NCSI: Machine Learning for Compressible Turbulent Mixing with Large-Eddy Simulations

Status	Completed
Mentor Name	Tulin Kaman
Mentor's XSEDE Affiliation	Education Allocation
Mentor Has Been in XSEDE Community	3-4 years
Project Title	Machine Learning for Compressible Turbulent Mixing with Large-Eddy Simulations
Summary	The undergraduate student will work with the mentor to investigate the machine learning models for compressible turbulent mixing simulations to predict the fluctuations between Direct Numerical Simulation (DNS) and Large Eddy Simulation (LES). With the power of today's HPC systems, resolving all turbulence length scales are handled by DNS. In LES, the unresolved smaller scale motions are modeled by subgrid scale model. The student will conduct compressible multi fluid simulations to build up a database to train the neural network, and apply machine learning techniques to explore and prepare data for an optimized neural network model.
Job Description	The student's primary goal is to learn how to use the high performance computing systems to perform efficient compressible multifluid numerical simulations on XSEDE. The student will work to develop skills in the field of high performance computing systems. The mentor and the student will meet weekly in Summer 2021 for 10 weeks, and the students will put efforts to work 30 hours per week on developing necessary skills on running efficient simulations, analyzing and visualizing large data set obtained from the compressible multi fluid numerical simulations.
Computational Resources	The student will use XSEDE online training materials and work on the usage of machine learning techniques for turbulent mixing simulations on XSEDE.
Contribution to Community	We will work on deep neural network models that can be used in applications of computational fluid dynamics.
Position Type	Apprentice
Training Plan	The training plan includes an introduction to the XSEDE resources, distributed memory programming, debugging, profiling, analyzing and visualizing large data set.
Student Prerequisites/Conditions/Qualifications	The student working on this project must be an undergraduate student in good-standing. The student engaged or interested in data science and scientific computing for computational fluid dynamics is eligible for this position.
Duration	Summer
Start Date	05/24/2021
End Date	07/30/2021