A Closed-Loop Reaction-Curve Method for Controller Tuning
Journal
Chemical Engineering Communications
Journal Volume
104
Journal Issue
1-3
Pages
87-100
Date Issued
1991
Author(s)
Abstract
This article proposes a closed-loop, reaction-curve method to provide a procedure for establishing controller settings on the classical three-mode (P1D) controllers. lt uses the partial extreme data available during the early period of step set-point change under porportional control mode to determine the ultimate period and the ultimate gain of the open-loop process. The resulting ultimate data are critical for subsequent controller settings according to Ziegler-Nichols rules. Simulation examples and test on a pilot-scale stirred-tank reactor demonstrate that the method can provide adequate critical process data, and that it is applicable to processes that are underdamped, characterized bv large dead-times. ooen-looD unstable or with inverse-response. ? 1991, Taylor & Francis Group, LLC. All rights reserved.
Subjects
Controller tuning
Identification
Reaction curve
Type
journal article