Interaction and Structure Concepts for Large-Scale Systems
In this chapter, the authors present the various concepts used and techniques proposed to deal with the structure selection and interaction analysis problems. They describe some state-space-based measures for assessing interactions. The authors present some of the more popular methods of characterizing the interactions among the manipulated and controlled variables in multivariable systems. Grosdidier et al. develop a set of theorems (assuming that an integral part is present) that characterize important closed-loop properties such as stability, sensor and actuator failure tolerance, robustness with respect to modeling errors, and the feasibility of decentralized control structures. Y. Arkun presents a formulation of dynamic block relative gain for systems using model reference control schemes and its use in analyzing the stability and performance of decentralized control structures. The identification of interaction and noninteraction to enable the decomposition of the control design has been a primary objective of most of the techniques reviewed.