engleski

Introduction to neural networks and fuzzy reasoning for process control

Modeling for control of biprocesses by fuzzy reasoning and neural networks is presented. The principle of modeling from systems point of view is reviewed. Analytical, fuzzy logic and neural network models are discussed based on perception of a process complexity, a priori knowledge about a process, and availability of information. A general structure of modeling by fuzzy logic models applied for process control is introduced. Considered are feedforward neural networks based on nonlinear autoregression with moving averages for general modeling of dynamical systems. Analyzed are the following neural network control structures: inverse control, internal model control, and feedback with parallel connection of PID and neural network model. Comparison of the fuzzy logic, neural networks and classical PID control is discussed. Presented are results of computer simulation and from industrial practice of neural network models.

process control, bioreactors, neural networks, fuzzy logic

nije evidentirano