Revisiting the impulse response creates an improved PID autotuner.

Area of Science:

• Control Engineering
• Process Identification
• Automation

Background:

• Industrial controller tuning faces challenges due to uncertain process information and avoidance of comprehensive identification.
• Changes in process properties or operating conditions often lead to ignored controller re-tuning, causing reduced productivity.

Purpose of the Study:

• To introduce a novel methodology for extracting relevant process information from short, minimally rich datasets.
• To enable automatic and robust tuning of Proportional-Integral-Derivative (PID) controllers.

Main Methods:

• A sine test, with duration related to loop settling time, is superimposed on nominal process inputs.
• The method estimates process impulse response coefficients, providing a band-limited frequency response.
• Applicable under both open-loop and closed-loop operating conditions.

Main Results:

• The methodology successfully extracts essential process information from limited data.
• Automatic robust PID controller tuning is achieved based on the extracted frequency response.
• Validation through numerical examples and experimental tests on integrating, non-minimum phase, and poorly damped systems.

Conclusions:

• The proposed method offers an effective way to tune controllers in industrial settings with uncertain or changing process dynamics.
• It provides a practical solution for improving long-term productivity by addressing controller re-tuning issues.
• The approach demonstrates superior relevance compared to widely used autotuning methods.