Research

Summary

The current problem of interest to the Center for Compressible Multiphase Turbulence (CCMT) is the simulation of explosive dispersal of metal particles.

The simulations employ several approximations and models to predict the behavior of the billions of particles present. To ensure that the simulations still capture the relevant physics, comparison must be made to experiments. This process is formally known as validation.

Kyle focuses on detailed understanding of the experimental details, quantification of uncertain inputs and metrics, surrogate construction, propagation of uncertainty through the simulations, and construction of validation metrics. The subsequent reduction of uncertainty in the simulation allows for identification of model error.

Validation Hierarchy

The experiments considered in my research are to demonstrate the predictive ability of the code developed by CCMT. It can be extremely challenging to validate a complex flow such as an explosive dispersal of particles. To perform a validation exercise in this case, it can be beneficial to break the problem into a set of smaller problems that can be validated. Oberkampf [1] recommended breaking validation experiments into pieces that attempt to decouple the relevant physics as much as possible. Accordingly, at CCMT the philosophy has been to break the problem into regimes designated by the number of particles present. The lower regimes are investigated to provide insight into the greater regimes and slowly build the problem to full complexity. There are currently two sets of microscale experiments and set of mesoscale experiments already completed. A set of mesoscale experiments and a macroscale experiments is planned for Summer 2017.