Berkeley Fluids Seminar

CANCELLED

Progress and challenges in computation of turbulent flows

Monday, March 9, 2020

12:00-13:00, 3110 Etcheverry Hall

Prof. Parviz Moin

( Department of Mechanical Engineering
Center for Turbulence Research
Stanford University )

Abstract: With the exponential growth in computer power, high fidelity numerical simulations have provided unprecedented comprehensive datasets for the study of the mechanics of turbulent flows. A brief review of fundamental research on the structure of wall-bounded turbulent flows will be presented with particular emphasis on the role played by large-scale numerical simulations in conducting controlled experiments of discovery and the insights gained as a result. On the other hand, the early promise of direct numerical simulations to provide data to improve engineering models has not been fulfilled. I will expand on this in the context of lessons learned of the limitations of data science in the development of physical models. Over the past decade there has been considerable progress in high fidelity simulation of multi-physics turbulent flows at reduced computational cost. This has been realized owing to improvements in computer power, advances in numerical methods, scalable mesh generation, and reduced order models to account for unresolved motions. A review of modern large eddy simulation technique (LES) for computation of complex turbulent flows will be presented. Applications range from prediction of supersonic jet noise, combustion in realistic gas turbine engines, two-phase flows, and high-lift aircraft. I will emphasize the critical importance of numerical methods used for LES. Methods with low numerical dissipation are essential for credible LES computations. In the absence of numerical dissipation, non-linear stability and robustness is achieved by enforcing higher order discrete conservation principles. The future outlook for computer architectures and routine use of LES in industrial settings will be discussed.