Closed-loop Frequency Data-driven PID Retuning

Abstract

This paper is about a data-driven frequency domain PID retuning from the initial controller parameters. It uses data collected from a specific closed-loop experiment to estimate frequency response points. The data-driven retuning method is applied to make a process matches with the defined reference model using a convex optimization technique that minimizes modified cost function. The proposed strategy is demonstrated in laboratory scale thermal plant based on Peltier effect, using crossover and critical frequencies estimations.