The shock-tube tests presented in Falle (1998) were used to
test the MHD Riemann solver. These tests consist of
the Brio (1988) test problem and a number of other tests which
focus on a specific MHD wave, as listed in Table 2.2. All
tests use
except the Brio and Wu test with ,
and all are dimensionless and run on a domain of unit length. Note
the Alfvén wave test is not a discontinuity so the left and right
states are the same. The wave is set up to have width 0.3 and the
velocity and magnetic fields rotate by in the plane over
this length. This test also uses periodic boundary conditions to
advect the wave across the domain five times. Other tests use
zero-gradient boundary conditions and only the small startup waves
leave the domain before the end of the simulation. The results are
shown in Fig. 2.2. All of the results are very
similar to those shown in Falle (1998) showing that our MHD
approximate Riemann solver correctly captures all of the waves
generated. Note that Falle (1998) evolve the Alfvén wave
three times its width for their test whereas we have evolved it 5
times across the domain; this is why our results have degraded
significantly more than theirs.
Table 2.2:
MHD Shock Tube Test Problems from Falle (1998).
State variables are
. BW is
the Brio & Wu test problem, AW is an Alfvén wave, FS and SS are
fast and slow switch-on shocks, FR and SR are fast and slow
rarefaction waves.
Test
Left State
Right State
BW
0.12
0.5
AW
5.0
0.65
FS
0.4
0.3
SS
0.5
0.3
FR
0.1
0.5
SR
0.3
0.5
Figure 2.2:
1D shock tube results for adiabatic MHD. The solid
line shows the results using 10,000 grid cells, and the
points show results using 200 cells. The figures show results for
the tests in Table 2.2 and are from top to bottom: BW,
AW, FS, SS, FR, SR. Note for the Alfvén wave the solid black
line is the initial state and the cyan symbols show the high
resolution results.