Determination of the Critical Anaerobic HRT to Maintain Settleability at Low DO

Introduction and Background
Low dissolved oxygen (DO) operation (0.3 — 0.6 mg/L) enables biological nutrient removal (BNR) facilities to reduce energy, but concerns about process performance, specifically sludge settleability, limit full-scale adoption. Low DO can lead to filament overgrowth, causing bulking sludge and increased clarifier pumping needs, reducing clarifier efficiency and raising costs (Jenkins et al., 2003).

In the recent Water Research Foundation Project 5083, two parallel pilot wastewater treatment trains were operated at high DO (1.5-1.8 mg/L) and low DO (0.3-0.5 mg/L) with two different configurations (A2/O and A/O). While both configurations had anaerobic zones that are needed for biological nutrient removal, specifically for the selection of phosphate accumulating organisms (PAOs), the A/O had an extended anaerobic hydraulic residence time (1 hour) compared to the A2/O train (0.5 hours). Selector zones upstream of aerobic basins in plug flow BNR systems, such as anaerobic or anoxic zones, can mitigate bulking sludge, as they promote the growth of desirable floc-forming microorganisms over filamentous bacteria by creating a competitive advantage for readily biodegradable carbon. Therefore, limiting the COD entering the aerobic zone that can result in filament growth by increasing the anaerobic HRT would result in better settling sludge.

The WRF study showed that excellent sludge settleability can be achieved at low DO concentrations of 0.3 — 0.5 mg/L. At this low DO, the A/O train maintained good settleability with SVI30 consistently below 110 mL/g. The SVI5/SVI30 ratio was 1.62, which was slightly worse than the 1.25 ratio at the higher DO. In contrast, the A2/O settleability worsened at the low DO condition, with an average SVI30 of 243 ± 63. In the final phase of the experiment where low DO was maintained while the anaerobic HRT increased by 150%, the SVI30 for the A2/O and A/O was very similar at 107 ± 12 and 107 ± 8, respectively. The A2/O settleability improved with a SVI5/SVI30 of 1.35 ± 4.1 and the A/O settleability ratio was maintained at 1.55.

For consistent full-scale implementation of low DO BNR, it is critical to determine the necessary residence time for anaerobic selector zones, especially for the more common A2/O configuration. This work will systematically test the relationship of anaerobic HRT and settleability at low DO concentrations, by understanding the COD availability, EPS content, particle size, and floc vs filament structure and abundance.

Objective:
Optimize anaerobic selector zone sizing for effective sludge settleability at low DO and determine the corresponding F:M, carbon uptake and particle sizes in the anaerobic zone

Methods
A 200-L pilot-scale reactor was operated as an A2/O process, described in WRF5130 Report (Figure 1). Previous results showed that A2/O configuration experienced filament growth higher SVIs until an anaerobic HRT calculated with RAS exceeded 1 hour. Additionally supplemental carbon as MicroC 2000 was needed, in order to increase the C:N ratio of 9.8 ± 3.2 (as Total COD:Total N), to achieve biological nutrient removal. This C:N ratio will be maintained for the following:
Anaerobic HRT (min) - 20 40 60 80 100 120 140
After 3 SRTs are reached at each anaerobic HRT, samples for microbial community analysis are taken. COD profiles are taken across the entire reactor.
Additionally, SVI, EPS (loosely and tightly bound), particle size, filament and floc imaging of anaerobic samples are completed.

Results
Initial low DO results when the anaerobic HRT was 40 mins, the A2/O configuration had an average SVI30 of 190 mL/g with an SVI5/SVI30 ratio of 2.03 ± 0.48 (Table 1). Images of the floc structure after low DO conditions were established show that larger floc structures greater than 212 µm were not obtained. Once the anaerobic HRT was increased and sufficient carbon was added to the system, the SVI30 decreased to 43 ±± 9 mL/g. The ratio of SVI5/SVI30 was 1.16 ± 0.1, indicating granular sludge formed with stable operation. Additionally, biological phosphorus removal and denitrification was achieved, as effluent phosphate was between 0.1 to 0.2 mg-P/L and effluent TIN was 2.8 ± 1.8 mg-N/L. Overall, the results show that efficient nitrification, denitrification, and phosphorus removal were achieved in at low DO when carbon was available. The performance far exceeds the required Kansas NPDES performance with consistent effluent TN below 5 mg-N/L and TP below 0.5 mg-P/L. A chemical profile across each treatment process shows that sCOD was consistently removed in the first anaerobic zone (Figure 2). This is a critical function of an anaerobic selector since the breakthrough of sCOD to aerobic zones is known to select filamentous organisms. However, there was a clear change in settleability between the lower and higher anaerobic HRT, motivating the need to dive deeper into the carbon dynamics present in low DO systems.

Conclusions
Increasing the anaerobic HRT in the A2O train achieved good sludge settleability. The anaerobic selector effectively removed all soluble COD, with an anaerobic F:M greater than 0.8 mg COD/mg VSS-d as soluble COD, sufficient for sludge densification. With sufficient carbon availability, granule formation improves, and very low and stable SVIs are achievable. These findings will provide a design guide for more confident full-scale implementation of low DO BNR systems.