CFD and physical models of PM separation for urban drainage hydrodynamic unit operations. (1st May 2019)

• Record Type: Journal Article
• Authors: Liu, Hua; Sansalone, John
• Is Part Of: Water research. Volume 154(2019)
• Page Start: 258
• Page End: 266
• Publication Date: 2019-05-01
• Subjects: Best management practice -- Computational fluid dynamics -- Stormwater -- Particle size distribution; Water -- Pollution -- Research -- Periodicals; 363.7394
• DOI: 10.1016/j.watres.2019.01.057
• Languages: English
• ISSNs: 0043-1354

Abstract

Abstract: A hydrodynamic separator (HS) is a common pre-treatment unit operation (UO) for separating coarser particulate matter (PM). While HS designs have disparate internal geometries, surface areas, and volumes; for a given PM granulometry this study hypothesizes there are limits to PM separation. In this study, a large group of full-scale HS units were tested with physical and computational fluid dynamics (CFD) models to evaluate HS functionality based on surface overflow rate (SOR), HS geometrics (diameter, height, baffling), PM granulometry, particle size distribution (PSD), turnover volume, hydraulic capacity, and short-circuiting. Models were loaded with a hetero-disperse PSD, a common testing metric. PM separation results show that PSD variability greater than ±2% from this metric are different (α = 0.05). Comparing PM separation using the PSD metric across the SOR range: (1) all HS results were below the ideal SOR model, (2) above Hazen's tanks-in-series model with N = 1, (3) above plain tank designs, and (4) declined logarithmically in a ±10% range. Integrated across a triangular loadgraph, the SOR model significantly over-estimated PM separation for all HS units. The SOR model deviated from the validated CFD model, as PSD loadings became increasingly mono-disperse. A validated CFD model is shown to be a valuable tool to examine HS design and functionality. Graphical abstract: Image 1 Highlights: Physical and CFD modeling of hydrodynamic separators (HS) was conducted. Surface overflow rate (SOR), Hazen, CFD were used to model physical data. All HS units generated a narrow PM separation band as a function of SOR. CFD and physical models matched and were below ideal SOR limit. Particle size distribution selection and deviation thereof were critical factors.