sandbox/zaleski/bump2D-vdw-1D.c
Bouncing Saint-Venant bump with the Mc Cormack scheme in 1D.
This test case is a restriction to 1D of bump2D-vdw.c for debugging purposes
#include "grid/cartesian1D.h"
#include "vdw.h"We start with initial conditions etc… as when using the standard bump2D.c .
#define LEVEL 8
double P0(double x)
{
return 0.5*x*x;
}
int main()
{
origin (-0.5, -0.5);
init_grid (1 << LEVEL);
DT = 1e-4;
run();
}
event init (i = 0)
{
lambda=0.;
foreach()
{
mu[] = 1.e-3;
rho[] = 0.1 + exp(-200.*(x*x + y*y)/(L0*L0));
q.x[] = q.y[] = 0.;
}
}
event logfile (i++) {
stats s = statsf (rho);
fprintf (stderr, "%g %d %g %g %g %.8f\n", t, i, dt, s.min, s.max, s.sum);
}
event outputfile (t <= 0.4; t += 0.2) {
foreach()
{
fprintf (stdout, "%g %g %g\n", x, rho[],q.x[]);
//if (rho[] < 0.)
//return 1; // stops when rho becomes unphysical.
}
fprintf (stderr, "\n");
}Same test with Saint-Venant solver
Results
plot 'out' u 1:2 w l t "rho", 'out' u 1:3 w l t "q.x"
Evolution of the density and momentum. The results are comparable to that obtained when running with the Saint-Venant solver but the MacCormack scheme has oscillations. (script)
