An object-oriented framework for adaptive fuzzy circuit control

A class library for process control is under development as a cooperative effort between the US Bureau of Mines and the University of Alabama. In this paper, some essential ingredients of such a library are described. In particular, an adaptive fuzzy controller class and a minerals beneficiation circuit class are presented. The fuzzy controller is a software object of type fuzzyController communicating with another software object of type controllableObject for the purpose of controlling the latter. A circuit is an interconnected set of software objects of types flow and process. The fuzzy controller model and the circuit model are unified by means of more specific derived classes which inherit properties as needed from the classes controllableObject and process. An example is presented in which a model of an operating minerals beneficiation circuit was constructed, tuned with a genetic algorithm, and brought under control by fuzzy controllers in two different ways: (a) by independently controlling the processes in the circuit, and (b) by directly controlling the entire circuit. The study indicated that controlling the entire circuit was somewhat more effective than independently controlling the processes in the circuit