Model-based evaluation of two BNR processes--UCT and A2N

Anoxic/oxic treatment without biomass recycle

The Anoxic/Oxic (A/O) process involves recirculating mixed liquor between its A and O tanks so that nitrate produced in the O tank can be used to for denitrification with influent COD in the A tank. Because biomass is recirculated along with nitrate, A/O operation leads to similar microbial communities in the A and O tanks, which may decrease the rates of denitrification and nitrification in each tank. Here, bench-scale experiments simulated this aspect of the A/O process by exchanging biomass between an anoxic flask and an oxic cylinder at exchange ratios of 0%, 20%, 30%, and 50%. Nitrification and denitrification rates were only 40% and 19% for 50% biomass exchange of that for no biomass exchange. Phylogenetic analysis documented that the microbial communities became much more similar with biomass exchange, and the finding was consistent with community composition in a full-scale A/O process in a municipal wastewater treatment plant. A two-stage vertical baffled bioreactor (VBBR) realized efficient total‑nitrogen removal in recirculation without biomass exchange. Average removals of COD and TN were respectively 6% and 22% higher for the two-stage VBBR than the conventional A/O process, but its hydraulic retention time (HRT) was 55% to 70% of the volume of a conventional A/O process treating the same influent wastewater. The VBBR was more efficient because its anoxic biofilm was enriched in denitrifying bacteria, while its oxic biofilm was enriched in nitrifying bacteria. For example, the phylum Chloroflexi was greater in the An-VBBR, while the phylum Proteobacteria was greater in the Ox-VBBR.

A novel metabolic-ASM model for full-scale biological nutrient removal systems

This study demonstrates that META-ASM, a new integrated metabolic activated sludge model, provides an overall platform to describe the activity of the key organisms and processes relevant to biological nutrient removal (BNR) systems with a robust single-set of default parameters. This model overcomes various shortcomings of existing enhanced biological phosphorous removal (EBPR) models studied over the last twenty years. The model has been tested against 34 data sets from enriched lab polyphosphate accumulating organism (PAO)-glycogen accumulating organism (GAO) cultures and experiments with full-scale sludge from five water resource recovery facilities (WRRFs) with two different process configurations: three stage Phoredox (A2/O) and adapted Biodenitro™ combined with a return sludge sidestream hydrolysis tank (RSS). Special attention is given to the operational conditions affecting the competition between PAOs and GAOs, capability of PAOs and GAOs to denitrify, metabolic shifts as a function of storage polymer concentrations, as well as the role of these polymers in endogenous processes and fermentation. The overall good correlations obtained between the predicted versus measured EBPR profiles from different data sets support that this new model, which is based on in-depth understanding of EBPR, reduces calibration efforts. On the other hand, the performance comparison between META-ASM and literature models demonstrates that existing literature models require extensive parameter changes and have limited predictive power, especially in the prediction of long-term EBPR performance. The development of such a model able to describe in detail the microbial and chemical transformations of BNR systems with minimal adjustment to parameters suggests that the META-ASM model is a powerful tool to predict and mitigate EBPR upsets, optimise EBPR performance and to evaluate new process designs.

Cost effectiveness of phosphorus removal processes in municipal wastewater treatment

Meeting stringent phosphorus (P) discharge standards remains one of the major challenges for wastewater utilities due to increased economic burdens associated with advanced (i.e., secondary, tertiary) treatment processes. In a trade-off between higher treatment cost and enhanced P removal, it is critical for the treatment plants to be able to select the most appropriate technology. To this end, established/emerging high performing P removal/recovery technologies (e.g., Modified University of Cape Towne process, Bardenpho process, membrane bioreactors, IFAS-EBPR, struvite recovery, tertiary reactive media filtration) were identified and full-scale treatment plant designs were developed. Using advanced mathematical modeling techniques, six different treatment configurations were evaluated in terms of performance and cost effectiveness ($/lb of P removed). Results show that the unit cost for P removal in different treatment alternatives range from $42.22 to $60.88 per lb of P removed. The MUCT BNR + tertiary reactive media filtration proved to be one of the most cost effective configurations ($44.04/lb P removed) delivering an effluent with total P (TP) concentration of only 0.05 mg/L. Although struvite recovery resulted in significant reduction in biosolids P, the decrease in effluent TP was not sufficient to meet very stringent discharge standards.

Efficient model calibration method based on phase experiments for anaerobic-anoxic/nitrifying (A2N) two-sludge process

A systematic calibration and validation procedure for the complex mechanistic modeling of anaerobic-anoxic/nitrifying (A2N) two-sludge system is needed. An efficient method based on phase experiments, sensitivity analysis, and genetic algorithm is proposed here for model calibration. Phase experiments (anaerobic phosphorus release, aerobic nitrification, and anoxic denitrifying phosphate accumulation) in an A2N sequencing batch reactor (SBR) were performed to reflect the process conditions accurately and improve the model calibration efficiency. The calibrated model was further validated using 30 batch experiments and 3-month dynamic continuous flow (CF) experiments for A2N-SBR and CF-A2N process, respectively. Several statistical criteria were conducted to evaluate the accuracy of model predications, including the average relative deviation (ARD), mean absolute error (MAE), root mean square error (RMSE), and Janus coefficient. Visual comparisons and statistical analyses indicated that the calibrated model could provide accurate predictions for the effluent chemical oxygen demand (COD), ammonia nitrogen (NH₄⁺-N), total nitrogen (TN), and total phosphorus (TP), with only one iteration.

Nitrate levels modulate the abundance of Paracoccus sp. in a biofilm community

Conditions required to enhance a particular species efficient in degradative capabilities is very useful in wastewater treatment processes. Paracoccus sp. is known to efficiently reduce nitrogen oxides (NOx) due to the branched denitrification pathway. Individual-based simulations showed that the relative fitness of Paracoccus sp. to Pseudomonas sp. increased significantly with nitrate levels above 5 mM. Spatial structure of the biofilm showed substantially less nitrite levels in the areas of Paracoccus sp. dominance. The simulation was validated in a laboratory reactor harboring biofilm community by fluorescent in situ hybridization, which showed that increasing nitrate levels enhanced the abundance of Paracoccus sp. Different levels of NOx did not display any significant effect on biofilm formation of Paracoccus sp., unlike several other bacteria. This study shows that the attribute of Paracoccus sp. to tolerate and efficiently reduce NOx is conferring a fitness payoff to the organism at high concentrations of nitrate in a multispecies biofilm community.

Incorporating microbial ecology into the metabolic modelling of polyphosphate accumulating organisms and glycogen accumulating organisms

In the enhanced biological phosphorus removal (EBPR) process, the competition between polyphosphate accumulating organisms (PAO) and glycogen accumulating organisms (GAO) has been studied intensively in recent years by both microbiologists and engineers, due to its important effects on phosphorus removal performance and efficiency. This study addresses the impact of microbial ecology on assessing the PAO-GAO competition through metabolic modelling, focussing on reviewing recent developments, discussion of how the results from molecular studies can impact the way we model the process, and offering perspectives for future research opportunities based on unanswered questions concerning PAO and GAO metabolism. Indeed, numerous findings that are seemingly contradictory could in fact be explained by the metabolic behaviour of different sub-groups of PAOs and/or GAOs exposed to different environmental and operational conditions. Some examples include the glycolysis pathway (i.e. Embden-Meyerhof-Parnas (EMP) vs. Entner-Doudoroff (ED)), denitrification capacity, anaerobic tricarboxylic acid (TCA) cycle activity and PAOs' ability to adjust their metabolism to e.g. a GAO-like metabolism. Metabolic modelling may further yield far-reaching influences on practical applications as well, and serves as a bridge between molecular/biochemical research studies and the optimisation of wastewater treatment plant operation.

Nitrogen and Phosphorous Removal in Municipal Wastewater Treatment Plants in China: A Review

Surface water environment in China was degraded rapidly in the last two decades, resulting in increasingly tighten criteria issued for municipal wastewater treatment plants (WWTPs). This paper reviewed the recent advances of process design and operational optimization for nutrients removal. Three major processes, as anaerobic-anoxic-oxic (AAO) process, oxidation ditch (OD), and sequencing batch reactor (SBR) occupied 65% of WWTPs amounts and 54% of treatment volumes of China in 2006. However conservative process designs and operational faults often impaired the process performances and energy efficiency. Therefore, typical processes were modified, combined, and innovated to meet the requirements of the diverse influent characteristics and lower energy consumptions. Furthermore, operational optimization techniques by modeling, simulation, and real-time control were also developed and applied in China to improve the process operation. Although great efforts had been contributed to improve the WWTPs performances in China, attentions should be continuously paid to the introduction, instruction, and implementation of advanced techniques. At last, the technical demands and appropriated techniques of WWTPs in China were briefly discussed.

Enhanced biological nutrient removal using MUCT-MBR system

Biological nutrient removal was investigated in a combined modified University of Cape Town and membrane bioreactor system. When the influent nutrient mass ratio (COD/TN/TP) was 28.5/5.1/1 to 28.5/7.2/1, average removal efficiencies of COD, TN and TP were 90%, 81.6%, 75.2%. Obvious denitrifying phosphorus removal occurred with C/N ratio 3.98. When nitrite was the main electron acceptor, the ratio of denitrifying phosphate uptake to the total phosphate uptake were 99.8% and the sludge yield was 0.28kg VSS/kg COD; when nitrate was the main electron acceptor, the ratio was 92% and the yield was 0.32kg VSS/kg COD. In case of nitrite, the system not only kept TP and TN removal at 89.1% and 82.2%, but also ensured less sludge production. Batch tests showed that the proportion of denitrifying phosphorus-accumulating organisms in the total phosphorus-accumulating organisms in the system was higher than 80%.

Model-based evaluation of competition between polyphosphate- and glycogen-accumulating organisms

Many studies show that glycogen-accumulating non-polyphosphate organisms (GAOs) can compete with polyphosphate-accumulating organisms (PAOs) for organic substrate under anaerobic conditions and may indeed cause the deterioration of enhanced biological phosphorus removal (EBPR) systems. Understanding their behaviors in an anaerobic/aerobic (A/O) system at different operational conditions is essential in developing control strategies that ensure EBPR. A model-based evaluation of competition between PAOs and GAOs under different operational conditions was presented in this study. At 30 degrees C and a 10-day sludge age, the dominance of GAOs in the A/O sequencing batch reactor (SBR) was strongly dependent upon their considerable kinetic advantage in anaerobic acetate uptake. At 20 degrees C and a 10-day sludge age, the kinetic advantage of GAOs in anaerobic acetate uptake could be less, compared to that at 30 degrees C and a 10-day sludge age, leading to the relative dominance of PAOs and a stable phosphorus removal in the A/O system. At 30 degrees C and a 3-day sludge age, the parameters responsible for determining the aerobic distribution of anaerobically stored X(PHA) for both PAOs and GAOs, other than kinetic parameters of anaerobic acetate uptake, are important for them being dominant in the A/O SBR. In a situation when the q(PHA,P) value is lower than q(PHA,G) but comparable, PAOs may still be dominant in the A/O SBR, presumably because their aerobic conversion fraction of biomass production from PHA was higher than that of the GAOs.

Aerobic phosphorus release linked to acetate uptake in bio-P sludge: Process modeling using oxygen uptake rate

The main processes involved in enhanced biological phosphorus removal (EBPR) under anaerobic and subsequently aerobic conditions are widely described in the literature. Polyphosphate accumulating organisms (PAO) are the organisms responsible for this process. However, the mechanisms of PAO are not fully established yet under conditions that differ from the classical anaerobic/aerobic conditions. In this work, we made a comparison between the behavior of PAO under classical EBPR conditions and its behavior when consuming substrate under only aerobic conditions. In addition, oxygen uptake rate (OUR) was measured in the set of experiments under aerobic conditions to improve the characterization of the process. A kinetic and stoichiometric model based on Activated Sludge Model No.2 (ASM2) and including glycogen economy (AnOx model), calibrated for classical anaerobic/aerobic conditions, was not able to describe the experimental data since it underestimated the acetate consumption, the PHB storage, and the OUR. Two different hypotheses for describing the experimental measurements were proposed and modeled. Both hypotheses considered that PAO, under aerobic conditions, uptake acetate coupled to PHB storage, glycogen degradation, and phosphorus release as in anaerobic conditions. Moreover, the first hypothesis (PAO-hypothesis) considered that PAO were able to store acetate as PHB linked to oxygen consumption and the second one (OHO hypothesis) considered that this storage was due to ordinary heterotrophic organisms (OHO). Both hypotheses were evaluated by simulation extending the AnOx model with additional equations. The main differences observed were the predictions for PHB degradation during the famine phase and the OUR profile during both feast and famine phases. The OHO hypothesis described the experimental profiles more accurately than the PAO hypothesis.

Metabolic modelling of full-scale biological nitrogen and phosphorus removing wwtp's

This paper evaluates the experiences with modelling full-scale biological phosphorus and nitrogen removing wastewater treatment plants. For the simulation, we used a metabolic phosphorus model integrated in ASM2d, further referred to as TUDP-model. It was found that the metabolic model for bio-P removal can be applied for modelling full-scale wwtp's, without extensive parameter adjustments. A stepwise modelling approach was proposed. Only three specific parameters were calibrated. Two parameters, the inert fraction in the influent and the actual anoxic sludge fraction will need calibration for all systems. Parameter sensitivity analyses showed that the sensitivity of operational data, often considered as known input data, is high. The model kinetics were found less sensitive. This will in general be the case for all low-loaded wwtp's recorded in their pseudo-steady state. Based on mass balance calculations, operational data and measurements were evaluated. Since all terms on a phosphorus balance can be measured, errors in operational data (e.g. SRT and flow rates) become very apparent in P-removal models (ASM2d, TUDP). We suggest using the P-balance in general for wwtp modelling, as a check on data consistency. This study showed that considering operational data per definition as known input data is not justified. Therefore, operational data should be evaluated, or considered in model calibration over the use of kinetic and stoichiometric parameters.