Start
Dec 14, 2007 - 11:00
End
Dec 14, 2007 - 12:00
Venue
CDEEP Seminar Hall KRESIT (3rd floor)
Event Type
Speaker
Prof. Lorenz T. Biegler
Title
Treatment of Unknown Information for Steady State Optimization
Design under unknown information needs to consider two forms variability of process operation over the life cycle of the process as well as uncertainty in the process model. These two sources affect the process differently and require a different treatment at the design stage. The optimal solution of these design problems also requires the application of large-scale process optimization strategies. This talk describes a two stage optimization strategy that includes both sources of uncertainty. The strategy consists of a discretization of an uncertainty description that leads to a multiperiod nonlinear program. Following this a feasibility test is introduced which checks for additional critical points. Uncertainty and Variability enter in different ways in this strategy. Finally simple case studies are presented that illustrate this approach.