A series of facultative lagoons operated by Thames Water treating industrial wastewater in Thailand were found to be performing poorly, particularly with respect to the removal of biological oxygen demand (BOD). A review of the design parameters for the site found that all the lagoons are of a sufficient area for the flow and BOD load. However, observations of the lagoons suggested that there may be significant hydraulic short-circuiting. Computational fluid dynamics (CFD) modelling was therefore carried out on one of the lagoons to establish the hydraulic regime. Two consecutive simulations were carried out, both with and without baffles; the first to establish steady flow conditions, and the second using a chemical species transport model to obtain the residence time distribution (RTD). The results of the modelling indicate that the lagoons do currently suffer from significant short-circuiting, and large dead-zones are present. The installation of baffles in the CFD model improved the plug-flow characteristics of the lagoons, substantially reducing the short-circuiting and the size of the dead-zones. It has therefore been concluded that the installation of baffles in the lagoons will lead to an improvement in their performance, by increasing the retention time of the system.

A Eulerian-Eulerian multiphase CFD model is employed for the air/water flow. A 3D structure grid is used to incorporate the air nozzle and tank geometry. The fixed frictionless wall boundary approximating the free surface acts as a sink to allow the air bubbles to escape. The air/water volume fraction in the flotation tank is evaluated to determine the effective air/water fluid density. The floc particle is then introduced and is tracked in the air/water fluid using a disperse Lagrangean model. Fate of these flocs depends on their sizes and density. Flocs therefore can either escape through the top water surface, settles in the main tank or breakthrough under the outlet weir. The CFD model is developed for a full scale DAF tank to predict the flow dynamic, particle removal and settled solid profile. The general flow pattern is compared with flow visualisation using the underwater camera. Comparison of average fluid velocities is carried out using acoustic Doppler velocimetry ADV measurement.

A laboratory model is used to measure the generic flow patterns in dissolved air flotation (DAF). The Perspex model used in this study allows the use of laser Doppler velocimetry (LDV), a non-invasive, high-resolution (+/- 2 mm s-1) laser technique of flow velocity measurement. Measurement of flow velocity in the single-phase situation was first carried out. Air-saturated water was then supplied to the tank and measurements of bubble velocity in the two-phase system were made. Vertical flow re-circulation was observed in the flotation zone. In the bottom of the flotation zone (near the riser) secondary flow re-circulation was observed, but only in the two-phase system. Another phenomenon was the apparent movement of flow across the tank width, which may be due to lateral dispersion of the bubble cloud. Data from preliminary computational fluid dynamics (CFD) models were compared against this measured data in the case of the single-phase system. The CFD model incorporating a k-epsilon model of turbulence was found to give closer agreement with the measured data than the corresponding laminar flow model. The measured velocity data will be used to verify two-phase computational fluid dynamics (CFD) models of DAF.

Over the past fifty years, considerable research in waste stabilization pond operation has led to the development of a number of models used to describe the hydraulic regime and predict treatment efficiency. Models range in complexity from plug or completely mixed simplifications to computational fluid dynamics (CFD) models which are able to predict flow hydraulics at a local level. Information about the exit age of pond effluent can be used to estimate pollutant decay. However, a mechanistic approach to understanding pond operation highlights the importance of knowing both the time and spatial history of pond effluent. A CFD model of a large pond system was constructed to demonstrate various hydraulic scenarios under different boundary conditions. Two scenarios were compared to visually demonstrate the effects of differing hydraulic conditions. Typical mechanistic models were applied to each condition to quantify biological differences. This simple example indicates that integrating biological and localised flow models will lead to a more holistic understanding of pond operation and treatment efficiency.