Proportional-Integral-Derivative Controllers Tuning for Unstable and Integral Processes Using Genetic Algorithms

Abstract

Abstract. During the last years the use of intelligent strategies for tuning Proportional-Integral-Derivative (PID) controllers has been growing. The evolutionary strategies have won an important place thanks to their flexibility. In this paper, the automatic tuning of systems with scarce initial information and integrative and unstable dynamics, through a genetic approach is presented. The advantages of the proposed approach were highlighted through the comparison with the Ziegler-Nichols modified closed loop method, and the Visioli genetic approach. The proposed methodology goal is to expand the intelligent tuning application to a wider range of processes (covering systems with oscillatory or unstable modes).Keywords: PID Controllers, Genetic Algorithms.The first attempts to automate the tuning of Proportional-Integral-Derivative (PID) controllers were based on the time response of a process, but this approach had the drawback of requiring a lot of user interaction. A very important advance was made when it was decided to use the frequency response of a process instead of its time response, in this way a bigger degree of automation was attained. The first industrial PID controller based on processes' frequency response, was the Novatune [1], which initialized a PID controller automatically.Applications based on PID controllers using frequency response were very successful [1] [2] [3], however the range of their applicability is restricted to asymptotically stable systems, due to the use of the relay feedback. By using M.