Improving a Denitrification System
Methanol dosing system meets all performance requirements
Federal and state requirements for improved effluent quality from wastewater treatment works have led to the development of tertiary treatment processes for the removal of nitrate-nitrogen and suspended solids. Deep-bed denitrification filters, the combination of denitrification and solids removal in a deep-bed filter process patented in 1979, have proven to be a highly effective treatment technology used by wastewater plants to meet low total nitrogen limits.
But when feeding a carbon source, such as methanol, for tertiary denitrification, dosage rates that are too small can lead to excess nitrate-nitrogen levels. Higher methanol to nitrate-nitrogen chemical dosage ratios can be helpful in ensuring that high levels of denitrification are achieved. However, if methanol is overfed, it will result in elevated effluent biochemical oxygen demand (BOD) concentrations. For every extra ppm of methanol overfeed, the BOD is increased by 1.5 ppm.
Stringent Methanol and Backwashing Limitations
At the Lower Reedy Wastewater Treatment Plant (WWTP) in Simpsonville, S.C., a suburb of Greenville, DeepBed Denite filters from Severn Trent Services were installed to reduce the effluent total suspended solids prior to ultraviolet disinfection. Additionally, the filters would provide the capability for increased denitrification capacity if more stringent future total nitrogen discharge limits were imposed.
As part of the Lower Reedy plant’s commissioning requirements, a rigorous performance test was conducted to assess the filters’ ability to meet the required effluent nitrate-nitrogen + nitrite-nitrogen (NO2-N + NO3-N, or NOx-N) concentrations without using an excess of methanol. To meet the limitation on methanol usage, the plant used a control system that included Severn Trent Service’s TetraPace automatic dosing control system for chemical feed control. A strict limitation on backwash frequency also was incorporated into the performance testing.
The design criteria for the Lower Reedy denitrification filters calls for removal of NOx-N to concentrations of less than 1.0 mg/L, while limiting the methanol feed. In addition, the backwashing rate of 6 gpm/ft2 affects the energy required to operate the system, and increased backwashing translates to higher spent backwash water return flows to the main plant. These methanol and backwashing limitations were some of the most stringent requirements ever to be specified for this type of system.
The performance test was conducted during winter months—February and March—to prove that the filters could provide the required treatment under cold weather conditions. After approximately six weeks of preliminary operation to establish the biomass, the test was conducted over a week-long period. Because the Lower Reedy WWTP was not equipped with a methanol feed system, a temporary system, including the feed pumps and analyzer, was supplied by Severn Trent Services to execute the performance test. Additionally, the filter influent wastewater was supplemented with sodium nitrate to simulate the design NOx-N loading conditions.
The automatic dosing control system was used to regulate the methanol feed rate throughout the performance test. During the biological denitrification process, wastewater is forced to flow around nitrogen gas bubbles that accumulate in media voids in the filtration vessel, improving biomass contact and filtration efficiency.
Effective removal of nitrate-nitrogen is undertaken by introducing methanol using TetraPace. This control system uses the filter influent flow rate and the influent and effluent NOx-N concentrations to attain an operator-inputted setpoint value for effluent NOx-N concentration. This is done by continuous, automatic adjustments of the methanol dosage rate.
An alternative to this system could incorporate a flow-paced or feed-forward or feedback system. But TetraPace is far more efficient. The advantages of tighter methanol control can be significant if the plant has a stringent biochemical oxygen demand (BOD) limit in combination with a low total nitrogen limit. Under these conditions, the tighter control and reduced risk can be critical components in ensuring the plant meets limits reliably.
The accuracy of the proprietary algorithm used to feed methanol during the denitrification process enables the automatic dosing control system to yield savings of up to 30% in methanol consumption costs, while guaranteeing effluent quality with a “no net TOC pickup across the filter” guarantee.
The filters achieved a backwash percentage of less than 2% of the treated flow when operating at an average loading rate of 2.2 gpm/ft2. All performance requirements were met, illustrating the effectiveness of the process and methanol control system. This superior performance was achieved on the first and only performance test. All goals and guarantees were met to the satisfaction of the Western Carolina Regional Sewer Authority, the owner of the Lower Reedy WWTP, and Black & Veatch, the project’s architect and engineer.
Several operations issues were observed:
- The methanol dosing ratio of 3.5 lb methanol/lb NOx-N removed was slightly higher than the typical dosage ratio of 3:3.2, but this resulted from the very high dissolved oxygen concentrations present in the secondary effluent feed to the filters.
- Once the filters were placed in operation, a noticeable change in the upstream activated sludge biological nutrient removal (BNR) process was observed. As the result of additional solids generated in the filters and returned to the plant, solids inventories in the BNR process increased, impacting the required wasting rate more than was anticipated.
- The methanol control system worked very well and was the key to achieving the required performance results. However, the methanol control system algorithm is based on nitrate readings. For a few days at the beginning of the testing, elevated NO2-N concentrations were observed in the secondary effluent feed to the filters. Since NO2-N was not part of the control algorithm, the system appeared to slightly underdose methanol. The methanol dosing was increased through an adjustment to the control system setpoint, which ultimately improved overall performance. Therefore, NO2-N concentrations must be monitored periodically to ensure the NO3-N-based control system is responding as desired.