Dynamically predicting nitrous oxide emissions in a full-scale industrial activated sludge reactor under multiple aeration patterns and COD/N ratios

The use of digital tools has become essential for quantifying and predicting greenhouse gas (GHG) emissions in urban wastewater treatment plants (WWTPs), enabling the development of operational regimes with a high probability of achieving net-zero targets. However, comprehensive studies documenting validation of model predictions-such as effluent quality, process economics, and emission factors-remain scarce within full-scale industrial settings. This paper aims to develop a decision support tool (DST) for (dynamically) predicting nitrous oxide (N2O) emissions in full-scale industrial activated sludge reactors (ASRs) and suggesting mitigation strategies. The DST, incorporating both biological and physico-chemical processes, was developed to address the unique characteristics of industrial wastewater. Specialized Gas-Liquid (G-L) mass transfer routines were also formulated to account for alternating anoxic and aerobic conditions in covered reactors. The proposed approach was validated using full-scale data collected at varying frequencies (from daily to minute intervals) during different campaigns at the largest industrial wastewater treatment system in Northern Europe. The DST was further tested across multiple aeration patterns and influent COD/N ratios. Results show that DST simulations can reproduce (daily) biological COD and nitrogen removal, sulfur transformations, and the physico-chemical precipitation of phosphorus with aluminum, achieving a deviation of 8.6 % over a six-week period. High-frequency (minute-level) dynamics for multiple nitrogen species (NHx, NO2-, NO3-, dissolved and gaseous N2O), dissolved oxygen (DO), and airflow were captured with a NRMSE of 0.16, 0.14 and 0.11 for three evaluated operational strategies (Baseline, Scenario #1 and #2), respectively. Both plant data and DST predictions indicate that the correlation (R2 up to 0.9) between emission factors (EFs) and influent COD/N ratios is significantly influenced by: i) oxygen supply dynamics (fast/slow) and ii) the duration of aeration periods. These EFs range from 0.2 % to 1.4 %. Analysis of derivatives identifies the denitrification (DEN) pathway as the primary contributor to N2O production, peaking at the anoxic phases, with the nitrifier-denitrification (ND) pathway contributing to a lesser extent at the end of aeration. Additionally, the DST generated response surfaces illustrating the key performance indicator (KPI) variations in EFs, nitrification capacity, effluent quality, and aeration energy consumption as functions of different aeration setpoints (DO and NO2-) across varying COD/N loads. The DST provided optimized strategies targeting those KPIs, which were successfully applied on site with improvements of most of the KPIs, achieving up to 71 % reductions of N2O emission (1.4 % to 0.4 %), potentially mitigating >15,000 tons CO2-e per year. These results demonstrate the DST's potential for broader applications in wastewater treatment processes.