Multiscale modeling

Delivering new modeling methods, discovering insights into material physics, and building next-generation material models to help solve complex engineering problems of national interest

Figure illustrating how characterization and performance feed data science, statistics, and machine learning to determine characterization features.
We utilize extensive and varied material characterization data and apply computer vision and machine learning algorithms to develop microstructure–performance correlation models.

Our group researches and models energetic materials—including high explosives (HEs)—and associated polymeric and metallic/ceramic materials over a wide range of length and time scales, including atomistic, meso-, and continuum scales. Our varied projects employ diverse simulation techniques such as:

  • Exploring femtosecond excited-state chemistry via time-dependent density functional theory.
  • Uncovering atomic-level details of hot-spot formation in HEs through grand-scale billion-atom molecular dynamics simulations.
  • Providing critical insights into initiation and detonation processes with continuum-level magneto-hydrodynamic simulations.
  • Developing chemically aware multi-physics models of viscoelastic response to support the design of polymeric components with superior structural, mechanical, and aging properties.

We also develop grain-scale HE material strength models and use shock simulations and detonation physics to streamline HE initiator design by determining the relationship between design parameters and performance. By explicitly incorporating electromagnetic physics at the continuum level, our approach produces fully coupled simulations from electrical stimulus through HE initiation.

Our group uses modern computer vision and machine learning (ML) methods to extract and/or identify structure–property relationships from non-destructive characterization data (for example, x-ray computed tomography) and employs statistical ML to analyze initiation sensitivity from binary response experiments.

Temperature profile showing hot spot formation in TATB.

Temperature profile from massive molecular dynamics simulations to fully resolve the multiscale features of hot spot formation in the energetic material TATB. The simulations reveal that hot spots are regions with rich structural variations, whose formation and evolution depend on the coupled physics of equation of state, mechanical strength, phase transformations, thermal diffusion, and chemistry.

Our research areas of interest include:

  • Equations of state and phase transitions
  • Reactive transport and chemistry
  • Mechanical strength
  • Aging and chemical degradation
  • Shockwaves and high-rate processes
  • Microstructure–performance correlation
  • HE initiation sensitivity

Researchers

Maiti, Amitesh
Kosiba, Graham D.
Graham D. Kosiba
Liesen, Nicholas Thomas