DEVELOPMENT OF ADAPTIVE NEURAL NETWORKS FOR FLEXIBLE CONTROL OF BATCHPROCESSES

This paper deals with the application of a neural controller for tempe rature control of a batch reactor. The term ''neural controller'' is u sed to refer to a multilayer neural network which computes the control values to be applied to the process. We present the design and the de velopment of the neural network: architecture, learning database and l earning procedure. In a first step, the learning phase consists in tea ching the neural network to map the dynamics of a classical adaptive c ontroller (generalized predictive control with double model reference) implemented on the process. Although the neural controller performanc e is good for operating conditions included in the learning set (inter polation), it exhibits limitations on extrapolation. In this work, two methods for the on-line adaptation of the network's weights are devel oped: one of them is the ''specialized'' learning technique, whereas t he other uses another neural network in order to model the reactor dyn amics. Several results are shown and prove the good capacities of neur al networks for controlling batch processes.