Skip to main content
NASA Advanced Supercomputing (NAS) Division World-class modeling, simulation, and supercomputing in support of the NASA mission

Towards Efficient and Reliable Model Reduction of Many-Query CFD Problems

AMS Seminar Series

Speaker: Masayuki Yano, University of Toronto Institute for Aerospace Studies

Presentation

Video Coming Soon

Abstract

Many engineering tasks, such as optimization, uncertainty quantification, and optimal control, require CFD simulation for many different configurations and/or in real time. In this talk, we consider projection-based model reduction of parametrized CFD problems to accelerate the solution of many-query/real-time problems by several orders of magnitude, while providing error estimates in predictive settings. The key ingredients are the following: an adaptive high-order discontinuous Galerkin method, which provides efficient full-order models (FOM) for high-fidelity analyses; reduced basis spaces, which provide rapidly convergent approximation spaces for reduced-order models (ROMs); the empirical quadrature procedure, which provides sparse quadrature for nonlinear PDE residual evaluation to yield hyperreduced ROMs (HROMs); the dual-weighted residual method, which provides effective error estimates for quantities of interest; and adaptive training algorithms, which simultaneously train FOMs and HROMs to meet the user-specified error tolerance in an automated manner. We demonstrate the framework for parametrized aerodynamics problems modeled by the compressible Euler and Reynolds-averaged Navier-Stokes equations, with applications to parameter sweep, uncertainty quantification, and data assimilation.

Biography

Photo of Masayuki Yano
Masayuki Yano is an associate professor in the Institute for Aerospace Studies at the University of Toronto. He obtained B.S. in Aerospace Engineering from Georgia Tech, M.S. in Computation for Design and Optimization from MIT, and Ph.D. in Aerospace Computational Engineering from MIT. His research interests lie in the development and assessment of numerical methods for PDEs with emphasis on adaptive high-order methods, error estimation, model reduction, and data assimilation.