Step input equivalent for system identification app

Hello

I have been trying to use the system identication app for step input experiments.

For a step input of 9v I've been using an array of ones multiplied by 9.

I've also tried to keep the first few entries 0 to comply with the transient response warning.

When I use the compare function on the resulting linear function and of my iddatas i get a very accurate fit,

however when I try the step response it's completely different.

Sometimes the system that reaches its steady state within a couple of seconds at most, takes 18000 secs for the step response.

I'd be grateful if someone were to explain what I'm doing wrong.

Answer 1

[Moksh]

Hi Youssef,

You can try using the 'createStep' function in MATLAB for creating step responses. This function provides with faster frequency responses and might be able to help you with reaching steady states in a more regular fashion.

Here is an example code for implementing a step response using this function.

input = frest.createStep('StepTime',5,'StepSize',9);
plot(input);

For further understanding of this function please refer to the following document

https://in.mathworks.com/help/slcontrol/ug/frest.createstep.html

Hope this helps!