Flowsheet Optimizer - Technical Information

The Flowsheet Optimizer unit operation either maximizes or minimizes an objective function by varying one or more flowsheet variables. The objective function may be an operational criterion, such as maximum recovery or minimum loss, or an economic criterion, such as maximum profit or minimum cost. You may specify only one objective function. In order to optimize an economic function, you must first include a calculator in the flowsheet in order to define the profit or cost. Then, the Optimizer can be used to minimize or maximize the calculator result. Optionally, you may define one or more performance specifications. You may also set minimum and/or maximum constraints on flowsheet parameter values other than the Optimizer specifications or variables.

Calculation Method

The Flowsheet Optimizer uses an iterative search technique to find the value of each optimization variable necessary to satisfy the objective function. The objective function should change continuously and monotonically as the optimization variables are changed. If the objective function changes discontinuously, exhibits several maxima or minima, or does not change as the optimization variable changes, then the optimizer is likely to experience convergence problems.

Placement of the Flowsheet Optimizer

If the Flowsheet Optimizer is placed inside a flowsheet recycle loop, PRO/II will calculate the Flowsheet Optimizer as part of the recycle loop. However, the user may change the calculation sequence to force the Flowsheet Optimizer to be calculated after the flowsheet recycle loop converges.

The Flowsheet Optimizer always creates another calculation recycle loop. Therefore it is possible to increase the calculation time of the simulation significantly. Also it is necessary to be sure that inner recycle loops have tighter tolerances than outer recycle loops.

Reasons for Failure

Under the conditions listed below, the Flowsheet Optimizer will fail to find values of all control variables which satisfy all the specification variables:

Any optimization variable reaches a maximum or minimum before all the specifications are satisfied.

The maximum number of cycles is reached before the relative change of any optimization variable is less than the maximum error allowed in the optimizer.

One or more specification variables do not change as the optimization variables are changed.

$image\SML_LINE.gif$