Lsim function returns an array filled with Nan

Im using lsim to simulate a MVE, for some reason the response y2 is filled with NaN when It should have the system response.

R=1;
C1=1*10^-6;
L=1*10^-3;

A= [1/C1 -1/C1 0; -2*(R/L) R/L -1/L; R/L (-5/2)*(1/L) 1/L];
B= [0 0; 1/L 0; 0 -1/(2*L)];
C= [0 1/2 0; 0 0 1]; 
D= [0 1/(2*R); 0 0];

sys = ss(A,B,C,D);

t2 = 0:0.01:20;
u1 = 5*heaviside(t2-3);
u2 = 10*heaviside(t2-6);
u = [u1(:), u2(:)];

[y2,t2]=lsim(sys,u,t2);
y2

y2 = 2001×2

NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN

Answer 1

[Paul]

Hi Gabriel,

Recheck the model. As written it has two unstable poles way out in the right half plane. That will be very difficult to simulate.

R=1;
C1=1*10^-6;
L=1*10^-3;
A= [1/C1 -1/C1 0; -2*(R/L) R/L -1/L; R/L (-5/2)*(1/L) 1/L];
B= [0 0; 1/L 0; 0 -1/(2*L)];
C= [0 1/2 0; 0 0 1]; 
D= [0 1/(2*R); 0 0];
sys = ss(A,B,C,D);
format short e
eig(sys)

ans = 3×1

1.0e+00 * 1.0020e+06 1.5801e+03 -1.5791e+03

Also, be careful using heaviside. If using the default sympref, then we have

heaviside(0)

ans = 
   5.0000e-01

You can change that 1 using sympref, which is probably what you want. Or define a new function like

unitstep = @(t) 0.5*(t==0) + heaviside(t);
unitstep(0)

ans = 
     1