For many decades AI has come with really high expectations and has often failed to deliver, also in the water sector. With the advent of the Googles of this earth, AI has made big inroads in our lives. It turns out that the wastewater field is now ready for this digital revolution too. Based on full-scale data, digital twins based on machine learning (and mechanistic models as the rebar for strength in hybrid models) are being readied for use in operational support (and even as part of real-time control systems). The contribution will present some of the recent successes in the US, Canada, and France and will discuss the key factors of success.
Speaker Bio:
Peter Vanrolleghem has a MSc in Bio-engineering (1987) and a PhD in Applied Biological Sciences – Environmental Technology (1994), both from Ghent University (Belgium). His PhD dealt with monitoring, modelling and control of activated sludge wastewater treatment processes, with special emphasis on model calibration and experimental design. He became an associate professor at Ghent University in 1997, initiating the BIOMATH research team that by 2006 had grown to 25 researchers focusing on modeling methodologies for bioprocesses, mainly in environmental systems.
Since 2006 he holds the Canada Research Chair on Water Quality Modeling, allowing him to start up the modelEAU research team at Université Laval in Quebec City (Canada). This group focuses on modeling, monitoring and control of water systems. modelEAU quickly built momentum with numerous research assistants, post-docs, PhD students and master students.
These large research teams have helped him develop his publication record of over 375 peer-reviewed papers. He is very active within IWA (IWA Fellow and member of IWA’s Board of Directors), is getting increasingly involved in WEF (as chairman of MEGA, the Modelling Expert Group of the Americas), and is also active within Réseau Environnement and the Canadian Association on Water Quality (CAWQ).