The dynamic on-line monitoring system (DMS) based on on-line training of ar
tificial neural network (ANN) has been successfully implemented in a refine
ry for real-time, on-line and dynamic estimation of quality indexes of frac
tionator side draw. To improve the predicted accuracy of DMS and fit the ne
w production cases, a novel ANN training method, on-line training method, h
as been proposed. It can save a lot of time and effort in contrast to the c
onventional off-line training for ANN. It also improves the reliability of
DMS. (C) 2000 Elsevier Science Ltd. All rights reserved.