Application of Gradient Booting regression toward the Computational Fluid Dynamics in the Manufacturing industry

Abstract

A faster calculation of MPS (Moving Particle Semi-implicit) method which is a computational fluid dynamics in the Manufacturing industry is proposed. Proposed method surrogates the semi-implicit part of MPS by the Gradient boosting regression trees using 10 original features as inputs. Finally, we confirmed that the qualitative properties of the proposed method coincide with the conventional MPS method by simulations of the dam-break problem.