Lee, Chengpeng

Ph.D. candidate in Environmental Engineering at Northwestern University with a focus on sustainable nitrogen removal and energy-efficient wastewater...

Titles from this speaker

Description: Addressing Controllability Challenges in Aeration Systems: Insights from Blue...

Addressing Controllability Challenges in Aeration Systems: Insights from Blue Plains

Abstract

Introduction
Advanced aeration control strategies, including DO control, ABAC, and AvN control, are critical for optimizing biological nutrient removal systems1. Despite their potential, confidence in these systems is often lost due to limitations and complexity2. Nearly 50% of wastewater facilities still operate in manual mode, bypassing basic PID control loops3. This highlights the need for online evaluation tools to identify issues related to tuning, probe maintenance, or control logic performance. This study evaluates DO control, ABAC with black-box and mechanistic feedforward approaches, and introduces Nitrate-Based Aeration Control (NBAC), aiming to establish a framework for real-time assessment and adjustment of aeration controls.

Methodology
Blue Plains operates 12 nitrification reactors with a daily flow of 300 MGD. Odd-side reactors use DO control, while Even-side reactors use ABAC (Fig.1). ABAC and DO control were evaluated over 267 days and 153 days, respectively. Weekly sampling assessed probe reliability and maintenance demands. Control performance was evaluated using setpoint accuracy, air valve limitations, and laboratory corrections. Setpoint accuracy measured the percentage of time setpoints were met within error thresholds. Physical limitations were identified by valve positions outside controllable ranges, and probe offsets were analyzed through laboratory correction. An integrated scoring metric reflected the percentage of time each criterion was met.

Results and Discussion
1.Overall Performance of the Control Systems
ABAC, operating with a setpoint of 2.0 mgN/L for NH4, demonstrated better performance compared to DO control, maintaining error margins within 30% thresholds (Fig.2). These margins were calculated as probe readings within 30% of the setpoint, detailed in Fig.2. In contrast, DO control systems, with setpoints ranging from 1.5 to 2.5 mgO2</Sub>/L, exhibited variability across zones, except for Zone 2B (Fig.2). Although NH4 concentrations remained below 0.5 mgN/L following DO control, the observed inconsistency suggests reliance on airflow and air valve control may not achieve precise control under dynamic conditions. Notably, Zone 2B displayed enhanced controllability following air valve maintenance in mid-Sep., highlighting the role of equipment upkeep in ensuring system performance.

2.Control System Performance Within Operational Limits
Physical limitations play a crucial role in achieving system performance. Positions below the controllable range are uncontrollable due to insufficient airflow, while those above the inflection point yield diminishing returns, reducing control precision. The DO control system demonstrated a lower percentage of time within the controllable range compared to ABAC, with Zones 1A and 2A exceeding the upper limits 77.8% and 63.2% of the time, indicating a reliance on fully open valves to meet airflow demands (Table1). In contrast, Zone 2B in the DO control system achieved setpoints more effectively when valve positions remained within the controllable range, as evidenced by the green data points in Fig.2D. The ABAC system outperformed DO control overall (Table1); however, instances where valve positions exceeded the upper limit (Fig.2A) hindered ABAC's ability to meet NH4 setpoints.

3.Control performance Considering Lab Correction
The reliability and maintenance demands of NH4 probes are critical for achieving setpoint accuracy in ABAC. Over the 267-day, ABAC required significant maintenance, including 14 matrix adjustments and electrode replacements (Fig.3A). While ABAC generally outperformed DO control in meeting setpoints, lab corrections revealed fluctuations due to probe inaccuracies (Fig.2A&3B). Scoring metrics showed 74% accuracy within a 10% error margin, rising to 82% within controllable airflow limits but falling to 29% with probe offsets (Fig.4). In contrast, DO control showed lower setpoint achievement due to physical limitations (Fig.4). Despite these constraints, DO probes demonstrated higher accuracy compared to NH4probes, with potential accuracy improvements from 24% to 72% within a 10% margin if physical limitations are resolved. Integrated scoring tools highlighted discrepancies between hourly and daily ABAC data, mainly driven by secondary effluent flow rates and air valve control (Fig.4&5). Fig. 5 illustrates Black Box ABAC control performance at Blue Plains, where valve positions were well-controlled between 0 to 11 AM, adjusting to NH4 loading. However, during high flows, the system often exceeded its capacity to open valves further, resulting in compromised setpoints. Comparisons between hourly and daily results in ABAC indicate that hourly real-time data reveal the need for improvements to handle flow rate dynamics (Fig.4A&B).

To address these challenges, novel aeration controls have been proposed. Mechanistic Feedforward ABAC integrates predictive modeling with feedback controls to anticipate system behaviors and adjust aeration proactively, reducing reliance on error-prone probe feedback5. Additionally, due to the unreliability of NH4 probes, a novel NBAC has been developed, incorporating a high-accuracy Wet Chemical Analyzer for NH4 feedforward signals and a UV NO3probe for feedback.

Conclusion
This study evaluates controllability to identify key constraints in aeration control and proposes strategies and solutions. The full paper will present proof-of-principle NBAC.

This paper was presented at WEFTEC 2025, held September 27-October 1, 2025 in Chicago, Illinois.

Presentation time

09:30:00

09:45:00

Session time

08:30:00

10:00:00

SessionOvercoming Challenges to Implement Advanced Aeration

Session locationMcCormick Place, Chicago, Illinois, USA

TopicLiquid Stream Treatment - Nutrient Removal and Recovery

Author(s)

Lee, Chengpeng, Hatcher, Jacob, Ngo, Khoa Nam, Fofana, Rahil, Wells, George, De Clippeleir, Haydee

Author(s)C. Lee1, J. Hatcher2, K. Ngo3, R. Fofana3, G. Wells1, H. De Clippeleir3

Author affiliation(s)Northwestern Univeristy1, The George Washington University2, DC Water and Sewer Authority3

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Sep 2025

DOI10.2175/193864718825160023

Volume / Issue

Content sourceWEFTEC

Word count11

An Evaluation of Dual External Carbon Source Strategies for the Full Denitrification Process

Abstract

Introduction Achieving low TIN levels in wastewater effluent often necessitates the addition of an electron donor (Fu et al., 2022). The selection of an appropriate carbon source is a critical decision, influenced by considerations of cost, availability, and effectiveness in terms of biomass yield (Peng et al., 2007). In the United States, methanol is a preferred choice due to its economic viability and low yield (WRF, 2019). However, recent literature suggests fermentate as a potential alternative (Ladipo-Obasa et al., 2022). One limitation of fermentate is its contribution of ammonium to the effluent. This study proposes an intermediate strategy wherein a portion of methanol is replaced with fermentate, aimed at reducing operational costs while easing the transition towards PdNA methodologies. The primary goal is to elucidate the denitrification stoichiometries under a dual carbon source, emphasising is placed on formulating operational strategies conducive to achieving desired effluent standards, along with an evaluation of reductions in methanol consumption. Methodologies A 360 L mainstream pilot system was operated for three phases. During the initial phase, methanol was utilized as the sole carbon source, with its dosing governed by a feedforward control. In the next phase, fermentate, derived from batch fermentation of primary sludge from full-scale primary clarifiers. The third phase introduced a dual carbon source strategy to facilitate a FdN process. Fermentate was dosed continuously, while methanol, serving as an additional carbon source, compensated for the remaining chemical oxygen demand necessary. Results and Discussion Phase 1: Baseline condition with methanol as external carbon source Over 153 days in this pilot operation the effluent targets were effectively met (1.91 mg N/L,Table 1). A stoichiometry factor of 4.9 g sCOD added/g NO3-N removed and a yield coefficient of 0.42 g COD/g COD were observed (Fig.1a,c); these are similar to expected literature values for methanol-driven denitrification (Mokhayeri et al., 2009). Phase 2: Nitrate removal with available primary sludge fermentate Fermentate dosed was of high quality and reached a yield of 0.21 g sCOD/g VSS (Table 2), which was on the higher end of previously reported yields (Ali et al., 2021). Throughout the operation of 125 days, a nitrate removal of 2.6 mg N/L was achieved (Table 1). The stoichiometry and yield coefficients of fermentate, recorded at 5.4 g sCOD added/g NO3-N removed and 0.47 g COD/ g COD, respectively (Fig.1a,c). Phase 3: Denitrification with dual methanol and fermentate addition a.Overall performance and impact on overall methanol needs Over a 137-day period utilizing dual external carbon sources, a feedforward control strategy was effective at achieving the average effluent TIN at only 1.91 mgN/L. (Table1). Despite significant variations in fermentate quality, particularly in terms of soluble COD due to primary sludge solids concentration variability the strategic incorporation of methanol to address these variations ensured the attainment of favorable TIN levels in the effluent (Table 1&2). It was observed that employing both carbon sources concurrently did not modify their individual behaviour. This was evident by the predicted nitrate removal using the stoichiometries from phase 1 and 2 being similar to the observed nitrate removal (Fig.1a). Even though the predicted nitrate removal showed a slightly improved efficiency of carbon use in phase 3, this was not statistically significant, as shown in (Fig.1b). The amount of external MeOH needed in this phase was 3.2+/-1.4 g sCOD added/gNO3-N removed or showed a MeOH saving of 35% compared to phase 1. b.Microbial selection and functionality under dual substrate addition Even though overall N removal was as expected under dual substrate dosing, a change in kinetic behavior was observed (Fig.2). Nitrite accumulation was observed in denitrification activity tests, irrespective of whether fermentate or methanol was the carbon source. Nitrite accumulation was never observed when doing activity tests with full-scale MeOH adapted biomass or during phase 1 (Fig.2a,d,e). The nitrite accumulation behaviour was also confirmed in the reactor by profiling tests (Fig. 3). Notably, the PdN efficiency was higher with fermentate at 76% and methanol at 52 % (Fig. 2b,c), mirroring findings from the pilot profiling experiment, which reported a PdN efficiency of 47%, as shown in (Fig.3). This result was surprising as methylotrophs do not tend to accumulate nitrite easily. The extended 262-day period of employing fermentate as a carbon source in the reactor potentially enhanced the diversity of nirS type denitrifiers, consequently altering the community-level dynamics in response to different electron donors (Hallin et al., 2006). Nitrate levels needed to be pushed below 1.97 mg N/L to avoid nitrite coming out of the denitrification zone. Conclusions This study at the Blue Plains AWTP underscores the efficacy of dual carbon sources-methanol and fermentate-in enhancing TIN removal in wastewater treatment. Employing a sequential methodology, the research highlighted the stoichiometric and operational efficiency of combining these carbon sources. These findings provide a strategic pathway for the implementation PdNA processes, signifying a substantial advancement in cost-effective and efficient nitrogen removal strategies.

This study assesses the effectiveness of using methanol and fermentate as dual carbon sources for Total Inorganic Nitrogen (TIN) removal in wastewater treatment. Findings show that combining these sources enhances nitrate removal efficiency and reduces methanol consumption, facilitating the transition towards Partial Denitrification- Anammox (PdNA) methodologies.

SpeakerLee, Chengpeng

Presentation time

15:30:00

16:00:00

Session time

15:30:00

17:00:00

SessionEnhancing Nitrogen Removal: Insights Into Carbon Sources and Mechanisms

Session number420

Session locationRoom 340

TopicAdvanced Level, Facility Operations and Maintenance, Municipal Wastewater Treatment Design, Nutrients, Research and Innovation

Author(s)

Lee, Chengpeng, Ngo, Nam, Islam, M.A. Sadikul, Hatcher, Jacob, Riffat, Rumana, Azam, Hossain, Wells, George, De Clippeleir, Haydee

Author(s)C. Lee1, N. Ngo2, M. Islam3, J. Hatcher4, R. Riffat5, H.M. Azam6, G. Wells7, H. De Clippeleir8

Author affiliation(s)1Northwestern University, VA, 2DC Water, VA, 3University of the District of Columbia, 4The George Washington University, MD, 5George Washington University, VA, 6University of the District of Columbia, DC, 7, IL, 8DC Water & Sewer Authority, VA

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2024

DOI10.2175/193864718825159619

Volume / Issue

Content sourceWEFTEC

Word count14

Description: Developing Approaches for Achieving Biological Phosphorous Removal and Improved...

Developing Approaches for Achieving Biological Phosphorous Removal and Improved Settleability in High-rate Activated Sludge Systems

Abstract

This study focuses on determining the minimum SRT maintaining significant bio-P activity, quantifying phosphorus removal methods, and evaluating sludge settling. This paper offers practical guidance on the implementation of HRAS systems for determining the actual bio-P phosphate removal through parallel contact stabilization assimilation reactors.

This work demonstrates that bio-P activity washed out of aerobic SRT of 1.2-1.8 day. Overall, about 56% of P removal was achieved through bio-P while the rest was achieved through assimilation. This study showed the need for developing strategies for SRT uncoupling to selectively retain PAOs in high-rate system to potentially achieve P removal and improved overall capacity of high-rate system through improving settling.

SpeakerNgo, Nam

Presentation time

11:30:00

12:00:00

Session time

10:30:00

12:00:00

SessionHealthy As A Horse? Assessing Bio-P Stability

Session locationRoom S403a - Level 4

TopicFacility Operations and Maintenance, Intermediate Level, Municipal Wastewater Treatment Design, Nutrients

Author(s)

Ngo, Khoa Nam

Author(s)K. Ngo 1; C. Lee 2 ; H. Truong 3; E. Kin 2; S. Fuentes 4; C, Chen 3 ; G. Wells 5; H. De Clippeleir 6; C. Lee 2;

Author affiliation(s)District of Columbia Water and Sewer Authority 1; District of Columbia Water and Sewer Authority, Northwestern University 2 ; District of Columbia Water and Sewer Authority, Northwestern University 3; District of Columbia Water and Sewer Authority, Northwestern University 2; District of Columbia Water and Sewer Authority 4; District of Columbia Water and Sewer Authority 3 ; District of Columbia Water and Sewer Authority 5; 6; 2;

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2023

DOI10.2175/193864718825159176

Volume / Issue

Content sourceWEFTEC

Word count16

Description: Optimizing Partial-Denitrification/Anammox (PdNA) Startup in IFAS system with Low...

Optimizing Partial-Denitrification/Anammox (PdNA) Startup in IFAS system with Low TIN Discharge Targets

Abstract

Introduction
To advance PdNA implementation, it is important to better understand PdNA startup strategies and timelines. Previous PdNA startups showed detection of anammox activity within three months for both methanol and glycerol as carbon sources, and MBBR, filters and IFAS configurations (Justin et al., 2022; Bachmann et al., 2024). At Blue Plains, PdNA integration is intended within the nitrification reactors, leaving a full denitrification polishing zone with MeOH downstream of the PdNA IFAS zone. This also means that the mixed liquor will maintain a full denitrification methylotroph population while PdN is targeted within the PdN zone. Previous work showed reduced PdN efficiencies under those conditions (Ladipo-Obasa et al., 2022) compared to only dosing MeOH with PdNA zone (Bachmann et al., 2024; Fofana et al., 2023) and also showed the importance of AnAOB presence to increase PdN efficiency by competing for nitrite (Ladipo-Obasa et al., 2022). The question remains however how one can startup a PdNA IFAS system when the MLSS maintains a healthy full denitrification methylotroph population and no initial AnAOB mass is present to compete for nitrite. The latter conditions are tested in this paper and the results of an IFAS PdNA system with MeOH for PdNA and downstream full denitrification is discussed.

Materials and Methods
A mainstream nitrogen removal pilot was operated at Blue Plains Advanced Wastewater Treatment Plant. The schematic for the pilot is shown in Fig.1. The pilot was initially inoculated with nitrification-denitrification sludge from the full-scale biological nutrient removal system.

Results and Discussions

1. Permit Compliance and Nitrogen Removal
As shown in Table.1, the effluent TIN concentration averaged 2.59 ± 1.2mg N/L, successfully complying with the discharge permit requirements throughout the entire operational period (< 3 mg N/L).

2 PdN selection
PdN was successfully established within seven days and gradually improved until day 42, reaching an PdN efficiency of 20%, and confirmed by activity tests (Fig.2). This result aligns with expectation of observing lower PdN efficiencies with MeOH in the absence of AnAOB (Ladipo-Obasa et al., 2022). However, PdN efficiency declined to below 10% between days 70 and 112, which coincided with suboptimal methanol dosing. Operational constraints, such as understaffing, also hindered the maintenance of conditions essential for optimizing PdN selection during this period. Fluctuations in PdN efficiency persisted throughout the first 100 operational days due to limitation in nitrate removal rates not allowing for observable nitrite accumulation (Fig.2). This limitation was corrected on day 115, by increasing nitrate removal rates. Also, SRT was reduced from 28±0.4 days to 21±4 days, minimizing endogenous decay. Those actions led to establishing PdN efficiencies back to 20% (Fig. 2A). An improvement in PdN efficiency was observed after the enrichment of AnAOB up to PdN efficiencies of 44.03±10.01%, providing results close the previously reported PdN efficiencies (Ladipo-Obasa et al., 2022). This reinforced the critical role of AnAOB in enhancing PdN performance with methanol.

3 AnAOB enrichment
The first measurable AnAOB activity was detected after 90 days (Fig.2D). However, operational inefficiencies delayed significant progress until corrective actions were implemented after day 115 (discussed above) to allow for nitrite levels of 0.82 ± 0.51 mg N/L to be maintained (Fig.2C). To ensure sufficient substrate availability, ammonium concentrations were maintained at ≥2.3 mg N/L (Fig.2C). After establishing nitrite levels > 0.5 mg N/L from day 131 onwards it took until day 153 to see consistent ammonium removal of at least 0.55±0.11 mg N/L or 22.38±10.71 mg N/m2/d (Fig 2D). For that point on, ammonium removal increased exponentially and reached 2.23±0.57 mg N/L or 151.48±48.8 mg N/m2/d at the end of the experiment (Fig.2D).

The observed anammox growth rate was calculated as 0.036 ± 0.004 d-1, which aligns with those reported in polishing filters and MBBR systems using glycerol as a carbon source (Fofana et al., 2022; Schoepflin et al., 2022) but is lower than growth rates in suspended PdNA systems (Le et al., 2021). This discrepancy may be attributed to diffusion limitations in biofilm systems or competition with heterotrophic denitrifiers for nitrite (Stewart Philip, 2003). Additionally, this study confirms that it is important to generate a limited amount of nitrite accumulation together with ammonium availability to enrich AnAOB.

The timing of the first measurable AnAOB activity (day 90) aligns with prior findings (Bachmann et al., 2024; Kocamemi et al., 2018; Verma et al., 2021), further validating the feasibility of these approaches under mainstream PdNA conditions. we believe earlier optimization of nitrate removal to maintain nitrite could have reduced startup time. During this experiment about 60 days were lost.

Conclusion
This study successfully demonstrated AnAOB enrichment in PdNA IFAS system with downstream MeOH based polishing.

This paper was presented at WEFTEC 2025, held September 27-October 1, 2025 in Chicago, Illinois.

Presentation time

08:30:00

09:00:00

Session time

08:30:00

10:00:00

SessionPdNA: Insights into Pilot and Full-scale Implementations

Session locationMcCormick Place, Chicago, Illinois, USA

TopicLiquid Stream Treatment - Nutrient Removal and Recovery

Author(s)

ISLAM, MD AL SADIKUL, Ngo, Khoa Nam, Lee, Chengpeng, Hatcher, Jacob, Riffat, Rumana, Azam, Hossain, Wells, George, De Clippeleir, Haydee

Author(s)M. ISLAM1, K. Ngo1, C. Lee1, J. Hatcher1, R. Riffat4, H. Azam3, G. Wells2, H. De Clippeleir1

Author affiliation(s)DC Water and Sewer Authority1, Northwestern University2, University of the District of Columbia3, The George Washington University4

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2025

DOI10.2175/193864718825160028

Volume / Issue

Content sourceWEFTEC

Word count13