Dynamic modeling of a continuously moving bed bioreactor performing tertiary nitrification and filtration
Sustainable process-product development & green chemistry
Sustainable & Clean Technologies - Ib: Extraction & Remediation (T1-4b)
Keywords: ASTRASAND, biofilm modeling, moving bed bio-filter, nitrification, tertiary treatment
In this study, the development of a dynamic model for a process called ASTRASAND® – a continuously moving bed sand filter performing a tertiary treatment of nitrogen (mainly nitrification) and suspended solids removal, is presented. The wastewater treatment in general involves three complementary steps: primary treatment aiming at solids removal, secondary treatment that focuses on mainly biological but sometimes chemical removal of carbonaceous matter, nitrogen and phosphorus from the wastewater and tertiary treatment that polishes the effluent before discharging into receiving bodies, e.g. rivers, surface waters or ground water. The interest into tertiary treatment within wastewater treatment plants has increased recently due to strict effluent legislation introduced by European Community directives within river basin and urban water management, simply because the secondary treatment doesn’t suffice alone to comply with stringent effluent discharge limits.
The ASTRASAND system is mainly a sand filter that removes suspended solids physically through sand-filtration and performs nitrification by specialized bacteria enriched in the biofilm developed on the sand particles. Finally, the sand particles are recirculated continuously in the reactor bed using an airlift concept to control clogging (typically occurring) in the sand filter. This system has been successfully applied in full-scale for tertiary nitrification in the UK and tertiary denitrification in Belgium and the Netherlands in both industrial and domestic wastewater treatment plants. Design, optimal operation and control of such systems rely on process engineering and empirical/experimental knowledge. To gain more insight and better understand this rather complex treatment configuration, we developed a dynamic model to mainly describe nitrification and suspended solids removal, with the focus placed on the former.
To this end, the WEST® (MostForWater, NV, Kortrijk, Belgium) modeling and simulation platform is used to implement the integrated modeling framework, necessary to describe the hydrodynamic, the biofilm and the microbial transformation processes occurring in the ASTRASAND system. In the hydrodynamic part of the model, the mixing in bulk water is described using tanks-in-series approach and the airlift recirculation of the sand particles is described using a dedicated algorithm written in WEST®. For the selection of an appropriate mathematical model for the biofilm matrix, we have used the criteria of Eberl et al. (2005) and used the 1-D numerical dynamical model of Rauch et al., (1999). Last, the microbial transformations (mainly nitrification but also decay and growth of heterotrophic organisms as well) taking place in the biofilm matrix are described using ASM1 (Henze et al., 2000).
We have evaluated the integrated model on a pilot-scale ASTRASAND plant located in Nether Stowey, UK. This pilot plant has a capacity of 1.4 m3 bed-volume and 0.7 m2 surface area. Three months data of influent and effluent ammonium and solids load have been used for the calibration of the model. The model was first calibrated to match the steady-state performance by using the Monte Carlo calibration method of Sin et al. (2007). Using the initial biofilm composition determined in the previous step, i.e. the initial biofilm thickness, the model was then calibrated using a two weeks long dynamic set of ammonium data and 6 tank-in-series to describe plug-flow behavior of the bulk water in the filter. The following parameters were adjusted during the dynamic calibration: the attachment, detachment, maximum growth rate (A) and yield parameters (YA) of the nitrification process. The calibration model was then validated using a data set different than the one used in the calibration. The Janus coefficient was found close to 1 (i.e. 0.9) confirming statistically that the model predictions remained valid in the validation period. The mean absolute error (MAE) between the model and the data was found 0.8 mgNH4-N/l. This is an acceptable bias considering the many assumptions and simplifications that were required for adequate modeling complexity– inherent to almost any deterministic modeling study. Finally, the calibrated and validated model can now be used for several applications ranging from scenario analysis to process operation and optimization purposes.
References:
Eberl, H., Morgenroth, E., Noguera, D., Picioreanu, C., Rittman, B., van Loosdrecht, M and Wanner, O. (2006) Mathematical Modelling of Biofilms. STR No. 18. IWA Publishing, London, UK.
Henze M., Gujer W., Mino T. and van Loosdrecht M.C.M. (2000) Activated Sludge Models ASM1, ASM2, ASM2d and ASM3. IWA Scientific and Technical Report 9 IWA. London.
Sin G., De Pauw D., Weijers S. and Vanrolleghem P.A. (2007) A Monte-Carlo based global calibration methodology for calibration of Activated Sludge Models (ASM). Submitted to Water Research.
Rauch, W., Vanhooren, H. and Vanrolleghem, P.A. (1999) A simplified mixed-culture biofilm model. Water Research, 33, pp. 2148-2162.
Presented Wednesday 19, 16:20 to 16:40, in session Sustainable & Clean Technologies - Ib: Extraction-Remediation (T1-4b).
