Design, Construction, Start-Up, and Operation of a Full-Scale Separate Stage Moving Bed Biofilm Reactor Nitrification Process

Utilization of biomass carriers for upgrading existing plants has been receiving increased attention in recent years and has resulted in a growing body of literature. In this study, data sets generated from operational integrated fixed film/activated sludge (IFAS) and fluidized fixed film systems (FFFR) systems reported in the literature were used to assess a model's representation of fundamental outcomes. Three plants were employed for the study: the Waterdown STP (IFAS), the Moorhead WWTF (FFFR) and Royal Roads University Pilot Plant (FFFR).The impact of process variables on performance was evaluated by modifying operational and influent parameters in the model such as SRT, temperature, dissolved oxygen, BOD and TKN (using default model parameters) and comparing the results to reported findings. Similarly, the effect of calibration parameters was evaluated by modifying each parameter considered and recording output nitrification rates.The model accurately predicted the average improvement in nitrification efficiency observed following an upgrade to IFAS in the Waterdown STP at default model parameters. Model predictions of effluent ammonia in a separate stage nitrification FFFR were also within two standard deviations of reported average summer and winter monthly conditions. Trends such as increases in nitrification rate with D.O. and decreases in nitrification rate with reduced temperature and elevated organic loading rate were established by the model. The three calibration parameters tested: L max ; reduction in diffusion in the biofilm; and liquid film thickness; all had significant consequences for nitrification rates. Influent characterization was found to have the potential to significantly affect nitrification rates in FFFR's.The predicted results (employing default parameters) corresponded well to available published results, considering the variability in the operational parameters. The model predicted nitrification rates to within 2-9% of the observed means. The model tended to predict lower than observed impacts of D.O. and organic surface loading rate on nitrification rates. Model results for L max less than the default value of 0.50 mm tended to further reduce the dependence of nitrification rates on the above noted variables and underestimate nitrification rates. Further research to refine default calibration parameters that can both simulate a reasonable biomass in the system and match observations would therefore prove useful. 6266

Section: Methodsmentioning

confidence: 99%

“…The FFFR consists of a single basin equipped with floating coarse bubble aeration grids, a 32% fill fraction by volume of Hydroxyl-PAC biomass carriers, and an effluent screen for retention of the carriers. A detailed plant description, including data references used below, can be found in Zimmerman et al (2004).…”

Section: Moorhead Wwtfmentioning

confidence: 99%

Modeling as a Tool for Investigating Ifas Processes

Maas¹,

Legge²,

Parker³

2005

“…A schematic diagram of the biofilm carrier system is shown in Figure 2 and design criteria are provided in Table 1. Additional system information has been previously described by Zimmerman et al (2004). Sophisticated modeling of biological wastewater treatment has become common in recent years due to the widespread use of powerful computers and availability of user-friendly software programs.…”

Section: Introductionmentioning

confidence: 99%

Simulation of Nitrification Performance in a Biofilm Carrier System Under Near-Steady State and Dynamic Operating Conditions

Zimmerman¹,

Maas²,

Richard³

2007

Self Cite

Three kinetic models were used to simulate, under near-steady state and dynamic operating conditions, the performance of a full-scale biofilm carrier process performing tertiary nitrification. Two of the models were based on simple Monod-type kinetic relationships. One incorporated measured biomass concentrations with kinetic relationships based on the maximum specific growth rate (μ m ) for the nitrifying biomass. The second assumed that attached growth nitrification is a surface-related phenomenon, and therefore independent of biomass considerations, with kinetic relationships based on the maximum substrate utilization rate (r′ m ). The Monod-type models were fit to dynamic operating data to determine the corresponding kinetic rates (μ m and r′ m ). The third model, a commercially available dynamic simulation package (GPS-X TM ), was calibrated with measured attached growth total solids (AGTS) concentrations and effluent data based on the user input maximum biofilm thickness (L max ).The models were used to evaluate the relative significance of process variables (temperature, pH, DO, and biomass concentration) independently and in combination. All of the models were capable of predicting performance under near-steady state conditions, though the dynamic simulator was highly dependent on the correct assumption of the L max . Under dynamic operating conditions, the models were also able to predict trends in performance reasonably well. Of the two Monod-type models, the best-fit model was that excluding biomass concentrations and including adjustments to r′ m for DO only. Data from the dynamic simulation package confirmed the results from Monod-type models with the best-fit model excluding temperature effects and including only limited DO effects. This may suggest that the biofilm is highly adaptive, increasing the mass of nitrifiers during cold temperatures and resulting in relatively high winter nitrification rates despite reduced nitrifier growth rates.The utility of the models developed in this study is dependent on the ability to predict (and in the case of GPS-X, calibrate) attached growth biomass concentrations in response to seasonal variations and/or changes in influent NH 3 -N loading. For that reason, the Monod-type model that excludes biomass concentrations is anticipated to be the simplest model to use in practice as it does not require prediction of AGTS.

“…Although the influent loading has been variable and consistently exceeded the design loading, the system has achieved the design effluent concentration during the months when permit limits are applicable. Performance was only slightly reduced during colder weather months when permit limits are not in effect (Zimmerman et al, 2004). Influent and effluent ammonia-nitrogen as measured across the MBBR basin since start up in 2003 are shown in Figure 3 and 4, respectively.…”

Section: Introductionmentioning

confidence: 99%

Is Your Moving Bed Biofilm Reactor (Mbbr) Running on All Cylinders?

Zimmerman¹,

Richard²,

Lynne³

et al. 2005

Self Cite

Evaluation of a full-scale MBBR process for nitrification is reported. Performance of the MBBR was found to be highly dependent on the hydraulic characteristics of the basin which could be optimized with improved influent flow distribution at the basin inlet. Installing a baffle on the influent pipe or introducing influent at multiple points located laterally across the entire influent end of the basin may provide this improvement. Higher aeration (mixing) rates alone did not improve influent distribution, however, increased mixing intensity focused at the inlet was not considered.Laboratory bench-scale analyses were conducted to evaluate the maximum substrate utilization rate, r' m , for the nitrifying biomass in the full-scale MBBR process. Assuming a half-saturation constant, K n , of 1 mg/L, r' m was estimated by statistically fitting the observed data to an attached-growth Monod-type rate expression. Although the data provided an excellent fit, r' m values (20ºC, 8.5 pH) exhibited considerable variation, ranging from 2.024 to 4.418 g/(m 2 -d) with an average of 3.370 g/(m 2 -d). Accurate determination of r' m may be affected by conditions in the full-scale MBBR (e.g. NH 3 -N loading, DO versus NH 3 -N rate limitation, etc.) and/or the assumption that the quantity of active nitrifying biomass is directly related to the media surface area provided. In order to develop a further understanding of the MBBR process, a standardized technique for the determination of r' m is needed.Results of hydraulic analyses indicated that the full-scale MBBR can modeled as two CFSTRs in series with a bypass flow representing hydraulic deficiencies. Model simulations indicated that hydraulic improvement in the full-scale MBBR could significantly increase the design influent NH 3 -N concentration and capacity under NH 3 -N rate limiting conditions. For an MBBR maximized with respect to hydraulic considerations, process performance was found to be highly dependent on r' m . With significant hydraulic deficiencies, process performance is not as sensitive to nitrification kinetics, but rather, is governed by the hydraulic characteristics of the basin.