Application of fuzzy logic based self-tuning control
John A. Dunlop, Keith J. Burnham, D.J.G. James, P.J. King · 2002
A novel form of fuzzy self-tuning control is proposed which is based around the concept of two coupled sub-algorithms for plant estimation and control. Whilst model based approaches have been applied to a variety of industrial systems, the analytical techniques employed in such approaches may be inappropriate for plant that exhibits nonlinear and rapidly varying dynamic response characteristics and where one is compelled to make use of high sampling speeds. This problem is significantly compounded when dealing with systems that exhibit varying time delays. With this in mind, the proposed controller utilises the more durable strategies of fuzzy estimation and control in order to provide a more robust controller strategy. In particular, this area of research is aimed towards accommodating the desirable features derived from methods based on Smith prediction, where delays may only otherwise be accounted for by reducing the controller gain. Results obtained when use is made of the proposed fuzzy controller configured to accommodate the time delay are compared to those corresponding to a PID controller with no a priori knowledge of the delay. Issues regarding the implementation and commissioning of the proposed scheme are also discussed.>