Yang, Cheng

Cheng Yang is a speaker at WEFTEC 2026.

Titles from this speaker

Agua Nueva WRF: Piloting Experience of Machine Learning Hybrid Nutrient Controller

Abstract

Water Research Foundation Project 5121 is titled Development of Innovative Predictive Control Strategies for Nutrient Removal. The project is focused on developing and full-scale testing of a hybrid (machine learning + mechanistic model) nutrient management controller at four different WRRFs. The project team has named the controller ODIN: Operational Decision-making Information Network. The primary objectives of the ODIN pilot for ANWRF are to: Provide twice daily DO setpoint recommendations for the step feed bioreactor. Provide WAS rate recommendations for appropriate SRT control The implementation of data-driven dissolved oxygen control recommendations at the Agua Nueva WRF has resulted in an aeration savings of approximately 10 percent as compared to historical operation at the facility. These recommendations have resulted in operations gaining more insight into the operation of the facility, as well as provide the cost savings described.

SpeakerJohnson, Bruce

Presentation time

16:00:00

16:15:00

Session time

15:30:00

17:00:00

SessionCase Studies of Machine Learning in Full-Scale Nutrient Management Part II

Session locationRoom S505b - Level 5

TopicAdvanced Level, Facility Operations and Maintenance, Intelligent Water, Nutrients

Author(s)

Johnson, Bruce

Author(s)B. Johnson 1; B. Johnson 1 ; C. Yang ; J. Registe 4; A. Menniti 5; A. McClymont 6; R Abel 6; T Mason 8;

Author affiliation(s)Jacobs 1; Jacobs 1 ; Jacobs ; Maia Analytica 4; Jacobs 5; Jacobs 6; Jacobs 6; Jacobs 8;

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2023

DOI10.2175/193864718825159117

Volume / Issue

Content sourceWEFTEC

Word count12

Clean Water Services ODIN Pilot Case Study

Abstract

Water Research Foundation Project 5121 is titled Development of Innovative Predictive Control Strategies for Nutrient Removal. The project is focused on developing and full-scale testing of a hybrid (machine learning + mechanistic model) nutrient management controller at four different WRRFs. The project team has named the controller ODIN: Operational Decision-making Information Network. The primary objectives of the ODIN pilot for CWS are to: Predict the expected diurnally varying plant influent orthophosphate loading over the following 24 hours. Recommend an alum dosing set point to maintain a requested primary effluent orthophosphate load. A secondary goal is the additional information provided on parameters like influent ammonia. For the CWS application, the ODIN systems appears to be a promising opportunity to improve plant operation using digital twin technology.

A hybrid (machine learning + mechanistic modelling) controller called ODIN has been developed through WRF5121 and is currently being tested at the Clean Water Services Durham WRRF. The system recommends for primary alum addition setpoints based on forecasted phosphorus loads and target bioreactor ortho-phosphate mass loading targets. The ODIN system appears to be a promising opportunity to improve plant operation using digital twin technology. This paper summarizes the status of the CWS pilot.

SpeakerMenniti, Adrienne

Presentation time

15:35:00

15:50:00

Session time

15:30:00

17:00:00

SessionCase Studies of Machine Learning in Full-Scale Nutrient Management Part II

Session locationRoom S505b - Level 5

TopicAdvanced Level, Facility Operations and Maintenance, Intelligent Water, Nutrients

Author(s)

Menniti, Adrienne

Author(s)A. Menniti 1; A. Menniti 1 ; B. Johnson 2; K. Lesnik 3; C. Yang 4;

Author affiliation(s)Clean Water Services 1; Clean Water Services 1 ; Jacobs 2; Jacobs 3; 4;

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2023

DOI10.2175/193864718825159116

Volume / Issue

Content sourceWEFTEC

Word count8

Description: Dance of the Machines at AlexRenew WRRF: Piloting Experience of Machine Learning...

Dance of the Machines at AlexRenew WRRF: Piloting Experience of Machine Learning Hybrid Nutrient Controller

Abstract

A full-scale pilot of a hybrid (machine learning + mechanistic modelling) based nutrient controller is being implemented at Alexandria Renew Enterprises (AlexRenew), VA as part of The Water Research Foundation project 5121. The hybrid nutrient controller developed for this project has been named ODIN (Operational Decision-making Information Network). At AlexRenew, ODIN is used to recommend equalization discharge and return pumping setpoints with the goal of minimizing overall energy and chemical usage at the facility while maintaining a very low effluent total nitrogen. Recent historical and simulated nutrient removal performance is presented along with the metrics selected to evaluate the control period with ODIN suggestions being considered.

This paper was presented at WEFTEC 2023 in Chicago, IL.

SpeakerOristian, Monica

Presentation time

16:25:00

16:40:00

Session time

15:30:00

17:00:00

SessionCase Studies of Machine Learning in Full-Scale Nutrient Management Part II

Session locationRoom S505b - Level 5

TopicAdvanced Level, Facility Operations and Maintenance, Intelligent Water, Nutrients

Author(s)

Oristian, Monica

Author(s)M. Oristian 1; M. Oristian 1 ; B. Johnson ; C. Yang 3; J. Registe 5; H Stewart 6; K. Lesnik 7; A. Menniti 8;

Author affiliation(s)Alexandira Renew Enterprises (AlexRenew) 1; Alexandira Renew Enterprises (AlexRenew) 1 ; Jacobs ; Jacobs 3; Jacobs 5; Jacobs 6; Jacobs 7; Clean Water Services 8;

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2023

DOI10.2175/193864718825159118

Volume / Issue

Content sourceWEFTEC

Word count16

Development of Innovative Predictive Control Strategies for Nutrient Removal

Abstract

This paper describes the development and deployment of an innovative hybrid nutrient removal controller that integrates machine learning with mechanistic modeling. Automated data pipelines facilitate the ingestion of SCADA, laboratory (LIMS), and meteorological data. Rigorous data cleaning and imputation processes ensure the high quality of input data. A soft sensor, using an auto-calibrated SUMO model, estimates dynamic influent flow and concentration profiles. Machine learning forecasters extend these profiles by 24 hours, while emulators swiftly simulate responses for optimization. The controller was deployed at four facilities, including Clean Water Services, where primary effluent orthophosphate loads were predicted with a mean average percent error of 12.20%. The flexible and modular structure of the controller, which integrates mechanistic and machine learning elements, exemplifies its impressive predictive optimization capabilities.

SpeakerLesnik, Keaton

Presentation time

14:00:00

14:15:00

Session time

13:30:00

15:00:00

SessionApplications of Machine Learning in Full-Scale Nutrient Management Part I

Session locationRoom S504c - Level 5

TopicFacility Operations and Maintenance, Intelligent Water, Intermediate Level, Nutrients

Author(s)

Lesnik, Keaton

Author(s)K. Lesnik 1; K. Lesnik 1 ; J. Registe 2; K. Lesnik 3; B. Johnson 4; C. Yang 5; D Pienta 6;

Author affiliation(s)Maia Analytica 1; Maia Analytica 1 ; Jacobs 2; Maia Analytica 3; Jacobs 4; Jacobs 5; Jacobs 6;

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2023

DOI10.2175/193864718825159079

Volume / Issue

Content sourceWEFTEC

Word count10

Description: Eleven Full-Scale Implementations in Five Years: Lessons from Hybrid Digital Twin...

Eleven Full-Scale Implementations in Five Years: Lessons from Hybrid Digital Twin Applications for Operations and Design

Abstract

Microplastics (MPs) are increasingly reported in surface waters and wastewater-derived streams, raising concern because these particles persist in the environment, can act as carriers for co-contaminants, and may contribute to downstream exposure when effluents impact receiving waters or when reclaimed/recycled water is produced (Carnevale Miino et al., 2024; Iyare et al., 2020; Koelmans et al., 2016; Pérez-Reverón et al., 2022; Tumwesigye et al., 2023). While conventional treatment processes can reduce larger plastic debris, smaller MPs and especially nano-sized plastics are more difficult to capture due to stable suspension and transport behaviors that limit retention by gravity-driven separation and some polishing steps (Chen et al., 2023; Tang & Hadibarata, 2021). Carbon-based media are promising for advanced MP mitigation because they are chemically stable in water and can be produced from renewable feedstocks (e.g., biochar/hydrochar), while their high surface area and pore/particle networks support MP removal via combined surface interactions and physical interception/filtration processes (Amirah Mohd Napi et al., 2023; Anuwa-Amarh et al., 2024; Essien et al., 2025; Zheng et al., 2024). The objective of this study is to optimize MP removal using carbon-based materials and to generate quantitative, engineering-relevant performance metrics that support media selection and process design. Five adsorbents were evaluated under identical laboratory conditions: (1) raw biochar, (2) iron-modified biochar (Fe-biochar), (3) thermally activated biochar produced at 850 °C (850-biochar), (4) activated carbon felt (ACF), and (5) activated carbon (AC). As a model sub-micron MP system, a 0.2 μm fluorescently labeled polystyrene MP suspension was used. Approximately 1 g of each adsorbent was contacted with the MP suspension under controlled mixing. Samples were collected periodically and MP concentration was quantified using a fluorescence spectrophotometer. Adsorption progress was evaluated using time-dependent uptake profiles, and performance was summarized using (i) time required to reach near-equilibrium and (ii) apparent maximum adsorption capacity (QMAX) based on the observed uptake behavior. Tests were replicated to quantify variability and ensure data comparability across media. Among the tested materials, ACF exhibited the strongest overall performance for sub-micron MP capture. ACF reached 95% of its adsorption capacity within 3.17 hours and achieved an apparent maximum adsorption capacity of QMAX = 17.74 ± 3.35 mg MPs g−1 ACF under the tested conditions. The superior behavior of ACF is consistent with its fibrous, highly porous architecture, which increases accessible retention sites and can enhance particle interception/entanglement in addition to surface-based interactions. The remaining carbon-based media (raw and modified biochars and granular activated carbon) exhibited uptake from 1.70 to 3.10 MPs g−1 under the same test conditions, indicating that material architecture and accessible surface features strongly influence both adsorption kinetics and capacity for sub-micron MPs. These results provide a practical, side-by-side benchmark for selecting carbon-based materials to mitigate sub-micron MPs in advanced treatment trains. The rapid kinetics and high capacity observed for ACF suggest potential for reduced required contact time and smaller unit footprint compared with conventional granular media, which is directly relevant to tertiary retrofits at utilities. The comparative dataset also supports a defensible pathway for implementation: rapid laboratory screening to identify high-performing media, followed by column-scale validation (breakthrough behavior) and matrix-sensitive testing in real wastewaters where dissolved organic matter and ionic strength may influence MP retention. Overall, this work advances the evidence base needed to translate emerging MP concerns into actionable treatment strategies using scalable carbon-based material.

This paper was presented at WEFTEC 2026 in New Orleans, Louisiana.

Presentation time

13:30:00

13:45:00

Session time

13:30:00

15:00:00

SessionSuccessful hybrid digital twin adoption in full-scale facilities

Session locationErnest N. Morial Convention Center

TopicFacility Operations and Maintenance, Municipal Wastewater Treatment Design, Research and Innovation

Author(s)

Yang, Cheng, Johnson, Bruce, Stewart, Heather, Rieger, Leiv, Schraa, Oliver, Registe, Joshua, Pienta, Drew, Printz, Katie, Zaveri, Jaydev, Miletic, Ivan, Emaminejad, Aryan

Author(s)C. Yang1, B. Johnson1, H. Stewart1, L. Rieger1, O. Schraa1, J. Registe1, D. Pienta1, K. Printz1, J. Zaveri1, I. Miletic1, A. Emaminejad1

Author affiliation(s)Jacobs, 1Jacobs, 1Jacobs, 1Jacobs Solutions Inc., 1Jacobs, 1Jacobs Solutions Inc., 1Jacobs, 1Jacobs, 1Jacobs, 1Jacobs, 1Jacobs Engineering Headquarters Office, 1

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Sep 2026

DOI10.2175/193864718825160345

Volume / Issue

Content sourceWEFTEC

Word count17

Description: Looking Under the Hood: Live Digital Twin Supports Operational Decisions for Stable...

Looking Under the Hood: Live Digital Twin Supports Operational Decisions for Stable Biological Phosphorus Removal

Abstract

INTRODUCTION Biological Phosphorus Removal (BioP) offers advantages such as reduced metal salt use, lower chemical sludge solids, and improved settleability. However, its complex biochemical relationships are sensitive to operational changes and prone to upsets that threaten effluent quality. Operators at the City of Ann Arbor Water Resource Recovery Facility (WRRF) rely on lab data, SCADA, and qualitative observations to optimize performance. In 2025, the City partnered with Jacobs Engineering to implement a Hybrid Optimizer (a live digital twin) for insight into BioP upsets and to provide operational decision support. To the authors' knowledge, this is the first plant-wide live hybrid digital twin for BioP. The Ann Arbor WRRF is a 29.5 MGD (112 MLD) advanced secondary treatment facility (recent annual average 13.7 MGD). Liquids treatment includes headworks, primary clarification, anoxic/anaerobic/aerobic bioreactors, final clarification, filtration, and UV disinfection before discharge to the Huron River. Solids treatment includes lime addition for phosphorus precipitation to mixed PS and TWAS before centrifuging. The Hybrid Optimizer combines machine learning and mechanistic modeling for full-scale nutrient management, developed under WRF Project 5121 (Johnson et al., 2024). Following project completion, the tool has undergone continuous refinement and has been applied at additional facilities, including Ann Arbor. This paper summarizes the facility's interaction with a live digital twin and highlights key operational insights from the Hybrid Optimizer deployment at Ann Arbor. METHODOLOGY Deployment steps mirrored previous live applications, outlined in the WRF 5121 Final Report and in previous publications (Yang et al. 2024, Yang et al. 2023; Registe et al. 2023; Menniti et al. 2023; Johnson et al. 2023; Oristian et al. 2023). The critical role of the facility is highlighted here. [b]Data Ingestion & Preparation:[/b] Continuous collaboration with staff ensured robust data pipelines, accurate tag interpretation, and reliable metadata. [b]Model Development:[/b] Seasonal load changes from the local university require frequent changes of basins in service. As basins often exhibit different performance, this required modeling individual trains in Sumo24©. Staff input was critical for dynamic solids operations and translating facility variability into the mechanistic model. [b]Machine Learning & Recommendations:[/b] Staff prioritized KPIs such as primary sludge blanket age and secondary loading when selecting information to daylight from the digital twin. Staff identified daily emails as the most useful way to receive insights into these KPIs and what operational changes are recommended as a result (see Figure 1 for example email). Influent forecasting is in progress. [b]User Interface:[/b] Jacobs and staff designed a GUI to display key soft sensor results and KPIs (Figure 3). Staff also opted for downloaded soft sensor data into a Sumo© model for offline 'what-if' analysis, saving many hours of manual input and enabling high-frequency simulation of costly-to-measure parameters. RESULTS AND DISCUSSION [b]Soft Sensor Predicts Process Upsets:[/b] The hybrid optimizer employed autocalibration of key parameters such as alpha_sat and hydrolysis to track the influent strength and effluent phosphorus performance long term. The system was able to predict influent laboratory data and effectively anticipate both biological phosphorus removal (BioP) and nitrification upsets, as shown in dashboard visualizations during these events (Figure 2). The model predicts slightly earlier and more extreme upsets, which is preferable as a warning signal for operators. [b]User Interface Daylights Unmeasurable BioP Indicators:[/b] The Hybrid Optimizer tracks key indicators such as PHA uptake, VFA-to-phosphorus feed ratio, and GAO/PAO balance to assess BioP stability. These mechanistic and soft-sensor insights reveal how primary fermentation and hydrolysis drive VFA production, which sustains PAO activity and prevent GAO dominance. Upsets in the simulated and measured secondary effluent OP spiked beyond the same threshold for these indicators, implying the model is capturing real BioP health (Figure 3). Staff have observed qualitative indicators of fermentation like primary clarifier sheen correlate with target indicator ranges. Unlike traditional monitoring, these predictive indicators enable proactive control of primary sludge blanket SRT and upset avoidance. Operational recommendations began in December 2025; secondary SRT guidance is under development. CONCLUSION Mechanistic and machine-learning calibration of hydrolysis rates to match BioP performance successfully produced the first live digital twin of its kind for biological phosphorus removal. The Ann Arbor Hybrid Optimizer pilot transforms advanced biokinetic modeling into practical, actionable insights for WRRF staff by leveraging a complex model of six independent trains to capture loading and wasting variability. Soft sensors provided critical data that cannot be measured directly, while developed KPIs and primary wasting recommendations were validated against both operator experience and real-time BioP upset events. The Ann Arbor Hybrid Optimizer pilot is ongoing; additional results will be provided in final paper.

This paper was presented at WEFTEC 2026 in New Orleans, Louisiana.

Presentation time

14:00:00

14:15:00

Session time

13:30:00

15:00:00

SessionDigital Twins for Operational Support

Session locationErnest N. Morial Convention Center

TopicFacility Operations and Maintenance, Research and Innovation, Utility Management and Leadership

Author(s)

Stewart, Heather, Printz, Katie, Johnson, Bruce, Yang, Cheng, Jaworski, Nicholas, Sanders, Keith, Zaveri, Jaydev, Pienta, Drew, Emaminejad, Aryan, GELDERLOOS, ALLEN

Author(s)H. Stewart1, K. Printz1, B. Johnson1, C. Yang1, N. Jaworski, K. Sanders2, J. Zaveri1, D. Pienta1, A. Emaminejad1, A. GELDERLOOS1

Author affiliation(s)Jacobs, 1Jacobs, 1Jacobs, 1Jacobs Engineering Group, 1City of Ann Arbor WWTP, 2Jacobs, 1Jacobs, 1Jacobs Engineering Headquarters Office, 1Jacobs, 1

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Sep 2026

DOI10.2175/193864718825160469

Volume / Issue

Content sourceWEFTEC

Word count16

Mechanistic Modelling Within a Hybrid Controller

Abstract

The Water Research Foundation project 5121, Development of Innovative Predictive Control Strategies for Nutrient Removal aims at developing full-scale hybrid nutrient management controllers by employing a hybrid approach that combines both machine learning and mechanistic models. This paper specifically examines the role of mechanistic models within the context of hybrid modeling, focusing on their integration and interactions with machine learning components. One crucial aspect of the hybrid optimizer is its ability to automate the simulation and calibration of mechanistic models, which transforms mechanistic models into soft sensors. The soft sensor provides benefits facilitated by the mechanistic models including monitoring the state of every process, enabling plant-wide control, generating adequate data for machine learning models and etc. Integrating the mechanistic model into the hybrid optimizer has resulted in valuable practical knowledge and experience, which significantly contribute to the ongoing discussions on hybrid modeling practice.

This paper specifically examines the role of mechanistic models within the context of hybrid modeling, focusing on their integration and interactions with machine learning components in full-scale hybrid nutrient management controller development. A key advancement is automating the simulation and calibration of mechanistic models. Knowledge and experience from integrating the mechanistic model into the hybrid optimizer significantly contribute to the discussions on hybrid modeling.

SpeakerYang, Cheng

Presentation time

14:25:00

14:40:00

Session time

13:30:00

15:00:00

SessionApplications of Machine Learning in Full-Scale Nutrient Management Part I

Session locationRoom S504c - Level 5

TopicFacility Operations and Maintenance, Intelligent Water, Intermediate Level, Nutrients

Author(s)

Yang, Cheng

Author(s)C. Yang 1; C. Yang 1 ; H. Stewart 2;

Author affiliation(s)Jacobs 1; Jacobs 1 ; Jacobs 2;

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2023

DOI10.2175/193864718825159080

Volume / Issue

Content sourceWEFTEC

Word count7

Description: Reducing Capital Cost in Process Design with Digital Twins: Two Full-Scale Case...

Reducing Capital Cost in Process Design with Digital Twins: Two Full-Scale Case Studies

Abstract

INTRODUCTION
Digital twins (DTs) are emerging as powerful technologies in the digital transformation of wastewater treatment facilities (WWTFs). While most DT applications currently focus on optimizing operations using live data and integrated modeling, to the authors knowledge no studies have explored their potential in the design phase of a project. Currently, due to insufficient data, process design assumptions often include multiple overlapping conservative parameters, leading to higher capital costs, such as assuming the coldest temperature, peak loadings, and peak flows occur simultaneously. By integrating commonly available data, DTs can provide high-quality, high-resolution dynamic data that were not previously available to reduce design uncertainty, mitigate risks, and reduce capital costs. Yang et al. proposed using DTs for soft sensing dynamic influent profiles, which addresses this data insufficiency challenge for design. High-resolution dynamics improve control design and equipment sizing accuracy. This paper investigates the benefits of using DTs in the process design phase, presenting the two full-scale case studies on incorporating DTs into process design:
1. Marine Park WWTF (City of Vancouver, Washington, US): improving existing aeration basin performance with advanced aeration control and mixed liquor recirculation to address increasing influent loads and sludge popping issues in the clarifiers.
2. Davyhulme WwTW (Greater Manchester, UK): generating a long-term influent dynamic profile that can be used by the design teams to minimize conservatisms and expansion costs for a number of planned expansion projects across the liquid streams.

METHODOLOGY
1. Dynamic Influent Profiles Estimation. Historical data were fed into the digital-twin-based Soft Sensor to generate the 15-min dynamic influent profiles along with biological kinetics, stoichiometry, and fractionation. The estimated influent profiles were collapsed into daily composite values and compared with measurements to verify their accuracy.
2. Design Option Analysis with Selected Periods. Depending on design objectives, relevant periods were selected to test design options.

RESULTS AND DISCUSSION
The measurements from the soft sensor were compared with those from both physical sensors and laboratory tests, and both comparisons showed a good match. The Davyhulme WwTW has two regularly maintained ammonia probes at the primary effluent, which allows for a direct comparison between the soft sensor and physical sensors, as shown in Figure 1 with performance metrics listed in Table 1. It is observed that the shape and magnitude of the signals (i.e., the timing of peaks and valleys, inclines and slopes) match well. With Table 1, it can be inferred that the soft sensor should be able to reproduce similar accuracy results as physical sensors. Comparison with lab measurements is provided in Figure 2 and Table 2. Additionally, the soft sensor generated dynamic biological kinetics, stoichiometry, and fractionation that can be used in design (Figure 2). The Marine Park WWTP soft sensor has a similar performance.

The dynamic influent profiles were used to refine process design by reducing uncertainty, optimizing equipment selection and sizing, and ultimately saving significant capital costs. The Marine Park case study selected four periods to analyze design options for 7-day dynamic maximum, average, minimum loading and wet weather conditions (Figure 4). Finding and design implications are summarized below:
- Influent peak loading, peak flow, and coldest temperature periods do not occur simultaneously, which can lower safety factor to reduce capital costs.
- Ammonia-based-aeration-control is proved to be a preferred advanced aeration control mechanism that is flexible for different aeration objectives with dynamic simulations.
- With dynamic influent profiles under various scenarios, the mixed liquor recirculation (MLR) design was evaluated to avoid oversizing the pumps and thus save equipment costs and energy.
- Investigations on aeration demand during minimum loading conditions eliminate the installation of aerobic zone mechanical mixers originally defined in the project scope, saving unnecessary equipment costs.

With these insights, the primary design was further refined, and the cost savings are detailed in Table 3. The reduction in MLR pump sizing and the removal of mechanical mixers, as suggested and verified by the DT and soft sensor, resulted in a total capital cost savings of $7,367,000 for this project.

CONCLUSIONS This paper presents two case studies utilizing digital twin technology to refine process design by reducing conservativeness, optimizing equipment selection and sizing, and ultimately saving significant capital costs. By leveraging a digital twin soft sensor, high-quality and high-resolution influent profiles were developed from available historical data that reduces design uncertainty. This approach unlocks process design possibilities previously restricted by insufficient dynamic data. It is the first attempt to integrate digital twin tools in supporting process design and has proven to be practical. Its most significant contribution lies in re-envisioning the use of digital twins beyond operation improvement and encouraging further exploration of their potential benefits.

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

Presentation time

16:00:00

16:10:00

Session time

15:30:00

17:00:00

SessionInnovative Approaches to Design and Optimization

Session locationMcCormick Place, Chicago, Illinois, USA

TopicProcess Control and Modeling

Author(s)

Yang, Cheng, Johnson, Bruce, Zhang, Miaomiao, Noesen, Matthew, Klibert, Corey, Pinochet Troncoso, Ivette, Dick, Frank, Taylor, Christopher

Author(s)C. Yang1, B. Johnson1, M. Zhang1, M. Noesen1, C. Klibert1, I. Pinochet Troncoso1, F. Dick2, C. Taylor3

Author affiliation(s)Jacobs1, City of Vancouver Public Works2, United Utilities3

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Sep 2025

DOI10.2175/193864718825159921

Volume / Issue

Content sourceWEFTEC

Word count14

Description: Soft Sensing Influent Concentrations with Airflow Rates by Digital Twins: A Real...

Soft Sensing Influent Concentrations with Airflow Rates by Digital Twins: A Real Case Study from Oldham WwTW

Abstract

With the abundance of data availability across treatment facilities, artificial intelligence is paving the way for unique optimization opportunities. The examples provided encompass the Agua Nueva Water Reclamation Facility (WRF), which concentrates on minimizing aeration energy expenses and enhancing nutrient management, as well as the Wilmington WWTF, emphasizing the optimization of disinfection chemicals. Several artificial intelligence (AI) and Bayesian modeling frameworks were developed with both sites using AI algorithms with mean absolute percentage errors < 10% for both forecasting of future conditions at a facility, as well as causal inferencing for predicting water quality. Post deployment, both sites are regularly achieving anywhere from 10-30% savings in energy or chemical usage. These case-study demonstrates significant progress in the successful implementation of AI at a treatment facility and the ability to aid in the empowerment of treatment plant operators while improving efficiencies in wastewater treatment.

United Utilities initiated a case study using already available data to predict the real-time influent concentrations to optimize chemical addition for phosphorus removal without installing physical sensors. This paper proposes a novel soft sensor mechanism that successfully estimates the high-resolution influent profiles with a digital twin model plus regular measurements (e. g. airflows, flows and lab measurements). The soft sensor overcomes the lack of dynamic profiles in DT applications.

SpeakerYang, Cheng

Presentation time

14:00:00

14:20:00

Session time

13:30:00

15:00:00

SessionPlanning and Process Improvement Case Studies: Winning with Twinning

Session locationRoom S505b - Level 5

TopicAdvanced Level, Energy Production, Conservation, and Management, Municipal Wastewater Treatment Design, Nutrients, Research and Innovation

Author(s)

Yang, Cheng

Author(s)C. Yang 1; B. Johnson 2 ; J. Registe 3; T. Johnson 4; A. Rahman 5; J. Kenyon 6; C. Yang 1;

Author affiliation(s)Jacobs 1; Jacobs 2 ; Jacobs 3; Jacobs 4; 5; 6; Jacobs 1;

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2023

DOI10.2175/193864718825159227

Volume / Issue

Content sourceWEFTEC

Word count18

WRF 5121 Final Report Overview: Long-Term Evaluation of Hybrid Digital Twins at Three Full-Scale WRRFs

Abstract

INTRODUCTION The Water Research Foundation (WRF) project 5121, Development of Innovative Predictive Control Strategies for Nutrient Removal, aims at developing full-scale hybrid nutrient management controllers (named here as Hybrid Optimizer, HO) by employing a hybrid approach that combines both machine learning and mechanistic models. Three full-scale pilots are deployed and evaluated for this project to demonstrate long-term controller abilities. The concept, design and preliminary results have been shared previously [1-7], which triggered wide interest and heated discussion. Over the past year, refinements were made to improve performance and evaluation approaches were standardized. Currently, all three pilots are fully deployed and operational. This paper provides an overview of three case studies. The control for three pilots is advisory only with no direct control functions due to cybersecurity concerns. Instead, operators receive daily email notifications outlining recommended operations and/or access a web-based user interface (UI), guiding their actions. The full-scale pilots are: 1. Clean Water Services Durham WRRF (CWS): Primary clarifier alum addition setpoints once a day, aiming to stabilize orthophosphate load to the downstream biological treatment system. 2. Agua Nueva WRF (AN): Dissolved oxygen setpoints twice daily for the bioreactor systems, aiming to minimize overall aeration energy consumption. 3.AlexRenew WRF (AR): HO recommends three equalization setpoints once a day to equalize the influent loadings to the bioreactor to minimize energy, chemicals, and blower staging. METHODOLOGY As shown in Figure 1, Hybrid Optimizer consists of seven major components. Descriptions of each component can be found in [1-3]. In this paper, the performance of Soft Sensor, Influent Forecaster and emulator/optimizer is mainly presented and discussed. Four different types of standard graphs are used to evaluate the performance of the HO. (1)Control Chart. Directly conveys the validity of the model, based on which warnings and alarms are derived. It integrates the measured and the modelled values into a dimensionless metric (t_score) and allows its comparison with predefined bounds. (2)Process Plot. The process plot presents the measured and the modelled values. (3)Unscaled Error Plot. The Error plot presents the residuals between the modelled and the measured. (4)XY Plot. The XY Plot presents the Modelled v. s. the Measured and the prediction intervals. (not shown here for brevity) RESULTS AND DISCUSSION The components and timelines of the three pilots are listed in Table 1. The Graphic User Interface (GUI) for two pilots are in progress. Evaluation results are available for all three, however this abstract presents only partial results. The soft sensor can be considered a full digital twin of the modelled components. Thus, it provides dynamic 15-minute frequency data on all wastewater components in the model scope. For CWS the primary influent and effluent ortho-phosphate data was used to support primary alum addition recommendations, without an actual sensor. For AN the focus was ammonia. The soft sensor demonstrates alignment with measured data (Figure 2).

*In the Control Chart, the t score falls within the warning and alarm bounds (set at 95% and 99% confidence intervals), which indicates soft sensor performance. The t-score is the normalized metric of a given set of measurements.

*The Process and Residual Plots are more intuitive for process engineers to evaluate. The modelled and measured values align closely, and their parallel trends indicate good soft sensor performance. Figure 3 shows the AN performance for primary influent and effluent ammonia. The top graph shows a good match between the composite samples. However, the primary effluent in the middle and bottom graphs shows a poor match. It was found that high influent sulfide was biasing the measurements high. The bottom graph showing the dynamic concentration variation, other than the bias, shows a good variability match. Flow forecasting is critical to predictive control. The Forecaster uses the available weather forecast and other variables to estimate plant influent flow. Figure 4 compares influent flow (forecasted vs actual) as well as the forecasted versus actual rainfall. A close examination shows that much of the influent forecast error can be attributed to the error in the precipitation forecast. The AR optimizer (Figure 5) aims to equalize ammonia load, which helps stabilize the aeration demand and optimization of aeration and chemicals. The flow and loading prediction from the HO allows the optimizer to recommend three equalization setpoints for the next 24 hours. In the top graph, the green line is the target daily average loading, red and blue show the load with and without the recommendations. CONCLUSIONS This abstract presents abbreviated results of the long-term performance of HO during three full-scale pilot studies. The HO appears to be a promising opportunity to improve plant operation using digital twin technology by giving both optimized operational recommendations and increased insight into the facility's status. The ongoing assessment of the HO performance is anticipated to yield further valuable practical experiences and insights, including the benefits and costs of such applications.

This paper was presented at the WEFTEC 2024 conference in New Orleans, LA October 5-9.

SpeakerYang, Cheng

Presentation time

13:30:00

13:50:00

Session time

13:30:00

15:00:00

SessionLeveraging Automation and Analytics for Better Situational Awareness and Optimization: Part I

Session number205

Session locationRoom 350

TopicEnergy Production, Conservation, and Management, Facility Operations and Maintenance, Intelligent Water, Intermediate Level, Nutrients

Author(s)

Yang, Cheng, Johnson, Bruce, Lesnik, Keaton, Registe, Joshua, Rieger, Leiv, Stewart, Heather, Miletic, Ivan, Menniti, Adrienne, Oristian, Monica, Pienta, Drew, Mason, Tim

Author(s)C. Yang1, B.R. Johnson1, K. Lesnik2, J. Registe3, L.P. Rieger4, H.A. Stewart5, I. Miletic6, A. Menniti7, M.K. Oristian8, D.J. Pienta9, T. Mason10

Author affiliation(s)1Jacobs, CO, 2Maia Analytica, OR, 3Jacobs Engineering, NJ, 4Jacobs, NL, 5Jacobs, PA, 6inCTRL Solutions Inc., ON, 7Clean Water Services, OR, 8Alexandria Renew Enterprise, VA, 9Jacobs, 10Jacobs Eng, AZ

SourceProceedings of the Water Environment Federation

Document typeConference Paper

PublisherWater Environment Federation

Print publication date Oct 2024

DOI10.2175/193864718825159642

Volume / Issue

Content sourceWEFTEC

Word count16