Combining Metabolic, Kinetic and Physical Selection to Achieve Full-Scale Continuous Flow Densified Activated Sludge and Nutrient Removal at Robert W. Hite Treatment Facility

Introduction: Metro Water Recovery's (Metro) Robert W. Hite Treatment Facility (Hite) in Denver, CO is permitted for a maximum month (MM) flow of 220 mgd and has two parallel liquids treatment trains including the North Secondary (NSEC). The NSEC utilizes a modified Ludzack Ettinger (MLE) process with a sidestream anaerobic reactor (SAR) to achieve nutrient removal. In 2018, Metro's facility plan identified a need for increased aerobic solids retention time (aSRT) to meet future potential total nitrogen (TN) and total phosphorus (TP) regulations (Colorado Regulation 31). To avoid costly expansion, Metro commissioned a full-scale demonstration train (NAB2) to trial densified activated sludge (DAS). This work documents findings from the full-scale demonstration facility and discusses how Metro is translating DAS into a full-scale solution at the RWHTF. Specifically, this paper will discuss factors that have led to the development and long-term operation of DAS with biological and physical selection, as well as provide insights into how particle size distribution (PSD) may be impacted by operational parameters and influence nutrient removal. Materials and Methods: DAS facility: The DAS demonstration facility (denoted as NAB2) consisted of an isolated, paired aeration basin and secondary clarifier from NSEC. Description of the facility is provided elsewhere (Avila et al., 2021). Metro operated NAB2 to test multiple distinct metabolic and kinetic selector configurations, with two primary configurations, Fig 1. A 'control' train was also operated without physical and biological selection. Routine influent/eff concentrations of organics, solids and nutrient content were determined using Standard Methods for the Examination of Water and Wastewater. PSD of the MLSS, hydrocyclone UF, and hydrocyclone OF were routinely characterized. Settling Column Testing: Settling column testing of bulk DAS ML was conducted on multiple occasions to determine the Vesilind equation settling coefficients (V0 & k) as well as the PSD. Additionally, settling properties for particle size fractions: <200µm, 200-600µm, and >600µm were characterized across a concentration range of approximately 1,500 mg/L to 6,000 mg/L. Activity Testing: Batch scale activity testing of bulk mixed liquor and three particle size fractions was used to evaluate nutrient removal rates of activated sludge in 5-gallon reactors. Analytes (NH4, NO3, NO2, and PO4) were measured with HACH TNT kits. Hydrocyclone Nozzle Testing: Hydrocyclone UF nozzle sizes ranging from 15 to 25 mm were individually tested at a target pressure of 30 psi. Flow measurements from the hydrocyclone feed and OF were recorded via flow meter. Total suspended solids (TSS) and PSD samples were collected from the RAS, OF, and UF. Process and Capacity Modeling - Data from NAB2 testing were used to develop, calibrate, and validate AB2 specific and a whole plant process model in BioWinTM and computational fluid dynamics (CFD) models of the NSEC secondary clarifiers. These models establish system capacity of existing infrastructure and design criteria for full-scale implementation. Results and Discussion: Impact of biological selection on densification: PSD of NAB2 biomass during configuration 1 (SAR-MLE) mode) stabilized into a uni-modal distribution, with median size increasing from 100 to 200 um respectively (Fig 2A). The average SVI30 for the NAB2 versus control basins were 102 mL/g and 153 mL/g respectively. PSD of NAB2 biomass during configuration 2 (SAR-A2O) mode reflected a bi-model distribution (Fig 2B). The average SVI30 for the test versus control basins were 81 mL/g and 173 mL/g respectively. Cumulatively, results suggest that improved settleability is observed when the mass fraction of granules (particles >212 um) in the sludge exceeds 15%. This was observed on multiple occasions when NAB2 operated in SAR-A2O mode with anaerobic food to microorganism ratio (F/M) being sustained between 1.0 and 3.0 lb sCOD/lb MLVSS-day. Impacts on PSD on Settling Performance and Biomass Activity: Settling velocity and compressibility were observed to increase as the particle size fraction increased (Table 1). Biomass activity results indicated that larger particle size fractions had lower denitrification and nitrification rates versus smaller particle size fractions while Bio-P activity was highest in particles greater than 200 um. These results indicate that minimizing particles sizes greater than 600 um may allow Metro to exploit high nitrification/denitrification/Bio-P rates and with superior settling and compressibility (Quoc et al., 2021). Influence of Physical Selective Pressures on PSD in NAB2 (Fig 3): Results indicated that mass flow through the UF could be reduced when the nozzle size decreases (Fig 2). This phenomenon can potentially be exploited to craft a wasting strategy focused on developing and keeping a target PSD. DAS impacts on Capacity: Stress testing indicated that NAB2 could be loaded at surface overflow rates (SORs) >700 gpd/sf and solids loading rates (SLRs) >60 pounds per day per square foot (ppd/sf) while still maintaining reliable blanket control and stable effluent TSS (Fig 4). Whole plant process and CFD modeling indicated that the densified scenario could allow Metro to successfully treat 2040 flows without needing to invest in additional aeration basins or secondary clarifiers, allowing for savings of $50 million dollars.