Neural Network for Sub-grid scale Turbulence Modeling in Large Eddy Simulations of the Jet

To study the jet, there are lots of methods such as hydraulic model and field experiments, and numerical methods are used in this study since it relatively consumes less space, time, and labor costs. The representative numerical methods include Direct Numerical Simulation (DNS), Large Eddy Simulation (LES), and Reynolds Averaged Navier Stokes (RANS). Reynolds Averaged Navier Stokes (RANS) yields only time-averaged flow properties, and it has a limitation to reproduce the complexity of jet. So, this study focuses on DNS and LES models. DNS does not require turbulence modeling because it resolves all scales of eddies directly. However, it requires the large number of grids and spends much time. LES resolves the larger-scale eddies directly and requires the model for the smaller-scale eddies. But it has slightly different results depending on the sub-grid model. To solve this problem, this study proposes a method to use the results of DNS learned by deep learning without using the existing sub-grid models of LES. This study proposed the proper neural network model to research the turbulent jet. LES model is not needed to calculate the sub-grid eddy viscosity, and it uses DNS data in the sub-grid scale. This study is aimed to pursue two objectives; (1) Comparing the LES and DNS energy spectrum and finding the wavenumber of sub-grid scale in LES; (2) Optimizing the algorithm for the turbulent jet flow and combining LES and algorithm.