Developments in industrial applications of self-tuning control

Abstract

Control strategies based on auto-tuning and self-tuning principles are now emerging to provide a realistic option for a wide range of industrial processes and plant. This is particularly the case when considering batch and fed-batch process control systems in which assumptions on slowly varying process dynamics permit the use of such adaptive schemes. However, there are many industrial systems where commercial ‘off the shelf’ controllers do not provide an adequate solution. Consequently there is much on-going research activity in the refining and tailoring of algorithms in order to provide effective implementation on specific industrial applications. With consideration given to two industrial case-study applications, the paper describes extensions and refinements that have been introduced in order to accommodate for improved estimation and control law design algorithms when dealing with systems exhibiting fast dynamic behaviour and non-linear response characteristics respectively. The requirement for improved estimation schemes arises when one is compelled to make use of low-order linear model structures whereas when dealing with non-linear systems, which exhibit slowly varying dynamic characteristics, one can afford to invest in improvements in control law design.