Proceedings of the 13th Asian Congress of Fluid Mechanics

17-21 December 2010, Dhaka, Bangladesh

Numerical Simulation of Transonic Flow past a Cropped Delta Wing

D. K. Marthandrao1, Sharanappa V. Sajjan2 and P. Honguntikar1
1
Poojya Doddappa Appa College of Engineering, Gulbarga – 585102, INDIA
2
Computational and Theoretical Fluid Dynamics Division, Council of Scientific and Industrial Research
National Aerospace Laboratories, Bangalore – 560017, INDIA

*E-mail of presenting author: [email protected]

 Proceedings of the 13th Asian Congress of Fluid Mechanics
17-21 December 2010, Dhaka, Bangladesh

Abstract Computational fluid dynamics analysis of flow over

a 65 - degree sweep cropped delta wing with a round leading
edge is carried out using Implicit Reynolds-averaged Navier- II. Numerical method
Stokes solver IMPRANS. The solver is based on an implicit The Reynolds-averaged Navier-Stokes equations that
finite volume nodal point spatial discretization scheme. The
govern three-dimensional compressible viscous flows can
algebraic eddy viscosity model due to Baldwin and Lomax is
used for the turbulence closure. The steady flow computations be written in non-dimensional conservative form as
have been performed for the of Mach number 0.85 and angles of
attack (10 - 30°). The results are obtained in the form of U E F G
aerodynamic coefficients and surface pressure distributions. The    =0 (1)
t x y z
sectional surface pressure distribution compares well with the
experimental data for the case considered. Here, U is the vector of conserved variables, E, F and G are
flux vectors, (x, y, z) is the Cartesian coordinate system and
Keywords Cropped delta wing, transonic flow, finite volume t is the time variable.
method, viscous flow, RANS solver Applying Euler’s implicit time differencing formula
n 1
 U 
 t  O  t 
I. Introduction U n  U n 1   2
(2)
 t 
Delta wing at high angle of attack in transonic or
supersonic region creates very complex flow field to equations (1), we obtain
n 1
involving flow separation, shock wave generation and also  E F G 
U n  t     0 (3)
formation and growth of two large scale vortices on the  x y z 
suction surface of the wing. These vortices lie roughly
Here U  U  t   U  nt  is the solution vector at time
n
along a straight line from the apex to the trailing edge. The
angle of attack at which these vortices first form is level n and U   U  U  is the change in U n over
n n 1 n

primarily a function of the wing sweep angle [1].

Computation of these flow features is a challenging work. time step t. In order to facilitate the finite volume
With the availability of advanced numerical techniques to formulation, the above equations are written in the integral
solve the full Reynolds-averaged Navier-Stokes (RANS) form as
equations, it is now possible to obtain detailed flow field
around the wing, which can provide valuable input to  
U n dxdydz +t   E n 1e x + Fn 1e y  G n 1e z  .n dS=0

(4)
analyze and enhance the performance of the aircrafts. where  is any three-dimensional flow domain  is the
boundary surface and n is the unit vector normal to the
The present work describes the application of an implicit surface bounding the control volume in the outward
Reynolds-averaged Navier-Stokes code (IMPRANS), direction. In the nodal point finite volume approach [2-7],
developed at the Computational and Theoretical Fluid the flow variables are associated with each mesh point (i, j,
Dynamics division of National Aerospace Laboratories k) of the grid and the integral conservative equations are
(NAL), Bangalore to the cropped delta wing for a range of applied to each control volume Ωijk obtained by joining the
Mach numbers and at different angles of attack. The centroids of the eight neighbouring hexahedron cells
RANS solver [2, 3] is based on an implicit finite volume surrounding the nodal point. Application of nodal point
nodal point scheme wherein a control volume is formed by spatial discretization to equations (4) leads to
joining the centroids of the neighbouring cells around a
nodal point in the computational domain. The numerical t m  6 
  E I  EV  m Smx   FI  FV  m Smy   G I  G V  m Smz   0,
n 1 n 1 n 1
scheme for solving the two- and three-dimensional Uijkn  (5)
ijk m 1 
Reynolds-averaged Navier-Stokes equations governing
steady and unsteady viscous compressible flow has been where Smx, Smy and Smz are the x, y and z components of the
derived indigenously by using Euler's implicit time surface vector corresponding to the m-th surface of the
differencing formula with nodal point spatial control volume. Linearizing the changes in flux vectors
discretization. Certain basic ideas from the implicit finite using Taylor’s series expansions in time and assuming
difference scheme of Beam and Warming [4], the nodal locally constant transport properties, equations (5) can be
point schemes of Ni and Hall [5], the Runge Kutta time- simplified to, after omitting superscript n,
stepping scheme of Jameson [6] and the cell-centered
schemes due to Hollanders and Viviand [7] have been U ijkn 
t 6   E R    F    G   
  A  x  U  Smx   B  yS  U Smy   C  zT  U  Smz 
combined efficiently to evolve the present method. The ijk m1   m   m   m 
t 6
(6)
code has been applied successfully for simulating    EI  E V  m Smx   FI  FV  m Smy   G I  G V  m Smz 
ijk m 1 
compressible flow around stationary or moving bodies like
aerofoils, wings, helicopter rotor and wind turbine blades Here A = ∂EI / ∂U, B = ∂FI / ∂U, C = ∂GI / ∂U, ER = ∂EV1/
[2, 3]. The algebraic eddy viscosity model due to Baldwin ∂Ux, FS = ∂FV2/ ∂Uy and GT = ∂GV3/ ∂Uz are the Jacobian
and Lomax is used for the turbulence closure [8].
matrices, EI, FI and GI are the inviscid flux vectors and EV,
FV and GV are the viscous flux vectors.
It may be noted here that the surface integrals in equations
(4) are evaluated by summing up the contributions due to
the flux terms over the six faces of the computational cell.
The terms containing inviscid flux vectors are calculated
by using the flow variables at the six neighbouring points
and the derivatives in the viscous flux terms are discretized
directly in the physical plane using Taylor's series
expansions. The resulting block tridiagonal system of
equations is solved by using a suitable block tridiagonal
solution algorithm and proper initial and boundary
conditions. A blend of second and fourth order artificial
dissipation terms is added explicitly to ensure convergence
and to suppress oscillations near shock waves. Implicit
second order dissipation terms are also added to improve
the practical stability bound of the implicit scheme.

1. Results and discussion

The present work describes the CFD analysis of flow over a

65 - degree sweep cropped delta wing with round leading
edge. The computations have been carried out using
Implicit Reynolds-averaged Navier-Stokes solver Fig. 1: Surface grid around the cropped delta wing
IMPRANS for the flow over a cropped delta wing for (167 × 61)
different Mach numbers and at different angles of attack.
The C-H type of grids around the semi-span of the wing of To allow the validation of computational results,
size 247  65  75 are used. The distribution of grid points comparisons were made with the experimental surface
on the wing surface is shown in figure 1. The cases pressure co-efficient distributions at three different
considered here are corresponding to the experiments of spanwise locations on the wing surface z/s = 0.3, 0.6 and
Hess et al. [1]. The present cases are selected to study the 0.8.
effect of different parameters on the lift and drag in the
transonic flow conditions and also on the flow field around Table 1: Aerodynamic parameters used for computation
the wing. The aerodynamic parameters used for the present
computations are given in table 1.

Test Mach Reynolds α deg.

cases number number
M Re× 106
Case 1 0.85 2.38 10
. Case 2 0.85 2.38 20
Case 3 0.85 2.38 30

The steady surface pressure distribution over the wing is

shown in figures 2 and 3, along with the experimental
values at different spanwise stations for α = 10 and at M ∞ =
0.85 and for α = 20 and at M ∞ = 0.85. The overall
comparison of surface pressure distribution at all the
spanwise sections, in all the cases computed show good
agreement between experimental values and the present
computations.
 Fig. 4: particle traces at different cross flow planes of
cropped delta wing at 30, 60, 80, 95, 110% of the root
chord for α = 100 and M∞=0.85, Re = 2.38*106

Fig. 2: Comparison of computed and experimental

coefficient of surface pressure distribution at different
spanwise stations at α = 3.04° and M∞ = 0.404

Fig. 5: Pressure contours on the upper surface of cropped

delta wing at α = 10° and M∞ = 0.85, Re = 2.38*106
Table 2: The lift and drag coefficients obtained for the
different cases

The figure 4 shows the particle traces at different cross

flow planes of cropped delta wing at 30, 60, 80, 95, 110%
of the root chord for α = 100 and M∞=0.85, Re = 2.38*106.
This case being a higher angle of attack and transonic
Mach number, the flow field is with shock and separation.
From the surface properties it can be seen that there is
decrease in pressure and density near the leading edge
from root to tip this could be due to the formation of two
rolling vortices on suction surface of the wing. Fig 5 shows
Pressure contours on the upper surface of cropped delta
wing at α = 10° and M∞ = 0.85, Re = 2.38*106 Here,
leading edge vortex formed on the upper surface is seen
clearly at different sections of the span.

Fig. 3: Comparison of computed and experimental

coefficient of surface pressure distribution at different
spanwise stations at α = 2.99° and M∞ = 0.90
 observed on the upper surface of the wing. The lift and
drag coefficients for all the cases are listed in Table 2.
Comparing the various cases in the table the Lift is
Cases α M Re× 106 Cl Cd
increased with increase in angle of attack for the given
Case1 0.05 0.399 9.35 0.001094 0.00199 Mach number.

Case2 0.05 0.899 9.77 0.001421 0.00235 2. Conclusions

Case3 0.00 1.120 9.68 0.000001 0.01003 CFD analysis of steady flow over a cropped delta wing has
Case4 3.04 0.404 9.25 0.061523 0.00125
been performed to determine the aerodynamic lift, drag
and surface pressure coefficients over a wide range of
Case5 5.04 0.403 9.22 0.102999 0.00043 Mach numbers and angles of attack. The comparison of
computed surface pressure coefficients in all the cases are
Case6 2.99 0.900 9.80 0.076107 0.00151 in good agreement with the experimental data. A leading
Case7 4.24 0.901 10.08 0.108580 0.00107
edge vortex became evident from at around 3° angle of
attack.

Acknowledgements
Authors are gratefully acknowledge to Dr. P. K. Dutta and
Dr. Vimala Dutta for their constant encouragement and
support during the work.

References
[1] Hess, R. W., Cazier, F. W. and Wynne, E. C., “Steady
and Unsteady Transonic Pressure Measurements on a
Clipped Delta Wing for Pitching and Control-Surface
Oscillations”, NASA TP-2594, Oct. 1986.
[2] Dutta V., Dutta P. K. and Sharanappa, “An Implicit
RANS Solver for Unsteady Compressible Flow
Computations”, NAL SP 0301, NAL, Bangalore, pp.
65-86, March 5, 2003.
[3] Dutta V., Sharanappa and Dutta, P. K., “Navier-Stokes
Computations for a Helicopter Rotor Blade in
Hover”, Proc. 8th Annual CFD Symp., CFD Division
of AESI, Bangalore, August 11-13, 2005.
[4] Beam, R.M., and Warming. R.F., “An Implicit
Factored Scheme for the Compressible Navier-Stokes
Equations”, AIAA J., vol.16, no.4, pp.393-402, 1978.
[5] Hall, M.G., “Cell Vertex Multigrid Scheme for
Solution of the Euler Equations”, Proc. of Conf. on
Num. Methods for Fluid Dyn., pp. 303-345, 1985.
[6] Jameson, A., Schmidt, W. and Turkel, E., “Numerical
Solution of Euler Equations by Finite Volume
Methods Using Runge-Kutta Time Stepping
Schemes”, AIAA Paper No. 81-1259, 1981.
[7] Hollanders, H,, Lerat, A. and Peyret, R., “Three-
Dimensional Calculation of Transonic Viscous Flows
by an Implicit Method”, AIAA J., vol. 23, no.11, 1985.
[8] Baldwin, B.S., and Lomax, H., “Thin Layer
Approximation and Algebraic Model for Separated
Turbulent Flows”, AIAA Paper No. 78-257, 1978.
Fig. 6: Mach number contour plots at different spanwise
sections on a cropped delta wing at α = 4.24° and M ∞ = 0.901
The Mach number contours are plotted in figure 6, for α =
4.24 and M∞ = 0.901, Due to the transonic Mach number
and higher angle of attack, there is a strong shock wave is