This paper summarises a 2-year industrial investigation into the applicatio
n of artificial neural networks in the area of process monitoring and contr
ol. The investigation was a collaborative programme between the University
of Newcastle-upon-Tyne, EDS and 24 UK based international companies. Descri
ptions of the major activities undertaken in this programme, which included
the application of neural networks for fault detection in a vitrification
process and the model based predictive control of a gasoline engine are pro
vided. The paper also describes some of the practical difficulties that wer
e experienced while applying neural networks and lists the important lesson
s that were learned through the completion of this project. The main conclu
sion from the work was that neural networks are capable of improving indust
rial process monitoring and control systems. However. the level of improvem
ent must be analysed on a problem specific basis and in many applications t
he use of neural networks may not be justified. (C) 2001 Elsevier Science L
td. All rights reserved.