V3I1P81

Flow Separation and Reattachment over Backward-Facing Steps

Otuami Obiga¹*

Abstract

This study investigated flow separation and reattachment characteristics over a backward-facing step under turbulent flow conditions using a numerical approach based on the Reynolds-averaged Navier–Stokes equations. The objective was to examine the influence of Reynolds number on reattachment length, pressure recovery, and turbulence behavior downstream of the step, as well as to assess the predictive capability of an advanced turbulence model. The computational domain consisted of a channel with a sudden expansion, and simulations were performed using the SST k–ω turbulence model due to its suitability for separated flows with adverse pressure gradients. Grid independence and validation against recent benchmark experimental data were conducted to ensure numerical accuracy and reliability. The results showed that increasing Reynolds number led to elongation of the recirculation zone and delayed flow reattachment, accompanied by enhanced shear layer instability and increased turbulence intensity. Pressure coefficient distributions revealed a sharp pressure drop immediately after the step, followed by gradual recovery downstream of the reattachment point, with higher Reynolds numbers improving pressure recovery but intensifying turbulence production. Turbulence kinetic energy was found to peak near the reattachment region, confirming its dominant role in momentum exchange and energy dissipation. Although the SST k–ω model captured the main mean-flow features and overall trends satisfactorily, limitations were observed in resolving unsteady and three-dimensional turbulence effects at high Reynolds numbers. The findings provide valuable insight into separated turbulent flows and contribute to improved understanding and modeling of backward-facing step configurations relevant to engineering applications.

Keywords:

backward-facing step; flow separation; flow reattachment; Reynolds number; pressure coefficient; turbulence modeling; SST k–ω model.