Neural network control of cold flow circulating fluidized bed

Circulating fluidized bed (CFB) technology is an efficient method of forcing chemical reactions to occur and has been widely accepted in a wide variety of fields, including catalytic cracking, power generation, mineral processing and many other processes. The recycle nature of CFB technology allows for a better process, but also making the tasks of modeling and controller design many times more difficult. The plant under consideration is a cold-flow circulating fluidized bed (CF-CFB), meaning there is no combustion component in it. Previous attempts have successful in making a good model for the CF-CFB. In this paper we describe a Neural Network (NN) controller for the CF-CFB. It has been shown that a NN can be used effectively for the identification and control of nonlinear dynamical processes. Results are presented.