Properties of Generalized Predictive Control

Predictive control using impulse response models

Application of Generalized Predictive Control

The rapid progress in the field of computer technology during recent years has led to an increasing use of process computers to analyse, supervise and control technical processes. Process computers provide the opportunity to sample and store measuring data in a short time, and to process them immediately for better process operation (e.g. with optimum efficiency, best product quality, optimum time behaviour etc.). An important condition for optimum process operation is the knowledge of the process behaviour in the past, at present and in the future. The use of computers allows a permanent analysation or identification of the process dynamics by evaluating the input and output signals of the system. The result of such a process identification is usually a mathematical model, by which the static and dynamic behaviour can be estimated or predicted.

Introduction to System Identification Using Parameter Estimation Methods

The widespread demand for increased product quality and the growing use of high-throughput energy-intensive plant means that alternatives to multi-loop PID controllers must be investigated. One significant approach, compatible with current powerful microcomputers, is multistep or longrange predictive control. This uses a process model and an optimization algorithm to determine the best set of future controls for achieving a desired closed-loop perfomance. The design of these controllers depends on the choices of model, optimization method, and performance index, but they have been applied successfully to a wide range of processes, including multi input/output, constrained, dead-time, and multi-modal plant.

J. Richalet

Papers

Self-adapting IDCOM