Past Projects

2024

Goal: Develop and improve the adjoint capability in the Sandia Parallel Aerodynamics and Reentry Code in support of inverse problems and design optimization.
Sandia collaborators: Travis Fisher (PI), Eric Phipps, Jared Crean
Research topics: adjoint methods; inverse problems, high performance computing
Funding source: Sandia National Laboratories’ Advanced Simulation and Computing Physics and Engineering Models program.

Goal: This project will improve projection-based reduced-order models (P-ROMs) for forward uncertainty quantification (UQ) and facilitate their adoption by application analysts via integrated and automated workflows. It will achieve this by focusing on the coupling of Pressio, a model reduction software ecosystem, and Dakota, a widely used software for optimization and UQ.
Sandia Collaborators: Eric Parish (co-PI), Elizabeth Krath, Gianluca Geraci, Michael Eldred
External Collaborator: Francesco Rizzi (NexGen, Pressio lead developer)
Research topics: nonlinear model reduction; Uncertainty Quantification; high performance computing
Funding source: Sandia National Laboratories’ Advanced Simulation and Computing Verification and Validation program.

Goal: Develop scientific machine learning approaches to infer the pressure distribution on a hypersonic flight vehicle given internal strain gauge and thermocouple readings
Sandia Collaborator: Bryan Morreale
External collaborators (selected): Noel Clemens (UT Austin, PI), Karen Willcox (UT Austin), Julie Pham (UT Austin), Carlos Cesnik (U Michigan)
Research topics: scientific machine learning; inverse problems; model order reduction; aerothermodynamics modeling
Funding source: AFOSR/NASA University Leadership Initiative
Project website: https://fast.ae.utexas.edu/

Goal: This project will develop methods for solving inverse problems with UQ and ROMs via an integrative approach that teams with the relevant subject matter experts and integrates these research tools with the InverseOpt toolkit in Sierra, thus providing gradient-based optimization tools with UQ and ROM functionality
Sandia Collaborators: Tim Walsh (PI), Vicente Romero, Drew Kouri, John Tencer
Research topics: nonlinear model reduction; inverse problems; error-estimation techniques
Funding source: Sandia National Laboratories’ Advanced Simulation and Computing Verification and Validation program.

Goal: This project aims to enable parallel, scalable, and performant projection-based model reduction capabilities to be adopted by any C++ application in a minimally intrusive manner with Pressio, an open-source C++11 header-only library.
Sandia collaborator: Eric Parish.
External collaborators: Francesco Rizzi (NexGen, lead developer), Mikolaj Zuzek (NexGen).
Research topics: nonlinear model reduction; high performance computing
Funding source: Sandia National Laboratories’ Advanced Simulation and Computing Verification and Validation program.
Project Website: https://pressio.github.io/

Goal: This project aims to develop novel model reduction methods for nonlinear computational simulations.
Sandia collaborators: Eric Parish (PI), Elizabeth Krath, Chi Hoang, Yuki Shimizu.
Research topics: nonlinear model reduction; error estimation
Funding source: Sandia National Laboratories’ Advanced Simulation and Computing Verification and Validation program.

Goal: Implement an adjoint capability in the Sandia Parallel Aerodynamics and Reentry Code in support of inverse problems and design optimization.
Sandia collaborators: Eric Phipps (PI), Jaideep Ray, Kathryn Maupin, Denis Ridzal
Research topics: adjoint methods; inverse problems, high performance computing
Funding source: Sandia National Laboratories’ Advanced Simulation and Computing Advanced Technology Development and Mitigation program.

Goal: This project aims to enable high-fidelity aerothermal simulations of hypersonic vehicles to be employed (1) to generate large databases with quantified uncertainties and (2) for rapid interactive simulation.
Sandia collaborators: Marco Arienti, David Ching, Jeff Fike, Micah Howard.
External collaborators: Francesco Rizzi (NexGen), Karen Willcox (UT Austin).
Research topics: nonlinear model reduction; uncertainty quantification; hypersonic vehicles
Funding source: Sandia National Laboratories’ Laboratory-Directed Research & Development.

Goal: This project aims to enable rapid design evolution, concept exploration, and prototyping of complex system components while (1) ensuring designs satisfy all system-level requirements and (2) rigorously accounting for underlying uncertainties.
Sandia collaborators: John Tencer (PI), Marco Arienti, Erin Mussoni, Chi Hoang.
Research topics: nonlinear model reduction; uncertainty quantification; domain decomposition; component design
Funding source: Sandia National Laboratories’ Laboratory-Directed Research & Development.

Goal: This project aims to extend the applicability of the PRIME epidemic forecasting model to better capture dynamics in multiple peak pandemic recovery scenarios.
Sandia collaborators: Cosmin Safta (PI), Jaideep Ray.
Research topics: Markov Chain Monte Carlo (MCMC); COVID-19 Forecasting.
Funding source: Sandia National Laboratories’ Laboratory-Directed Research & Development.
Project website: https://sandialabs.github.io/PRIME/

Goal: This project aims to integrate reduced-order model methods within a multifidelity uncertainty quantification framework and to demonstrate the greater efficiency and generality of this approach for several test problems with respect to their state-of-the-art counterparts.
Sandia collaborators: Gianluca Geraci (Co-PI), Mike Eldred, Kevin Carlberg.
External collaborators: Francesco Rizzi (NexGen).
Research topics: nonlinear model reduction; uncertainty quantification
Funding source: Sandia National Laboratories’ Laboratory-Directed Research & Development.