Simulink simulation error: solution not finite

I encountered the following simulation error:

I tried reducing the step size in the solver but it doesn't work. How should I fix this?

[Walter Roberson]

Sometimes there is just a singularity, and you need to either change the blocks (so you are doing something different), or else you need to change the initial or boundary conditions in a way that avoids the singularity.

I recommend logging (scoping) the output of the signal to see what the trend is.

[Randy Chen]

Hi, here is what happened to the scope of my angle:

It seems like at T=3.5 s, the angle just blew up to an astronomical value. I wonder how I should fix this error?

[Randy Chen]

It doesn't look like I can change the initial conditions for my 'theta' block, because I'm modelling an inverted airfoil, and the initial angle measured from the vertical is pi, and that's consistent with how I derived my equations of motion

[Walter Roberson]

I suggest that set the stop time to 3.5 and look at the signal. Flat on the 10^190 scale might be pretty bumpy locally.

[Randy Chen]

It turns out that 3.5 seconds also yields a non-finite result. I changed the stop time to 2.3 seconds, and here is what happened:

\

Does this mean I have to tune my PID gains now?

[Walter Roberson]

I am a little confused, in that your original post the error message had a singularity about 1.88 seconds

Sorry, I have never looked at airfoils at all.

Answer 1

[Aqib Memon]

Are you using a fixed time step in your simulation settings? if yes, then try setting it to variable time step and see if it works.

[Randy Chen]

I am using variable time step right now