Qualitative Trend Analysis: An artificial intelligence approach to information extraction
Event details
| Date | 28.10.2014 |
| Hour | 16:15 › 17:15 |
| Speaker | Dr Kris Villez, SPIKE (Sensor management, Process Intelligence and Control), Process Engineering (Eng), EAWAG Duebendorf |
| Location | |
| Category | Conferences - Seminars |
Abstract:
One of the most critical elements for automation of wastewater treatment processes is measurement of the key variables involved. In practice, many such key variables cannot be measured directly. Fortunately, indirect measurements can be of great assistance to approximately assess the changing condition of wastewater treatment plants. Evaluated indirect measurements include macroscopic variables such as the pH and oxidation-reduction potential and images of sludge settling experiments. To infer the state of a wastewater treatment plant from such indirect measurements, specialized algorithms for information extraction are however necessary.
Currently, qualitative trend analysis (QTA) is one of the strategies selected to do so. QTA permits to extract coarse grained information about the process dynamics by focusing on the signs of signal derivatives rather than their exact values. Currently, these algorithms are developed further and fine-tuned for an ever increasing diversity of applications. Early successes were achieved for reaction end-point detection in a sequencing batch reactor for nitrogen and phosphorus removal from communal wastewater. In the meantime, newly developed QTA algorithms have also been applied successfully for batch process diagnosis and control of the Hard wastewater treatment plant in Winterthur (Switzerland). Currently, the QTA algorithms are tested for analysis of images of sludge settling experiments. In this talk, basic concepts, challenges and opportunities regarding QTA will be covered.
Short biography:
Kris Villez obtained his Ph.D. in Applied Biological Sciences in 2007 at Ghent University (Belgium). After postdoctoral work at Université Laval (Québec, QC, Canada) and Purdue University (West Lafayette, IN, USA) he now works at Eawag (Dübendorf, Switzerland) since 2012 in a tenure-track group leader position. In his research, he focuses on monitoring, diagnosis, and automation of wastewater treatment processes.
One of the most critical elements for automation of wastewater treatment processes is measurement of the key variables involved. In practice, many such key variables cannot be measured directly. Fortunately, indirect measurements can be of great assistance to approximately assess the changing condition of wastewater treatment plants. Evaluated indirect measurements include macroscopic variables such as the pH and oxidation-reduction potential and images of sludge settling experiments. To infer the state of a wastewater treatment plant from such indirect measurements, specialized algorithms for information extraction are however necessary.
Currently, qualitative trend analysis (QTA) is one of the strategies selected to do so. QTA permits to extract coarse grained information about the process dynamics by focusing on the signs of signal derivatives rather than their exact values. Currently, these algorithms are developed further and fine-tuned for an ever increasing diversity of applications. Early successes were achieved for reaction end-point detection in a sequencing batch reactor for nitrogen and phosphorus removal from communal wastewater. In the meantime, newly developed QTA algorithms have also been applied successfully for batch process diagnosis and control of the Hard wastewater treatment plant in Winterthur (Switzerland). Currently, the QTA algorithms are tested for analysis of images of sludge settling experiments. In this talk, basic concepts, challenges and opportunities regarding QTA will be covered.
Short biography:
Kris Villez obtained his Ph.D. in Applied Biological Sciences in 2007 at Ghent University (Belgium). After postdoctoral work at Université Laval (Québec, QC, Canada) and Purdue University (West Lafayette, IN, USA) he now works at Eawag (Dübendorf, Switzerland) since 2012 in a tenure-track group leader position. In his research, he focuses on monitoring, diagnosis, and automation of wastewater treatment processes.
Practical information
- General public
- Free
- This event is internal
Organizer
- EESS - IIE
Contact
- Prof. Samuel Arey, LMCE