The application of aerobic granular sludge (AGS) technology for treating real, low-concentration sewage (0.89 ± 0.26 kg COD/m3·day; COD: 195.6 ± 57.8 mg/L; C/N 9:1 C/N 9:1) at a pilot scale was investigated. Two sequencing batch reactors (SBRs) were started up with no inoculum and operated with different cycle configurations. SBR1 operated on a 4-h cycle with a 40-min non-aerated phase followed by 180 min of aeration, while SBR2 had a 15-min non-aerated phase followed by 205 min of aeration, treating real sewage to determine how feeding phase duration influences granulation, EPS dynamics, and nutrient removal under high-temperature and low-strength conditions. AGS formation occurred within 80 and 48 days in SBR1 and SBR2, respectively. Both reactors achieved COD and ammonia removal rates of over 70% and 80%, respectively. SBR1 showed superior COD removal (83 ± 8% vs 74 ± 15%) due to higher biomass retention, while SBR2 exhibited faster granulation and higher average EPS (203 ± 109 vs 157 ± 137 mg/g VSS). Nitrite accumulation in SBR1 (8.2 ± 5.3 mg/L) reflected free ammonia inhibition of NOB, whereas SBR2 produced more nitrate (4.4 ± 4.2 mg/L). EPS declined sharply to 33-49 mg/g VSS under low F/M ratios (<0.3 kg COD/kg VSS·d), indicating endogenous metabolism and structural maintenance. Microbial analysis revealed EPS-producing genera (Mesorhizobium 25.9% in SBR1; Paracoccus 24.2% and Devosia 10.3% in SBR2) that enhanced stability but limited complete denitrification under the tested conditions. These findings demonstrate that shorter non-aerated feeding accelerates granulation but can also reduce biomass retention, while longer feeding enhances pollutant removal but results in slower startup, these results provide critical design parameters for tropical AGS implementation with real sewage.
Sales et al. (Fri,) studied this question.