Mainstream Partial Denitrification-Anammox Application With Raw Fermentate: Concept Development For Blue Plains

Abstract
This paper discusses specific application needs for implementation of partial denitrification anammox (PdNA) technology when using fermentate as a carbon source. The drivers and practical implementation guidelines are discussed in detail.
Learning Objective: Understanding the specific application needs for implementation of partial denitrification anammox technology when using fermentate as a carbon source. Introduction Mainstream partial denitrification – anammox (PdNA) is rapidly being implemented in full-scale applications (Klaus et al., 2021; Wang et al., 2021). This short-cut nitrogen removal process has proven to be simple and reliable over a broad range of conditions and carbon sources (Le et al., 2019a, b; Campolong et al., 2019; Cao et al., 2013; Du et al., 2017). As a matter of fact, previous studies have successfully demonstrated its application with recovered carbon, i.e. fermentate (Ali et al., 2020; Forney et al.,2020). Although, the PdN efficiencies associated with fermentate (30%-70%) are lower than those observed with acetate and glycerol (80%-100%), about 17-27% of incoming nitrogen was removed via the PdNA pathway. Thus, every nitrate going through the PdNA route, results in the carbon-free removal of an additional nitrogen in the form of ammonium, and thus a decrease in external carbon needs. The objective of this study was to assess and translate learnings from previous pilot work into a viable concept for Blue Plains WWRF by focusing on determining control levels and setpoints needed, achievable MeOH savings and impact carbon footprint.
Methodology A 360 L mainstream nitrogen removal pilot system was operated at Blue Plains Advanced Wastewater Treatment Plant of DC Water, as shown in Figure 1 (Le et al., 2019, Ali et al, 2020). Fermentation occurred accidentally in the full-scale primary sludge gravity thickeners without full conversion to optimized fermenters and was collected at the centrifuges before thermal hydrolysis.
Results and Discussion A new pilot run, adjusted based on previous learnings, was operated for 13 weeks so far and this paper shows the main findings on conditions needed to enhance PdN selection and PdNA rates. Additional data collected will be added to the final paper. These findings inform implementation approach for Blue Plains as well as potential impact on MeOH savings and carbon footprint. Guidelines for PdN selection using fermentate as carbon source Previous studies have emphasized that carbon sources, such as acetate and glycerol are associated with higher PdN efficiencies (80%-100%) compared to fermentate (30%-70%). (Le et al., 2019, Campolong et al., 2019, Ali et al 2020, Ladipo-Obasa et al, In review). The lower PdN efficiencies for fermentate lead to lower PdNA contributions to TN removal (17-27%) and thus less potential methanol savings and/or capacity enhancements. However, batch tests have demonstrated that fermentate could attain PdN efficiencies of up to 93% (Ali et al., 2020). This increased potential in PdN efficiency is dependent on fermentate's composition and time of the year, especially at Blue Plains, where reliance on accidental fermentation in gravity thickeners might be limited. This new pilot run explored enhancement of PdN selection by increasing nitrate residual. The nitrate residual concentration was identified before as a driver for PdN selection (Le et al., 2019) and increased residual in previous pilot runs seemed to overcome composition limitation of fermentate and avoid loss of PdN selection (Ladipo-Obasa et al., In Review). Figure 2A shows the impact of nitrate residual on the PdN efficiency achieved when using fermentate with variable soluble COD (500-3500 mg COD/L). The fermentate composition differences were equally spread over the nitrate residuals tested and did not directly impact PdN selection (Figure 2B). Thus, it was concluded that when operating with a nitrate residual concentration above 4 mg NO3-N/L, higher PdN efficiencies could be achieved. Moreover, PdN selection when using fermentate was not limited to the previously reported 50-60% PdN efficiency but could attain similar levels as acetate or glycerol examples. PdNA rate followed the same trends and were thus driven by PdN efficiency (Figure 2C). However, PdNA rates did become limited when nitrate removal rates were too low due to nitrate limitations (Figure 2D). Lastly, ammonium levels up to 0.5-1 mg N/L did not limit PdNA rates and thus ammonium could be driven low enough as long as PdN efficiency is high. Nitrite accumulation was not observed in any of these cases and anammox capacity was at least 2 times higher than the operational rates. This high AnAOB capacity can explain why one can operate with very low nitrite (< 0.5 mg N/L) and low ammonium concentrations. Implementation of PdNA within existing treatment schemes WRRFs are legally required to meet effluent limits. Therefore, operation with fermentate at higher nitrate residual concentrations would impact the placement of the PdNA zone. For plants with stringent effluent limits, such as Blue Plains, the promising configuration would involve integrating PdNA upstream of polishing aeration and anoxic zones (Figure 1). There will be a clear need for full denitrification polishing to decrease nitrate further after the PdNA zone. With the current guideline of a nitrate residual concentration of 4-5 mg N/L and PdN efficiencies between 60-80%, desired ammonium setpoints for upstream ammonium-based aeration control were estimated to be 3-9 mg N/L to achieve a minimum ammonium of 0.5-1 mg N/L at the end of the PdNA zone. This would result in 35-70% of TN removal driven by PdNA and MeOH saving for Blue Plains were estimated at 58-100%. About 56-82% of those MeOH savings is directly driven by PdNA, the other fraction is driven by full denitrification using fermentate instead of MeOH. The latter forms a clear driver for full-scale implementation. The implementation of PdNA in the nitrification zones (Figure 1) would directly make use of the decreased need for nitrification volume as ammonium is treated through the PdNA pathway. In addition, from a sustainability standpoint, assuming PdNA does not result in significant changes in nitrous oxide emissions (given nitrite does not accumulate), a decrease of our direct CO2 emission of 10-13% was estimated. Currently, the proposed ABAC and PdN control targets are implemented in the pilot. Data will be collected through the winter to provide a basis for a business case proposal for implementation at Blue Plains WWRF. The full paper will include the final results of the proposed concept and will detail how we envision implementation will occur full-scale.
Conclusion This paper provided guidelines for upstream aeration and PdN control setpoint needed to maximize PdNA success when using fermentate as carbon source. This study served as a direct simulation for implementation at the Blue Plains WWRF.