Page 338 - Design of Simple and Robust Process Plants

324 Chapter 8 Instrumentation, Automation of Operation and Control
around exothermal reactions, might result in a unstable open loop design this can
be made stable by adapting the heat removal systems. Sometimes these situations
are difficult to prevent, in which case the hazard is often minimized by changing
the operational conditions to limit the reactants inventory and its concentrations. A
typical example is the redesign of a batch reactor into a continuous-fed batch reactor
design.

8.4.4
Divide the Process into Separate Sections

The intention is to reduce the control design problem. The subdivision should be
done in such a way that relevant recycle streams are included, such as a recycle from
the finishing back to the reactor. It would not be relevant if an incremental recycle
simply replaced a small part of a fresh feed stream. In case the recycle would contain
some impurities which might impact on the reaction or the finishing train and
would build-up, this needs to be considered. A similar situation might occur with
heat integration. When this is provided with a back-up (as in Figure 4.29 in Chap-
ter 4), it is not considered relevant for the separation sections (although it will still
appear as disturbance). The disturbance must be absorbed, as was shown in the
example of the heat exchanger with heat balance control.

8.4.5
Determine the Degrees of Freedom

The degree of freedom analysis is the first step before the selection of the variables.
The following equation has been derived:

N DOF =N Variables ±N Equations
where N DOF is the number of degrees of freedom, N Variables is the number of pro-
cess variables, and N Equations is the number of independent equations that describe
the process.
The number of manipulated variables of is less than the number of DOFs, as
some variables are externally defined:

N DOF =N manipulated +N External Defined
where N manipulated is the number of manipulated variables, and N External Defined is
the number of external defined variables.
Now, the number of independent manipulated variables is:

N manipulated =N Variables ±N External Defined ±N Equations
The number of independent manipulated variables is preferably equal to the
number of controlled variables that are controlled. When a manipulated variable is
paired with a controlled variable (control loop), the DOF is transferred to the set
point