Digital Twin of a Rotating Machine: Model Order Reduction and Artificial Intelligence for Hydroelectricity Production

Hydroelectric power units are more solicited today than in the past when they were usually operated at baseload. Because of this, their operating conditions include many starts and stops, and partial load conditions, resulting in premature and unpredictable failures. With a digital twin, these units can be properly monitored and accurately modeled in order to schedule maintenance and optimize usage to minimize wear.
The aim of my project is to create a digital twin for a vertical axis rotating machine (VARM) with similar dynamical properties to those of the hydroelectric turbine-alternator shaft line. This will be accomplished by building the experimental setup and developing the mathematical reduced-order model of the VARM using Proper Generalized Decomposition. We see the significance of this project in the fact that it will serve as a proof of concept, enabling companies and developers to overcome the same challenges required to develop digital twins for real industrial equipment using sensor data and physical modeling.

Faculty Supervisor:

Frederick Gosselin

Student:

Partner:

Arts et Métiers Sciences et Technologies

Discipline:

Engineering

Sector:

Advanced Computing; Artificial Intelligence; Green/Alternative Energy

University:

Polytechnique Montréal

Program: