Numerical Methods and Advanced Computing

Selected Publications

figure iDEM: An impulse-based discrete element method for fast granular dynamics
https://doi.org/10.1002/nme.4923
This work introduces an impulse-based discrete element method (iDEM) for efficient simulation of granular materials. By replacing contact-force calculations with collision impulses and directly updating particle velocities, the method bypasses acceleration integration while preserving fidelity. The approach is numerically stable and achieves speedups approaching two orders of magnitude over conventional DEM, enabling large-scale simulations on accessible computing hardware.

figure Image-based 3D modeling-to-simulation of single-wythe masonry structure via reverse descriptive geometry
https://doi.org/10.1016/j.jobe.2023.107125
This study develops an image-based 3D modeling-to-simulation framework for rapid vulnerability assessment of single-wythe masonry structures. Using reverse descriptive geometry, 2D wall images are transformed into 3D discrete element models for seismic analysis. The framework combines automated geometry reconstruction, impulse-based dynamic simulation, and high-fidelity masonry modeling to enable efficient, accurate, and computationally scalable hazard assessment.

figure BRITE Pivot: DEMIAN - Discrete Element Method Infused with Artificial Neural computations
Sponsor: National Science Foundation (PI: Seung Jae Lee)
This project develops DEMIAN (Discrete Element Method Infused with Artificial Neural computations), a next-generation discrete element simulation framework that integrates AI-driven computations to enable real-time, high-fidelity simulation of granular materials at unprecedented scale.

figure QUAD: Quantum Computing-Accelerated Discrete Element Method
Sponsor: FIU Office of the Provost (PI: Seung Jae Lee)
This project aims to develop QUAD, a quantum computing–accelerated discrete element method for simulating granular materials. By reformulating DEM computational bottlenecks to exploit quantum computing, the project seeks to dramatically accelerate large-scale particle simulations beyond the limits of conventional and high-performance computing, enabling transformative advances in granular mechanics, hazard prediction, and engineering design.