Self-tuning control of a titanium dioxide kiln

Abstract

This paper presents an application of the self-tuning regulator to the control of a TiO 2 kiln. The regulator is used in a feedback loop around a kiln already controlled by a discrete regulator designed on minimum-variance principles. The self-tuning regulator has only two parameters. A study of the asymptotic properties shows that the Self-tuning regulator preserves optimality if the minimum-variance regulator is already optimal. The equilibrium points of the regulator are established for delays of one and two steps. Experimental results showed an increase in "in-spec" time over previous regulators. Grade-change results are also given, leading to marked improvement over previous practice.