Optimizing Nutrient Removal at a 75 mgd ENR Facility – From Process Modeling to Reality

Authors:

  • Katya Bilyk PE, David Wankmuller PE, Paul Pitt PE, PhD, Dana Fredericks, Ron Latimer, James Grandstaff, Perry Green, Joshua Irby

The existing liquid and solid treatment processes are depicted above. Influent BOD has been steadily decreasing since 2006, requiring the facility to use a significant amount of supplemental carbon (glycerin) for denitrification to meet their effluent TN limit.

Extensive historical data was reviewed, and a year-long dynamic BioWin model was calibrated using data from a detailed special sampling event. As a result of the study, an SND pilot was implemented in August 2015. One tank was used for the SND pilot, and an identical tank was used for the control.

The most effective DO control pattern in the SND tank was keeping the DO lowest in Cell 6 (the first aerobic zone) and gradually increasing DO downstream.

Operating in an SND mode whenever possible could save the facility on the order of $500,000 per year in glycerin costs, while still meeting the low effluent TN limit of 5.0 mg/L. Because of this significant savings potential, the County piloted this mode of operation and was able to achieve a 40 percent reduction in carbon use without sacrificing nitrogen removal performance.

Objectives
This paper summarizes the results from a nutrient optimization project for the Henrico County Water Reclamation Facility (HCWRF) that led to operational changes and reduced operating costs. The project included updating an existing BioWin model and developing different scenarios for optimization of the existing nutrient removal process. One of the most promising optimization configurations identified was operating the enhanced nutrient removal (ENR) tanks in a simultaneous nitrification denitrification (SND) mode. Operating in SND mode at current conditions offers the plant substantial operational savings on the order of $500,000 per year.

Background
The HCWRF has a design flow of 75 mgd with a current annual average flow of 41.8 mgd. The ENR system is comprised of six small tanks and two large tanks each configured as a six-stage process. The allowable effluent TN and TP concentrations corresponding to mass allocations at the design flow of 75 mgd are 5.0 and 0.5 mg/L, respectively. Influent BOD has been steadily decreasing since 2006, requiring the facility to use a significant amount of supplemental carbon (glycerin) for denitrification to meet their effluent TN limit.

Methodology
Extensive historical data was reviewed, and a year-long dynamic BioWin model was calibrated using data from a detailed special sampling event. Multiple optimization scenarios were analyzed using the calibrated model, and include: optimizing the NRCY rate based on primary effluent (PE) cBOD:TKN ratios, optimal glycerin split to the first and second anoxic zone, optimal RAS rate for nitrogen removal, and the benefit of SND.

As a result of the study, an SND pilot was implemented in August 2015. One tank was used for the SND pilot, and an identical tank was used for the control. The control tank was operated with the typical DO strategy of maintaining a 2.0 mg/L DO in the aerobic zone and tapering oxygen to 1 mg/L in the final aerobic cell.

The SND tank received less flow than the control to compensate for the slower nitrification kinetics. In full-scale implementation, the plant would increase the SRT in the system to compensate but this was not possible in a pilot as there is no dedicated clarifier for the SND tank.

Two process control strategies were developed and programmed into the existing SCADA system to control the SND tank: ammonia-based and a timer-based DO control mode. During the pilot the ammonia-based DO control mode was implemented, where DO in each zone was controlled within a user-specified range to achieve an ammonia setpoint at the end of the SND tank aerobic zone. Provisions were made so that if the target airflow is below the minimum mixing airflow for more than 10 minutes, there was a 3-minute purge program for resuspension of the MLSS.

A total of four Hach probes were used for controlling and monitoring the SND pilot and control tank. An ammonia probe was placed in cell 9 of the SND pilot to monitor ammonia leaving the aerobic zone. The probe was placed in Cell 9, rather than Cell 10 (the end of the aerobic zone), since ion selective ammonia probes cannot reliably read below 1.0 mg/L. It was assumed that a value near 1.5 mg/L in Cell 9 would correspond to full nitrification in Cell 10. Nitrate at the end of the aerobic zone was used to monitor nitrate load entering the second anoxic zone. Nitrate probes were also placed at the end of the SND tank and a control tank in order to compare the denitrification performance of each tank. The goal of the pilot was to have no ammonia and equal or less nitrate leaving the SND tank compared to the control while using less carbon.

Findings
The most effective DO control pattern in the SND tank was keeping the DO lowest in Cell 6 (the first aerobic zone) and gradually increasing DO downstream. At the beginning of the pilot the effluent nitrate from the SND tank is lower than the control, but this is due to incomplete nitrification—the ammonia probe setpoint in Cell 9 was 2.5 mg/L, and the ammonia concentration in Cell 10 was not negligible. The ammonia setpoint was reduced to 1.5 mg/L in Period 2, slightly increasing the SND tank effluent nitrate, but keeping it comparable to the control tank. During Period 3, carbon dose to the SND tank was reduced to see if additional carbon savings were achievable. During this time the SND tank effluent nitrate increased suggesting that there was insufficient carbon for comparable denitrification, so in Period 4, carbon dose was increased back to Period 2 levels. Late in Period 3 and throughout Period 4, the ammonia probe signal drifted upwards (ammonia was not as high as the probe indicated), resulting in higher DO concentrations for long periods of time as opposed to the typical DO cycling pattern in Periods 1 and 2. Until maintenance was performed, the SND tank effluent nitrate was higher than typical.

Nitrification was maintained in the SND tank throughout the pilot and on average the SND tank used 40% less carbon than the control tank. The 40% savings in carbon accounts for the SND tank receiving less flow than the control tank.

Summary
Results from the process optimization report confirmed outstanding operation of the facility and identified operational changes that may result in operational savings. A useful tool was developed for the facility to estimate glycerin dose as a function of the PE cBOD:TKN ratio. The operational modification with the greatest potential cost savings was to operate in an SND mode whenever possible. This configuration could save the facility on the order of $500,000 per year in glycerin costs, while still meeting their low effluent TN limit of 5.0 mg/L. Because of this significant savings potential, the County piloted this mode of operation and was able to achieve a 40 percent reduction in carbon use without sacrificing nitrogen removal performance.

For more information, please contact the author at kbilyk@hazenandsawyer.com.

Hear about new publications with our email newsletter

We will never share your details with anyone else.

Horizons

Newsletter Newsletter

Horizons Fall 2017 (pdf)

Horizons showcases significant water, wastewater, reuse, and stormwater projects and innovations that help our clients to achieve their goals, and can help you achieve yours. Articles are written by top engineers and process group leaders, demonstrating and explaining the beneficial application of a variety of technologies and tools.

View previous issues »