Newswise — Tongji University researchers conducted a study titled “Moving Bed Biofilm Reactor for Blackwater Treatment: Insights into Pollutant Removal, Microbial Community, and Water Quality Prediction Using Machine Learning.” This research was published in Frontiers of Environmental Science & Engineering, Volume 19, Issue 8.
Blackwater is a major source of domestic wastewater pollutants, including high concentrations of COD, TN, and TP. Traditional treatment technologies such as septic tanks have limitations such as low efficiency, occupying large amounts of land, and low resistance to environmental changes. The objective of this research was to develop an efficient and robust blackwater treatment technology. Researchers established a new two-stage anoxic moving bed biofilm reactor (A/O-MBBR), operated it for 82 days, and investigated its pollutant removal performance under different hydraulic retention times (HRTs) and low temperature conditions. They also developed an Extreme Gradient Boosting (XGBoost) machine learning model to analyze microbial community structure and predict contaminant concentrations within nuclear reactors. Key findings include that the reactor achieved 94.4% COD, 99.7% NH3-N, and 74.6% TP removal when HRT was 25.5 h and maintained stable performance under shortened HRT and low temperature (8–15°C). Attached biofilms play an important role in nitrification, with major microbial genera such as Thiothrix, Azospira, Acinetobacter, and Tauera contributing to nutrient removal. The XGBoost model showed high predictive accuracy (R² > 0.9) for pollutant concentrations, with temperature and HRT identified as the main influencing factors. This study not only demonstrates the superior performance of two-stage A/O-MBBR in blackwater treatment but also provides a data-driven approach for process optimization.
For more information, the full paper is available at https://doi.org/10.1007/s11783-025-2022-7.
