GPU-ACCELERATED SIMULATION OF VON KARMAN FLOW AROUND A CYLINDER

Abstract

© 2025, Begell House Inc.. All rights reserved.This study presents a numerical investigation of Von Karm an flow at a Reynolds number of 200, induced byflow past a single cylinder. The Von Karm an vortex street, characterized by alternating vortices shed froma bluff body, is a fundamental phenomenon in fluid dynamics with significant applications in engineering systems. We employ the Finite Volume Method (FVM) to discretize the governing Navier-Stokes equations, capturing the intricate interactions within the velocity field surrounding the cylinder. Computations are performed using Graphics Processing Units (GPUs) to leverage their parallel processing capabilities. The GPUaccelerated FVM achieves a computational speedup of 50 times compared to traditional calculations on an Intel i9 CPU. This substantial acceleration enables high-resolution simulations that provide deeper insights into the flow structures of the system. The results demonstrate the effectiveness of GPU computing in solving complex fluid dynamics problems and highlight its potential to advance research in computational fluid dynamics.

Keywords