Download presentation
Presentation is loading. Please wait.
Published byHelena Bailey Modified over 9 years ago
1
© Saab AB 2008 www.saabgroup.com 1 Transonic store separation studies on the SAAB Gripen aircraft using CFD Ingemar Persson and Anders Lindberg Stockholm, October 18, 2010
2
© Saab AB 2008 www.saabgroup.com 2 Outline of presentation Introduction Models Computations (CFD and 6-DOF) Results Conclusions
3
© Saab AB 2008 www.saabgroup.com 3 Introduction External stores must be released in a safe and well predicted manner 6-DOF simulation model to predict store trajectories SSM includes free-flying store aerodynamics as well as interference aerodynamics Data collected from complicated and extensive WT-tests using a two- sting-rig Time and budget limitations lead to alternative ways in achieving store release predictions Advances in CFD and low cost computer power make computational aerodynamics an interesting alternative
4
© Saab AB 2008 www.saabgroup.com 4 Introduction SAAB Gripen Flight condition M=0.9, AoA=1.9 degrees Payload: Four 227 kg Mk82LD at 2L/2R and 3L/3R WT: centerline DT300 droptank FT: centerline FUNK camera pod
5
© Saab AB 2008 www.saabgroup.com 5 Models – CAD models CATIA v.4 models, imported to ICEM CFD using direct CAD interface Slight modifications of surface geometry to improve prismatic boundary layer grid Discrepancy: Pylon 4 not present in CAD/CFD models Different levels of modelling complexity tried on pylons Complexity of sway braces Suspension lugs present or not Gap distance between payload and pylon uncertain, used 8.7 mm Both WT and FT configurations modelled
6
© Saab AB 2008 www.saabgroup.com 6 Models – Discrete CFD models Grid generation using ICEM Tetra/Prism Captive position and 6 subsequent vertical positions Tetrahedral grids approximately 4.5-6.5 Mnodes Mixed tetrahedral / prismatic grids approximately 19-22 Mnodes Boundary layer grid holding 40 prismatic layers y + =1 grid, Initial cell height 2e-5 m, expansion factor 1.2 Far field positioned 10 a/c lengths away from aircraft
7
© Saab AB 2008 www.saabgroup.com 7 Grid generation – Detail of mixed tetrahedral / prismatic grid around payload Mk82LD in captive position Detail of grid around the bomb fins
8
© Saab AB 2008 www.saabgroup.com 8 Computations - General CFD solution obtained using the EDGE v.4.1.0 fluid flow solver Inviscid computations utilising a central scheme using JST art.visc. Viscous computations performed with both central and upwind scheme Thin shear layer NS equations solved Turbulence model is Menter SST k-w RK time integration with agglomorated FAS multigrid conv.acc. Upwind computations always initiated with 1st order scheme and later switch to 2nd order using a Roe flux difference splitting employing a minmod limiter
9
© Saab AB 2008 www.saabgroup.com 9 Computations – Boundary conditions Farfield – Riemann invariants Solid surface – slip / no slip Engine inlet / outlet – flow through surfaces with prescribed flow ECS inlet / outlet – flow through surfaces with prescribed flow Example of RANS computation
10
© Saab AB 2008 www.saabgroup.com 10 Computations – 6-DOF simulation Store relative motion depends on Store free flight aerodynamics Mass and inertial data Aircraft interference aerodynamics Aircraft motion during separation ERU force on the store ERU module consists of a gas dynamic model ODE system solved by RK-Merson Solution visualised in SAAB system ICARUS Example from ICARUS
11
© Saab AB 2008 www.saabgroup.com 11 Flight test – Separation as seen from chase a/c
12
© Saab AB 2008 www.saabgroup.com 12 The first simulation based on WT data from a configuration with a DT300 attached to pylon 5
13
© Saab AB 2008 www.saabgroup.com 13 Surface pressure field, Euler simulation Note the difference in shock strength. The drop tank results in a large under pressure on the bomb fins which gives a large yawing moment.
14
© Saab AB 2008 www.saabgroup.com 14 Results – Corrective techniques
15
© Saab AB 2008 www.saabgroup.com 15 Results – Corrective techniques
16
© Saab AB 2008 www.saabgroup.com 16 Results – Corrective techniques
17
© Saab AB 2008 www.saabgroup.com 17 Results – Simulation with captive corrections
18
© Saab AB 2008 www.saabgroup.com 18
19
© Saab AB 2008 www.saabgroup.com 19 Results – Based on CFD alone Grid based approach Captive and subsequent vertical positions computed (0.0625, 0.125, 0.25, 0.5, 1.0, 2.0 m) Different complexity of pylon attachment investigated (sway bracer realisation, suspension lugs etc.) Euler and Navier-Stokes Different numerical schemes (central and upwind) Captive absolute condition hard to capture
20
© Saab AB 2008 www.saabgroup.com 20 Example of RANS computation with 0.25 m vertical drop of RHS Mk82LD
21
© Saab AB 2008 www.saabgroup.com 21 Results – Based on CFD alone
22
© Saab AB 2008 www.saabgroup.com 22 Results – Based on CFD alone
23
© Saab AB 2008 www.saabgroup.com 23 Results – Based on CFD alone
24
© Saab AB 2008 www.saabgroup.com 24
25
© Saab AB 2008 www.saabgroup.com 25 Comparison between simulations based on the different settings
26
© Saab AB 2008 www.saabgroup.com 26 Conclusions Computational aerodynamics is a useful tool when used in a corrective manner Achieving the correct captive aerodynamic load is of vital importance Computational aerodynamics as a sole contributor of aero data was not as accurate but can be used as an indicative method For this case, inviscid physics was ”accurate enough”. Viscous computations did not improve results to motivate the increased work load
27
© Saab AB 2008 www.saabgroup.com 27 www.saabgroup.com
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.