Watermatex 2019 Plant-wide optimization through dynamic compartmental modelling and analysis of limiting processes. C. De Mulder*, U. Rehman**, T. Flameling***, S. Weijers***, Y. Amerlinck* and I. Nopens* * BIOMATH, Department of Data Analysis and Mathematical Modelling, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, 9000 Gent, Belgium (Email: chaim.demulder@ugent.be, ingmar.nopens@ugent.be) ** AM-TEAM, Oktrooiplein 1, Box 601, 9000 Gent, Belgium (Email: usman.rehman@am-team.com) *** Waterschap De Dommel, PO Box 10.001, 5280 DA Boxtel, The Netherlands (Email: tflameling@dommel.nl, sweijers@dommel.nl) Abstract: The current Tanks in Series (TIS) approach used to model the hydrodynamics of Water and Resource Recovery Facilities lacks in spatial detail to address most modelling goals, including reactor design and predictions during wet weather. On the other hand, advanced CFD simulations to reach these goals are often too time-consuming. This paper investigates an intermediate model structure, named Compartmental Model (CM), in terms of prediction capacity compared to the classic TIS approach. It does so in the context of a plant-wide model, by comparing online data with simulations from a TIS model and two types of CMs, static and dynamic. Results show that this new model structure indeed improves the prediction quality on several fronts. Further improvements, both in model performance and compartimentalisation methodology, are still challenging. Keywords: Compartmental Modeling; plant-wide modeling; process limitations Introduction Hydrodynamic modelling of Water & Resource Recovery Facilities (WRRFs) is, at least within flow-sheet tools, currently mostly based on a Tanks-In-Series (TIS) approach. This approach conceptually divides a bioreactor into multiple completely mixed tanks in only one dimension, i.e. the direction of the bulk flow. Variations in the other two dimensions are not taken into account, although these do occur in reality, causing for example mass transfer-limited zones or short circuiting. This makes the TIS approach oversimplified; it does not contain the hydrodynamic detail needed to achieve several common modelling goals such as reactor design evaluation, the development of precise control strategies or model development for both dry and wet weather, i.e. without the need for recalibration. A solution to this problem can come in the form of Computational Fluid Dynamics (CFD) simulations, modelling hydrodynamics in a very detailed way, even including biokinetics. However, the use of CFD as a mainstream WRRF modelling tool is limited, mainly due to its high computational demands and the increased need for detailed validation data. An intermediate modelling approach with sufficient hydrodynamic detail and manageable calculation times seems appropriate. Such an intermediate model structure, named a Compartmental Model (CM), has been the topic of previous research already (Alex et al., 2002; Rehman, 2016). Rehman (2016) developed a method that uses a detailed CFD(-biokinetic) model to construct a Compartmental Model that describes a reactor as a conceptual network of spatially localized compartments, connected through convective and exchange fluxes. Relatively new in the context of WRRF modelling, this Compartmental Model structure has only been applied in a very limited number of cases, made use of fixed compartment volumes and fluxes and has never been validated on a full scale WRRF (Gresch et al., 2009; Le Moullec et al., 2011). This paper deepens the knowledge on CMs by applying the model structure to an available plant-wide model and by making the compartment volumes dynamically dependent on factors well-known for affecting the hydrodynamic patterns (e.g. influent CORE Metadata, citation and similar papers at core.ac.uk Provided by Ghent University Academic Bibliography