U S Department of Energy Sustainable Wastewater Infrastructure

U. S. Department of Energy - Sustainable Wastewater Infrastructure of the Future (SWIFt) Webinar February 28 th, 2018 Ammonia-based Aeration Control Practical Implications Leiv RIEGER, Ph. D. , P. Eng. in. CTRL Solutions Inc. Oakville ON, Canada

Overview ØContext ØAeration Control v. Nitrification fundamentals v. Aeration control strategies v. Control fundamentals v. Constraints ØMaintenance, Quality Control, and Safety Nets ØCosts versus Savings ØTake Home Message 2

Context • High variability of incoming load and temperature • Fixed reactor volumes • WWTP design based on max load Unused capacities Nitrification is the rate-limiting process and therefore the primary target of BNR aeration control strategies 3

Context Control opportunities ØBenefits § Make use of all available capacity § Increases efficiency/robustness ØCosts § § Additional infrastructure (sensors, actuators, PLC, etc. ) Design Implementation Maintenance ØRisks § Operates plant at its limits increased risk 4

Aeration Control v. Nitrification fundamentals v. Aeration control strategies v. Control fundamentals v. Constraints 5

Nitrification Fundamentals Nitrification requirements Ø Sufficient provision of dissolved oxygen Ø Ammonia as substrate (+ essential nutrients) Ø Sufficiently long aerobic sludge retention time Ø Sufficient mass of nitrifiers Autotroph 6

Nitrification Fundamentals DO constraints – nitrification kinetics At 2 mg DO/L: ca. 80% of max. rate More air does not necessarily help the process 7

Nitrification Fundamentals Ammonia as substrate Typical ammonia profile from fully aerated plant 8

Nitrification Fundamentals Ammonia effluent variations Typical ammonia variation from fully aerated plant • Last aerated reactor (blue) • Effluent (red) à SRTaerob = 8 days à Average ≈ 0. 4 mg. N/L What is the reason for the ammonia breakthroughs? 9

Nitrification Fundamentals Nitrifier mass • The mass of nitrifiers changes slowly • The total mass depends on average ammonia load and SRT • The influent ammonia load may vary substantially over a day Ammonia break-through often: due to limited mass of nitrifiers not a problem of insufficient oxygen (or other limiting components) 10

Aeration Control Strategies Ammonia-based aeration control Objectives Ø 1) Limiting aeration: Reduce energy consumption, increase denitrification, improve bio-P performance Ø 2) Reducing effluent ammonia peaks: Reduce the extent of effluent ammonia peaks Clearly define control goals 11

Aeration Control Strategies 1) Limiting aeration: DO vs. NHx control PST BOD removal Denitrification Nitrification FST Control handle: Aeration Ø DO control aims for optimal DO for all aerobic processes Ø ABAC optimizes nitrification process 12

Aeration Control Strategies 1) Limiting aeration: Cascaded NHx–DO control Aeration intensity control (or intermittent aeration) Measured variable (Actual value) DO Controller (setpoint) Manipulated variable Pressurized air O 2 13 Reference variable M NH 4 controller DO f(NH 4) NH 4

Aeration Control Strategies 2) Reducing ammonia effluent peaks Ø Intensity control: Manipulate aeration intensity early to create buffer for incoming peak Limited impact Ø Volume control: Change aerated volume by switching on/off swing zones High impact 14

Control Fundamentals Feedback vs. Feedforward control Feedback control Setpoint Reference variable / Setpoint u Disturbances z ε Controller Measured variable r y Final Control Element Controlled variable x Measuring Device Covers most applications in wastewater treatment 15 Process Target variable Measure process answer

Control Fundamentals PID controller Fast Remove steady Simple prediction of compensation state error process output Off-set Control signal 16 Error u – x (setpoint – meas. variable) Proportional component Integral component Derivative component

Control Fundamentals Feedback vs. Feedforward control Measure process disturbance Disturbances z r System model Reference variable / Setpoint u 17 ε Controller y Measuring Device Final Control Element Process Controlled variable x Fast reaction before disturbance hits the plant Process model required Must be complemented by feedback signal More sensors required Use only if required

Control Fundamentals Feedback and Feedforward control Maximumcriteria Feedforward Controller Ref. variable Measured variable NH 4 Q DO Controller Manipulated variable Press. air O 2 M NH 4 controller DO f(NH 4) NH 4 Coordinate controllers 18

Control Fundamentals ABAC-SRT control Measured NHx set point Measured DO Ammonia Controller DO set point Desired Average DO Concentration DO Controller Airflow set point Supervisory SRT set point Controller Calculated Dynamic SRT Controller MLSS Setpoint Measured MLSS Controller Control SRT so that ABAC can work unrestricted 19 WAS Flow Rate Patent-pending, Schraa et al. , 2017

Control Fundamentals Cascade control Qair Controller DO Controller 10, 000 PI 2. 5 mg DOhigh/L PI 0. 5 mg DO/L scfmlow 2, 000 Qair scfmhigh scfm 7, 800 NHx-N Controller PID 2. 0 mg N/L 0. 5 mg DOlow/L 0. 65 7, 897 NHx Setpoint mg DO/L 1. 75 mg N/L scfm Valvemax % 80 Valvemin % 30 Mvalve 75 % Valve Position AB 1 Protect against controller windup 20

Control Fundamentals «Direct» ABAC Direct ABAC NHx-N Controller Qair Controller 10, 000 PI NHx Setpoint Logic scfm 7, 800 2, 000 Qair scfmhigh scfmlow 1. 35 7, 897 5. 0 mg N/L mg DO/L 1. 75 mg NHx-N/L scfm Valvemax % 80 Valvemin % 30 Mvalve 75 % Valve Position AB 1 Not recommended 21 • How to limit DO at high load? kinetics • External DO controller required (manipulating upper airflow bound) • Tuning more difficult • What NHx setpoints to select for multiple aeration grid control?

Constraints q Regulatory constraints Effluent permit (concentration, mass, removal efficiency, time frame) q Equipment constraints § § blower maximum capacity and turn-down capabilities air flow requirements for mixing air flow per diffuser valve or piping constraints q Process constraints § mass of specific organisms (e. g. nitrifiers) not sufficient at high load Sludge Retention Time § oxygen transfer limitations q Organizational constraints § § budget for investing in new equipment budget for maintenance and quality control experience level of the plant personnel incentive system is missing Know your constraints 22

Constraints Analyzing technical constraints using dynamic simulation 23 Dynamic simulators are a great tool to design control solutions

Constraints Analyzing technical constraints using dynamic simulation Constant header pressure setpoint DO Concentrations in Basin 1 Valve Positions Make sure you can get the air to the right place 24

Maintenance, Quality Control, Safety Nets 25

Maintenance and Quality Control Sensors Base Calibration Achieve match between standards and measurements Maintenance Guarantee proper measurement conditions Quality Control Quantify the accuracy of a sensor over time 26

Maintenance and Quality Control Sensors Base Calibration At sensor installation Maintenance According to manufacturer specifications Quality Control For control: Reference measurement: 1/week Signal behavior: every measurement 27 Rule of thumb: Maintenance and QA/QC per sensor: 1 -2 hours/week

Maintenance and Quality Control On-line Data Quality Assessment & Control ØCheck data quality as you measure ØIntegrate multiple on-line and off-line signals into smart tools ØTrigger Corrective Actions before data quality is impacted ØCombine Data Quality Control with Maintenance Planning ØAutomate data quality control so that the data is ready at all times No Data Reconciliation required 28

Maintenance and Quality Control in. CTRL’s Advanced QA/QC Concept Increased Effort to Collect Information On-line Methods Sensor Infos Models Auto Experiments Off-line Methods Manual Experiments Single signal Univariate methods Sensor status Meta data Multivariate Models Process knowledge Triggered events Comparison to lab measurements Fault Detection 1 st Level Fault Detection 1 st Level Warning Level Fault Detection 2 nd Level Alarm Level • • Sensor evaluation Additional Reference measurements • • • Stop usage / Switch to replacement signal In-depth analysis of sensor Corrective actions 29 29 Sensor characteristics (response time, calibration or interference experiments) Fault Detection 1 st Level

Maintenance and Quality Control Sensor Initialization Procedure Investing into proper sensor initialization saves time during sensor operation 30

Maintenance and Quality Control Sensor Integrity Mounting of TSS probe in aeration basin © Tracy Doane-Weideman Endress+Hauser Sensor Installation can have significant impact on data quality 31

Maintenance and Quality Control Sensor Integrity Automatic cleaning cycle Ø Check for regular patterns, e. g. : Cleaning, Calibration, Pumping schedules Ø Remove before feeding to controller 32

Maintenance and Quality Control Application Integrity Sensor initialization is essential for QA/QC procedure 33

Maintenance and Quality Control Off-line Fault Detection (based on reference measurements) Direct comparison but only when reference measurements available 34

Maintenance and Quality Control On-line Fault Detection q Deviation is not detected unless variables are combined q Use of correlation structure between variables Sensor or Process Fault ? 35

Maintenance and Quality Control Data Gaps: Soft Sensors Short-term replacement of sensor signals: Time Series Model Signal Generation for Variables difficult to measure: PLS Model Soft Sensors can also be used in Fault Detection and differentiation between sensor and process faults. 36

Maintenance and Quality Controller tuning © Geisenhoff, 2014 Controllers need maintenance as well 37

Safety Nets Fall-back strategies Qair Controller DO Controller NHx-N Controller NHx Setpoint PI Qair scfm 7, 800 PI 2. 0 mg DO/L 2. 10 7, 897 mg DO/L PID Na. N 2. 0 mg N/L scfm Mvalve 75 % Valve Position AB 1 Automatically switch to underlying control loop if sensor fails Triggered by quality control system 38

Safety Nets Safety net Alarm signal from combined effluent © Bott, 2018 Independent measurement to trigger alarm 39

Costs versus Savings 40

Costs vs. Savings Aeration costs 41

Costs vs. Savings Ammonia-based aeration control WRRF Morgental 35, 000 PE 3. 5 mgd WRRF Thunersee 130, 000 PE 10 mgd WRRF Werdhoelzli 600, 000 PE 50 mgd simulation full-scale simulation Aeration Energy -30% -20% -30% -16. 5% -25% TN removal +48% +40% +60% +40% +32% $ 360’ 000 $ 1’ 200’ 000 Annual net savings $ 53’ 000 Rieger et al. , WER 2012 42

Costs vs. Savings HRSD Nansemond Treatment Plant (last webinar) § Energy consumption: -5% § Energy cost: -10% § Supplemental carbon: -50% Review Paper (Amand et al. , 2013) § Energy consumption: -5 -30% 43 Benefits include: Ø Energy savings Ø Improved TN removal Ø Reduction in supplemental carbon addition Ø More stable aeration control

Costs vs. Savings Annual Sensor costs Maintenance and QA/QC: 1 hr/(sensor and week) Annual Sensor Costs Service and QA/QC: Personnel 42. 08% 44 (including taxes) 13. 32% Yearly investment costs (including taxes) 18. 92% Yearly consumable costs 34. 79% Yearly consumable costs for QA/QC 4. 21% Maintenance and QA/QC: 2 hr/(sensor and week) Yearly Annual Sensor Costsinvestment costs Service and QA/QC: Personnel 59. 23% Yearly consumable costs 24. 49% Yearly consumable costs for QA/QC 2. 96% Maintenance and QA/QC have highest impact on cost The sensors with the lowest maintenance requirements are the cheapest

Costs vs. Savings TOTEX HRSD Nansemond Treatment Plant as example (last webinar) § Energy consumption: -5% § Energy cost: -10% § Supplemental carbon: -50% Estimated Net Benefit Sensor Maintenance 45 Energy reduction 10% Energy Reduction 15% Reduction in Suppl. Carbon 2. 0 hr/(sensor&week) Net negative ± 0 Net positive 1. 0 hr/sensor&week) Net negative Net positive 0. 5 hr/sensor&week) ± 0 Net positive The net benefit is driven by: Ø sensor maintenance Ø number of sensors Ø saving potential (energy, suppl. carbon, water quality cost, …)

Take Home Message Ø Real-time control gets more out of your plant Ø Understand the system (e. g. by using models) v Influent, process and system dynamics v Process constraints / control authority v Equipment constraints Ø Cost-benefit analysis on annual costs Ø Mitigate risk Ø Plan and check your Data Quality Ø Make sure your operators are happy 46

Presenter Contact Information Leiv Rieger Ph. D. , P. Eng. in. CTRL Solutions Inc. Oakville ON, Canada Email: rieger@in. CTRL. ca Website: www. in. CTRL. ca 47