Why Does Proper Simulation Matter In The Acronym Soup of Biofilm Systems? AGS, IFAS, MBBR

INTRODUCTION
Advanced nutrient management approaches are increasingly relying on biofilm systems to achieve their goals. These include such processes as aerobic granular sludge (AGS), both batch and continuous flow, the mobile carrier (MOBâ„¢) process that attempts to simplify AGS, moving bed biofilm reactor (MBBR) technologies for Anammox based systems in both main and sidestream processes, and IFAS technologies for low temperature or intensified nitrification. Many practitioners do not have available the knowledge or mentoring needed to properly model such biofilm systems, and thus often rely on empirical approaches or vendor provided sizing for such systems. These empirical approaches can result in either oversizing or under sizing biofilm-based systems, both of which can have negative impacts upon the adoption and use of biofilm systems to benefit the public and environment. This paper discusses the capabilities, considerations, and implications of advanced biofilm system simulation with the goal of providing readers with the information they need to understand what biofilm simulation results actually mean.
METHODS
Simulations of the above-described biofilm systems were completed in Dynamita's Sumo21® to illustrate the impacts of various modelling settings and structure upon the simulation results. For continuous flow systems an identical primary and MLE bioreactor system was used to compare various approaches to biofilm modelling. For sequencing batch reactor (SBR) systems, the base Sumo AGS reactor was used, as well as Jacobs SBR AGS model. The model structures of these approaches are compared against each other and the structures used by Envirosim's Biowin.
RESULTS AND BENEFITS
All currently implemented biofilm models in commercial simulators use a layer approach where a biofilm is modeled as a variable number of CSTRs vertically along the depth of the biofilm acting as different layers (Figure 1), with the needed diffusion and mixing between layers. The geometry of these layers is generally rectangular, i.e. the area of the biofilm is the same on the bottom as it is on the top, but with the recent interest in granular activated sludge, the geometry can also be implemented as spherical, where the area of the bottom layer is smaller than the top. Two of the key parameters in biofilm modelling is the selected number of layers, as well as the mass transfer boundary layer (MTBL) thickness. The MTBL thickness is variable depending on the shear rate present in the reactor being modelled, generally the MTBL is considered to be thinner in aerated zones and thicker in the non-aerated. The selection of the number of layers to use in the model is very dependent on the desired accuracy of the results balanced against the speed of the simulation. Figure 2 shows the impact of the number of modeled biofilm layers on the ammonia concentration in the IFAS tank, with all other parameters being held constant. Another significant consideration in modelling biofilm systems is the physical geometry of the biofilm reactor. Almost all current simulators use a MBBR model where the biofilm is retained within the CSTR and only suspended growth transfers between reactors. This assumption works relatively well in traditional biofilm reactor systems like IFAS and MBBRs, but if there is a plug flow through the biofilm bed (as in AGS SBRs) or that the biofilm transfers between CSTRs, as in continuous flow AGS systems, this assumption is no longer correct. To illustrate this point, a continuous flow AGS system was modeled as a series of MBBRs with retained biofilms, and another was modeled with the mobile carrier approach where biofilm granules are allowed to flow freely throughout the process. The effluent ammonia concentrations produced by a typical diurnal variation is shown in Figure 3. It can be seen that the modeling results of these two approaches provide significantly different results, with the mobile carrier model structure showing improved ammonia removal. This is a result of the mobile carrier having a higher nitrifier biomass fraction than the MBBR reactor in the last aerobic zone. The only commercial simulators that allow transfer of biofilms between zones are Sumo and SIMBA# by ifak. Sumo uses an inert spherical core and biofilm model (Mobile Carrier) with a variable number of fixed thickness layers. This core is allowed to flow between units, and thus transfer the biofilm between CSTRs. SIMBA# uses a detailed geometry model to adapt the model to any biofilm type and shape with a variable thickness biofilm model. A special Mobile Carrier Biofilm Reactor (MCBR) block library has been developed for modeling movement of carriers within a plant. AGS can be modeled with a specific AGS reactor block or with the mobile carrier library. The current intense industry interest in aerobic granular sludge has resulted in all the simulator providers developing AGS reactor models. It is important to understand the differences between these models from a functional perspective. Three are compared in Figure 4: 1) Jacobs AGS Model: Full hydrodynamic model with effluent breakthrough and bed growth calculations and mobile carrier model. 2) Biowin AGS Model: Single MBBR AGS model followed by a plug flow clear liquid reactor. Does not capture plug flow through reactor bed. Will result in differences from actuality, and 3) Sumo AGS Model: Single MBBR AGS Model, with specified effluent and sludge TSS. Approximates effluent quality by fixing the decant at the concentrations present at the start of feed/decant cycle. Each of these reactors have varying levels of functionality that are appropriate for different modelling goals.
CONCLUSIONS
Biofilm based systems are rapidly becoming core to achieving intensification and efficiency goals in our industry and proper simulation of them will encourage their use and capabilities. The structure and setup of the biofilm simulation to achieve the simulation goals is of critical importance and often requires the users to understand the deeper model structures for proper usage. This paper will compare and contrast the various model structures to illustrate the impacts of these decisions for engineers and utilities.