VOL. 24, NO. 12, DECEMBER 1986 AIAA JOURNAL 1931
Artificial Dissipation Models for the Euler Equations
Thomas H. Pulliam*
NASA Ames Research Center, Moffett Field, California
Various artificial dissipation models that are used with central difference algorithms for the Euler equations
are analyzed for their effect on accuracy, stability, and convergence rates. In particular, linear and nonlinear
models are investigated using an implicit approximate factorization code (ARC2D) for transonic airfoils. Fully
implicit application of the dissipation models is shown to improve robustness and convergence rates. The treat-
ment of dissipation models at boundaries will be examined. It will be shown that accurate, error free solutions
with sharp shocks can be obtained using a central difference algorithm coupled with an appropriate nonlinear
artificial dissipation model.
I. Introduction
T
HE solution of the Euler equations using numerical tech-
niques requires the use of either a differencing method
with inherent dissipation or the addition of dissipation terms
to a nondissipative scheme. This is because the Euler equa-
tions do not provide any natural dissipation mechanism (such
as viscosity in the Navier-Stokes equations) that would
eliminate high frequencies which are caused by nonlinearities
and especially shocks. A variety of numerical algorithms and
computer codes for the Euler equations have been developed.
Methods such as MacCormack's
1
explicit scheme and
Steger's
2
application of the Beam and Warming
3
implicit
algorithm are in wide use. Some notable recent developments
based on explicit Runge-Kutta schemes are the work of
Jameson et al.
4
and Rizzi and Eriksson.
5
The time integration
scheme, boundary condition treatment, and other details are
different from method to method. These all have one thing in
common: The use of a basically central difference approxima-
tion to the spatial derivatives and the addition of some form of
artificial dissipation. In contrast there is a currently popular
class of schemes, [monotone, total variation diminishing
(TVD), flux split, flux difference, lambda] that employ some
form of upwind differencing under the assumptions of
characteristic theory and wave propagation. The work of
Steger and Warming,
6
Roe,
7
Van Leer,
8
Osher and Chak-
ravarthy,
9
and Marten's TVD methods
10
all fall in this
category. Although it will not be shown here for every case,
these schemes are all equivalent to a central differencing scheme
plus some form of dissipation.
The addition of artificial dissipation to central differenc-
ing will be the focus of this paper. It is added for two main
reasons: first, to control the odd-even uncoupling of grid
points typical of central differencing and, second, to control
strong nonlinear effects such as shocks. The paper in-
vestigates various forms of artificial dissipation employed in
the Euler codes. The particular difference forms and theory
behind the choice of coefficient will be examined here, as
well as some stability and accuracy arguments when possible.
Results for flow over airfoils with shocks will be used as
test cases. A recent version of Steger's
2
two-dimensional im-
Presented as Paper 85-0438 at the AIAA 23rd Aerospace Sciences
Meeting, Reno, NV, Jan. 14-17, 1985; received May 10, 1985; revi-
sion received March 17, 1986. Copyright © American Institute of
Aeronautics and Astronautics, Inc. No copyright is asserted in the
United States under Title 17, U.S. Code. The U.S. Government has
a royalty-free license to exercise all rights under the copyright claimed
herein for Governmental purposes. All other rights are reserved by the
copyright owner.
* Research Scientist, Computational Fluid Dynamics Branch.
Member AIAA.
plicit central difference scheme described in Refs. 11 and 12
will be used as the test code. The dissipation models dis-
cussed here have been applied in viscous computations, but
will not be addressed in this paper. In particular, the interest
is in evaluating the effect of the artificial dissipation on the
accuracy and stability of fluid dynamic results.
II. Two-Dimensional Euler Equations
Much of the development of the equations and algorithm
used here is available from other sources (see Refs. 2, 3, 11,
and 12. Just the aspects important to the current develop-
ments are presented here.
The two-dimensional Euler equations can be transformed
from Cartesian coordinates to general curvilinear coordinates
where
r = t t = Z(x,y,t) ri = ri(x,y,t) (1)
The Euler equations written in generalized coordinates are
=o (2)
P
pu
pv
e
E = J
PU
F=J
pV
pvV+ri
y
p
V(e+p)-T)
t
p _
where
are the contravariant velocities. Pressure is related to the
conservative flow variables Q by the equation of state
p=(y-l)[e-V2
P
(u
2
(3)
where y is the ratio of specific heats, generally taken as 1.4.
The speed of sound a =\lyp/p. The ( " ) is dropped for
simplicity of notation except where noted.