This paper presents a strategy to represent and classify process data for d
etection of abnormal operating conditions, focusing on a novel approach in
de-noising of process data using wavelet coefficients. A case study feature
s two continuous stirred tank reactors (CSTRs) in series that uses an inter
mediate mixer for the second feed. (C) 2000 Published by Elsevier Science L
td. All rights reserved