Page 411 - Design of Simple and Robust Process Plants

398 Chapter 9 Operation Optimization
9.6.13
Implement CLO: Step 13

The updated rigorous model with adjusted parameters is used for the optimization
with the selected optimization routines. The final result of the optimization must
pass the criteria set in decider-4. Initially, deciders-1, -2 and -3 need to be true, and
the optimization run needs to be converged. Some processes lift the process stability
criteria, as set in decider-1 for data reconciliation and parameter estimation, for the
implementation of the optimization results. The DOFs used for the optimization
need still be available for set-point implementation; if not, then the optimization
run must be restarted with the available DOFs. Another criterion is that the differ-
ence between simulated performance and optimized performance should be outside
the standard deviation of the simulated results in order to avoid unnecessary process
disturbances. The results of the optimization provide new set-point available for
downloading to BC and MBC. These set-points are implemented through the set-
point implementation module. The implementation of a set-point can be done gra-
dually, by applying set-point rampers, or as a step change. The implementation of
all set-points will follow a defined scheme

9.6.14
Evaluate Project and Build-up Maintenance Structure: Step 14

The evaluation of an OO project starts from the moment the project begins. During
the feasibility stage, cost and the savings of the project are estimated. Cost can be
tracked during the project; external costs can easily be determined, but internal
costs requires book-keeping of internal hours. The savings are more difficult to
determine, particular as most suppliers of this technology do not have a real interest
in measuring process performance. This was one of the main reasons why the per-
formance meter was introduced. Another reason was to develop a tool for validation
of the optimization model. The savings are determined by measuring the process
performance before the implementation of a project. The data of the mass balance
reconciliation and the operating margin are collected ongoing, and can be used for
the evaluation for different price sets at a later date, during and after the project.
The performance meter can also be used for the evaluation of other project
aspects, such as hardware, and also the contribution of control projects such as the
control part of the OO project. Savings can also be determined to some extent after
the outage of an OO. The results are somewhat debatable, as operation also learns
from the optimizer. However, with a performance meter in line a good indication of
the savings can be achieved:

Maintenance The maintenance of an OO project is not trivial, and the benefits can
only be continuously achieved by ongoing maintenance. This has several different
aspects, including instrumentation, control, model updating and ongoing operation
attention.