As DARROW enters its final year, we are not only highlighting the project’s technical results, but also the people behind them. Throughout this interview series, DARROW partners reflect on the project’s outcomes, the challenges they faced, and the lessons learned along the way.
In this interview, Elena Torfs, Assistant Professor at Ghent University, shares her perspective as a modelling expert working on digital twins and advanced optimisation for wastewater treatment. She reflects on the challenges of achieving holistic optimisation in wastewater treatment through stable yet adaptable full-plant models in a changing operational environment.
Tell us a bit about yourself and your organization. What role have you played in DARROW?
My name is Elena Torfs. I am an assistant professor working at Université Laval in Québec, Canada and Ghent University in Belgium. My expertise is on using models to build digital twins and advanced optimisation and control systems for (waste)water treatment systems.
Our team was responsible for building the process model of the Tilburg plant. This process model was then used to train reinforcement learning and reduced order models. We have also worked closely with IMEC to develop the reinforcement learning control algorithms and are developing a MOOC to explain the concepts of the DARROW project to a broader audience.
In your view, what is the most valuable innovation or tool that DARROW has developed?
I think the innovation lies in connecting the different tools to create a framework to go from data to actual multi-objective optimisation of treatment plants. Showcasing that it is possible to optimise water quality, energy consumption, process efficiency and greenhouse gas emissions simultaneously in an automated yet operator-inclusive way.
If you had to describe DARROW in one sentence, what would it be?
DARROW focuses on turning raw data into holistic, multi-objective decision support for treatment plants.
What is something you have contributed to DARROW that you are especially proud of?
We have developed and applied an automated model calibration strategy for the full-plant simulators. This was very useful to automatically recalibrate different versions of the model as new datasets became available or when new versions of the model were needed to train different data-driven applications.
What was one of the biggest challenges you or your team faced, and how did you overcome it?
As with many digitalisation projects, data, data quality and data access are a major challenge. Despite having an amazing utility partner onboard (De Dommel) who was open to share a lot of data and information, it still took a lot of work from many people to define a stable and high-quality dataset for model development and calibration.
What advice would you give to a future project team taking on something as ambitious as DARROW?
Have a back-up plan in the form of a clearly defined fallback case study. Updates to the Tilburg treatment plant presented a big challenge in the project since we struggled to achieve a representative dataset of stable operation. A fallback case study would have been helpful.
What kind of change do you think DARROW can bring to how we manage water and resources in the future?
I believe that the wastewater treatment sector is ready to transition from local, process-based optimisation to a holistic, system-wide, multi-objective optimisation. I think that the DARROW project showcases how to achieve this transition.
Elena clearly highlights how essential rigorous modelling is to achieving holistic optimisation in wastewater treatment. By keeping models reliable, adaptable, and aligned with plant reality, DARROW builds a pathway toward more holistic and data-driven plant management.