Mickael Ruben Kaiway, Joni, Agustinus Giai, Samuel Parlindungan Siregar, Enos Tambing, Obia Pius

Uncertainty-Quantified Grid-Convergence Analysis of RANS Turbulence Models for 2-D Incompressible Backward-Facing Step Flow in OpenFOAM

Introduction

Uncertainty-quantified grid-convergence analysis of rans turbulence models for 2-d incompressible backward-facing step flow in openfoam. Grid-convergence analysis of RANS turbulence models for 2D incompressible backward-facing step flow in OpenFOAM. Quantifies uncertainty for mesh optimization & reliable turbulence modeling.

Abstract

A concise evaluation of Reynolds-Averaged Navier–Stokes (RANS) turbulence modeling for two-dimensional, incompressible, steady backward-facing step (BFS) flow at Re = 1000–3000 was conducted using OpenFOAM’s SimpleFoam solver with the standard k–ε model. A tri-level mesh enhancement (coarse, medium and fine) was implemented, and ambiguity was measured utilizing the Convergence Ratio (CR) and Grid Convergence Index (GCI). The fine grid (CR = 0.54; GCI = 0.0059%) was the only configuration exhibiting monotonic convergence, ensuring valid GCI estimation. Results showed reattachment length increasing from 0.11 m to 0.12 m, with stronger vortical structures and steeper shear gradients at higher Re. This study uniquely integrates RANS model validation with grid-uncertainty quantification, providing guidance for mesh optimization and reliable turbulence modeling in BFS simulations.

Review

This paper presents a focused and methodologically rigorous evaluation of Reynolds-Averaged Navier–Stokes (RANS) turbulence modeling for two-dimensional incompressible backward-facing step (BFS) flow using OpenFOAM. The authors effectively integrate grid-convergence analysis with uncertainty quantification, a crucial step often overlooked in computational fluid dynamics studies. By systematically assessing the influence of mesh resolution on simulation outcomes for Reynolds numbers ranging from 1000 to 3000, the study provides valuable insights into the reliability and accuracy of RANS models for this canonical flow problem, thereby enhancing confidence in numerical predictions. A key strength of this work lies in its rigorous application of the Convergence Ratio (CR) and Grid Convergence Index (GCI) metrics across a tri-level mesh refinement strategy. The finding that only the fine grid configuration exhibited monotonic convergence (CR = 0.54; GCI = 0.0059%) is particularly significant, as it underscores the importance of such analyses for ensuring the validity of GCI estimations and the overall reliability of simulation results. The reported physical findings, such as the increase in reattachment length from 0.11 m to 0.12 m and the observation of stronger vortical structures and steeper shear gradients at higher Reynolds numbers, align well with expected flow physics for BFS and provide a solid basis for RANS model validation. The study makes a commendable contribution by uniquely integrating RANS model validation with a comprehensive grid-uncertainty quantification framework. This approach not only enhances the confidence in the reported simulation results but also offers practical guidance for researchers aiming to optimize mesh designs and establish more reliable turbulence modeling practices in complex flow simulations, particularly those involving separation and reattachment phenomena like the backward-facing step. The findings are highly relevant for the broader CFD community, emphasizing the critical need for robust verification and validation procedures in computational studies. Overall, this paper provides a valuable and timely contribution to the field, promoting best practices for high-fidelity numerical simulations.

Full Text

Comments

You need to be logged in to post a comment.