https://www.youtube.com/watch?v=ONDp4Rdj0OY
Matlab program with the explicit Lax method for the advection equation
%Square-wave test for the lax method to solve the advection equation
clear;
%Parameters needed to Implement the advection
Lmax = 1.0; % Maxinum length
Tmax = 1.; % Maximum time
c = 1.0; % Advection velocity
%Parameters needed to Implement the Lax Method
Nt = 300; % Number of time steps
Dt = Tmax / Nt; % Time Step
Nx = 300; % Number of space steps
Dx = Lmax / Nx; %Space step
b = c * Dt/(2.*Dx); % beta parameter in the finite-difference imp
%Remember that the lax method is stable for b =< 1/2
%But it gets diffussed unless abs(b) = 1/2
%Initial Condition
Nfront = round(Nx / 2); % Square wave
for i = 1:(Nx + 1)
if i < Nfront
u(i,1) = 1.;
else
u(i,1) = 0.;
end
x(i) = (i-1)*Dx; % We also define vector x, due to the space discretization
end
%Boundary Condition
for k = 1:Nt + 1
u(1,k) = 1.;
u(Nx + k) = 0.;
t(k) = (k-1)*Dt; % We also define vector t
end
%Implement of the lax method
for k = 1:Nt
for i = 2:Nx
u(i,k+1) = 0.5*(u(i+1,k)+u(i-1,k))+b*(u(i+1,k)-u(i-1,k));
end
end
plot(x,u(:,1),'-b'); %time = 1
plot(x,u(:,round(Nt/10)),'--g'); %time = Nt / 10
plot(x,u(:,round(Nt/3)),':b'); %time = Nt / 3
plot(x,u(:,Nt),'-.r'); %time = Nt
xlabel('X');
ylabel('Amp(X)');
首先Nt = 300,这样b刚好等于0.5,介于稳定和不稳定之间
Nt = 400,这样b就小于0.5,可以说是稳定了
Nt = 290,b就大于0.5,不稳定