CITIES invites to PhD Defence by Peter Stentoft: “Stochastic Predictive Control of Wastewater Treatment Processes”

The water we use in modern societies is disposed through the sewer system wherefrom it is led to water resource recovery facilities. These facilities are extremely important structures as they protect local ecosystems from large nutrient loads and reduce human health risks related to exposure to faecal matter. However, the operation of a water resource recovery facility is very resource-demanding in terms of electricity, chemicals, maintenance etc. and thus control is essential. In addition, the use of electricity and the biological processes at the plant are related to large amounts of greenhouse gas emissions. The most electricity demanding process at a plant is the aeration. Aeration, or ”adding air”, happens in the biological stage of a water resource recovery facility. The reason is that adding air activates some special biological processes which in return remove nutrients from the water phase.

This thesis investigates new ways of modelling and controlling the aeration process. The idea behind is that using online data from the various nutrient sensors and machinery at a plant, models that use both physical knowledge and adapts by using data can be useful in predicting plant performance. Now, using these predictions, future control of aeration can be planned in an optimal way to minimize different objectives. One possibility is to plan aeration using online electricity prices and thereby aim to minimize costs. Another possibility is to use forecasts of greenhouse gas emissions related to the electricity mix and thereby try to minimize emissions. In the thesis, different models and objectives are investigated both qualitatively to show the dynamics and quantitatively using data from Danish plants and electricity markets. Results show, the potential for reducing costs by 9-45% compared to current control, and around 35% of greenhouse gas emissions if this is prioritized. Finally, the strategies are tested full-scale on different plants and some initial results are discussed. Consequently, this study is considered a large step towards better wastewater treatment for future smart societies.

For participation please register no later than 10 November at:

https://www.conferencemanager.dk/phddefensepas

Supervisors:

• Associate Professor Jan Kloppenborg Møller, DTU Compute
• Co-supervisor: Professor Henrik Madsen, DTU Compute
• Co-supervisor: Professor Peter Steen Mikkelsen, DTU Environment
• Co-supervisor: Senior Process Engineer Thomas Munk-Nielsen, Krüger A/S, VeoliaWater Technologies
• Co-supervisor: Development Engineer Luca Vezzaro, Krüger A/S, Veolia WaterTechnologies

Examiners:

• Associate Professor Uffe Høgsbro Thygesen, DTU Compute (chairman)
• Professor Peter Vanrolleghem, Université Laval, Québec
• Director of projects Pernille Ingildsen, Hillerød Forsyning