ODE45 and dsolve result discrepency

I'm having a weird problem. I'm trying to solve a 2nd order ode with both ode45 and dsolve. The results are fine as long as I have non-zero initial conditions but they don't match when the equation has zero initial condition. Any idea why this is happening? Also, which one would be the right choice as I am supposed to implement it in a larger piece of code.

Here is my script:

clc
%clear all
m=3.4e6;
 k=3.51e10;
 c=13.8e6;
 f=7.2578;
[t1,x]=ode45(@pend,[0 5],[0 0] );
j=1;
t2=0:0.01:5;
l=length(t2);
disp2=zeros(l,1);
vel2=zeros(l,1);
for t=0:0.01:5
  sol_disp2=dsolve('m*D2x+c*Dx+k*x=f','x(0)=0,Dx(0)=0'); 
  sol_vel2=diff(sol_disp2);
  disp2(j)=vpa(subs(sol_disp2));
  vel2(j)=vpa(subs(sol_vel2));
  j=j+1;
end

subplot(2,2,1)
plot(t1,x(:,1))
subplot(2,2,2)
plot(t1,x(:,2))
subplot(2,2,3)
plot(t2,disp2)
subplot(2,2,4)
plot(t2,vel2)

and the ode45 function:

function dxdt = pend(t,x)
m=3.4e6;
 k=3.51e10;
 c=13.8e6;
 f=7.2578;
x1=x(1);
x2=x(2);
% fun=@(x) sin(x)/z2;
dxdt=[x2; (f-c*x2-k*x1)/m];
end

thanks in advance

[Torsten]

Just insert your symbolic solution into

m*D2x+c*Dx+k*x=f, x(0)=x'(0)=0

to see whether it's correct or not.

Best wishes

Torsten.

Answer 1

[Nitin Khola]

Hey Ridwad,

I understand you are facing discrepancies in solutions from "dsolve" and "ode45" for zero initial conditions. It appears that the system has faster dynamics compared to the default tolerances in "ode45". You can set the absolute and relative tolerances to smaller values using "odeset" as follows:

 >> options = odeset('RelTol', 1e-10, 'AbsTol', 1e-12);
 >> [t1,x]=ode45(@pend,[0 5],[0 0],options);

Setting these tolerances to appropriate values get the solutions from the two solvers to match as shown below. Hope this helps.

[Ridwan Hossain]

Hi Nitin, Thank you very much. It solves the problem.

Ridwan