Name

Joint UK efforts for tackling N₂O process emissions with knowledge-based AI and machine learning

Speakers

Jose Porro, Ruth McConnell

Authors

Jose Porro, Yuge Qiu and Julia Porro, Cobalt Water Global, USA
Joshua Williams, Adam Fairman and Shaveen de Almeida, Dwr Cymru Welsh Water, UK
Hamish Todd, Ruth McConnell and Elise Cartmell, Scottish Water, UK

Description

Dwr Cymru Welsh Water and Scottish Water collaborated and shared results from long-term N2O monitoring campaigns at four WwTws in which AI and ML was used to continuously assess the data and provide insights on mitigation opportunities and test machine learning for monitoring N2O. A joint meeting was then held to share and discuss the results and a report prepared to document the conclusions, consensus gained and lessons learned from the effort. This presentation will highlight the key takeaways from the joint meeting and report including the following:

-Risk correlates closely with measured N2O
-Mainly Low DO risk is the cause of N2O peaks, although High DO risk is seen to correlate with N2O at times
-N2O varies throughout the year
-Lower risk generally corresponds with lower N2O, which validates recommendations to adjust DO to lower risk
-Sites that may be limited in aeration capacity will be susceptible to higher N2O emissions during higher loading due to low DO conditions.   

Other key outcomes from this work include a QA/QC plan to address sensor placement, data issues, machine learning model use and performance, and a validated method for estimating N2O emissions and prioritizing which assets to monitor and mitigate first.

Time

2:05 PM - 2:30 PM

Location Name

Room 1

Track

10:05 - 17:30 Process Emissions