Application of Deep Galerkin Method to Solve Compressible Navier-Stokes Equations

Recently, the application of a deep-learning technique to fluid analysis has been suggested. Additionally, deep-learning-based method called Deep Galerkin Method (DGM) suggested for solving partial differential equation. In DGM, loss function training deep neural network is formulated so that operators, boundary conditions, and initial conditions targeted equation are satisfied. This study aims extend apply DGM compressible Navier-Stokes equations examine feasibility using analysis. this paper, applied two-dimensional Burgers with periodic one-dimensional shock tube problem, supersonic flow around blunt body. The approximate solutions obtained show generally good agreement finite difference method.