![](/ii/tfec2024.jpg)
ISSN Online: 2379-1748
9th Thermal and Fluids Engineering Conference (TFEC)
GPU-ACCELERATED RANS SIMULATIONS IN COMPUTATIONAL FLUID DYNAMICS
Аннотация
This study introduces NEMESYS, a novel algorithm designed to exploit the parallel processing capabilities of
Graphics Processing Units (GPUs) to significantly enhance computational efficiency in fluid dynamics
simulations. NEMESYS integrates the Reynolds-Averaged Navier-Stokes (RANS) equations with the k−ε
turbulence model, and its efficacy is validated through simulations of two classical flow scenarios: laminar flow
around a cylinder, exhibiting von Kármán vortex shedding, and turbulent flow over a backward-facing step. The
algorithm's performance is critically assessed against established benchmarks from scientific literature and leading
commercial Computational Fluid Dynamics (CFD) software. Key performance metrics include the Strouhal
number for the cylinder flow and reattachment length for the backward-facing step flow. Results demonstrate
remarkable accuracy and reliability of NEMESYS, with a notable reduction in computation time − up to 99.5% faster than traditional CPU-based software. This substantial reduction in computational effort presents significant cost savings and opens new avenues for real-time analysis and accelerated design processes. The implications of this advancement are far-reaching, offering transformative potential for various engineering domains, such as automotive, civil, and environmental engineering, thereby redefining approaches to fluid dynamics analysis and
design.