From the control system toolbox, I plotted the transfer function response using step() function and got a response with overshoot. However, I got another response for the same transfer function but without overshoot in the Simulink environment. After research, I found that, in the control system toolbox, the derivative gain in the PID controller that I set was pure (without filter), and when I set the derivative gain with filter (N = 1x10^18) in the Simulink environment, I get approximately the same response as I get with the control system toolbox. Therefore, I have three questions in mind here:
I am looking forward to seeing the answers to these questions.
The way i understand your question, i would like ro answer question no.2 and question no.3
Reply to question 2.
Filter is associated only with derivative function since derivative function amplify the noise which comes with feedback signal in the loop.so this filter filter out the amplified noise to get ripple free signal under steady state and less overshoot in dynamic response.
Reply to question 3.
Derivative block belongs time domain. Function 'S' belongs frequency domain.both can not be mixed up in the circuit of either time domain or frequency domain.
Viswanatha v.
Thank you for your comment. Although, I am not so satisfy with your answers to my question 2 and 3. For your answer to Q2, when I apply a filter to the derivative block in Simulink, I get the response with overshoot, but without filter no overshoot. For your answer to Q3, as I know that the Simulink environment holds a frequency domain, so why derivative block is in time domain while integrator is in frequency domain?
Meanwhile refer my article which speaks about question nos.2 and 3.
https://publons.com/publon/25822886/
Hi, sometimes I've got the same problem using Simulink, however, always I've realized it was a mistake in the implementation. Regarding question 2, the main reason you need the filter is that once you include the D component, your closed-loop system becomes more sensible with respect to the noise of the output measurements, then you would like to filter it. Another reason is that a PID controller itself is a noncausal system, then you make it causal adding the filter. Regarding question 3, I don't remember quite well whether you can or not add only 's', but what you can do instead of using the PID block is to incorporate the PID controller transfer function using the transfer function blocks.
I agree with the response of Jorge Humberto Urrea-Quintero.
The additional filter for component D from (PID controller) has the role of attenuation of disturbances, but first of all it ensures the causality of the system (physical realization) : the degree of numerator it's less or equal with the degree of denominator of the transfer function of PID. In simulations this can be hidden and therefore generates errors.
Jorge Humberto Urrea-Quintero
Thank you for your reply, although in Simulink, I cannot add a transfer function of the controller since the degree of numerator must be less or equal to the degree of denominator.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Control Theory
- Control Systems