Optimization of intermittent aeration for increased nitrogen removal efficiency and improved settling
Fredericks, Dana Kathleen
MetadataShow full item record
Nitrogen, when present in excess, can cause eutrophication in waterways, which may result in hypoxia and the desertion or death of aquatic life. As nitrogen continues to pollute our water, wastewater discharge limits are becoming more stringent with effluent limits based on preserving receiving waters. This project took place at the Hampton Roads Sanitation District's, Chesapeake-Elizabeth Wastewater Treatment Plant; a High-Rate Activated Sludge (HRAS) plant with no primary clarifiers operating at an SRT of 1.5" 2 days without biological nitrogen removal (BNR). BNR is considered more cost-effective than comparable chemical and physical processes, but it requires considerable resources to meet increasingly strict discharge limits. As these limits decrease, the resource requirement increases, making them no longer cost-effective. By 2021 HRSD anticipates the plant will be included in a bubble permit, resulting in a total nitrogen (TN) effluent target of approximately 5-8 mg/L. Conventional BNR plants remove carbon and nitrogen simultaneously, which requires both increased volume (capital costs) and aeration energy demand (operating costs). As an alternative, HRSD is pilot testing an A/B process; a two-sludge system comprised of a carbon removal stage followed by a nitrogen removal stage. The very high rate, low dissolved oxygen (DO) A-stage could reduce the organic load, allowing the B-stage to perform BNR within the existing reactor volume and eliminating the need for primary clarifiers. However, improper control of the carbon removal system can lead to carbon and alkalinity deficiencies, which results in poor nitrogen removal. This is mediated by employing a short-cut nitrogen removal technology. A novel aeration strategy based on set-points for reactor ammonia, nitrite and nitrate concentrations with the aim of maintaining equal effluent ammonia and nitrate + nitrite (NOx) concentrations was successfully employed. The goal was to inhibit nitrite-oxidizing bacteria (NOB) so the nitrification process stopped at nitrite. This helps promote an effluent with equal parts ammonia and nitrite, which is amenable to anammox polishing to achieve low effluent nitrogen concentrations. NOB suppression has been successfully applied in sidestream anaerobic digestion waste streams because NOB out-selection is favored in warm, nitrogen-rich conditions. However, the cold, dilute conditions of continuous mainstream processes are not favorable to NOB out-selection. The mechanisms employed to achieve sidestream NOB out-selection are not reasonable for mainstream applications. This study employed operational and process control strategies to aggressively out-select NOB based on optimizing the chemical oxygen demand (COD) input, imposing transient anoxia, aggressive solids retention time (SRT) operation approaching ammonia oxidizing bacteria (AOB) washout, and a dissolved oxygen concentration (DO) of 1.5 mg O2/L during aeration. This pilot-scale study demonstrated that when run aggressively, the proposed online aeration control is able to out-select NOB in mainstream conditions and provide relatively high nitrogen removal without supplemental carbon and alkalinity at a low hydraulic retention time (HRT). Successful full-scale implementation would promote improved water quality that is economically sustainable. The ability of two different process configurations (full intermittent aeration and Modified Ludzak-Ettinger [MLE]) to achieve high nitrite accumulation and nitrogen removal efficiencies in four equal volume tanks in series followed by a cone-bottom clarifier in a pilot scale biological nitrogen removal (BNR) process (V=0.61 m3) was evaluated. All four biological reactors were equipped with a variable speed mixer, a 17.7 cm membrane disc diffuser, and a Hach LDO probe. Aeration capacity in all four tanks allowed the system to be operated with or without a defined anoxic zone. Both processes utilized a novel aeration strategy based on set-points for reactor ammonia, nitrite and nitrate concentrations with the aim of maintaining equal effluent ammonia and NOx concentrations. The B-stage had a variable HRT (2-7 hours) and a variable influent flow rate. When operating in the MLE configuration, an internal mixed liquor recycle (IMLR) line returned nitrified mixed liquor from the last aerobic reactor to the anoxic reactor using a peristaltic pump at a rate between 200-450% of the influent flow. When IMLR was used the first tank was not aerated. RAS from the clarifier was returned to the anoxic zone at 100% of the influent flow. SRT was controlled by wasting solids from the last aerobic tank. The wasting was automated to maintain desired SRT. The nitrite accumulation ratio (NAR), NO2- -N/(NO2- -N+ NO3- -N), was best under full intermittent aeration, achieving 0.43+0.10 at a 3 hour HRT and influent carbon to ammonia ratio (COD/NH4+-N) of 7.9+1.4. As an MLE, the NAR decreased with increasing internal mixed liquor return (IMLR); at IMLR of 200%, 325% and 450%, the NAR was 0.20+0.04, 0.17+0 and 0.14+0.03, respectively. The MLE did, however, improve the overall TIN removal efficiency compared to operation where all reactors were intermittently aerated. The TIN removal efficiency was best under MLE operation, increasing as the IMLR and influent COD/NH4+-N increased. When the IMLR was 200%, 325% and 450%, the TIN removal efficiencies were 76.4+4.0%, 80.2+0% and 86.3+5.0%, respectively, which corresponded to an influent COD/NH4+-N and HRT of 9.2+0.8 and 4 hr, 9.8+0.4 and 6 hr, and 10.3+1.2 and 6 hr, respectively. In addition to process operation, key issues of filamentous bulking were assessed. Concrete solutions to this continual issue are not available as the unique features of each plants influent and process dynamics prohibit the formulation of a universal solution. Filaments observed throughout this study included Type 0041, Type 0675, Type 0803, Nocardia, Thiothrix I and Thiothrix II. Type 0041 and Type 067 were observed throughout the study and are typical of BNR systems; they arguably do not contribute to settling issues. Type 0803 filaments are linked to low F/M, high SRT systems. It was present at the start of the experiment and then no longer detected. Nocardia made a brief appearance on day 72 causing temporary foaming issues. This was fixed by vacuuming the surface of the clarifier daily and may be attributed to the high surface area to volume ratio present in pilot-scale systems. Thiothrix I and Thiothrix II were observed after day 93, however, never as the dominant species. Thiothrix related bulking was observed in the A-stage (Miller et al, 2012), which was attributed to high sulfide and organic acids in the influent raw wastewater during high temperature periods and carryover of sulfide and Thiothrix from the over-sized A-stage clarifier. The goals of this evaluation were to identify favorable parameters of common filaments and establish their impacts on the system. Typically an SVI of 150 mL/g indicates good settling. Overall the study experienced good settling (128.3+36.3 mL/g), indicating that operating under different influent substrate concentrations and process configurations did not result in poor settling.
- Masters Theses