Degrémont is a global leader in the design, build and operation of water and wastewater treatment plants all over the world. These plants employ a number of components standardized by Degrémont but since each plant is unique, bespoke design is necessary in both new-builds and retrofits when it comes to the large treatment structures.
Together, all of the components that form a treatment plant intend to change the physical and chemical characteristics of the water to suit the delivery needs (potabilisation and/or pollutants removal by filters, settlers, biological aerated tanks, to name a few). In physical terms, for example, the aforementioned steps equate to free-surface, two-phase, transient, fluidised bed, porous medium, chemical mixing and reactors, flow control and repartitions. The analysis and control of the flow fields is thus critical to ensure a robust and economic operation of these plants. The large-scale nature of these components further contributes to the challenge.
Usually, when correlation data or reduced model analysis is not enough, either experimental or computational analysis have to be undertaken. Experimental analysis are sometimes too expensive in regards of the stakes, or time consuming, while computational analysis are generally more flexible but less convincing in the eyes of the stakeholders. Nevertheless, the numerical analysis presents itself as a suitable methodology for the reduction of uncertainty when it comes to the detailed studies.
In this paper we present a range of examples where CFD was used to ensure such a goal: Circular and non-Circular Distribution Chamber cases, Channel Distribution Case, Sector Control Valve Case, UV Distribution Case and a Head Loss Determination Case. We comment on the process that was required to perform the simulation and on the accuracy of final results. Furthermore, for the Sector Control Valve Case, we show how the CFD results were used to enable a critical component in the definition of the control automation of the overall plant, and for the Non-Circular Distribution Case, we show the cross validation with the existing reduced models. We finish with expected new application areas and desired modeling capabilities.