Extensive Sludge Characterization Allows Large New Secondary Clarifiers At The Columbia Boulevard Plant in Oregon To Be Optimized Via 3-D Computational Fluid Dynamics Analysis

Introduction
Secondary treatment capacity of the 450,000 m[sup]3[/sup]/day (120 mgd) City of Portland, OR, Columbia Boulevard Wastewater Treatment Plant is limited by its eight 38.1 m (125 foot) square 'squircle' secondary clarifiers. Peak secondary treatment capacity is being expanded by adding two circular clarifiers using limited available area on the plant site. Two clarifiers with the sidewater depth of 6.5 m (21.3 ft) and the diameter of 44.2 m (145 ft) were designed to achieve surface overflow rate (SOR) up to 73.3 m/day (1,800 gpd/sf). These dimensions best fit the available area, considering subsurface conditions related to excavation and foundation design. Preliminary computational fluid dynamic (CFD) modeling identified that, due to the high SOR desired, characterization of the concentration and settling/flocculation characteristics of dispersed solids in the field was essential to properly evaluate optimizing the clarifier configuration. This differs from most secondary clarifier modeling and designs because the objective was to maximize the peak secondary treatment capacity of the expanded system. Consequently, extensive characterization of activated sludge settling characteristics with three dimensional (3D) CFD analysis was conducted to optimize design of these new clarifiers to capture improved operation capacity. Three-dimensional CFD analysis was performed to evaluate different types of energy dissipation inlets (EDIs), center wells, and peripheral baffle shapes to optimize clarifier geometry given the footprint of the clarifier (Figure 1). Further CFD analysis was conducted to understand the sensitivity of clarifier performance to a transient mixed liquor (ML) flow rate pattern during a high flow event. These results will be of interest to those seeking to optimize wet weather treatment capacity of treatment systems, especially water resource recovery facilities that are expecting increased peak wet weather flows due to climate change.
Methods
User-defined expressions were added to a commercial general-purpose CFD solver software package, ANSYS CFX, to model hydraulics and solids transport in the clarifier in 3D. The formulation of the model is generally based on the one proposed by Griborio (2004), which models discrete settling, flocculation settling, hindered settling, compression settling, shear flocculation, and differential flocculation. In the model, large, medium, and small flocs settle independently when the solids concentration is lower than the threshold concentration for the discrete settling (Figure 2). In preliminary 3D CFD simulations, the characteristics of the three floc fractions were assumed based on values from literature and previous projects that considered solids with medium to low sludge volume index (SVI). Results from the preliminary simulations indicated that the performance of the clarifier model was sensitive to the characteristics of the floc fractions. Further characterization of the flocs was necessary given the high SVI of the ML and the high clarifier SOR objective. A batch-type test for flocculent settling (e.g., Metcalf and Eddy, 2014) was performed to measure the settling velocity and mass fraction of large and medium flocs. The height and diameter of the column were 1.8 m (6.0 ft) and 0.15 m (6 in), respectively, and five sampling ports were provided on the column at one-foot intervals (Figure 3). The ML was diluted with secondary effluent to provide a solids concentration of 500 mg/L or less. The samples were extracted from the ports every three minutes and measured for total suspended solids (TSS). Testing was conducted on two different days. SVI, dispersed suspended solids (DSS), and flocculated suspended solids (FSS) tests characterized hindered settling, mass fraction of small flocs, and non-settleable solids, respectively. Floc parameters were updated in the model. Subsequent 3D CFD simulations demonstrated the sensitivity of clarifier performance to different EDIs, sizes of the center well, the shapes of peripheral baffle, and ML flow rate patterns. A number of clarifier configuration were evaluated to improve the peak capacity of the clarifier, including: - Existing clarifier EDI - Los Angeles-EDI (LA-EDI) - center well dimensions of 16 m (53 ft) diameter and the depth of 3.3 m (11 ft) - shallower center well with 2.4 m (8.0 ft) depth - smaller center well with 12 m (40 ft) diameter - McKinney baffle - Stamford baffle - McKinney baffle lengthened by 0.76 m (2.5 ft). The analysis evaluated different configurations under steady-state conditions. The best performing configuration was evaluated under the following sequential transient conditions: 3 hours at 75 percent solids flux capacity, 3 hours at 124 percent capacity, and 3 hours at 100 percent capacity.
Results and Discussion
Results from diluted flocculent settling tests showed that a larger fraction of flocs settled faster than assumed in preliminary simulations. The slopes of the contour lines between 50 and 250 mg/L TSS were similar to each other in the time history of TSS over the depth of the column (Figure 4). This means that the majority of solids settled at a similar velocity. Solids settling velocity as a function of solids removal rate was calculated from the time history of TSS (Figure 5). The distributions of settling velocity from the two tests agreed reasonably well, considering the uncertainty in the field tests. On average, 90 percent of large and medium flocs settled at 9 m/hr (5,300 gpd/sf) and the rest settled at 5 m/hr (2,900 gpd/sf). In the preliminary simulations, it was assumed that 70 percent of large and medium flocs settled at 6.5 m/hr (3,800 gpd/sf) and the rest settled at 5 m/hr (2,900 gpd/sf) based on the literature and previous projects that considered solids with medium to low SVI. This difference between assumed and measured solids characteristics agrees with general understanding of the difference between ML with lower and higher SVI. The preliminary simulations with assumed solids characteristics were overly conservative. The results from the CFD simulation showed that the optimal clarifier configuration consists of an LA-EDI, a 12 m (40 ft) diameter and 3.3 m (11 ft) deep center well, and a McKinney baffle. During the 9-hour transient simulation using the optimal clarifier geometry, the sludge blanket was below the peripheral baffle when the solids flux was at 75 percent capacity, reached the peripheral baffle at 124 percent capacity for 3 hours, and remained at the same height when the solids flux decreased to 100 percent (Figure 6).
Conclusions
Assessment of not only zone settling characteristics but also of floc size distribution and settling velocity is essential when designing secondary clarifiers for higher sustained peak overflow rates to accommodate peak flows. Field measurements coupled with 3D CFD analysis can improve secondary clarifier design and achieve improved performance at high surface overflow rates. The optimized design from this analysis maximized the capital investment with increased capacity within a limited footprint.