ABSTRACT
Department of Electrical and Electronic Engineering, Ashby Building, The Queen’s University of Belfast, BT9 5AH, UK
2.1. INTRODUCTION
Reduced cost of computer hardware, and the availability of a software literate work force have led to significant increases in the use of digital control in many manufacturing and process industries. As confidence is gained in the reliability of embedded computer systems, engineers and managers are beginning to realise the potential for better performance through advanced control. Conventional control technology cannot cope well with changing plant conditions and dynamics. PID controllers must therefore be constantly re-tuned to provide good control at a nominal setpoint. Novel control strategies can deal with plant non-linearity, and ageing of plant components, automatically. This is carried out by a learning process, whereby the controller detects changes in plant operating conditions, and updates the control appropriately. Self-tuning control consists of two tasks; selection of feedback gains, and modelling the plant. For computational simplicity, initial studies have been made using linear modelling techniques. The use of non-linear models is now being considered, together with artificial intelligence techniques for modelling and control.
