Development of a novel model to quantify nitrous oxide emissions in the biological nutrient removal process of wastewater treatment plants

Abstract

This thesis aimed to develop a novel mathematical tool for the mitigation of nitrous oxide (N2O) emissions in wastewater treatment plants (WWTPs). Considering that N2O is a greenhouse gas with a grave global warming impact, tools for the prediction of N2O production in WWTPs are essential to accurately estimate the emissions and effectively reduce them. The first chapter reviewed past studies focusing on the N2O generation in WWTPs. The major findings underlined the need to optimise the applied WWTP processes and use models that consider multiple N2O production pathways and changes of majorly influencing operational factors (e.g. dissolved oxygen, DO). The second chapter presented the development of an N2O model following the widely accepted International Water Association (IWA) Activated Sludge Model (ASM) structure that described the operation of a full-scale anaerobic-anoxic-aerobic municipal WWTP. The simulation results showed that low-aeration strategies require optimisation to avoid unstable nitrification and increased N2O production via the nitrification-related pathways. The third chapter introduced an ASM-type N2O model for the operation of a real full-scale municipal WWTP that provided data for the model calibration. The simulation results indicated that lower DO setpoints than those documented during the monitoring campaign can be applied to decease energy requirements without observing higher N2O emission. The fourth chapter explored the development of an N2O model based on an alternative concept that describes the complex electron transfer processes of the bacterial populations involved in the N2O production. The developed model was adapted to the operation of a real full-scale municipal WWTP that provided data for the model calibration and validation. The results showed how important the applied aeration regime is while considering mitigation strategies. This last chapter emphasised how errors/inconsistencies in the sampling campaigns can lead to the development of inaccurate models if these data are used for calibration/validation purposes.