A Deep Learning Technique for Modeling Fluid Moments of Swimming Robots

Abstract

We investigate a novel application of Random Forest Regression to modeling errors in prediction of Reynolds Stress of a turbulent flow. In this context, the true solution is given by Direct Numerical Simulation (DNS) data, while the predictive model is a wall-modeled Large Eddy Simulation (LES). Turbulence is the most dominant characteristic of a turbulent flow. Therefore, successful modeling of turbulence can significantly improve the results of numerical simulation. Large Eddy Simulation (LES) computation of turbulent flows has been achieved great attention recently since post-processing of LES results yields information of both mean flow and statistics of resolved fluctuations which is unique to LES. The focus of this paper is on efforts to use data-driven deep learning to model the fluid moments of underwater robots. Increasing efficiency by accurate modeling is a key issue for underwater robots. The proposed model can be an alternative, less computationally expensive approach to resolve the fluid motion in underwater robot modeling.