Page 151 - Design of Simple and Robust Process Plants

136 Chapter 4 Process Synthesis and Design Optimization
From the RGA the control loops can be selected by pairing the controlled out-
puts with the manipulated inputs in such a way that the relative gains are
positive, and as close as possible to unity (meaning no interaction). When a
manipulated variable is paired with a controlled variable (control loop) the
DOF is transferred to the set point.
5. The results of the interaction analysis at the end of controllability step 1 are
used as input for screening of different flowsheet alternatives. In case an eco-
nomic interesting design alternative would be excluded based on the interac-
tion analysis results, it could be worthwhile to evaluate possibilities to de-
couple the interaction. A system which has considerable interaction might be
de-coupled easily by simple algebraic equations which can be implemented
at the basic control layer, see Chapter 8.
4.2.4.2 Dynamic analysis
Dynamic analysis (controllability step 2) runs for a large part of its activities in parallel
to reaction/distillation/integration step 2 activities. This concept has been chosen, as
the controllability studies are very time-consuming and would cause a delay on the pro-
ject. During step 2 a reduced set of consolidated flowsheets comes available for controll-
ability analysis, while the process integration alternatives are wider under discussion.
1. The split in process section might be revisited, as the integration might con-
nect units which before were considered separated.
2. At that point, RGA analyses must be repeated and used as input for evalua-
tion of flowsheet alternatives.
3. Dynamic simulations must be made for the selected process sections. These
will be based on reduced or short-cut models. Most design models are very
detailed, and are less relevant for the dynamics of the process. Often, the
reactor model can be simplified by the elimination of certain components, or
by converting it to a conversion model. Distillation models can be reduced by
using clustered components and trays. Time constants and response times
are initially estimated based on hold-up estimates and idealized flow pat-
terns. These will be updated as the result of layer 2 optimizations become
available, which include preliminary equipment sizing. In particular, it is the
dynamic simulation that is time consuming and which must be started early
in order to avoid project delay.
4. Determine the type and level of disturbances that the system is expected to
reject / absorb. The prevention of a disturbance requires considerable atten-
tion, as these should have a higher priority then rejecting them.
5. Determine dynamic controllability parameters such as: dynamic RGA, condi-
tion number, Morari resiliency number, disturbance gain matrix to judge the
controllability of the flowsheet (these are parameters described in detail the
literature). Low controllability performance should be approached by process
design and control design solutions to improve controllability. The purpose
of this step is not to design a control configuration, but to judge the controll-
ability and enable the selection of a feasible flowsheet for further design.