Comparison between CFD and Surface Overflow Rate Models to Predict Particulate Matter Separation in Unit Operations for Combined Sewer Overflows

Abstract

Modeling particulate matter (PM) separation has been a significant topic since unit operations (UOs) has been introduced for clarification of combined sewer overflows (CSOs). In contrast with surface overflow rate (SOR) model, long debated for the limits of its applicability, computational fluid dynamics (CFD) has been an emerging tool for accurately predicting hydrodynamics and PM transport in UOs. In this study an Eulerian-Lagrangian CFD model and a SOR model coupled with particle size distribution (PSD) are used to predict PM separation from three Type I settling UOs with same surface area (SA) but different inlet locations. UOs are loaded with hydraulic conditions and PM gradation measured from samples collected in the combined sewer system of the city of Cosenza, Italy. Specifically, the samples characterize the influent flow entering into a detention tank located downstream of the combined sewer system. The detention tank is used for treating the combined sewer by gravitational settling. The two models predict the overall PM separation efficiency and the separated PM percentage as a function of flow rate and particle size for three different UO configurations. The difference between results from the two models increases as a function of flow rate (differences ranging from about 5 up to 23%) for all three UO configurations. This is probably because of the fact the SOR does not account for the mixing phenomena because of turbulence generated by high hydraulic loadings. Under those conditions, CFD model, properly validated with field measurements, may represent a more accurate predictive tool for design of stormwater settling tanks (STs) than a SOR model.