Page 336 - Design of Simple and Robust Process Plants

322 Chapter 8 Instrumentation, Automation of Operation and Control
. self-optimizing control of units applied for; stand-alone units and at inter-
mediate time spans between implementation of closed loop steady state
operation optimization set points; and
. bases for model predictive control and closed loop process optimization.
To achieve robustness it is vital that any basic model applied for control is preferably
solved by an analytical technique. Iterations and optimization are to be avoided, as
these might fail in finding a solution unless sufficient provisions are made.
Control design requires a controllability analysis to select the most robust control
solution. The result of the analysis may include the need for process modifications
to enable control, reflecting interaction between process design and control design.
The procedure is based on the latest development in control design and includes the
work from Luyben et al. (on plant wide control), TyrØus (on selection of dominant
variables on partial (basic) control with a thermodynamic approach), self-optimizing
control of Skogestad (1999), and the controllability analysis procedure of Seider et
al. (1999).
A procedure for controllability analysis of a process plant has the following
sequential steps as presented by Seider et al. (1999):
1. Define control objectives.
2. Evaluate the open loop stability of the process based on static models.
3. Divide the process into separate process sections; be aware that a process sec-
tion must include a recycle stream from another section, as applied for reac-
tors with unconverted reactants.
4. Determine the degrees of freedom (DOFs).
5. Determine preliminarily the controlled, manipulated, measured and distur-
bance variables.
6. Determine feasible pairing options in respect of plant wide control and unit
control.
7. Evaluate static interaction of the selected pairing options.
8. Evaluate dynamic interaction of the reduced set of selected pairings. In case
of evaluation of the controllability during process synthesis this information
is input for the final flowsheet selection.
9. Establish the final pairing and design the controllers.
10. Tune and test the performance of the controller in a dynamic simulation.

An overview of the development activities for control design at the basic control
layer is presented in overall format in Figure 8.15.

8.4.2
Definition of Control Objectives

Definition of control objectives for a total process operation is; to maximize the prof-
its by converting raw materials into products which have to meet quality specifica-
tions while respecting safety and environmental requirements and operational con-
straints.