At the D-cubed Lab, we investigate and develop efficient modeling techniques for the simulation and design of complex engineering systems. Our research integrates computational mechanics, topology optimization, and data-driven modeling, with an emphasis on capturing mechanical and dynamical behavior across scales and applications.
A core theme of our work is the development of intrusive and nonintrusive data-driven methods that retain physical accuracy while enabling significant computational speed-ups. These tools are critical for simulating nonlinear, multiphysics phenomena efficiently—whether in structural dynamics, materials design, or energy systems.
Our contributions span:
Data-driven ROMs for systems with discontinuities and nonlinearities
Energy-based closure strategies for improved accuracy in low-dimensional models
Topology optimization frameworks for periodic and lattice structures with tailored dynamic properties
Integration of neural ODEs and machine learning into physics-based modeling pipelines
As a natural extension, we are increasingly applying these methods to support digital twin technologies, where real-time simulation and decision-making are essential. By enabling rapid yet accurate predictions in domains such as neutron transport, combustion, and manufacturing, our models help make digital twins a practical tool for engineering insight and control.
Our long-term goal is to push the limits of simulation-driven design, combining physics, computation, and data to tackle emerging challenges in modern engineering.