Improving Effluent Total Nitrogen Using Advanced Dynamic Control of Aerobic/Anoxic Phasing
INTRODUCTION
The 12.0 MGD North Cary WRF employs BioDenipho® BNR process consists of three trains of paired oxidation ditches where the active process volume is continuously alternated between aerobic and anoxic conditions to achieve biological nitrogen removal. The current influent flow rate is approximately 6.5 MGD, or 55% of design flow and one train remains offline. Each BioDenipho process train is followed by multi-stage post-denitrification reactors and a post-aeration zone. Faced with the prospect of increasingly stringent TMDLs for the Neuse River Drainage Basin, the facility evaluated upgrade alternatives and process modifications to further decrease effluent TN. One of several options investigated was employment of advanced process control through the use on-line nutrient analyzers.
Figure 1. North Cary WTF
BACKGROUND
Each biological process train at North Cary WRF consists of an anaerobic selector for biological phosphorous removal, a pair of oxidation ditches, and a secondary anoxic zone for nitrate trimming (Figure 1). Each of the oxidation ditches is equipped with four surface brush aerators (60 HP), one submersible aerator (40 HP) and two submersible mixers. In addition, each oxidation ditch is equipped with a dissolved oxygen (DO) probe for automatic DO control. Supplemental carbon addition to the secondary anoxic zone is currently not required to meet TN removal goals.
North Cary WRF currently does not have permit limits for TN in concentration units. Based on a design flow of 12 mgd, the WRF are required to be less than 4 mg/L TN leaving the plant discharge. The discharge is mass based and currently is 64,965 kg (143,246 pounds). But just meeting the limit is not the Town of Cary's goal. The Town of Cary strives to be the leader in environmental stewardship and are not satisfied with just meeting the limits. It is the town’s goal to explore other technologies to further lower nutrients discharged from the facility. The Town of Cary expects that only more stringent limits will be imposed in the future and want to investigate other alternatives that lower the TN values to an enviable status. The Town of Cary does not want to be behind the eight ball in terms of nutrient reduction, but in the forefront as leaders in the State of North Carolina
The BIO-DENIPHO® process accomplishes both nitrification and denitrification within a pair of ditches though a repeating cycle alternating anoxic and oxic phase pairs (Figure 2) without employing separate anoxic reactors or internal recycle streams. Typically, the phase lengths are manual inputs into the control system and remain fixed until changed by the plant operator. Changes in the lengths of the main phases change the effective aerobic volume between 40% and 100% of the total process volume with the remaining volume allocated to anoxic operation for denitrification. These “fixed” phase length are typically set conservative so as to ensure complete nitrification during periods of high loading and therefore not offer the most optimal process control for either effluent Total Nitrogen or energy consumption.
Typically only four out of the 10 available phases are used during Fixed Phase Control in a repeating sequence as shown in Figure 3. Other phases noted in Figure 2 may be activated during prolonged periods of high or low influent loading. Specific examples are low loads following new plant commissioning where the recent housing market crashes stalled planned development and a plant serving a major university that incorporated a “game” mode to accommodate high loads during football events.
Figure 2. Sequential phase pairs for phased ditch operation
Figure 3. Main operating phases of the BioDenipho process during Fixed Phase Control
Previous studies demonstrated the existence of an optimal operation cycle length for nitrogen removal (Nielsen, M.K. 2001). It has been shown that, for a given set of process conditions, an ideal cycle length exists when denitrification just comes to completion (i.e., nitrate is almost completely consumed) and the next aerated nitrification phase should be scheduled to begin. This corresponds to the cycle lengths, which minimize the concentration of nitrogen in the process effluent. Of course, process conditions such as influent loading and wastewater temperature are not constant and fixed phase length operation cannot provide optimal process performance for either effluent total nitrogen or energy consumption. Therefore, a control strategy has been developed in which the cycle length is automatically adjusted as process conditions vary (Potter, T.G. et al 1996). And now with the improvement of reliability, accuracy, and ease of maintenance for nutrient process analyzers, the information required to successfully implement dynamic variation of phase length (Thomsen, H.A. and Kisbye, K. 1996) is available for real time control.
Figure 4 shows a typically 24-hour dynamic in one of the ditches with a Fixed Phase Control strategy. Typically phase lengths are input by the operator to the control system and remain unless long term changes in influent loadings warrant doing so.
A 24-hour dynamic with a Fixed Phase Control. (North Cary WWTF)Black underline indicates wasted Nitrification time and Green underline indicates wasted Anoxic time
The sinusoidal pattern of the ammonia and nitrate is the result of phasing between oxic and anoxic conditions for nitrification/denitrification. The periods (Black and Green underlines), where either the ammonium or nitrate concentration remains near zero for extended periods compromise TN removal performance and result in wasted energy. Terminating a phase at an acceptable concentration of ammonia or nitrate just before the concentration reaches zero effectly clips off the rising peak of its counterpart parameter which continues to rise if the phase is not terminated. The time required to reduce the concentration to the termination point will vary with incoming load and temperature. Dynamic Phase Control overcomes the limitations imposed by Fixed Phase Length control by varying the operating time allocated to nitrification and denitrification in response to process requirements.
Figure 5 shows a 24-hour dynamic in one of the ditches with a Dynamic Phase Control strategy. As is clear of figure 4, that using a criteria for phase selection and shorter phase length, will result in a much lower overall concentration of ammonium and nitrate in the ditch.
Figure 5.A 24-hour dynamic with a dynamic phase control strategy. (North Cary WWTF)
More than 20 years of Full-scale experience performed at other plants have successfully demonstrated that Dynamic Phase Control improves overall plant performance resulting in lower effluent TN concentrations reduced aeration energy consumption, and potentially increasing treatment capacity (Nielsen, et. al. 1993) and (Thomsen, H.A. 2009). In the United States, a case study on Dynamic Phase Control performed in single oxidation ditch mode at the Kill Creek WWTF (Gardner, KS) reduced effluent TN from approximately 4 mg/L to 2.5 mg/L.
The North Cary WRF staff conducted a six-week study to evaluate the impact of dynamic phase control (STAC) on effluent total nitrogen and aeration energy consumption. Secondary goals of the study were:
- Incorporate the on-line nitrogen analyzer(s) into PLC-based control system and utilization of the data for process control
- Rather than a sequential progression of phases according to Figure 1, develop criteria for phase selection that optimizes process performance.
METHODOLOGY
The study was performed on one of the process trains over a six week period with one week operation under Dynamic Phase Control followed by one week operation under Fixed Phased Length Control. This two week operational sequence was repeated three times during the study. A Chemscan analyzer monitored the ammonia and nitrate concentrations in each ditch of the paired ditch configuration. An ultra filter was used for filtering the samples from each ditch.
The two ditches were operated under Fixed Phased Control with a predetermined phase length of 90 for the main phases B/G and 30 minutes for the intermediate phases E/J and a duration of one full cycle of 4 hours, which results in a total aerated volume of 62.5%
Daily secondary effluent flow composite samples were collected at the clarifiers and analyzed for ammonium and nitrate during the six-week study. In addition the samples from the secondary effluent were analyzed for total nitrogen in the last two week of the study. Additionally, in order to better evaluate the Dynamic Phase Control (STAC) performance before the Post-denitrification and Re-aeration zones, time composite filtrated samples were collected directly from the outlet of the oxidation ditches.
RESULTS AND DISCUSSION
Figure 6 shows data from the on-line analyzer collected over a period where Fixed Phase and Dynamic control were implemented sequentially. During Dynamic Phase Control, all 10 phases were available for selction by the control algorithem. Phase selection is based upon a set of criteria comparing nitrate and ammonia concentrations in both ditchs.
Flow composite data for secondary effluent nitrate are shown in Figure 7.
Figure 6. Dynamic of ammonium and nitrate in oxidation ditch 1 under FixedPhaseControl and Dynamic Phase Control
Figure 7. Comparison of effluent Nitrate during Dynamic Phase Control and Fixed Phase Control at North Cary WWTF, North Carolina
Figure 6 and 7 indicates a strong trend of effluent improvement by using the Dynamic Phase Control, when compared to Fixed Phase Control. Secondary Effluent nitrate removal was according to table 1 improved by 67 % by using Dynamic Phase Control averaging less than 0.5 mg/L compared to 1.2 mg/L under fixed time control. At the same time there was no detectable increase in the level of ammonium concentration, which during the six week study was constantly below 0.1 mg NH4-N/l. The time composite samples taken from the effluent of the ditch also showed a significant reduction in nitrate concentration 55% averaging 1 mg NO3-N/L with Dynamic Phase Control and 2.2 mg NO3-N/L with Fixed Phase Control.
Table 1 summarizes the average nitrate concentrations in the two sample points, directly from the outlet of the ditch (prior to the secondary anoxic zones) and in the secondary clarifier effluent during Fixed Phase Control and Dynamic Phase Control.
Table 1. The average Nitrate concentrations in the outlet of the ditch and in the outlet of the secondary clarifier
Figure 8 shows daily composite total nitrogen samples collected after the secondary clarifier and analyzed by a certified laboratory during a two-week period (7/16-8/5)
Figure 8. Comparison of effluent TN under Dynamic Phase Control and Fixed Phase Control at North Cary WWTF, North Carolina
The results in figure 8 suggest that total nitrogen has been reduced from an average of 3.6 mg TN/l to 2.2 mg TN/l, which is a reduction at approx. 40% of Total Nitrogen by switching to the Dynamic Phase Control.
The phase lengths shown by the SCADA system during Dynamic Phase Control suggest a typical aerobic operation time of less than 50%, while the Fixed Phase Control have an aerobic operational time of 62.5%. The shorter aerobic operational time implies less aerator runtime. This is consistent with the SCADA data obtained during the 6 weeks case study at North Cary WRF, which indicated an average 10-12% decrease in aerator runtime during the Dynamic Phase Control period.
The Cary North WRF has a yearly permit to discharge 64.965 kg TN/year, which at full capacity (12 MGD) equalize 3.9 mg TN/L. The effluent demand for ammonium is 1.0 mg N/L (Nov-Apr) and 0.5 mg N/L (May-Oct). The amount of total nitrogen discharged from North Cary WWTF in the year 2009-2010 is summarized in table 2. A conservative estimate of the nitrate reduction of only 60% has been used to estimate the total nitrogen discharged from the treatment plant with Dynamic Phase Control (STAC), even if the study indicated a reduction of nitrate up to 67%.
Table 2. Estimated Total Nitrogen discharge with Fixed and Dynamic Phase Control.
1. Other process improvements such as RAS addition to the secondary anoxic zone and better DO control implemented in 2010 achieved also significant reductions in effluent TN compared to Year 2009.
The plant has achieved good effluent quality using the phased oxidation ditch operation with a Fixed Phase Length (Table 3) averaging 4.0 mg TN/L in 2009. In 2010 other process improvements such as RAS addition to the Post Anoxic zone and an improved DO control decreased the average TN in the effluent to 2.9 mg TN/L. However, Dynamic Phase Control is predicted to be able to reduce the effluent TN even more to below 1.9 mg/L, allowing the plant to meet its goal for TMDL compliance. Combined with the other process modifications, the facility expects to achieve between a 50 and 60% reduction in effluent TN compared with 2009 without the use of supplemental carbon.
The Dynamic Phase Control System (STAC) has been implemented as the long term control method at another Full-scale BioDenipho Wastewater plant in the town of Highspire, Pennsylvania (2.0 MGD). Figure 9 shows the on-line data from the Chemscan unit with Fixed phase Control (40% DN) and with Dynamic Phase Control. The results obtained in the full scale BioDenipho plant in Highspire confirm all the results from the case study at North Cary WRF. Figure 9 indicates a strong trend of Nitrate effluent improvement by using the Dynamic Phase Control, when compared to Fixed Phase Control. Results from the Laboratory confirm that the nitrate removal in the effluent was improved by 78 % by using Dynamic Phase Control. The Nitrate in the effluent averaging less than 1.0 mg NO3-N/L compared to 4.5 mg NO3-N/L under Fixed Phase Control, with no detectable increase in the level of ammonium concentration. The SCADA data on rotor runtimes also confirms the Cary North data. The period were the Fixed Phase Control was very short (3 days), but the data implies a reduction of 15% in rotor runtime.
Figure 9. Dynamic of ammonium and nitrate in oxidation ditch 1 with Fixed Phase Control and Dynamic Phase Control at Highspire WWTP, Pennsylvania.
CONCLUSION
Although high effluent quality has been achieved by using predetermined phases and anoxic/oxic phase length at North Cary WRF, the process can be further optimized by incorporation of Dynamic Phase Control. The North Cary case study and the results from Highspire WWTP indicated a significant improvement in TN removal (40% without chemical addition) and reduced aerator runtime (10-12%) by incorporating Dynamic Phase Control. The control criteria strategy for phase selections developed for the control methodology was proven effective for optimization of the process performance. The benefits of Dynamic Phase Control can be significant energy savings and reduced supplemental carbon dosage for nitrogen removal in some WWTPs.
REFERENCES
Nielsen, M.K. (2001). Control of Wastewater Systems in Practice. IWA – ICA, Scientific and Technical Report, Part 3.
Potter, T.G., Koopman, B. and Sovoronos, S. (1996). Optimization of a Periodic process for Nitrogen Removal from Wastewater. WaterResearch, 30, 142-452
Thomsen, H.A. and Kisbye, K. (1996). N and P On-line Meters; Requirements, Maintenance and Stability. Water Science and Technology. 33, 147-157.
Thornberg, D. E.; Nielsen, M. K.; Andersen, K. L. (1993) Nutrient Removal: On-Line Measurements and Control Strategies. Water Science and Technology (G.B.), 28(11-12), 549.
Thomsen, H.A. and Ohnerth, T.B. (2009). Results and Benefits from Practical Application of ICA on more than 50 Wastewater Systems over a Period of 15 years. Proceedings of IWA-ICA, Cairns, Australian.