Optimization of nonlinear PI controllers for chemical processes Conference Paper uri icon

abstract

  • Linear PI controllers are the most popular method for controlling chemical processes due to their simplicity in implementation and understanding. As shown in [1], model-based controllers arising from linear first-order models are exactly linear PI. For nonlinear processes, the application and tuning of linear PI controllers can be problematic, if the nonlinearity is severe enough. The extension of linear PI concepts to develop a nonlinear PI control law can lead to many alternative forms, since the term "nonlinear PI" does not specify the form of the control law; it only specifies that the control law is not linear. Most prior work has suggested ad hoc modifications of the controller parameters as functions of the error [2], or parameter scheduling [3], or both [4]. More recently, [5] have shown mat nonlinear model-based controllers synthesized from nonlinear first-order systems are exactly nonlinear PI. The present work defines a large class of nonlinear PI controllers which can be either model-based or ad hoc and suggests approaches for their optimal tuning. 1998 AACC.

name of conference

  • Proceedings of the 1998 American Control Conference. ACC (IEEE Cat. No.98CH36207)

published proceedings

  • PROCEEDINGS OF THE 1998 AMERICAN CONTROL CONFERENCE, VOLS 1-6

author list (cited authors)

  • Wright, R. A., Kazantzis, N., & Kravaris, C.

citation count

  • 1

complete list of authors

  • Wright, RA||Kazantzis, N||Kravaris, C

publication date

  • January 1998