Nitrous oxide gas emissions estimated by liquid-phase measurements: robustness and financial opportunity in single and multi-point monitoring campaigns.

A liquid-phase nitrous oxide sensor can be used as a proxy to estimate the gas emissions. Experiments conducted in a pilot-scale Anammox reactor, at different degrees of aeration intermittency, indicate a predictive error in the range of 13.4-19.3% during the stripping phase, with a higher error range in the unaerated phases (23.4-62.8%). The total emissions not explained by the aerated model amounted to 14.1%. Only a negligible fraction (3.6%) of the total nitrous oxide emissions were not captured by the unaerated phase model, indicating thus a minor concern for full-scale application. A sensitivity analysis performed on the present study indicates that the quality of the nitrous oxide measurement is of extreme importance to decrease the load prediction uncertainty. Air flow measurement errors have lower impact on the overall load prediction. The financial attractivity of this monitoring approach is significant in completely mixed tank reactors. In presence of a multi-point analysis, and starting from two monitoring points, the financial interest deteriorates by the relatively short lifetime of the commercially available liquid-phase nitrous oxide sensor.