(555c) Model Predictive Control Strategy for Optimizing Biological Nitrogen Removal (BNR) Processes Accounting for Greenhouse Gas Emissions

Biological Nitrogen Removal (BNR) is a widely practiced and increasingly mandated process for treating wastewater. BNR involves the sequential oxidation of ammonia to nitrate (termed nitrification) and subsequent reduction of nitrate to nitrogen gas (denitrification).  Nitrification and denitrification are typically promoted under aerobic and anoxic conditions, respectively.   However, recent studies have shown that subjecting nitrifying and denitrifying bacteria to alternating aerobic and anoxic conditions- although useful for BNR, can lead to the release of copious amounts of nitrous oxide (N₂O).  N₂O is a greenhouse gas, 300 times as potent as CO₂ and is also the most powerful ozone depleting substance emitted during the 21^stcentury. Therefore, optimization, balancing and control of BNR processes is essential to: (i) improve the effluent quality, (ii) reduce the energy consumption, (iii) increase the amount of wastewater processed, and (iv) reduce the greenhouse gas emissions from wastewater plants.

In this work, a model predictive control strategy is utilized to control and optimize the BNR process.  A hybrid model of the BNR process is developed using the detailed studies conducted by Yu et al. on Nitrosomonas europaea (a widely studied nitrifying organism) under aerobic and anoxic conditions. The hybrid dynamic model is used to predict the effluent ammonia, nitrite, nitrous oxide and nitric oxide concentrations and fluxes produced in the BNR process for the given input conditions (namely, two substrates, influent ammonia and oxygen loadings). This model is used in a MPC framework to achieve the aforementioned objectives by manipulating the oxygen supply, rate of recycles in the system, and scheduling of aerobic and anoxic conditions in a modified BNR process.