The objective of plantwide control (the control of a complete chemical plant) can be formulated as:
- support (not guarantee) safety,
- realize the required conversion,
- minimize the operating costs.
It should be noted that the three aspects of the objective are mentioned in the order of economic importance. The support of safety is essential to avoid (economic) loss, the required conversion is a necessary condition to make profit and minimizing the operating costs leads to maximum profit.
There are at least two approaches to realize this objective. The first approach is traditional. This approach focuses on the required conversion and tries to realize this by control. If necessary or possible the support of safety and the minimization of the operating costs is also handled by control. An advantage of this approach is that it leads to control problems that can be solved easily real-time. A disadvantage is that it does not minimize the operating cost to the lowest possible level, so potential profit is lost. Another disadvantage is that the traditional approach only works well for continuous processes.
The second approach uses the objective to formulate a dynamic optimization problem. The optimization approach is attractive since it leads to economic optimal plantwide control. Furthermore it can not only handle continuous processes but also batch processes. However the typical size of the optimization problems involved is considerable (the number of variables and equations is ). And the effect of disturbances can only be taken into account by repeating the optimization real-time. The current research concentrates on the question how to solve large optimization problems real-time (in a plantwide control context). |