Abstract :
The main goal of this monograph is to deal with both problems in an integral way: given a process to be controlled, how should a simple model be identified and used to design a controller to fulfill some requirements? This book is neither a comprehensive handbook of process control design [1] nor a treatise on system identification and parameter estimation [2] but rather provides the tools to reach the final objective of designing a control system for an industrial process. The book has two parts: Part I: Process Identification (Chapters 1??6) and Part II: Control System Design (Chapters 7??12). One of the main difficulties in handling experimental data is the presence of measurement noise. Generally, experimental data are filtered and integrated to get a smoother set of data. Different identification algorithms are used according to the selected model structure, and uidelines to choose the involved parameters are suggested and discussed. A number of examples illustrate the practical use of these algorithms. In the second part of the book, these reduced-order models are used to design a variety of control structures, including single loop control, two-degree-of-freedom (2DOF) controllers, cascade control, multiloop control, decoupling control, and batch control. The main issues considered in the design are disturbances, noise rejection, robustness, and setpoint tracking, with special attention paid to the control of multivariable time-delayed plants. The content of the book is well structured, and each chapter deals with a specific topic. The scope and objective of process identification, the excitation signals commonly used for open-loop and closed-loop identification tests, and the model fitting criteria are summarized in Chapter 1.
