Projection-Based Reduced Order Model for Simulations of Nonlinear Flows with Multiple Moving Objects

TL;DR

This paper introduces a reduced order modeling approach using POD and Galerkin projection for efficient simulation of nonlinear flows with multiple moving objects, significantly decreasing computational time while maintaining accuracy.

Contribution

It develops a novel reduced order model incorporating moving boundary techniques for nonlinear flow simulations with multiple objects, enabling rapid and accurate transient analysis.

Findings

Simulation time reduced by over three orders of magnitude.

Accurate flow field and force predictions for oscillating cylinders.

Model effectively handles complex moving boundary conditions.

Abstract

This paper presents a reduced order approach for transient modeling of multiple moving objects in nonlinear crossflows. The Proper Orthogonal Decomposition method and the Galerkin projection are used to construct a reduced version of the nonlinear Navier-Stokes equations. The Galerkin projection implemented in OpenFOAM platform allows accurate impositions of arbitrary time-dependent boundary conditions at the moving boundaries. A modelling technique based on moving domain and immersed boundary techniques is proposed to overcome the challenge of handling moving boundaries due to movements of the multiple objects. The model is demonstrated capable to capture the complex flow fields past one and two oscillating cylinders and the forces acting on the cylinders. Simulation time could be reduced by more than three orders for a small case on a fine mesh as compared to an existing method and could be more for large cases. In general, the simulation time of the reduced model is of order of seconds as compared to hours of the full order Computational Fluid Dynamics models.