Recovery boiler intelligent control

Abstract

The idea of applying intelligent control using neural networks and fuzzy logic to actual plants, where modern control theory is not powerful enough, is becoming popular. This paper presents one example applied to a recovery boiler in paper pulp mill. Experts still operate recovery boilers because operating a plant must involves many factors, such as visual images. The main concept we discuss in this paper is what we call visual feedback control; if the condition of the visual image changes into a poor condition, the controller can change some actuators to recover the visual image. To perform the control algorithm, neural networks are used to recognize images, and fuzzy logic, to build a controller. We have implemented this algorithm and are now testing it in the field.