LAAR-1998: New faster Solution for the Steady State Back Off Problem

C. RASPANTI; A. BANDONI; L. BIEGLER

LATIN AMERICAN APPLIED RESEARCH

Universidad Nacional del Sur - CONICET

Lugar: Bahia Blabca; Año: 1998 vol. 28 p. 129 - 134

0327-0793

The optimal operation point of a chemical process can be calculated through an optimization problem by minimizing some economic function. This optimum normally lies at the intersection of several process constraints. In order to guarantee feasible operation this point must be moved into the feasible region to avoid constraint violation due to parameter uncertainties. Finding this always feasible operating point is called the steady-state back-off problem. In this paper we present the reformulation and two novel solution strategies for this problem. Two new algorithms are given based on the introduction of a constraint overestimation function and smoothing functions respectively. Three examples are presented where the total CPU is compared with previous algorithms showing a substantial decreasing in the resolution time. The examples also show that the percentage of time decreasing is larger when larger is the number of constraints and uncertain parameters of the original problem.