CONTENTS

Over the past 30 years, much have been written about advanced control; the underlying theory, implementation studies, statements about the benefits that its applications will bring and projections of future trends. During the 1960s, advanced control was taken to mean any algorithm or strategy that deviated from the classical three-term, Proportional-Integral-Derivative (PID), controller. The advent of process computers meant that algorithms that could not be realised using analog technology could now be applied. Feed forward control, multivariable control and optimal control philosophies became practicable alternatives. Indeed, the modern day proliferation of so called advanced control methodologies can only be attributed to the advances made in the electronics industry, especially in the development of low cost digital computational devices (circa 1970). Nowadays, advanced control is synonymous with the implementation of computer based technologies.

It has been recently reported that advanced control can improve product yield; reduce energy consumption; increase capacity; improve product quality and consistency; reduce product giveaway; increase responsiveness; improved process safety and reduce environmental emissions. By implementing advanced control, benefits ranging from 2% to 6% of operating costs have been quoted [Anderson, 1992]. These benefits are clearly enormous and are achieved by reducing process variability, hence allowing plants to be operated to their designed capacity.

What exactly is advanced control? Depending on an individual's background, advanced control may mean different things. It could be the implementation of feedforward or cascade control schemes; of time-delay compensators; of self-tuning or adaptive algorithms or of optimisation strategies. Here, the views of academics and practising engineers can differ significantly.

We prefer to regard advanced control as more than just the use of multi-processor computers or state-of-the-art software environments. Neither does it refer to the singular use of sophisticated control algorithms. It describes a practice which draws upon elements from many disciplines ranging from Control Engineering, Signal Processing, Statistics, Decision Theory, Artificial Intelligence to hardware and software engineering. Central to this philosophy is the requirement for an engineering appreciation of the problem, an understanding of process plant behaviour coupled with the judicious use of, not necessarily state-of-the art, control technologies.

This report restricts attention to control algorithms. Current approaches in this area rely heavily upon a study of system behaviour and the use of process models. Therefore this report will focus only on model based techniques. Although most of the methodologies to be described are applicable to a wide spectrum of systems, e.g. aerospace, robotics, radar tracking and vehicle guidance systems, only those pertinent to the process industries will be discussed.