Artificial neural networks and their application in process engineering

A typical neural network topology consists of highly interconnected `neuron´ like nodes. These nodes act as nonlinear processing elements, hence the attractive feature of the technique is that given an appropriate network topology, nonlinear functional relationships can be characterised. Thus, the methodology provides a generic, cost effective, nonlinear modelling philosophy which may be a valuable tool in alleviating current process engineering problems. This contribution introduces the concepts involved in the formulation of artificial neural networks. Their suitability for solving process engineering problems is discussed and illustrated using results from simulation studies and applications to industrial data