1303 Unmanned Ccombat Air Vehicle Flowfield Simulations and Comparison with Experimental Data

Abstract

In this work, high-order computations of the flowfield around a 1303 unmanned combat air vehicle configuration are performed and compared with recently collected experimental data obtained at Lehigh University. The computational approach used a high-order overset-grid flow solver developed in the Air Vehicles Directorate of the U.S. Air Force Research Laboratory that employs up-to-sixth-order compact finite differences and high-order, low-pass numerical filters to accurately resolve detailed flow features in a robust manner. The experimental data was collected via a particle image velocimetry technique in a free-surface water channel. Both quantitative and qualitative comparisons between computational and experimental results are done at a plane located at eight-tenths of the half-span for various Reynolds numbers and angles of attack, with the results comparing quite favorably for most flow conditions. Computational images of the flowfield are used to elucidate angle of attack and Reynolds number effects on this configuration, as well as to investigate the formation and evolution of the leading-edge and centerbody vortical structures and the impact that angle of attack has on their formation mechanisms. Nomenclature a, b = explicit compact differencing coefficients C root = wing root chord C = local wing chord at given spanwise locationviscous vector fluxes f i = explicit compact filter coefficients I = identity matrix I p , J p , K p = computational indices of interpolation donor point J = Jacobian of the coordinate transformation L mac = mean aerodynamic chord L ref = reference length (L ref L mac ) M = Mach number Pr = Prandtl number, 0.72 for air p = nondimensional static pressure Q = vector of dependent variables Re = reference Reynolds number, 1 u 1 L ref = 1 Re C root = reference Reynolds number based on wing root chord, 1 u 1 C root = 1 Re mac = reference Reynolds number based on mean aerodynamic chord, 1 u 1 L mac = 1 R , R , R = Laplace interpolation coefficients in , , and directions S = wing semispan T = nondimensional static temperature t = nondimensional time U, V, W = contravariant velocity components U 1 = freestream velocity u, v, w = nondimensional Cartesian velocity components in the x, y, and z directions u 1 , u 2 , u 3 = u, v, w hVi = time-averaged velocity vector x, y, z = nondimensional Cartesian coordinates in streamwise, normal, and spanwise directions based on wing orientation (nondimensionalized by the mean aerodynamic chord, L mac ) x LE = x coordinate of the local leading edge as measured from the apex x 1 , x 2 , x 3 = x, y, z x = nondimensional chordwise distance from the local leading edge, x x LE =C x root = nondimensional chordwise distance from the apex, x=C root = angle of attack f = implicit compact filter coefficient = implicit compact differencing coefficient = specific heat ratio, 1.4 for air , , = interpolation offsets in the , , and directions ij = Kronecker delta function 2 , 2 , 2 = second-order finite difference operators in , , 6 , 6 , 6 = sixth-order finite differe...

A Discontinuous Galerkin Chimera scheme

“…High-order schemes have proven to be particularly useful for Large Eddy Simulation calculations [11][12][13][14] and have the potential to reduce numerical discretization errors and reduce computation time for steady flow problems relative to 1st-and 2nd-order accurate methods [15]. However, high-order schemes, such as compact differencing [16,11] and WENO [17,18], require large stencils.…”

A discontinuous Galerkin scheme for Chimera overset viscous meshes on curved geometries

“…The use of structured meshes with the Chimera method has facilitated the application of higher-order schemes to complex geometries. High-order schemes have proven to be particularly useful for Large Eddy Simulation type calculations [4][5][6][7]. However, these high-order schemes, such as compact differencing [8,4] and WENO [9,10], generally require large stencils.…”