Design of a Process System with Fixed Tolerance by Mixed Integer Linear Programming

Abstract

During the operation of a process system, not only manipulated variables but also the operating pattern is adjusted for the uncertainties in the system. Then, the flexibility of the system is represented by a union of the tolerance regions for all operating patterns. In the present work, an iterative procedure to design the process system with fixed tolerance was developed for such systems. The two-stage approach is composed of the optimal design by mixed integer linear programming and the feasibility test of the design for the range of tolerance required by the designer. When the design does not have enough flexibility, the range of system inherent tolerance is enlarged by a design revision strategy. As an illustration, design problems of a steam and power system with fixed tolerance were studied under uncertainties in the demand and plant parameters.