VOL. 12, NO. 10, MAY 2017 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences © 2006-2017 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 3112 CFD ANALYSIS OF SD 7003 AIRFOIL AT LOW REYNOLDS NUMBER WITH A LAIMINAR KINETIC ENERGY BASED TRANSITION MODEL Mazharul Islam 1 , Jiří Fürst 2 and David Wood 3 1 Department of Mechanical Engineering, Kulliyah of Engineering, International Islamic University Malaysia, Kuala Lumpur, Malaysia 2 Department of Technical Mathematics, Faculty of Mechanical Engineering, Czech Technical University in Prague, Karlovonam Praha, Czech Republic 3 Department of Mechanical and Manufacturing Engineering, Schulich School of Engineering, University of Calgary, Canada E-Mail: mazharul@iium.edu.my ABSTRACT The characteristics of laminar separation bubbles (LSBs) on the SD 7003 airfoil have been extensively studied in the past at low Reynolds numbers. It has been found that the LSB is extensive, especially at airfoil at angle of attack (α) of 4º. To analyze separation, transition and reattachment of flow around SD 7003 airfoil effectively, Computational Fluid Dynamics (CFD) analysis can be performed with suitable transition models. In this article, a modified version of k-kL-ω transition model, originally proposed by Walter and Cokljat [1], has been used with open source CFD tool OpenFOAM for analyzing SD 7003 at Reynolds number (Re) of 60,000. The article investigated k-kL-ω transition model with two recently developed variants for analyzing SD7003 airfoil. These two variants are based on Pohlhausen and Falkner-Skan profiles to consider effect of pressure-gradient for natural transition. It has been found that both the variants under-predicted the lift coefficients and slightly over-predicted the drag coefficients. Both of the pressure-gradient sensitive variants gave better prediction of separation of the laminar BL. However, the reattachment locations were delayed significantly. Among the two variants, the Falkner-Skan based variant predicted the reattachment location slightly earlier than the Pohlhausen based variant and thus conforming better with different experimental and computational results. Keywords: SD7003, airfoil, LSB, transition modeling, CFD, k-kL-ω, OpenFOAM, pohlhausen, falkner-skan. ACRONYMS AND ABBREVIATIONS APG adverse pressure gradient BL boundary layer Cd drag coefficient CFD Computational Fluid Dynamics Cf skin friction coefficient Cl lift coefficient Cm moment coefficient Cp pressure coefficient DGM Discontinuous Galarkin Method DNS Direct Numerical Simulation ILES Implicit Large Eddy Simulation LES Large Eddy Simulation LSB Laminar Separation Bubble PIV Particle Image Velocimetry RANS Reynolds-average Navier-Stokes Equation Re Reynolds number Tu Turbulence intensities INTRODUCTION At low Re, flow over airfoils can exhibit LSBs which results from separation of the laminar boundary layer (BL) and subsequently reattachment as turbulent BL. The characteristic of LSB of the SD 7003 airfoil has been extensively studied in the past at low Reynolds numbers. Apart from different experimental methods [2], the SD 7003 has been investigated through diversified computational techniques [3]-[8] involving Direct Numerical Simulation (DNS), Large Eddy Simulation (LES), Reynolds-average Navier-Stokes Equation (RANS). Based on these previous research activities, transition points are approximately in the range of 40 to 60%. This fact necessitates that CFD analysis of SB 7003 should properly model or simulate the separation of the laminar BL and the subsequent development of the LSB for acceptable results. Though transition phenomena of airfoils can be fairly accurately simulated with DNS and wall-resolved LES techniques, however, the number of cells and computational time required for these techniques are often prohibitive (Choi & Moin 2011) [9]. Because of this reason DNS and LES techniques are usually applied for selected specials cases at low Re which can be used for validation of RANS based results. CFD analysis with RANS models requires substantially fewer cells and computational time. Over the year, diversified RANS-based models have been proposed by researchers to mimic characteristics of flows when they undergo through transition from laminar to turbulent. The main three classes of the transition models are: (i) linear stability theory based e N models [10]-[14], (ii) local correlation based transition models [15], [16], and (iii) laminar kinetic energy based transition models [1], [17]. Each class of model has its strengths and weaknesses. However, for CFD analysis, the first class of models (i.e. e N models) are considered impractical due to its need for non-local parameters. It should be noted that modern CFD codes are now-a-days used for unstructured grid through massively parallel operations in High Performance Computing (HPC) clusters. Due to this fact, now-a-days the local correlation based γ-Reθt [18] or laminar kinetic energy based k-kL-ω [1] models are usually used for modeling transition and turbulence in CFD analysis. To strengthen the modeling capacity of k-kL-ω model in flows with