Sulfur Modeling: Assessing the Efficacy of Liquid Treatments for Primary Clarifier Odor Control and Their Whole Plant Impacts at WRRFs

Hydrogen sulfide (H2S) has long been recognized as a dominant odorant at water resource recovery facilities (WRRFs), with primary clarifiers identified as a main source of H2S emissions (Calvo et al., 2018). Liquid treatments such as chemical addition have been used to reduce primary clarifier emissions, but the efficacy of this approach and the impact of these chemicals on other plant processes can be uncertain without full scale piloting. Whole treatment plant sulfur modeling can provide a comprehensive planning tool to address both concerns regarding treatment efficacy and whole plant impacts. Some recent studies have explored sulfur interactions with iron and phosphorus using whole plant modeling (Johnson et al, 2019, Hauduc et al., 2017, Flores-Alsina et., 2016), but integrated sulfur models for whole plant process optimization and design are still not well used in practice. A recent project in Texas provides an example of how this tool can be used to support whole plant assessments of liquid stream odor treatments when nutrient removal is also a concern. The Trinity River Authority of Texas' (TRA's) Central Regional Wastewater System Treatment Plant (CRWSTP) is a 189 mgd annual average flow facility includes headworks, primary and secondary treatment, filtration and disinfection. The plant employs an enhanced biological removal process (EBPR) for nutrient removal. Plant solids processes will soon be changing, as TRA moves from lime stabilization to thermal hydrolysis pretreatment (THP) with anaerobic digestion (AD) a move that will impact recycle streams at the plant. Whole plant sulfur modeling to screen potential liquid stream treatments at the CRWSTP to: Determine sulfide removal efficiencies at the primary clarifiers Assess impacts at the aeration basins, with a focus on EBPR impacts Define impacts of both current and future recycle streams on liquid stream treatments and downstream impacts. Specific treatments assessed included: Pure Oxygen Ferrous Sulfate (FeSO4) Magnesium Hydroxide (Mg(OH)2) Combined FeSO4 and Mg(OH)2 This paper provides an overview of the modeling effort and its findings. Methodology All whole plant simulations were completed utilizing the Sumo2S package, version 2019 (Dynamita Sumo). The modeling approach included an influent waste characterization, historical data review, model calibration, and special sampling. Waste characterization of the influent and recycle stream were completed based on the Methods for Wastewater Characterization in Activated Sludge Modeling (Melcer, 2003). Results and Discussion Model Development and Calibration A model of the influent headworks through primary treatment was developed with the current plant configuration. This model was used for model calibration to historical and field pilot study data. Figure 1 summarizes the general model structure with multiple tanks-in-series representing physical infrastructure including screening, lift station, pipeline and grit removal. The model was then expanded to include THP and AD. The model was calibrated with previous field pilot data, with the following results : Baseline Figure 2a summarizes the model calibration with no chemical addition for dissolved sulfide control. Pure oxygen calibration Figure 2b/c summarizes simulated versus field data of both dissolved oxygen and dissolved sulfide concentrations. FeSO4 calibration Figure 2d summarizes the calibration of FeSO4 addition and dissolved sulfide concentrations. Mg(OH)2 validation Dose ratios per flow unit were determined with bench scale testing (pilot field data was unavailable). Model efforts identified the following parameters as critical for calibration: Airflow rates within individual reactors Low airflow rates were introduced within reactors to represent physical structures such as weirs and areas of turbulence. Organisms present within the raw influent Relative fractions of ordinary heterotrophic, sulfate oxidizing, and sulfate reducing organisms. Precipitation and dissolution rate of iron sulfide Calibration to field data resulted in reduction of both rates. Whole Plant Simulations Current and future recycle streams modeling scenarios are presented in Table 1 and 2 respectively. Both model results indicate that FeSO4 and the dual feed of FeSO4 and Mg(OH)2 addition provides the highest level of dissolved sulfide removal in the primary clarifiers. Results of the current recycle stream model are presented in Figure 3 and compared against a baseline condition that assumes no chemical addition. Downstream impacts of chemicals addition included the following: Volatile fatty acids (VFAs) decreased by 15 % (10 mg/L VFAs) in the primary effluent. FeSO4 addition contributed to alkalinity consumption and reduce the primary effluent pH to 6.7. Mg(OH)2 addition did not impact EBPR process. When the new CRWSTP solids treatment facilities are completed, TRA plans to add ferric chloride (FeCl3) addition to the ADs, and so the whole plant impact of recycle streams with this addition was assessed. Modeling of future solids conditions with no FeCl3 addition to the ADs show that dissolved sulfide removal efficiency decreased by 10 % due to recycle stream. The modeling revealed the following with respect to plant impacts: Recycle stream ammonium loading from future solids facilities combined with FeSO4 addition to the headworks for odor control purposes had a significant impact to the liquids stream alkalinity and pH. High phosphorous fixation occurred under all scenarios. The addition of FeCl3 to solid and/or FeSO4 to the influent forms 10 percent more precipitates within the solid's facilities due to vivianite production. The addition of Mg(OH)2 results in a negligible increase to the relative precipitate formation. Overall results are presented in Figure 4. FeSO4 addition at both the headworks (for primary clarifier control) and digesters dropped nitrification reactors pH down to 6.6. Simulations showed a gas phase H2S concentration of 2,400 ppm for baseline condition. The addition of iron to the headworks or anaerobic digestion, results in low concentration levels of H2S in the biogas. Conclusion The evaluation of odor control processes using whole plant modeling with the inclusion of sulfur interactions provides valuable insights on both the effectiveness of liquid stream treatments for odor control and the impact of those treatment processes on whole plant impacts. While all treatments studied reduced wastewater sulfides to some extent, the modeling showed that adverse process impacts could be expected for most scenarios assessed, with super oxygenation resulting in the least process impacts (and reasonable sulfide reduction in the primary clarifiers).