Using Runge-Kutta algorithm to solve second order ODE

27 次查看(过去 30 天)
For this task I need to create a program which manually uses the RK algorithm to solve a second order ODE. I'm trying to get my head around it but I think I'm going at it in the wrong way.
My original function is with given values for m, F0 and w, so I have rearranged it to x'' = 50sin(8*pi*t)-x'-50x
I've then switched x' for y and rearranged to get 50sin(8*pi*t) = y' + y + 50x and this is what I'm using in the script. I need to plot x(t) and y(t) but at this point I'm just trying to find values for x and y over time.
The method I've used to get this far is below the code. Excuse my poor code. Thanks in advance.
PS I'm not allowed to use a function like ode45 to solve it for me.
close all; clear; clc;
t=0;x=0;y=0;
dy=myfunction(t,x,y)
x0=0;y0=0
h = 0.5;
for t=0:5.5
dx1 = h*y
dy1 = h*dy
dx2 = h*(y+(dy1/2))
dy2 = h*myfunction(t+(h/2),x+(dx1/2),y+(dy1/2))
dx3 = h*(y+(dy2/2))
dy3 = h*myfunction(t+(h/2),x+(dx2/2),y+(dy1/2))
dx4 = h*(y+dy3)
dy4 = h*myfunction(t+h,x+dx3,y+dy1)
deltax = (dx1+(2*dx2)+(2*dx3)+dx4)/6
deltay = (dy1+(2*dy2)+(2*dy3)+dy4)/6
xth = x+deltax
yth = y+deltay
x=xth
y=yth
t = t+h
end
function [dy] = myfunction(t,x,y)
dy=(50*sin(8*pi*t))-(50*x)-y;
end
Method used (I swapped v for y):

回答(1 个)

Torsten
Torsten 2022-4-22
Can you take it from here ?
%Solves y'' - (-exp(-B*t)-y+5*exp(-2*t)-2*exp(-(B+2)*t)+exp(-B*t)+t) = 0
% t in [0 1]
% y(0) = 1, y'(0) = -1
% B = 4
tstart = 0.0;
tend = 1.0;
h = (tend - tstart)/20;
T = (tstart:h:tend).';
Y0 = [1 -1];
B = 4;
f = @(t,y) [y(2) -exp(-B*t)-y(1)+5*exp(-2*t)-2*exp(-(B+2)*t)+exp(-B*t)+t];
Y = runge_kutta_RK4(f,T,Y0);
plot(T,Y)
function Y = runge_kutta_RK4(f,T,Y0)
N = numel(T);
n = numel(Y0);
Y = zeros(N,n);
Y(1,:) = Y0;
for i = 2:N
t = T(i-1);
y = Y(i-1,:);
h = T(i) - T(i-1);
k0 = f(t,y);
k1 = f(t+0.5*h,y+k0*0.5*h);
k2 = f(t+0.5*h,y+k1*0.5*h);
k3 = f(t+h,y+k2*h);
Y(i,:) = y + h/6*(k0+2*k1+2*k2+k3);
end
end
  2 个评论
Daniel Jordan
Daniel Jordan 2022-4-22
So can I change this line
f = @(t,y) [y(2) -exp(-B*t)-y(1)+5*exp(-2*t)-2*exp(-(B+2)*t)+exp(-B*t)+t];
into my own equation? Wouldn't it need to have t, x and y?
Also what do y(1) and y(2) mean in that line?
Torsten
Torsten 2022-4-22
编辑:Torsten 2022-4-22
For your function,
f = @(t,y) [y(2) F0/m*sin(omega*t)-c/m*y(2)-k/m*y(1)]
where
y(1) = y, y(2) = y'.
So you shouldn't work with x and y, but with a vector y=(y(1),y(2)) to make everything easier.
Of course, omega, F0, m and k have to be given values before.

请先登录,再进行评论。

类别

Help CenterFile Exchange 中查找有关 Ordinary Differential Equations 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by