Long-term simulation of a full-scale EBPR plant with a novel metabolic-ASM model and its use as a diagnostic tool.

This study evaluates the predictive capacity of the META-ASM model, a new integrated metabolic activated sludge model, in describing the long-term performance of a full-scale enhanced biological phosphorus removal (EBPR) system that suffers from inconsistent performance. In order to elucidate the causes of EBPR upsets and troubleshoot the process accordingly, the META-ASM model was tested as an operational diagnostic tool in a 1336-day long-term dynamic simulation, while its performance was compared with the ASM-inCTRL model, a version based on the Barker & Dold model. Overall, the predictions obtained with the META-ASM without changing default parameters were more reliable and effective at describing the active biomass of polyphosphate accumulating organisms (PAOs) and the dynamics of their storage polymers. The primary causes of the EBPR upsets were the high aerobic hydraulic retention times (HRTs) and low organic loading rates (OLRs) of the plant, which led to periods of starvation. The impact of these factors on EBPR performance were only identified with the META-ASM model. Furthermore, the first signs of process upsets were predicted by variations in the aerobic PAO maintenance rates, suggesting that the META-ASM model has potential to provide an early warning of process upset. The simulation of a new viable operational strategy indicated that troubleshooting the process could be achieved by reducing the aerated volume by switching off air in the first half of the aeration tank. In this new strategy, the META-ASM model predicted a simultaneous improvement in the biological phosphorus (P) and nitrogen (N) removal due to the enhancement of the hydrolysis and fermentation of the mixed liquor sludge in the new unaerated zone, which increased the availability of volatile fatty acids (VFAs) for PAOs. This study demonstrates that the META-ASM model is a powerful operational diagnostic tool for EBPR systems, capable of predicting and mitigating upsets, optimising performance and evaluating new process designs.