Most of the flow phenomena that are important in real-world applications involve turbulence. To understand and find practical solutions to common problems engineers needed to be able to predict quantitatively the behavior of such turbulent flows. With the rise of data-driven methods, new approaches have been developed to both solve and learn the Partial Differential Equations (PDEs) that govern turbulent flows.
In this work, I use Deep Learning methods to learn the turbulence models from simulated time-series data. This data is generated using open-source CFD solvers and illustrate solutions to classical turbulent flow problems.