THE APPLICATION OF NEURAL PREDICTION AND FUZZY REASONING IN THE OPTIMIZATION CONTROL OF CATALYTIC REFORM PROCESS

Some process plants are complex and nonlinear,whose control depends on man's intelligent experience.And this control can hardly be performed by traditional control strategies.In this paper,we use the idea of integration of fuzzy reasoning and neural learning to realize a controller for this kind of plants,and propose a hidden node value regulation (HNR) algorithm for neural network learning to assure this integration.An optimization control system based on this schema has been in operation for controlling the second catalytic reform process of maoming.petroleum and chemical industry corporation in Guangdong province.Operation data have shown its feasibility.