Robust and data-Efficient Learning for Industrial Control
Solutions developed
There is a potential in integrating machine learning in control design to overcome the complexity while satisfying safety constraints, as shown in robotics and automotive industry. However, IPCC indicated that ""The key challenge for making an assessment of the industry sector is the diversity in practices, which results in uncertainty, lack of comparability, incompleteness, and quality of data available in the public domain on process and technology specific energy use and costs"". The research question I will address in this project is if and how incorporating data-driven learning in design of control algorithms leads to improved environmental performance and safe operation of large-scale industrial networks."
Main results
Our life depends on heat, power and gas networks. The greening of these networks is crucial to Europe’s energy and resource efficiency targets. In this context, the EU-funded RELIC project will explore a holistic approach to how resources and energy are delivered to the industry via distribution networks. It will explore how incorporating data-driven learning in the design of control algorithms leads to improved environmental performance. Currently, timescales ranging from milliseconds to ensure the safe operation of pumps or generators to days or months make operation complicated. There is uncertainty in terms of the operating conditions and incomplete information. The project will develop new operating strategies for distribution networks.
Our website uses cookies to improve your browsing experience. Check our Privacy policy to be informed about your data.