Simple Models for Robust Control of Unknown or Badly-Defined Multivariable Systems

摘要

A class of unknown multivariable process plants is identified such that robust proportional plus integral control systems can be designed without encountering difficulties which are characteristic of self tuning controllers. Attention is focused primarily on the case of discrete sampled data systems with synchronous input and output sampling. Controller design is shown to be possible on the basis of a plant model that is a multivariable generalization of the classical notion of a first order lag deduced from open loop plant step response data. Fast response speeds, zero steady state errors and small loop interaction effects are obtained. Transient effects are also discussed. Numerical examples are given.