[Making optimal operation for a BNR process: modeling prediction and experimental verification].

Based on the process model of a BNR system (BCFS), the effects of operational parameters on the effluent quality were predicted by modeling, and were testified simultaneously by a lab-scale experiment, from which almost the same results in the modeling and the experiment were obtained. This means that modeling can be realizably applied to make the optimal operation schemes regardless of pilot-scale and/or full-scale experiments. Both the modeling and the experiment demonstrated that the bio-P removal performance was not influenced by the biomass amount in the anaerobic tank when the returned ratio (rA ) reached 1.5 and that rA had no significant correlation with COD and N removals. After the returned mixed liquor ratio (rB) increased over 2, the TN removal efficiency was not improved any more, and the COD and TP removals were not influenced by the variations of the rB. The returned mixed liquor ratio rC had almost no influences on the COD, TP and TN removals. Further, the COD and TP removals were not influenced when the dissolved oxygen (DO(R5)) in the aerobic tank was in the range of 1-2.5 mg x L(-1), but the effluent NH4+ -N increased over 1 mg x L(-1) when DO(R5 ) was below 2 mg x L(-1). So, the optimal operational parameters for the BCFS should be set at rA = 2, rB 2-2.5, rC = 0, DO(R5) 2-2.5 mg x L(-1).