Application 2. Bi-Objective Optimization of a MTBE Reactive Distillation Column

5 Application 2. Bi-Objective Optimization of a MTBE Reactive Distillation Column 

A multiobjective optimization problem is formulated for the MTBE RD column with respect to economic performance and exergy efficiency. The formulation includes the balance equations 8.1-8.5 and 8.29, the criteria definitions 8.42 and the optimization formulation 8.62. Constraints imposed by operating conditions and product specifications are included. For the sake of controlled built-up of optimization complexity the first approximation of this approach omits the response time constant as objective function. 

The resulting optimization problem is formulated as follows and is implemented and solved in gPROMS /gOPT , 

The planar view of the two normalized objective functions at different weighting factors is depicted in figure 8.9. The normalization of the objectives requires a priori knowledge of the minimum (/ ) and maximum values for each objective. These values 

This curve is termed Pareto curve cf. explanatory note 8.1) and clearly shows that one objective function can only be improved at the expense of the other objective function. This trade-off leads to significant deviations between optimal designs and the utopia point (up to 60%) for the pair /4on /exergy The decision maker should have in mind the importance of compromising both objectives at early stages of the design cycle. 

---