Enhanced process models for final settlement tanks
Burt, D.1, Wimshurst, A.1, Jarvis, S.2, 1Frazer-Nash Consultancy Ltd, 2Thames Water, UK
(free)
The use of process models to provide accurate prediction for waste water plant performance remains reliant on the definition of empirical constitutive relationships for sludge settlement velocity, sludge rheology and floc aggregation dynamics. Recent work to create a Generalised Settling Model (GSM) for Final Settlement Tanks (FSTs) acknowledges that only the hindered regime is well understood and has focussed on improving compression settling for a more complete representation of settlement in the lower regions of the tank at high sludge concentrations.
This model has recently been applied to the Riverside STW Final Settlement Tanks which feature a flat floor and a TowBro sludge removal mechanism. A large range of state points were investigated with both Mass Flux theory and the process model (implemented in the CFD code OpenFOAM). Comparisons were made for the current site configuration and improved influent options with particular attention given to sludge age and resilience to partial blockage of the TowBro. Sludge line settling tests were used to characterise both the hindered and compressive settling of activated sludge in order to create the final settling velocity model. When compressive settling was accounted for in the CFD model, the sludge concentration increased throughout the bulk of the sludge blanket, thus increasing the sludge blanket depth.
The CFD results for effluent solids concentration and sludge blanket depth, obtained for each state point, were plotted as a response surface. The use of a response surface has many parallels in other industries where it is desired to collapse the complex dynamic performance for a piece of process equipment into a single performance function. Although FSTs have 4 key process variables (SSVI, MLSS, Q and R), it was found the continuous surfaces could be produced as a function of 3 variables with Q/R made a single variable valid in the range of R = 0.5 to 1.0. Collating results in this way allows for future inspection and interpolation of all state points from a limited computed data set (typically 64 data points).
Keywords: Final Settlement Tank, Clarifier, Hindered and Compressive Settling, Generalised Settling Model
