S.A. Zulkeflee and N. Aziz (Malaysia)
Model identification; Batch esterification reactor; NARX
model; ARX model; State-space model.
In this work, identifications of batch esterification
process to produce Citronellyl laurate have been done
based on both linear and nonlinear approaches. A
state-space, ARX and NARX model are developed to
represent the batch esterification process. General first
principle is developed and used to generate input
output data of identified model. Three sets of data were
used in developing the empirical model; the first set of
data was used for estimation, the second and third data
sets were used for validation. The results show that the
performance of proposed NARX model is superior than
the linear models.