Tale of Two WRRFs: Optimizing Internal Carbon for Meeting Low Nutrient Limits

Introduction This investigation draws on comparative analyses from two neighboring WRRFs employing the 5-stage Bardenpho process. WSSC Water operates three other BNR facilities, encouraging sharing of valuable insights and strategies for broader use in the water resource recovery community. Accordingly, this paper delineates: 1) Operational approaches that leverage high Carbon-to-Nitrogen (C/N) ratio (~12) to enhance nutrient removal efficiency. 2) A comparative analysis of Nitrogen (N) and Phosphorus (P) removal mechanisms in two WRRFs sharing similar configurations but different operational approaches. 3) A conceptual framework to evaluate and optimize simultaneous nitrification-denitrification (SND), pre-anoxic, and post-anoxic denitrification through endogenous respiration and internally stored carbon. Methods Both WRRFs are in Montgomery County, MD, and eventually discharge into the Chesapeake Bay. The stringent nutrient limits are TN < 4 mg/L and TP < 0.27 mg/L annually. Fig. 1 shows the side-by-side comparison of BNR trains from the two facilities. Table 1 shows the key operational differences. The experimental plan included: - Reactor profiling for ammonia, orthophosphate, nitrate, and nitrite with morning and afternoon grabs; - Influent and effluent monitoring of carbon, N, and P species; - Batch testing to determine specific denitrification rates without supplemental carbon, sampling just before the post-anoxic zone; - Tracking airflow; - Collecting DNA samples for microbial analysis; and - Collecting Polyhydroxyalkanoates (PHA) and glycogen samples to understand carbon transformations in post-anoxic denitrification. Results and Discussion The Seneca WRRF has successfully implemented ammonia-based aeration control (ABAC) for about one year. Both facilities achieve excellent nutrient removal. The 90th percentile values are 2.5 mgN/L for TN and 0.18 mgP/L for TP at Seneca, while Damascus reported 2.1 mgN/L for TN and 0.13 mgP/L for TP. Notably, these outcomes are achieved without supplemental carbon. Furthermore, Seneca's ABAC system reduces aeration demand by 20% due to low dissolved oxygen (DO) operation compared to Damascus, which operates at high DO (Fig. 2). N Removal: SND represents the primary N removal mechanism, accounting for 69% of TN elimination at Seneca and 55% at Damascus (Fig. 3). Given the operational conditions-ABAC at Seneca and high DO at Damascus-the superior SND performance at Seneca aligns with expectations. Surprisingly, Damascus also exhibits substantial SND activity despite the higher DO. Both facilities demonstrate comparable pre-anoxic zone contributions to N removal. Nonetheless, Damascus depends more on post-anoxic denitrification, achieving 7.0 mgN/L vs. 3.5 mgN/L at Seneca, contributing to exceptional effluent TIN levels (Fig. 4). The post-anoxic zone makes up 20% of Damascus's total volume, double Seneca's 9% (Table 1). The maximum specific denitrification rates were also slightly higher at Damascus than at Seneca (Fig. 5). Additionally, denitrification rates in the absence of external carbon sources, measured at the end of the main aerobic zone-indicative of internal carbon storage-were higher at Damascus by ~0.3 mgN/gVSS/hr compared to Seneca. These measurements exceed the baseline endogenous denitrification rate of 0.35 mgN/gVSS/hr, suggesting a not-so-trivial role for internally stored carbon in the denitrification process. P Removal: Seneca and Damascus use both biological and chemical P removal to meet strict TP limits. Aluminum sulfate (alum) is added before the secondary clarifiers to precipitate residual P. Seneca relies more on bioP and uses half as much alum per million gallons per day (MGD) compared to Damascus (Fig. 6). Notably, recent operational shifts at both aim to reduce the alum dose (Seneca: 15 gallons of alum/MGD; Damascus: 35 gallons of alum/MGD). Enhanced P release in Seneca's pre-anoxic zone suggests more effective bioP activity than Damascus (Fig. 6). Additionally, a notable re-release of P is observed in Seneca's post-anoxic zone, likely due to a long SRT of ~25 days, which may lead to the decay of phosphorus-accumulating organisms. Despite a similar SRT, such a re-release is not evident at Damascus, potentially due to the inhibitory effects of high alum dosing on bioP processes (Fig. 7). With a strategic reduction in SRT and alum use, Seneca and Damascus are expected to rely on bioP to achieve the low TP limits. Particle Size Distribution and Settling: The extended SRT and high organic content of the influent contribute to a low anaerobic Food-to-Microorganism (F/M) ratio at both Seneca (1.8 gBOD/MLSS/d) and Damascus (1.4 gBOD/MLSS/d), which falls below the optimal ratio of 2.4 gBOD/MLSS/d recommended for favorable settling. Despite this, settling performance at both facilities remains commendably effective. Image analyses reveal a higher incidence of filamentous bacteria at Seneca than at Damascus; however, Seneca's biomass particles are larger (Fig. 8). Given that the lower SVI at Damascus than at Seneca, it can be inferred that filament abundance may significantly impact sludge settleability more than particle size.