MODEL-AIDED METHODOLOGY FOR SYSTEMATIC EXPLORATION OF THE OPTIMAL OPERATIONAL DOMAIN OF NUTRIENT REMOVING SBR SYSTEMS

In this study, a computer-aided systematic approach to determine the optimal operation strategy for nitrogen (N) and phosphorous (P) removal in sequencing batch reactors (SBRs) has been developed and evaluated at a pilot-scale SBR. The methodology developed is based on using a grid of possible scenarios to simulate the effect of the important degrees of freedom in SBR operation such as number and sequence of process phases, step feeding, dissolved oxygen set-point, etc. In the case study, the grid of scenarios is simulated using a calibrated ASM2dN model developed in a previous study. Effluent quality in combination with robustness criteria is used to select the best scenario. The implementation results of the best scenario showed that the SBR performance was improved by approx. 50% and 40% for total nitrogen and phosphorous removal respectively, which was better than what the model had predicted. However, the long-term SBR performance was found unstable. When confronted with reality observed under the new operating conditions, the model used for the optimisation of SBR operation was found to be invalid. Among others, the model was unable to predict the nitrite-build up provoked by the implemented best scenario. These results imply that drastically changing the operation of an SBR system using a model may significantly change the behaviour of the system beyond the (unknown) application domain of the model. In view of reaching the target of the model-based optimisation, the systematic methodology was therefore iterated a second time. To this end, a model update step was performed, i.e. a 2-step nitrification and 2-step denitrification version of the ASM2d model (ASM2d2N) was developed and used to find a new best scenario for optimal operation. Finally, to improve the systematic methodology it is proposed to explicitly involve expert knowledge during the decision making step in an attempt to make up for any inadequacy associated with modelling SBR systems, particularly the inability of models to predict settling under different operating conditions.