Simulation of long-term nutrient removal in a full-scale closed-loop bioreactor for sewage treatment: an example of Bayesian inference

Zheng Li,Rong Qi,Wei An,Takashi Mino,Tadashi Shoji,Willy Verstraete,Jian Gu,Shengtao Li,Shiwei Xu,Min Yang
DOI: https://doi.org/10.1007/s11783-014-0660-2
2014-02-22
Abstract:In this study, the performance of nitrogen and phosphorus removal in a full-scale closed-loop bioreactor (oxidation ditch) system was simulated using the ASM2d model. Routine data describing the process for two years were compiled for calibration and validation. To overcome the identifiability problem, the classic Bayesian inference approach was utilized for parameter estimation. The calibrated model could describe the long-term trend of nutrient removal and short-term variations of the process performance, showing that the Bayesian method was a reliable and useful tool for the parameter estimation of the activated sludge models. The anoxic phosphate uptake by polyphosphate accumulating organisms (PAO) contributed 71.2% of the total Poly-P storage, which reveals the dominance of denitrifying phosphorus removal process under the oxygen limiting conditions. It was found that 58.7% of the anoxic Poly-P storage and denitrification by PAO in the reactor was achieved in the aerated compartment, implying that the PAO’s anoxic activity was significantly stimulated by the low dissolved oxygen (DO) level in this compartment due to the oxygen gradient caused by brush aerator.
environmental sciences,engineering, environmental
What problem does this paper attempt to address?