Study And Application Of The Multi-AO Process Simulation System Based On ANN/ASM2D Model

Multistage anoxic-oxic(AO)process is a step-feeding wastewater treatment process that is developed based on the traditional anaerobic-anoxic-oxic(A~2O)process.This process has its own advantages,including high utilization rate of carbon sources,excellent nitrogen removal efficiency and low energy consumption.However,due to the intrinsic properties of the sewage treatment process,i.e.nonlinearity,complexity and uncertainty characteristics,there are rare studies on the simulation system for multistage AO process to instruct its stable operation and adjustment.In this study,a simulation system for the modified multistage AO process was established based on ASM2 D model and artificial neutral network(ANN)model with input of in-situ operation condition parameters monitoring and kinetics parameters measurement,which could real-time feedback the system maintenance,adujest system stability and optimize the process performance.Series of treatment performance was carried out in a lab-scale multistage AO process,to identify the key factors affecting the effluent quality that were deemed as the main input parameters for the simulation system.The operation condition parameters,including the bulk oxidation-reduction potential(ORP),dissolved oxygen(DO),temperature(T),mixed liquor suspended sludge(MLSS),and p H were found to be rather important as the input parameters for simulation system.The correlation analysis between the effluent qualities and abovemented paramters in every functional reactor were investigated to determine the monitoring point,and control zone.Accordingly,an in-situ parameters measurement and feedback system was established.Based on the on-line monitoring system of multistage AO process,the key parameters for the ANN system were identified,and four kinetics parameters and stoichemisty parameters with high sensitively to ASM2 D model,i.e.heterotrophic yield coefficient YH,autotrophic maximum specific growth rate ?ATU,heterotrophic maximum specific growth rate ?H,and endogenious coefficient of b H,were determined in batch-tests.By combination of ASM2 D model and ANN model that involved the continuous input of ORP,p H,DO,MLSS and T,a simulation system of multistage AO process was successfully established.By the nonlinear mapping relationship between the ASM2 D model output and water quality measurement,the multistage AO process performance could be online predicted.After the simulation system calibration,the system was used to regulate the multi-stage AO process performance in real-time by on-line adjust the operation condition parameters,during which the microbial community response under varied operation condition were also analyzed.The results showed that the system could effectively regulate the multistage AO system to achieve stabile effluent quality with energy saving and consumption reduction.