Abstract

Modern industrial process automation systems are organized as hierarchical systems, with at least three levels: held, direct control, and supervisory level, This paper presents applications of radial basis function (RBF) networks in each one of these levels. The first example illustrates the application of these networks to improve the performance of a control valve. The second example demonstrates the potential of an RBF network to deal with nonlinear processes as direct controller. The third example, illustrates the use of a supervisory system to tune proportional integral (PI) controllers, implemented in a distributed environment. These examples do not only present some applications of RBF technologies to process control, but also demonstrate, through practical examples, that this technology can be implemented in existing control systems.