Organisation
EEDT_DC
Research Group
The Flanders Make EEDT_DC core lab combines in-depth physical insightsinto electromechanics with insights obtained from data that becomesavailable in today industry 4.0 setting applying state-of-the-art andinnovative artificial intelligence routines (Hybrid AI). As such we contribute to building:
· Explainable and transferabledigital twins, and
· Explainable and trustworthyAI for the manufacturing industry.
Specific aspects of our research include:
· Development of:
o (multiplicative)hybrid data-physical models for explainable condition monitoring;
o (multiplicative)hybrid data-physical models for explainable system identification;
o Probabilistic decision-making routines that are adaptive, explainable andtransferable;
o Reinforced learning control enhanced with physics suitablefor the manufacturing industry;
o Low-probability eventestimations and predictions.
· Sensing &monitoring of (a fleet of) drivetrains;
· Control &decision making of (a fleet of) drivetrains;
· Virtual sensing of electricalmachines based on electrical signal analysis for condition monitoring;
· Stochastic and uncertaintymodelling to evaluate and improve the effectiveness of sensing and monitoringstrategies;
· Camera-based sensing and monitoring of oil film in anddegradation of bearings and gearboxes;
· Controller structures thatleverage on the well-known, safe and robust traditional control structuresenhanced with AI to include data driven learnings or (high fidelity) models toimprove performance
· Techniques to do automatic(e.g. context aware) PID tuning
· Distributed optimal controltechniques
· Use of AI techniques tocorrelate product-process features