18 - Fuzzy Neural Network Systems Techniques and Their Applications to Nonlinear Chemical Process Control Systems

Abstract

Human beings have the capability to acquire and manipulate symbol, pattern-based, heuristic, and fuzzy knowledge in an ingenious manner. This chapter discusses the unique characteristics of chemical process control, the use of neural networks for chemical process control, presents an introduction to fuzzy logic and its use in chemical process control, introduces the concept of fuzzy neural networks, the general structure and learning algorithms, and describes the application of fuzzy neural network model-based nonlinear process control. Chemical process control has unique characteristics that make it different from control in areas such as mechanical, electrical, or aeronautics applications. A fuzzy neural network in which fuzzy logic formalism is integrated with the learning ability of the network is one of the more promising approaches. However, even a qualitative knowledge that determines the structure of fuzzy neural networks is not available for most chemical processes, and therefore fuzzy neural networks have to be used in manner similar to that of the hybrid neural networks scheme for chemical process control.