Mathematical and Computational Foundations for Enabling Predictive Digital Twins at Scale

Explore the mathematical and computational foundations for enabling predictive digital twins at scale in this comprehensive lecture by Karen Willcox from the University of Texas at Austin. Delve into the concept of digital twins as coupled computational models representing unique physical assets, and discover the growing potential of this technology in revolutionizing decision-making across various fields. Examine a unifying mathematical formulation using probabilistic graphical models to create robust digital twin implementations at scale. Learn about the abstraction of asset-twin systems as coupled dynamical systems and their computational realization as dynamic decision networks. Gain insights into physics-based reduced-order models and their role in enabling digital twins. Follow the demonstration of a structural digital twin for an unmanned aerial vehicle, and understand the importance of combining data with physics-based models in creating effective digital twins. Engage with topics such as historical perspectives, digital twins in economics and aerospace, mathematical foundations, and the critical enablers for digital twin technology.

Introduction
Digital Twins
Historical Perspective
Digital Twins of Economies
Digital Twins of Aerospace
Mathematical Foundations
Data alone is not enough
Physicsbased models
What is a physicsbased model
The computational cost of physicsbased models
The critical enabler for digital twins
The digital twin
The digital state
Dynamic vision network
Dynamic decision network
Digital twin example
Structural model
Digital state
Create the digital twin
Measure the geometry
Point estimates
Tip displacement test
Update material properties
Initial condition test
Calibration
Example
Questions