Page 362 - Design of Simple and Robust Process Plants

Design of Simple and Robust Process Plants. J. L. Koolen 349
Copyright 2002 Wiley-VCH Verlag GmbH & Co. KGaA
ISBNs: 3-527-29784-7 (Hardback); 3-527-60047-7 (Electronic)

Chapter 9
Operation Optimization

9.1
Introduction

The overall objective of a simple and robust process plant says that, ªAn optimal de-
signed safe and reliable plant, operated hands-off at the most economical condi-
tionsº. This statement immediately points to operation optimization (OO), and as it
was formulated in the design philosophies as ªoperation optimization makes
moneyº. In this chapter, an historical overview is provided of developments in the
field of OO, together with an introduction about closed loop optimization (CLO)in
continuous processes.
The description of an on-line performance meter of the process plant is seen as a
key instrument to ongoing measure plant performance. The instrument can also be
used for validation of OO models and tracking the closed loop optimization (CLO)
performance. The accuracy and the design of such a meter is discussed.
The functioning and the design of CLO for continuous process plants is discussed
extensively with the implementation procedure.
There are two specific points to be mentioned to limit the complexity of OO. First,
the concept of self-optimizing control, which is a recent development to operate units
at the optimum. The approach is based on the evaluation and selection of control
variables with constant set-points to be implemented in a simple feedback loop to
operate satisfactorily (with an acceptable loss), subject to the expected disturbances.
Several examples have been reported in the literature to support the concept. This
approach makes the modeling less complex, and also strives for a simpler control
structure by avoiding MBC, such as for dual composition control in rectifying opera-
tions.
A pragmatic approach for OO. is explained. The ultimate CLO is not always the
optimal solution for OO. A qualitative approach is presented for a gradual imple-
mentation of OO up to an economic level. A step-wise approach is worked out,
using the following sequence of development/implementation: reactor modeling;
performance meter implementation; off-line optimization model; process stability;
and quality control, including self-optimizing control, eventual followed by con-
straint control and CLO.