This study describes the development of a modified activated sludge model No.1 framework to describe the organic substrate transformation in the high-rate activated sludge (HRAS) process. New process mechanisms for dual soluble substrate utilization, production of extracellular polymeric substances (EPS), absorption of soluble substrate (storage), and adsorption of colloidal substrate were included in the modified model. Data from two HRAS pilot plants were investigated to calibrate and to validate the proposed model for HRAS systems. A subdivision of readily biodegradable soluble substrate into a slow and fast fraction were included to allow accurate description of effluent soluble chemical oxygen demand (COD) in HRAS versus longer solids retention time (SRT) systems. The modified model incorporates production of EPS and storage polymers as part of the aerobic growth transformation process on the soluble substrate and transformation processes for flocculation of colloidal COD to particulate COD. The adsorbed organics are then converted through hydrolysis to the slowly biodegradable soluble fraction. Two soluble substrate models were evaluated during this study, i.e., the dual substrate and the diauxic models. Both models used two state variables for biodegradable soluble substrate (SBf and SBs) and a single biomass population. The A-stage pilot typically removed 63% of the soluble substrate (SB) at an SRT <0.13 d and 79% at SRT of 0.23 d. In comparison, the dual substrate model predicted 58% removal at the lower SRT and 78% at the higher SRT, with the diauxic model predicting 32% and 70% removals, respectively. Overall, the dual substrate model provided better results than the diauxic model and therefore it was adopted during this study. The dual substrate model successfully described the higher effluent soluble COD observed in the HRAS systems due to the partial removal of SBs, which is almost completely removed in higher SRT systems.

INTRODUCTION

The high-rate activated sludge (HRAS) process for carbon removal uses high food-to-microorganism ratios and low solids and hydraulic retention times (SRTs and HRTs) for the biological transformation and removal of wastewater organics (chemical oxygen demand (COD)). When an HRAS system is the first step in the adsorption/bio-oxidation (A/B) process (Böhnke & Diering 1986), the general objectives are to maximize the removal of organics through adsorption/absorption while minimizing the energy input required for treatment, and to produce large amounts of waste sludge that can be converted to biogas by anaerobic digestion (Schulze-Rettmer & Zuckut 1998). A key mechanism in the adsorption of colloidal and particulate COD (cCOD and pCOD) is the production of extracellular polymeric substances (EPS) produced as part of the aerobic growth transformation process. It is argued that the production of EPS is essential to sludge floc formation (Li & Yang 2007). A key mechanism in the absorption of soluble substrate is the production of cellular storage products. The literature (Beun et al. 2000; Third et al. 2003) suggests that the operating environment of HRAS systems, high slowly biodegradable soluble organics ratios and low dissolved oxygen (DO), is conducive to the production of storage products, which allows for the redirection of COD to downstream processes, e.g., anaerobic digestion for energy recovery. Hence, accurate modeling of this system is of importance to design, control, and optimize the performance of not only HRAS systems, but of the A/B process as a whole.

The modeling of the activated sludge process, particularly COD transformations, has significantly evolved toward fundamental principles in recent decades from simple single-substrate models to more complex multiple-substrate models involving the processes of oxidation, hydrolysis and storage (Dold et al. 1980; Sin et al. 2005). However, these models have evolved to describe COD removal in systems operating at long SRT (i.e. >3 days) where the biodegradable soluble organic substrate (SB) can be modeled as a single substrate with a single kinetic expression. However, full-scale and pilot-scale results from HRAS (Jimenez 2002; Haider et al. 2003; Miller et al. 2012) show that very low SRT (i.e. <1 day) may result in a selection of fast-growing bacteria, which are only able to biodegrade the most readily degradable organics in the process conditions of the A-stage (i.e., low SRT and short contact time). Haider et al. (2003) showed that the inert soluble COD (sCOD) fraction (SU) in the effluent from an HRAS with an SRT of 0.5 days was always higher than the same COD fraction from a system with an SRT of 20 days. Hence, they recommended that for modeling, the SB fraction of the wastewater should be split into two distinct biodegradable fractions. In addition, current models assume that flocculation and adsorption of colloidal and particulate substrate (CB and XB) is complete and instantaneous; hence, flocculation can be ignored in these models (Jimenez 2002; Haider et al. 2003). However, in HRAS systems such as those employed in the A/B process, these assumptions with respect to organic substrate flocculation are no longer applicable (Jimenez et al. 2005) and removal of XB and CB is only partial.

ASM1 model

The activated sludge model No.1 (ASM1) model framework as presented by Henze et al. (2000) was used as a first step to simulate the performance of the HRAS system, and the results are depicted in Figure 1. In summary, the ASM1 model does not properly predict the removal of organic substrate at low SRTs. The ASM1 model under-predicts the performance of the HRAS system with respect to effluent sCOD and does not address the higher effluent cCODs observed for the HRAS.

Figure 1

ASM1 prediction for the HRAS system (aerobic SRT = 0.25 d, DO = 0.2 mg/L, HRT = 30 minutes, μmax = 6.0 d−1; Ks = 20 mg/L; KO,H = 0.2 mg/L; bH = 0.62 d−1; YH = 0.67).

Figure 1

ASM1 prediction for the HRAS system (aerobic SRT = 0.25 d, DO = 0.2 mg/L, HRT = 30 minutes, μmax = 6.0 d−1; Ks = 20 mg/L; KO,H = 0.2 mg/L; bH = 0.62 d−1; YH = 0.67).

This study discusses a modeling approach that evaluates the organic substrate transformation as it pertains to HRAS systems. This approach uses the ASM1 (Henze et al. 2000) as the initial framework. The original framework was modified to describe the proper mechanisms required to accurately describe the performance of the HRAS system.

MATERIALS AND METHODS

Historical operating data from two pilot systems were evaluated to understand the organic substrate transformation mechanisms in HRAS and used to calibrate and validate the proposed process model. The data used during this study include operating data from an A-stage pilot plant owned and operated by the Hampton Roads Sanitation District (HRSD; Miller et al. 2013) and from an HRAS pilot plant operated at the University of New Orleans (Jimenez et al. unpublished data).

HRSD owns and operates an A/B pilot plant located at the Chesapeake–Elizabeth treatment plant in Virginia Beach, Virginia, USA. The pilot plant consists of an HRAS reactor for carbon removal followed by a B-stage for nitrogen removal (Figure 2). Currently, the HRAS A-stage includes three reactors (45 gal per reactor), operated at a 0.2 day SRT and 0.5 hour HRT, and is fed screened and degritted raw municipal wastewater at 4.5 gpm (24.53 m3/d).

Figure 2

HRSD A-stage pilot plant configuration.

Figure 2

HRSD A-stage pilot plant configuration.

The data set used to calibrate the modified models was collected by Jimenez (unpublished data) and is referred to as the New Orleans (NO) data set. The SRT and DO concentration were varied by Jimenez in order to evaluate the effect of these operating parameters on the production of EPS and the removal of organic substrate, and the same parameters were varied in the models presented in this paper during calibration using the NO data set.

A partial list of the kinetic parameter values established through model calibration is summarized in Table A3 in the Appendix (available online at http://www.iwaponline.com/wst/071/051.pdf). The kinetic parameters added represent the pathways incorporated in the modified models for soluble substrate, EPS production, adsorption/flocculation, and creation of storage polymers.

The NO activated sludge pilot plant comprised the following components: a rotating screen, an inlet mechanism (30 gal mixing tank), an aeration tank (40 gal), a mechanical flocculator, and a secondary clarifier (70 gal) (Figure 3). The unit was designed for a flow rate of 7.5 m3/d (2,000 gal/d) and an HRT in the aeration tank of 30 minutes.

Figure 3

University of New Orleans pilot plant configuration.

Figure 3

University of New Orleans pilot plant configuration.

The sampling plan for the NO pilot plant involved collecting grab samples several times per week. The sampling points include the effluent from the rotary screen (plant influent), the supernatant and mixed liquor suspended solids (MLSS) from the solids contact tank, the return activated sludge and secondary effluent. Samples were analysed for total COD, filtered COD using a 0.45 micron Hach No. 300 (Loveland, CO, USA) glass qualitative filter paper, dissolved COD using flocculated samples filtered using a 0.45 micron Hach No. 300 (Loveland, CO, USA) glass qualitative filter paper, and total and volatile suspended solids (TSS and VSS). EPS was extracted using the extraction method developed by Frolund et al. (1995), which is summarized as follows:

  • 300 mL of sludge were transferred to an extraction beaker with baffles and the cation exchange resin (CER) was added (70 g CER/g VS).

  • The suspension was stirred for 3 hours at 1,000 rpm.

  • The extracted EPS was harvested by centrifugation of a sample of the CER-sludge suspension for 1 minute at 12,200 rpm to remove the CER.

  • The supernatant was centrifuged twice for 15 minutes at 12,200 rpm in order to remove remaining floc components.

  • EPS was quantified by measuring the total organic carbon content of the sample using an Apollo 9000HS-TOC analyzer fabricated by Tekmar–Dohrmann (Teledyne Tekmar, Mason, OH, USA).

EPS was extracted at least three times per SRT. Triplicate values were averaged.

Dynamic simulations were run to calibrate and validate the HRAS model framework. The model was calibrated against the NO data set based on the removal efficiency of the soluble (SB) and cCOD considerations and the EPS production rate. These parameters were evaluated under variable SRT and variable DO (constant SRT) conditions. Model validation was based on the removal efficiency of the soluble substrate (SB) using a subset of the HRSD data set spanning a 4-week period. Weekly averages were used as input parameters generated from daily composite samples.

MODIFIED MODEL DESCRIPTION

To describe the behavior of the HRAS pilot plants, the ASM1 model was modified to incorporate non-steady-state material balance equations for dual soluble substrate (SBf, SBs) utilization, production of EPS (XEPS), production of storage products (XSTO), and adsorption of inert and biodegradable cCOD (CU and CB). The modifications were also influenced by the literature (Jimenez 2002; Laspidou & Rittmann 2002b; Miller et al. 2013).

Fate of soluble substrate

Conventionally, the method to quantify the non-biodegradable sCOD from an activated sludge plant is to operate a laboratory- or pilot-scale system at an SRT longer than 3 days (Ekama et al. 1986) and use the effluent sCOD as the non-biodegradable soluble fraction. However, at the low SRT (and low HRT) of the HRAS system, this method is no longer valid since there is a fraction of the effluent COD that is biodegradable in the higher SRT B-stage, but not biodegradable in the A-Stage. This led to the establishment of two state variables for SB designated as SBf (SB fast) and SBs (SB slow); SBf corresponds to the high affinity sCOD that is biodegradable in the HRAS system at low SRT and HRT (Haider et al. 2003; Pala-Ozkok et al. 2013). The high affinity substrate is defined as the raw influent volatile fatty acids (VFA) concentration. To further support this fractionation, in our model SBf is removed in the A-stage, which is validated by the experimental results from the HRSD A-stage pilot that shows, on average, 95% removal of influent SCVFAs. The SBs is the lower affinity fraction of the sCOD that is biodegradable at a slower rate in the HRAS system and thus only partially biodegraded in the A-stage. Two modified models were developed. In one model, SBf is biodegraded first, and it is only when SBf is fully utilized that biodegradation of SBs becomes significant. This is analogous to diauxic growth in which one substrate is biodegraded immediately by constitutive enzymes, and only when the first substrate runs out are enzymes induced for metabolism of the second substrate. The data do not show, or disprove, this mechanism, but this model is at least plausible mechanistically. This is referred to in this study as the diauxic model. The second model is a dual substrate model, where SBf and SBs are utilized simultaneously, with the growth on SBf occurring at a higher maximum specific substrate utilization rate than growth on SBs. This model is referred to as the dual substrate model in this study. The difference between the two model frameworks is the addition of an inhibition function in the diauxic model (KBf/(KBf + SBf)), which applies to heterotrophic growth on SBs. There are significant differences in the kinetics between each model, with the diauxic model including maximum growth rates and half-saturation coefficients for each soluble substrate fraction. The dual substrate model uses single maximum growth rates with specific half-saturation coefficients for each fraction. Using ASM1 as the template for the development of both models, several new state variables were added to the matrix; a list of the state variables added to both models is shown in Table 1. The stoichiometric and kinetic matrices were also modified with the changes included in the Appendix (available online at http://www.iwaponline.com/wst/071/051.pdf). Further details are discussed in the following sections.

Table 1

Partial list of state variables (g COD·m−3)

Symbol Name 
SU Soluble non-biodegradable organics 
SBf Rapidly biodegradable soluble organics 
SBs Slowly biodegradable soluble organics 
CU Colloidal non-biodegradable organics 
CB Colloidal biodegradable organics 
XU Particulate non-biodegradable organics 
XB Particulate biodegradable organics 
XOHO Active ordinary heterotrophic organisms 
XE Particulate non-biodegradable endogenous products 
XEPS Extracellular polymeric substances 
XSTO Intracellular storage polymeric substances 
Symbol Name 
SU Soluble non-biodegradable organics 
SBf Rapidly biodegradable soluble organics 
SBs Slowly biodegradable soluble organics 
CU Colloidal non-biodegradable organics 
CB Colloidal biodegradable organics 
XU Particulate non-biodegradable organics 
XB Particulate biodegradable organics 
XOHO Active ordinary heterotrophic organisms 
XE Particulate non-biodegradable endogenous products 
XEPS Extracellular polymeric substances 
XSTO Intracellular storage polymeric substances 

Adsorption of cCOD

For the purpose of this investigation, the pCOD consists of organic suspended solids (ssCOD) and organic cCOD in the wastewater (pCOD = ssCOD + cCOD). The sCOD excluding colloids is the truly soluble organic material in the wastewater and this was quantified by coagulation/flocculation followed by filtration (i.e., ffCOD; Mamais et al. 1993). The sCOD is the sum of the state variables SBf, SBs, and SU.

In the diauxic model, it must be noted that the growth rates from SBf and SBs will never both be significant at the same time. This is due to the model kinetic equations being such that SBs transformations will not be significant until SBf runs out (i.e., when SBf < KBf).

The colloidal substrate is distinguished from suspended material in the HRAS model since flocculation and enmeshment of colloidal solids may not be complete in low SRT and HRT HRAS systems. The colloidal fraction is separated into its biodegradable fraction (CB) and its non-biodegradable fraction (CU) and added as new state variables. The CB are enmeshed into XB via bio-flocculation (see process r5 in Table A1 of the Appendix, online at http://www.iwaponline.com/wst/071/051.pdf) and when this occurs in the model they become part of the XB, which can subsequently be hydrolyzed (process r4) to form SBs. The kinetic rate expression for flocculation is a first-order rate expression with respect to the colloidal concentration (see Table A2 in the Appendix, online at http://www.iwaponline.com/wst/071/051.pdf). The CB is flocculated onto the XB, becoming part of that category of organics. The adsorbed organics are then converted through hydrolysis to SBs, which can then be oxidized or converted to EPS or biomass by the microorganisms. The CU is flocculated onto the XU and removed from the system through wasting. The A-stage influent CU concentration is defined as the B-stage effluent CU concentration. Owing to flocculation and adsorption in the B-stage, this definition could result in the model under-estimating the influent CU concentration, just as current methods for estimating SU for ASM1 and ASM2 type-models may over-estimate SU since soluble microbial products are not explicitly accounted for because it is not practical to do so. This is the same reason we do not explicitly account for the bio-flocculative/adsorptive removal of influent CU. However, better methods for determining these COD fractions may be developed in future investigations, with the major difficulty being developing something still suitable for practical use (which the current method we propose is), as opposed to methods only possible in a handful of academic laboratories.

EPS production

Extracellular polymeric substances production impacts the bio-flocculation removal efficiency for particulate and colloidal substrate (Jimenez 2002). The EPS data produced by Jimenez et al. (unpublished data) were used as calibration data for the models. This data set shows a linear correlation between the substrate utilization rate and EPS production and an increase in EPS production with SRT over a range of 0.3–2.0 days. In addition, EPS increased with the DO concentration over the same range of SRT values. Laspidou & Rittmann (2002a) indicated that the net EPS concentration is a function of the portion of influent soluble substrate (substrate electron pool) shunted to EPS formation versus the EPS hydrolysis rate. Hence, the modified models incorporate EPS production as part of the aerobic growth process on SBf and SBs. The proportionality coefficient kEPS,PC (Equation (2)) quantifies the portion of influent electrons shunted to EPS formation. The portion of substrate electrons that are shunted to EPS formations (kEPS,PC) are then subtracted from the biomass yield coefficient YOHO,AER, i.e., YOHO,AER*(1-kEPS,PC)), reducing the electrons available for biomass synthesis. In the diauxic model, EPS formation is first driven by SBf during aerobic growth. EPS formation on SBs does not occur until SBf starts to run out. In contrast, the EPS formation in the dual substrate model occurs simultaneously on both soluble substrate fractions. Additional SBs becomes available through hydrolysis of XB (Carucci et al. 2001). The values for kEPS,MAX (maximum EPS production) and KO,EPS were estimated using a non-linear regression analysis of the EPS production data versus DO concentration data provided by Jimenez 2002. This analysis resulted in an estimated kEPS,MAX value of 0.25 (g CODEPS/g VSS) and KO,EPS value of 1.5 (). The proportionality coefficient kEPS,PC is calculated as shown in Equations (1) and (2). 
formula
1
The value kEPS,SC is the stoichiometric coefficient normalized to the biomass concentration (g CODEPS/g CODXOHO). The term iCB (1.48 g CODXOHO/gVSS) is a stoichiometric conversion factor that converts kEPS,MAX from units of g CODEPS/gVSS to g CODEPS/g CODXOHO. Since EPS is produced as a function of growth rate, and EPS hydrolysis is first order with respect to the amount of EPS available, EPS increases at very low SRTs; but as SRT increases, a point is reached where the EPS hydrolysis rate exceeds EPS production, and beyond that point EPS decreases with increasing SRT. EPS decreases with increasing SRT, which occurs at an SRT greater than 2.0 days for current models. 
formula
2

Production of storage products

The review of the literature suggests that in systems operated at low DO concentrations (<0.9 mg/L; according to Third et al. 2003) typical of an HRAS system, the microbial uptake of rapidly biodegradable soluble COD (SBf) could result in the formation of storage polymers. Third et al. (2003) found that by using acetate as the substrate for COD, the microbial uptake of acetate and its conversion to storage polymers was strictly oxygen dependent. At low DO, the flow of electrons was used for acetate uptake and production of storage polymers. Higher DO supply rates resulted in higher growth rates with the flow of electrons mainly going to biomass production with approximately 20% of the substrate being oxidized, independent of the DO concentration. The following expression was added to both the diauxic and dual substrate models to simulate the flow of electrons to storage as a function of DO concentration. 
formula
3
where fSTO represents the fraction of storage products in the active biomass, fShunt,max represents the maximum flow of electrons as a function of dissolved oxygen concentration, and KO,STO is the half-saturation coefficient for .

The diversion of substrate electrons to storage in the modified models is represented by the proportionality constant kSTO,PC. The portion of electrons that are shunted to kSTO,PC are also subtracted from the biomass yield coefficient YOHO,AER, i.e., (YOHO,AER*(1-kEPS,PCkSTO,PC)), for aerobic growth using SBf and SBs, further reducing the electrons available for biomass synthesis. Storage products are biodegraded in the model when both SBf and SBs are depleted; XSTO is hydrolyzed directly to SBs and used for aerobic growth in such cases.

RESULTS AND DISCUSSION

The diauxic and dual substrate models were analyzed using the process simulator software SUMO version 0.9.15.0 developed by Dynamita (Nyons, France). The experimental data sets from NO and HRSD were used to calibrate and validate the modified models.

Model calibration

The soluble substrate (SB) removal efficiency for the NO data set of 70% compared to 69% for the dual substrate model and 64% for the diauxic model indicates that, although both models were effective in predicting the removal efficiency, the dual substrate model results were slightly closer to the NO data. Figure 4 presents calibration curves comparing the dual substrate model results, as a function of SRT and DO, with the NO experimental data. The model fit the NO data reasonably well.

Figure 4

Dual substrate model calibration results.

Figure 4

Dual substrate model calibration results.

Colloidal substrate removal was calibrated by adjusting the adsorption rate parameter qADS (0.07 d−1) and the surface limitation parameter KSL (0.002) until the dual substrate model results trended well with the experimental data. Ongoing research includes analysis of independent data sets to validate these parameter values.

Modified model validation

The HRSD data set used for validation spanned a 4-week period where the pilot plant had reached steady-state operating conditions. Weekly averages, based on daily composite samples, were calculated for that period and the data reduced to a format compatible with the models. Influent values for select state variables are shown in Table 2; SU, SBs, CB, and CU were obtained by comparing the A-stage (low SRT) and B-stage (high SRT) effluent COD fractions, including 1.5 micron filtration, ffCOD (0.45 micron membrane filters plus coagulant) and VFAs. Dynamic input data for model validation also included weekly DO concentrations, and return and waste activated sludge (RAS and WAS) flow.

Table 2

Influent state variables for model validation

Time (week) Q (m3/d) SU (g/m3SBf (g/m3SBs (g/m3CU (g/m3CB (g/m3XU (g/m3XB (g/m3XOHO (g/m3
24.53 28 ± 1.2 53 ± 2.7 76 ± 3.0 4 ± 0.5 44 ± 4.5 31 ± 5.5 312 ± 34 20 ± 1.0 
24.53 19 ± 1.03 50 ± 3.12 89 ± 5.4 11 ± 1.5 40 ± 3.5 38 ± 6.5 278 ± 58.7 20 ± 1.34 
24.53 19 ± 0.22 51 ± 0.6 89 ± 1.01 10 ± 1.75 32 ± 4.0 53 ± 3 355 ± 2.89 40 ± 0.25 
24.53 27 ± 0.31 48 ± 0.72 83 ± 0.94 2 ± 0.2 53 ± 7.0 20 ± 1.5 362 ± 13.2 40 ± 0.25 
Time (week) Q (m3/d) SU (g/m3SBf (g/m3SBs (g/m3CU (g/m3CB (g/m3XU (g/m3XB (g/m3XOHO (g/m3
24.53 28 ± 1.2 53 ± 2.7 76 ± 3.0 4 ± 0.5 44 ± 4.5 31 ± 5.5 312 ± 34 20 ± 1.0 
24.53 19 ± 1.03 50 ± 3.12 89 ± 5.4 11 ± 1.5 40 ± 3.5 38 ± 6.5 278 ± 58.7 20 ± 1.34 
24.53 19 ± 0.22 51 ± 0.6 89 ± 1.01 10 ± 1.75 32 ± 4.0 53 ± 3 355 ± 2.89 40 ± 0.25 
24.53 27 ± 0.31 48 ± 0.72 83 ± 0.94 2 ± 0.2 53 ± 7.0 20 ± 1.5 362 ± 13.2 40 ± 0.25 

Ranges shown are ± one standard deviation.

Dynamic simulations were performed in order to compare the predicted effluent values with those of the HRSD pilot plant. Figure 5(a) presents a comparison of the effluent soluble biodegradable fraction (SBf + SBs) for both the diauxic and dual substrate models. Based on these results, the dual substrate model better predicts the performance of the HRSD pilot plant; however, this may be due to the specific process configuration (especially HRT) in combination with the specific value of the switching function parameter in the diauxic model. This will be further evaluated in the future. Figure 5(b) shows the dual substrate model predicted values for SBf and SBs. Based on the definition of the SBf and SBs fractions discussed previously, the model predicts almost full removal of SBf, whereas a significant part of the SBs fraction passes through the biological reactor operated at an SRT of approximately 0.2 days.

Figure 5

Validation results for the diauxic and dual substrate models. (a) Comparison of the effluent SB concentration from diauxic and dual substrate model, (b) dual substrate model validation for SBF and SBS concentrations.

Figure 5

Validation results for the diauxic and dual substrate models. (a) Comparison of the effluent SB concentration from diauxic and dual substrate model, (b) dual substrate model validation for SBF and SBS concentrations.

The A-stage typically removed 68–77% of the influent SBs, with the remainder being removed in the B-stage. The lower removal efficiencies/higher effluent SB in the A-stage during weeks 1 and 2 can be attributed to lower SRTs (0.086 days and 0.126 days, respectively) when compared to weeks 3 and 4 (0.236 days and 0.23 days, respectively).

Effect of influent biomass

The HRSD pilot operates at an SRT approaching the washout SRT condition based on the maximum growth rate for the heterotrophic biomass population. This results in a biomass population that varies significantly depending on the influent biomass concentration. Figure 6 shows the effect of influent biomass concentration at various SRTs on the reactor biomass concentration. The results show that as the SRT increases above 0.3 d, the biomass concentration in the reactor approaches a condition where the influent biomass concentration no longer determines the reactor biomass concentration. However, at the lower SRT (typical of the HRSD pilot), the biomass concentration varies significantly depending on the influent biomass concentration. Therefore, the influent biomass concentration is essential to the model's ability to predict the MLVSS and the removal of soluble COD. This is consistent with the results in Figure 5 where the influent biomass concentration increased for weeks 3 and 4 (Table 2).

Figure 6

Effect of influent biomass concentration.

Figure 6

Effect of influent biomass concentration.

CONCLUSIONS

A modified ASM1 model was developed to describe the organic substrate transformation in the high-HRAS process. Data from two HRAS pilot plants were used to calibrate and to validate the proposed model for HRAS systems. Two soluble substrate models were evaluated during this study, i.e., the dual substrate and the diauxic models. Both models used two state variables for biodegradable soluble substrate (SBf and SBs) and a single biomass population. The dual substrate model provided better results than the diauxic model and therefore it was adopted during this study. Overall:

  • The dual substrate model described successfully the higher effluent soluble COD observed in the HRAS systems due to the partial removal of SBs, which is almost completely removed in higher SRT systems.

  • The dual substrate model was able to accurately predict the elevated (compared to SRT > >1 day) effluent soluble COD for the HRSD A-stage.

  • The dual substrate model was more accurate than the diauxic model with respect to effluent soluble COD during the validation phase using HRSD A-stage data.

ACKNOWLEDGEMENT

This paper was presented at WWTmod2014 and the fruitful discussions are kindly acknowledged.

REFERENCES

REFERENCES
Beun
J. J.
Paletta
F.
Van Loosdrecht
M. C. M.
Heijnen
J. J.
2000
Stoichiometry and kinetics of poly-hydroxybutyrate metabolism in aerobic, slow growing, activated sludge cultures
.
Biotechnology and Bioengineering
67
(
4
),
379
389
.
doi: 10.1002/(SICI)1097-0290(20000220)67:4 < 379::AID-BIT1 > 3.0.CO;2-2
.
Böhnke
B.
Diering
B.
1986
United States Patent No. 4,568,462. U. S. P. a. T. Office
.
Carucci
A.
Dionisi
D.
Majone
M.
Rolle
E.
Smurra
P.
2001
Aerobic storage by activated sludge on real wastewater
.
Water Research
35
(
16
),
3833
3844
.
Dold
P. L.
Ekama
G. A.
Marais Van
G. R.
1980
A general model for the activated sludge process
.
Progress in Water Technology
12
(
6
),
47
77
.
Ekama
G. A.
Dold
P. L.
Marais Van
G. R.
1986
Procedure for determining influent COD fractions and the maximum specific growth rate of heterotrophs in activated sludge systems
.
Water Science and Technology
18
(
6
),
91
114
.
Frolund
B.
Griebe
T.
Nielsen
P. H.
1995
Enzymatic activity in the activated-sludge floc matrix
.
Applied Microbiology and Biotechnology
43
(
4
),
755
761
.
Haider
S.
Svardal
K.
Vanrolleghem
P. A.
Kroiss
H.
2003
The effect of low sludge age on wastewater fractionation (SS, SI)
.
Water Science and Technology
47
,
203
209
.
Henze
M.
Gujer
W.
Mino
T.
Van Loosdrecht
M.
2000
Activated Sludge Models ASM1, ASM2, ASM2d and ASM3
.
IWA Publishing
,
London
,
121
.
Jimenez
J. A.
2002
Kinetics of COD removal in the activated sludge process including bioflocculation, The Department of Civil and Environmental Engineering (PhD thesis), University of New Orleans, umi:3084238
.
Jimenez
J. A.
La Motta
E. J.
Parker
D. S.
2005
Kinetics of removal of particulate chemical oxygen demand in the activated-sludge process
.
Water Environment Research
77
(
5
),
437
446
.
Miller
M. W.
Regmi
P.
Bunce
R.
Hingley
D. M.
Kinnear
D.
Wett
B.
Murthy
S.
Bott
C. B.
2012
A/B Process Pilot Optimized for Nitrite Shunt: High Rate Carbon Removal Followed by BNR with Ammonia-Based Cyclic Aeration Control
.
Paper presented at the WEFTEC 2012
,
New Orleans, Louisiana, Water Environment Federation, Alexandria Virginia
.
Miller
M. W.
Jimenez
J.
Murthy
S.
Kinnear
D.
Wett
B.
Bott
C. B.
2013
Mechanisms of COD Removal in the Adsorption Stage of the A/B Process
.
Paper presented at the WEFTEC 2013
,
Chicago, Illinois, Water Environment Federation, Alexandria Virginia
.
Pala-Ozkok
I.
Rehman
A.
Kor-Bicakci
G.
Ural
A.
Schilhabel
M. B.
Ubay-Cokgor
E.
Jonas
D.
Orhon
D.
2013
Effect of sludge age on population dynamics and acetate utilization kinetics under aerobic conditions
.
Bioresource Technology
143
,
68
75
.
Schulze-Rettmer
R.
Zuckut
S. W.
1998
Treatment of textile dyeing wastewater by adsorption/bio-oxidation process
.
Textile Chemist and Colorist
30
(
5
),
19
23
.
Sin
G.
Guisasola
A.
De Pauw
D. J. W.
Baeza
J. A.
Carrera
J.
Vanrolleghem
P. A.
2005
A new approach for modelling simultaneous storage and growth processes for activated sludge systems under aerobic conditions
.
Biotechnology and Bioengineering
92
(
5
),
600
613
.
Third
K. A.
Newland
M.
Cord-Ruwisch
R.
2003
The effect of dissolved oxygen on PHB accumulation in activated sludge cultures
.
Biotechnology and Bioengineering
82
(
2
),
238
250
.

Supplementary data