Double command model-free hybrid control of a nonlinear CSTR

Moteza Mohammadzaheri, Hamidreza Atrinejad, Mehdi Kopaei, Fariba Behnia-Willison

    Research output: Contribution to conferencePaperpeer-review

    Abstract

    in this paper, a new methodology for feedforward-feedback control system design is proposed. Initially, the concept of control equilibrium point is introduced. Using this concept, the steady state control command is determined so as to maintain the desired situation of the system. Non-model-based feedforward control law is conducted on this basis using an artificial neural network. The feedback controller is a gain pushing the system towards the reference. In this article, the case study is the concentration control of a non-thermic Catalytic Stirred Tank Reactor (CSTR). Using the proposed control system, the value of feedback controller gain can be arbitrarily high with a guaranteed BIBO stability. The mathematical model of the system is used neither in design nor in stability analysis, and stability of the control system is addressed using some evident practical assumptions which can be extended to many other systems. In this case study, the level height of the reactor is not particularly subject to control but the control system is so designed that this variable never goes lower than a specified limit. The proposed method returns surprisingly good results in comparison with the results with a well-designed fuzzy control system.

    Original languageEnglish
    Pages227-232
    Number of pages6
    DOIs
    Publication statusPublished - 1 Dec 2010
    Event2nd International Conference on Computational Intelligence, Modelling and Simulation -
    Duration: 28 Sep 2010 → …

    Conference

    Conference2nd International Conference on Computational Intelligence, Modelling and Simulation
    Period28/09/10 → …

    Keywords

    • Artificial neural networks
    • CSTR
    • Feedforward control
    • Nonlinear control
    • Process control

    Fingerprint Dive into the research topics of 'Double command model-free hybrid control of a nonlinear CSTR'. Together they form a unique fingerprint.

    Cite this