Identification of critical points for the design and synthesis of flexible processes

• Published in: Computers & chemical engineering Vol. 32; no. 7; pp. 1603 - 1624
• Main Authors: Pintarič, Z. Novak, Kravanja, Z.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 24.07.2008
• Subjects: Critical point; Design; Flexibility; Nonvertex; Process; Synthesis; Two-level; Uncertain; Vertex; Design; Uncertain; Vertex; Synthesis; Process; Critical point; Nonvertex; Two-level; Flexibility
• ISSN: 0098-1354; 1873-4375
• DOI: 10.1016/j.compchemeng.2007.08.003

Optimization problems for the design and synthesis of flexible chemical processes are often associated with highly discretized models. The ultimate goal of this work is to significantly reduce the set of uncertain parameter points used in these problems. To accomplish the task, an approach was developed for identifying the minimum set of critical points needed for flexible design. Critical points in this work represent those values of uncertain parameters that determine optimal overdesign of process, so that feasible operation is assured within the specified domain of uncertain parameters. The proposed approach identifies critical values of uncertain parameters a-priori by the separate maximization of each design variable, together with simultaneous optimization of the economic objective function. During this procedure, uncertain parameters are transformed into continuous variables. Three alternative methods are proposed within this approach: the formulation based on Karush–Kuhn–Tucker (KKT) optimality conditions, the iterative two-level method, and the approximate one-level method. The identified critical points are then used for the discretization of infinite uncertain problems, in order to obtain the design with the optimum objective function and flexibility index at unity. All three methods can identify vertex or even nonvertex critical points, whose total number is less than or equal to the number of design variables, which represents a significant reduction in the problem's dimensionality. Some examples are presented illustrating the applicability and efficiency of the proposed approach, as well as the role of the critical points in the optimization of design problems under uncertainty.