الفهرس 
System Optimization and identification.Only 14 pages are availabe for public view
 |  | from | 153 |  |  |
| | | from | 153 | | |
Abstract
This thesis is mainly concerned with the
design o~ an algqrithm ~or static optimization o~
dynamic industrial The idea behind theprqcesses.
suggested algorithm is to represent the target syst~m by
a nonlinear model relating the per~ormance index to the
manipulating inputCs). Identi~ication of the model is
carried out by using the Extended Kalman Fi1ter CEKF).
The results o~ the identi~ication are then used to
calculate
the manipulatingopt.Imal values o~
the
inputCs).
The thesis includes a general introduction to
the subject. and a bibliographic research that aims to
analyze the existing static optimization methods and
nonlinear ~i1tering methods. The CEKF) is chosen as a
nonlinear identi~i cation algori t.hm to be used in the
development o~ the optimization algorithm. This research
work indicated that the developed algorithm is a new
static optimization industrialmethod o~
dynamic
processes.
The development o~ the algorithm is then
presented. The robustness o~ the suggested algorithm is
then investigated via simulation study. The simulation
results indicated that the algorithm is in general conver ging to the new oper ating condi tions but. under
certain conditions, may.be biased. This bias is due to
the identification technique.
An algorithm for optimal transfer of the
system 1’rom initial operating
conditions to
the
identi fi e.d.-opti mal
operating
conditions is
also
developed.
The developed static optimization algorithm is
applied to optimize the operation of a distillation
column existing
at ”Laboratoire d’Automatique de
Grenoble” (LAG) in France. The results indicated that
the algorithm is capable of predicting the optimal
values of inputs.