Prediction of Multiscale, Multiphysics Turbulent Flow Phenomena Using Unstructured Large Eddy Simulation

PI Name: 
Parviz Moin
PI Email:
Center for Turbulence Research Stanford University
Allocation Program: 
Allocation Hours at ALCF: 
60 Million
Research Domain: 

Parviz Moin, Stanford University, was awarded 60 million hours for large-eddy simulation of jet noise and validation of turbulence models in complex geometries.

Moin’s research aims to broaden the community of researchers able to fully utilize leadership computing resources to predict and quantify the uncertainty of multiscale, multiphysics turbulent flow phenomena using a one-of-its-kind, exascale-ready unstructured large-eddy simulation (LES) platform, charles, developed at the Center for Turbulence Research.

Building upon the success of their previous ALCC allocation for supersonic jet noise crackle in complex geometry nozzles, Moin’s team has expanded their research to include supersonic combustion in mixing layers with finite-rate chemistry and Grid-independent, explicitly filtered LES of a diffuser.

Each project stresses different aspects of the platform: aeroacoustics, combustion and turbulence modeling. The computational approach for each project depends upon the shared features of the simulation infrastructure.