K., Mahendran, B., Bagley, D.M., Leung, K.T., Liss, S.N., Liao, B.Q., 2009. Sludge properties and their effects on membrane fouling in submerged anaerobic membrane bioreactors (SAnMBRs). Water Res. 43, 3827–3837. Liu, R., Huang, X., Sun, Y.F., Qian, Y., 2003. Hydrodynamic effect on sludge accumulation over membrane surfaces in a submerged membrane bioreactor. Process Biochem. 39, 157–163. Lu, S.G., Imai, T., Ukita, M., Sekine, M., Higuchi, T., 2002. Modeling prediction of membrane bioreactor process with the concept of soluble microbial product. Water Sci. Technol. 46 (11–12), 63–70. Lu, S.G., Imai, T., Ukita, M., Sekine, M., Higuchi, T., Fukagawa, M., 2001. A model for membrane bioreactor process based on the concept of formation and degradation of soluble microbial products. Water Res. 35 (8), 2038–2048. Luedeking, R., Piret, E.C., 1959. A kinetic study of lactic acid fermentation batch process at controlled pH. J. Biochem. Microbiol. Technol. Eng. 1 (4), 393–412. Mannina, G., di Bella, G., Viviani, G., 2011. An integrated model for biological and physical process simulation in membrane bioreactors. J. Membr. Sci. 376, 56– 69. Meng, F., Chae, S., Drews, A., Kraume, M., Shin, H., Yang, F., 2009. Recent advances in membrane bioreactors (MBRs): membrane fouling and membrane material. Water Res. 43, 1489–1512. Meng, F., Zhang, H., Li, Y., Zhang, X., Yang, F., 2005. Application of fractal permeation model to investigate membrane fouling in membrane bioreactor. J. Membr. Sci. 262, 107–116. Namkung, E., Rittmann, B.E., 1986. Soluble microbial products (SMP) formation kinetics by biofilms. Water Res. 20 (6), 795–806. Ng, A.N.L., Kim, A.S., 2007. A mini-review of modeling studies on membrane bioreactor (MBR) treatment for municipal wastewaters. Desalination 212, 261– 281. Ngo, H.H., Guo, W., Xing, W., 2008. Evaluation of a novel sponge-submerged membrane bioreactor (SSMBR) for sustainable water reclamation. Bioresour. Technol. 99, 2729–2735. Nielsen, P.H., Jahn, A., Palmgren, R., 1997. Conceptual model for production and composition of exopolymers in biofilms. Water Sci. Technol. 36, 11–19. Oliveira-Esquerre, K.P., Narita, H., Yamato, N., Funamizu, N., Watanabe, Y., 2006. Incorporation of the concept of microbial product formation into ASM3 and the modeling of a membrane bioreactor for wastewater treatment. Braz. J. Chem. Eng. 23, 461–471. Orhon, D., Artan, N., Cimcit, Y., 1989. The concept of soluble residual product formation in the modelling of activated sludge. Water Sci. Technol. 21, 339–350. Pan, J.R., Su, Y., Huang, C., 2010. Characteristics of soluble microbial products in membrane bioreactor and its effect on membrane fouling. Desalination 250, 778–780. Parco, V., du Toit, G., Wentzel, M., Ekama, G., 2007. Biological nutrient removal in membrane bioreactors: denitrification and phosphorus removal kinetics. Water Sci. Technol. 56 (6), 125–134. Patsios, S.I., Karabelas, A.J., 2010. A review of modeling bioprocesses in membrane bioreactors (MBR) with emphasis on membrane fouling predictions. Desalin. Water Treat. 21, 189–201. Pendashteh, A.R., Fakhru’l, A.–R., Madaeni, S.S., Abdullah, L.C., Abidin, Z.Z., Biak, D., R.A., 2011. Membrane foulants characterization in a membrane bioreactor (MBR) treating hypersaline oily wastewater. Chem. Eng. J. 168, 140–150. Rosenberger, S., Laabs, C., Lesjean, B., Gnirss, R., Amy, G., Jekel, M., Schrotter, J.C., 2006. Impact of colloidal and soluble organic material on membrane performance in membrane bioreactors for municipal wastewater treatment. Water Res. 40 (4), 710–720. Spèrandio, M., Espinosa, M.C., 2008. Modelling an aerobic submerged membrane bioreactor with ASM models on a large range of sludge retention time. Desalination 231 (1–3), 82–90.

M.F.R. Zuthi et al. / Bioresource Technology 122 (2012) 119–129 Sun, F., Wang, X., Li, X., 2011. Effect of biopolymer clusters on the fouling property of sludge from a membrane bioreactor (MBR) and its control by ozonation. Process Biochem. 46, 162–167. Tchobanoglous, G., Burton, F.L., Stensel, H.D., 2003. Wastewater Engineering: Treatment and Reuse, fourth ed. McGraw-Hill, Boston. Trussell, R.S., Merlo, R.P., Hermanowicz, S.W., Jenkins, D., 2007. Influence of mixed liquor properties and aeration intensity on membrane fouling in a submerged membrane bioreactor at high mixed liquor suspended solids concentrations. Water Res. 41 (5), 947–958. Wang, X., Li, X., Huang, X., 2007. Membrane fouling in a submerged membrane bioreactor (SMBR): characterisation of the sludge cake and its high filtration resistance. Sep. Purif. Technol. 52, 439–445.

129

Wintgens, T., Rosen, J., Melin, T., Brepols, C., Drensla, K., Engelhardt, N., 2003. Modelling of a membrane bioreactor system for municipal wastewater treatment. J. Membr. Sci. 216 (1–2), 55–65. Xuan, W., Bin, Z., Zhiqiang, S., Zhigang, Q., Zhaoli, C., Min, J., 2010. The EPS characteristics of sludge in an aerobic granule membrane bioreactor. Bioresour. Technol. 101, 8046–8050. Zarragoitia, A.-G., Schetrite, S., Alliet, M., Jáuregui, U.-H., Albasi, C., 2008. Modelling of submerged membrane bioreactor: conceptual study about link between activated sludge biokinetics, aeration and fouling process. J. Membr. Sci. 325, 612–624. Zhang, J., Chua, H.C., Zhou, J., Fane, A.G., 2006. Factors affecting the membrane performance in submerged membrane bioreactors. J. Membr. Sci. 284 (1–2), 54–66.