Process Modeling and Aeration Control Design with ABAC for A/O SND Process with Densification

ABSTRACT: Boulder WRRF is under construction for upgrades for DAS with inDENSE™ and low DO A/O to promote Simultaneous Nitrification Denitrification (SND). Given the first-of-its-kind application of A/O SND in a cold climate with a daily effluent ammonia limit, detailed process modeling was conducted using BioWin, SUMO, and SIMBA#. Sensitivity analysis was conducted to assess the impact of kinetics on biological performance for the design. Aeration control using ABAC was modeled with feedback and feedforward strategies in BioWin and SIMBA#. Modeling showed this design can meet all the anticipated permit requirements for N and P. The ABAC strategy provided 3-13% energy saving compared with constant low DO operation. The models and aeration control approach will be validated with full scale data as part of BWRRF start up in early 2025.

PROCESS MODELING
BioWin Modeling: Calibrated and validated dynamic process models were developed in BioWin and SUMO to confirm performance. SIMBA# was used to support fine tuning of aeration control design. All platforms provided consistent results for the preliminary modeling efforts for using constant low DO approach. Table 1 compares the average effluent quality in BioWin and SUMO using default kinetics. Compared with BioWin, SUMO overpredicted denitrification and resulted in lower effluent TIN. For the permitting application, BioWin model (Figure 1) was used because it provided more reasonable performance. Key aeration design parameters for the DAS process, additional model scenarios and dynamic modeling graphs with BioWin and SUMO will be presented in the full paper. Process modeling was conducted using adjustments in kinetics (Table 2). For the model runs with adjusted kinetics, KDO of AOB and NOB were adjusted to 0.4 and 0.2 mg O2/L. The model with default kinetics provided slightly better N removal performance. Sensitivity analysis was conducted to assess the performance of the Aeration Basins (ABs) and Solids Contact Tanks (SCTs) with default and adjusted kinetics at high (1 mg/L) and low DO (0.3 mg/L) set points (DOSP) in SCTs. Figure 2 shows results from sensitivity analysis. With default kinetics, NOB concentrations drop significantly in SCTs (suggesting nitrite shunt) when reducing DO levels in SCTs while the NOB population remains the same with adjusted kinetics and provides more reasonable performance of the proposed process. Sensitivity analysis was also conducted in BioWin by including hydrocyclones with an assumed specified removal for Phosphate Accumulating Organism (PAO) of 1.5 to study the impact on EBPR (in full paper).

SIMBA# Modeling: Constant low DO and ABAC control strategies were used in SIMBA#. The ABAC strategy in SIMBA is presented in Figure 3. The location of the ammonia probes is shown in Figure 4. The scenario matrix used for SIMBA# modeling is presented in Table 3.

RESULTS Control Objectives: The design intent for aeration control included ABAC using conventional PI loops for feedback control with the following goals: Daily effluent ammonia limit of 1.0 mg N/L, TIN of 11 mg/L, and TP < 1 mg/L. Model calibration was conducted for the number and type of diffusers in each basin. Constant low DO strategy was also modeled in exploratory runs.

Modeling Results with SIMBA# ABAC Modeling with Feedback Control Strategy: Figures 5 and 6 compare air flows with constant low DO and ABAC for the existing system and the future A/O SND process (with valves and diffusers and new dual core blowers), respectively. Figures 7 and 8 compare the energy use with constant DO and ABAC for existing and future upgrades, respectively. Comparing the two control strategies, on average, ABAC consumes less energy than constant low DO strategy. Although differences in energy consumption are less than 5-6 % during high loading, the energy consumption under low loading can reach 13%. Energy consumption varies over a wider range for ABAC compared with low DO due to DO changes ranging from 0.1 to 3 mg/L. For all the high loading condition modeling scenarios, effluent ammonia and TP targets were met but effluent TIN was slightly elevated from target values at 11.3 C and 13.5 C. Under the low loading conditions, effluent targets for ammonia, TP, and TIN were met at all three temperatures.

ABAC Modeling with Feedback in BioWin: ABAC was modeled in BioWin using BioWin Controller using the distribution tool and default kinetics.

Comparison of Models for ABAC Modeling with Feedback Strategy: Though differences exist between the models in BioWin and SIMBA# for ABAC modeling with feedback strategy for DOSP in each aerobic zone, both showed that the proposed process can consistently meet the effluent daily ammonia goal of 1 mg N/L. Figure 9 compares the ammonia profile in the aerobic zones for BioWin and SIMBA with feedback control strategy. While BioWin uses a tapered aeration mode with the air distribution tool, SIMBA# uses the same DOSP in each of the aerobic zones. Figure 10 compares the DO profile with SIMBA# and BioWin.

ABAC Modeling with Feedback and Feedforward Step Gain in SIMBA#: A feedforward step gain (for DO) based on ammonia ranges in Zone 7 was developed during this project in SIMBA#, using default kinetics. The resulting DOSP was fed to all DO controllers in all basins. The advantages and disadvantages are discussed in the full paper.

SUMMARY: Modeling showed that the design can meet all the anticipated permit requirements.