Design for Operability: A Singular-Value Optimization Approach within a Multiple-Objective Framework

Abstract

Design-for-operability is a field of active research in process systems engineering because of the high economic impact of design on operability. Dynamic operability has been widely studied by means of the so-called (open-loop) controllability and resiliency indices, which are mostly based on linearized (Laplace/frequency domains) models of the dynamic systems. In particular, the minimum singular value of the transfer function matrix is a fair measure of resilience to disturbances. Among the plethora of approaches to design for dynamic operability, a multiple-objective formulation between cost and controllability naturally arises because of the conflicting characters of the two objectives. In this contribution the cost/minimum-singular-value multiple-objective design problem is solved for the meaningful reactor-separator-recycle system and two different control strategies. The generation of the noninferior solution set is performed here within the framework of an eigenvalue optimization approach.