Investigation of the vortex-induced vibration by convolution neural network

Abstract

: The focus of this paper is to use a convolutional neural network to predict the unsteady laminar flow field in 2D domains, specifically as it relates to vortex induced vibration. In this scenario, an incompressible fluid is flowing around a cylinder, creating a vortex street that leads to oscillation of the cylinder. The movement of the cylinder, which has only one degree of freedom, is modeled using a linear spring-mass-damper equation. The research examines the cylinder’s amplitude response for different spring stiffnesses and damping coefficients, and compares the results to those obtained from FSI computational fluid