Visit

Waste Deep – Using Deep Learning methods to improve Waste Water Treatment

Treatment of wastewater is a global challenge, where improper or no treatment leads to various environmental issues. Wastewater treatment plants generally aim at reducing the carbon content, and the content of nutrients such as nitrogen and phosphorous. When done properly, wastewater treatment not only prevents pollution but can also be used for energy and nutrient recovery. The most common steps in wastewater treatment can be classified as mechanical, chemical, and biological treatment and combinations of these.

The focus of this talk will be on the modelling and controlling of aerobic biological wastewater process which is incredibly complex as it is dependent on the incoming wastewater characteristics, the multitude of microorganisms that are removing the pollutants and the operating factors. Aerobic processes are energy intensive and thus reduction of nutrients and energy use is an optimization problem, which is the focus of our work.

Wastewater treatment occurs in volumetrically large continuous batches where the system exhibits dynamics that are coupled, nonlinear and varying along with varying time delays. This complicates the control problem significantly.

This talk will introduce the problem of modelling and controlling these processes and show some of our approaches in modeling the system using machine learning and in particular deep learning methods. The goal of our work is to optimize the system, where energy, economy and environment are the constraints in the optimization problem.

Speaker Bio

Associate Professor Petar Durdevic leads the Offshore Drones and Robots research group at Aalborg University. His research focus lies in Modelling and Control, Deep Learning, Deep Reinforcement Learning applied in process control and robotic systems. He has been working on water treatment for almost a decade, both off- and on-shore and is currently mainly focusing on onshore WWTPs.

In addition, he works on robot visual servo control and condition monitoring, predominantly on wind turbines. He has set up the Robotics lab at Aalborg University Esbjerg Campus. He is also a board member of the research centre Aalborg Robotics at AAU, which promotes research, innovation, education and dissemination within robotics, AAU.

Currently Petar is a Fulbright Fellow at Cornell university, hosted by professor Silvia Ferrari at the Sibley School of Mechanical and Aerospace Engineering.