PID Tuning for fixed settling time and minimum overshoot

Hi all I am a bit new to PID tuning and PIDs in general so kindly assist me

I have a PID that needs to be tuned for the operation of a motorized Propotional Valve working with a Vacuum Pump as a Plant to control pressure

Feedback to PID is the vacuum pressure output measured from the valve+pump system with input as the required vacuum pressure and output of the PID is the duty cycles in the form of a digital number restricted between 0 and 100

The pump and valve data is unavailable so I am using experimental data from the actual plant to create a transfer function based LTI plant for use in Tuning

The system needs to settle within a time of 10s or less with minumum overshoot

Unfortunately no matter whatever I try be it with using parts of the data or the full data the system has a settling time much longer than this

How do I go about solving this issue ?

[Sam Chak]

Hi @Anirudh

Can you post the identified transfer functions for the vacuum pump? We need them to examine whether a PID controller will suffice. Otherwise, consider a higher‑order dynamic compensator (PID is just a 2nd‑order special case).

Answer 1

[sneha]

Hello,

If the system cannot achieve a settling time under 10 seconds, check the following points:

1. Model validation: Ensure that the identified transfer function or LTI model accurately represents the plant’s real response. Compare simulated and measured step responses. If the model appears too sluggish, consider identifying a higher-order model or including time delays. Reference: https://www.mathworks.com/help/ident/gs/system-identification-workflow.html
2. Physical limitations: The plant may have inherent response constraints due to valve and pump characteristics. No controller can exceed the physical bandwidth of the system. Reference: https://www.mathworks.com/help/overview/control-systems.html
3. Controller aggressiveness: Increase the proportional gain to speed up the response and introduce derivative action to reduce overshoot, while monitoring stability and noise sensitivity. Reference: https://www.mathworks.com/help/control/ref/pidtuner-app.html
4. Actuator saturation: Verify that the controller output is not continuously saturating at 0% or 100%. Implement anti-windup protection to prevent integrator accumulation and improve recovery. Reference: https://www.mathworks.com/help/simulink/slref/anti-windup-control-using-a-pid-controller.html
5. Feedforward compensation: If nominal operating conditions are known, add a feedforward term to improve response speed and reduce controller effort. Reference: https://www.mathworks.com/matlabcentral/fileexchange/131329-simulation-of-feedforward-control-techniques