Hybrid Analytics for Sim2real Models
This talk will cover methods to create fast and precise predictive models for industrial assets with limited data, useful for control design, XiL models, and digital twins.
Ansys empowers the next generation of engineers
Students get free access to world-class simulation software.
Design your future
Connect with Ansys to explore how simulation can power your next breakthrough.
Students get free access to world-class simulation software.
Connect with Ansys to explore how simulation can power your next breakthrough.
This talk will cover methods to create fast and precise predictive models for industrial assets with limited data, useful for control design, XiL models, and digital twins.
Date/Time:
February 17, 2026
11 AM EST
Venue:
Virtual
In this talk, we will describe methods to develop predictive and accurate models for industrial assets in the regime of small measurement data. These models are fast, precise, and valuable for plant models for control design, XiL models, and digital twins.
Models are built by using machine learning techniques that augment measurements with simulations to learn system behavior from limited data. We demonstrate that such hybrid models exhibit superior extrapolation and learning efficiency, as illustrated by relevant case studies.