You are on page 1of 5

Artificial Viscosity

Daniel Topa

September 22, 2016

Abstract
A popular misconception about numerical solutions of partial di↵erential equations
is that smaller time steps produce better solutions; shrinking the time step can com-
pletely dissipate the propagating wave. The archetypical case is the advection equation
solved below using a Lax scheme. The averaging in the spatial derivative introduces
artificial viscosity; as the time step shrinks this viscosity damps out the solution. A
demonstration follows.

1 Problem statement
One of the most elementary hyperbolic di↵erential equations is the advection equation.
Given a function a : R2 7! R the spatial and temporal derivatives are related by the
wave velocity c:
@a @a
= c . (1)
@x @t
Given the initial condition
a(x, 0) = f0 (x) (2)
the analytic solution is
a(x, t) = f0 (x ct) . (3)
The wavefront travels without distortion; this provides a simple comparison with nu-
merical solutions.

2 Numerical implementation
The forward time, centered space (FTCS) discretization uses derivatives forward in
time
@a a(xj , tµ + ⌧ ) a(xj , tµ )
⇡ + O (⌧ ) , (4)
@t ⌧
and centered in space

@a a(xj + h, tµ ) a(xj h, tµ )
⇡ + O h2 . (5)
@x 2h

1
The numerical scheme of Lax advances the time step according to
1 µ c⌧ µ
aµ+1
j = a + aµj a aµj (6)
2 j+1 1
2h j+1 1

where the mesh points are represented by subscripts j = 1, 2, . . . , n and the time steps
are denoted by the superscript µ = 0, 1, 2, . . . .

3 Example
A toy problem demonstrates the e↵ect of the artificial viscosity in a problem with
periodic boundary conditions: a(x + L, t) = a (x, t). The input wave is a cosine mod-
ulated Gaussian pulse with periodic boundary conditions. The simulation parameters
are shown in table (1).

parameter value use


n 50 points in spatial mesh
⌧ 0.15 time step
L 1 domain length
c 1 wave velocity

Table 1: Input parameters for the simulation.

The derived parameters are the grid spacing

h = L/n = 0.02 (7)

and the Courant-Freidrichs-Lewy (CFL) parameter

h
tmax = = 0.02 (8)
c
The time step is ⌧ = 0.015 < tmax which will quickly damp out the signal.
Figure (1) shows the numerical computation (blue dots) against the analytic solu-
tion (gray line) at time steps 0 and 5; the damping of the wave is immediately apparent.
Figure (2) shows time steps 10 and 50 and strong damping of the signal.

4 Discussion
The artificial viscosity is an essential feature of the discretization of the problem whose
strength is inversely proportional to the magnitude of the time step ⌧ . When the time
step is too large (⌧ > tmax ) the damping does not have the strength to stabilize the
numeric problem. The figures below show that when the time step is too small the
damping overwhelms the signal.

2
References
[1] Randall J. LeVeque, Finite Di↵erence Methods for Ordinary and Partial Di↵er-
ential Equations, SIAM, Society for Industrial and Applied Mathematics (2007).
[2] Alejandro L. Garcia, Numerical Methods for Physics, Prentice Hall (2000).
[3] Michelle Schatzman, Numerical Analysis: a mathematical introduction, Oxford
University Press (2002).
[4] Mathematica notebook,
/Users/rditldmt/Dropbox/ nb/drc/AdH/advection/advection basics 03.nb (2016)

3
Time step 000

Time step 005

Figure 1: Early time evolution; the gray line is the analytic solution. The black dots are
the numerical solution.

4
Time step 035

Time step 065

Figure 2: Late time evolution.

You might also like