Operations of refuse incineration plants involve many uncertain factors. They include human factors, the complexity of refuse burning phenomenon and the physical properties of refuse. On the other side, fuzzy control has been widely put to practical use as a rulebased controller, which applies the experience of skilled operators and the knowledge of experts. This paper describes a fuzzy control system for the refuse combustion process, consisting of two elements; i.e., a fuzzy sensing system and a fuzzy controller.
The fuzzy sensing system, located between the process outputs and the fuzzy controller, infers the situation of the plant replacing human observation. This concept is considered useful when it is difficult to collect manipulation rules directly from the process variables.
In the fuzzy controller, we propose an ordinal structure model of control rules, with the aim of reflecting the operator's sense of priority among the control rules. In this method, the inference results of each control rule are compensated according to the difference of priority and the relative magnitude of the premise. The aim is to simplify the procedure for partitioning the input space of the controller and also to reduce the number of control rules.
Some test runs on an actual 200t/d incineration plant are also presented. From these results, it is concluded that the system proposed is effective and practical for plant automation.