| Abstrakt: |
In recent years, besides economic questions, environmental problems and subsequent regulatory measures have emerged as a driving force for the chemical industry. It has become essential to identify potential environmental problems as early as possible in the design process in order to avoid costly process changes later. However, in early design stages frequently high uncertainty prevails around possible values that economic, ecological, or technical parameters can assume. In this paper, we propose a new approach on how to select promising process alternatives while taking explicit account of uncertainties. The method is based on approximating flowsheets (as modeled in commercial flowsheet simulators) by response surfaces. Because those response surfaces exhibit a lower complexity than the original flowsheets, much larger design problems can be tackled. This is especially helpful when complex problems, such as multiobjective design problems under uncertainty, have to be solved. While the method we present is generally applicable, it is illustrated with a decision-making case study on selecting a production process for hydrocyanic acid in which more than 400 uncertain variables are treated. The applicability of the method is shown, and optimal process alternatives are identified. |
