Combined Sensitivity–Complementary Sensitivity Min–Max Approach for Load Disturbance–Setpoint Tradeoff Design
ABSTRACT An approach to proportionalintegrativederivative controller tuning based on a simple plant model description, first order
plus time delay, is presented. The approach is based on the formulation of an optimal approximation problem in the frequency
domain for the sensitivity transfer function of the closed loop. The inclusion of the sensitivity function allows for a disturbance
attenuation specification. The solution to the approximation problem provides a set of tuning rules that constitute a parameterized
set that is formulated in the same terms as in [1] and includes, a third parameter that determines the operating mode of the
controller. This factor allows one to determine a tuning either for step response or disturbance attenuation. The approach
can be seen as an implicit 2degreeoffreedom controller because by using one parameter, the operating mode (servo/regulation)
of the control system is determined as well as the appropriate tuning of the controller.
 Citations (15)

Cited In (0)

[Show abstract] [Hide abstract]
ABSTRACT: In this paper, a simple PI controller design method is proposed which can achieve userspecified gain and phase margins if they are suitable, or otherwise the automatically adjusted alternatives for performance enhancement. Unlike reducedorder modelbased tuning methods, exact margins, once specified or adjusted, can be accomplished regardless of the process order and damping nature. The response of the closedloop system using the proposed design is hence more predictable than those using modelbased methods when performance specifications are given in gain and phase margins. The proposed method is based on finding the intersections between two graphs that are plotted using the frequency response of the process. The given gain and phase specifications can be achieved if intersections can be found, and each intersection corresponds to one solution. Suggestions are provided on how to modify the specifications for achieving satisfactory responses for cases where there are no solution or when the specifications can be met but poor responses result. Simulation examples are given to show the usefulness of the method.Automatica 09/1998; DOI:10.1016/S00051098(98)800010 · 3.13 Impact Factor 
Article: A course in H 1 control theory
[Show abstract] [Hide abstract]
ABSTRACT: H ∞ denotes the Hardy space of complexvalued functions F(s) of a complex variable s which are analytic and bounded in the open right halfplane, Re s>0; bounded means that there is a real number b such that F(s)≤b, Re s>0. The least such bound b is the H ∞ norm of F, denoted ∥F∥ ∞ . The subset of H ∞ consisting of realrational functions is denoted by RH ∞ · This book is devoted to the solution of a model matching problem: given three matrices T i in RH ∞ , find a matrix Q in RH ∞ to minimize ∥T 1 T 2 QT 3 ∥ ∞ . If α denotes the mf {T 1 T 2 QT 3 ∥ ∞ : Q∈RH ∞ }, the problems concerned with are: (1) find α ; (2) find Q (when it exists) for which α is attained. Many problems in modern control synthesis and design can be stated as special cases of this quite general problem. The first problem of this type was proposed by Zames (author’s reference 1979) and since then, over the past seven or eight years this subject has become quite popular both for control theorists and operator theorists. The reason for this is that problems of this sort have their classical roots in the interpolation theory of Carathéodory, Nevanlinna, Pick and their modern operator theoretic ones in the work of Nehari, Adamjan, Arov and Krein, and Sarason. The author was uniquely qualified in writing this text. He was actively involved in all phases of the research on this problem. I would say he is probably the only one that mastered all the techniques and directions that this research involved. He also has an extremely clear and concise writing style that allows him to express nontrivial ideas in a very clear elementary fashion. This book is a significant contribution to the growing field of control theory and should be read by everyone interested in this area. The book begins with three introductionary chapters where the basic terminology and mathematical framework of the problem are described. Chapter four is concerned with deriving the Youla parametrization of all stabilizing compensators for a given plant. All results are given with corresponding statespace formulas for the actual computation of the compensators. This is a unique feature that is continued throughout the book. All solutions are given in state space form which allows straightforward computational algorithms. In Chapter 5 the author defines the notion 0+a Hankel operator, relating it to the systemtheoretic notions of controllability and observability Grammians. In Chapter 6 the main results of the book begin to appear. Here a sufficient condition for the existence of an optimal Q is given and a complete solution of the problem is given for the scalar case. An actual computational example is worked out in complete detail. The rest of the book is dedicated to the much more complicated matrix version of the problem. This uses such tools as canonical factorizations of the WienerHopf and InnerOuter type, the introduction of Krein spaces and the BallHelton theory of invariant subspaces Krein spaces. While these notions are quite technically complicated they are presented in a style which is so clear and compelling that it leaves one with the feeling of having undergone a great intellectual experience. The last chapter is dedicated to the analysis of some qualitative properties of the problem. To summarize, I highly recommend this book strongly as the best way to enter into an area of research which has been central in control theory during the last ten years. 
Industrial & Engineering Chemistry Process Design and Development 04/2002; 25(1). DOI:10.1021/i200032a041