Next-generation compartmental models for applications in Digital Twinning of WRRFs

Abstract

One of the bottlenecks of mechanistic plant-wide water resource recovery facility (WRRF) models commonly used in most of the available digital twin (DT) cases relates to the level of hydrodynamics details included in these models. They often neglect the detailed hydrodynamics of the reactors and simplify the mixing behaviour with the Tanks-In-Series (TIS) approach. In this study, a compartmental model (CM) based on an integrated hydrodynamic-biokinetic model was developed to describe the hydrodynamics of an anoxic reactor at a full-scale water resource recovery facility (WRRF). The results presented in this study highlight the importance of local density variations as a function of suspended solids to improve the model's predictive power and avoid unnecessary calibration of the model parameters. There are likely many plants that have a RAS returning from the bottom of a bioreactor, leading to stratification and, hence, non-ideal reactor performance. This will hamper DT development for these plants but can be solved by introducing a CM. The availability of a model with sufficient predictive power to respond to new scenarios and disturbances without frequent validation/(re)calibration is a necessity for the transition to DT applications and the use of models as a real-time decision support tool for the WRRFs. In addition, the computational demand of CFD models is a bottleneck for their application in the real-time operation of DTs and extensive scenario analysis. It is believed that compartmental modelling based on the methodology presented in this study can open a whole new array of advantages, from more prediction power than conventional TIS models to less computational demand than pure CFD models, which can form the new basis for WRRF models implemented in the context of DT.