Partial Denitrification Model Extension for Mainstream Shortcut Nitrogen Removal: Model Evaluation

As the development of mainstream partial denitrification and anammox (PdNA) process continues to evolve, and more applications are showing tangible benefits even at full-scale application [Klaus 2020], the modelling of such technology is becoming critical for its implementation at more broad scope. PdNA relies on controlling heterotrophic denitrification to produce nitrite needed by anaerobic ammonia oxidizers (AnAOB) and thus the avoidance of nitrite reduction [De Clippeleir, 2018]. While the PdN selection is still in the realm of research, few hypotheses supported by observations from bench, pilot and fullscale testing were formulated [Tri et al., 2019b; Campolong 2019, Al-Omari et al. 2020]. During WEFTEC 2020, we introduced the early development of the PdN extension to the ASM-based model Sumo2 to account for a mechanism associated with using acetate as carbon source and selecting PdN route based on nitrate concentration. The model was calibrated using data generated from a pilot testing with acetate. In this paper, the extended PdN model is evaluated via simulating implementation in post polishing MBBR. The process was operated using glycerol and methanol as supplemental carbon sources. Overall, the model predicted the performance of these processes when glycerol was used. When methanol is used, the model was underpredicting the nitrite accumulation by methylotrophs. The model parameters such as nitrite half saturation constant were significantly increased to allow nitrite to accumulate and match the measured profiles.