(285b) Optimal Design of Reliable Integrated Chemical Production Sites | AIChE

(285b) Optimal Design of Reliable Integrated Chemical Production Sites


Terrazas-Moreno, S. - Presenter, Carnegie Mellon University
Grossmann, I. E. - Presenter, Carnegie Mellon University
Wassick, J. - Presenter, The Dow Chemical Company

An integrated site consists of a network of plants or large-scale chemical processes. The overall objective of the integrated site is to transform materials supplied from outside of the network into a set of products for which there is an external demand. The network structure is determined so that intermediate products are produced by some processes and consumed by others. Uncertain events affect the performance of an integrated site. Some of these are discrete such as process failures while others are continuous such as variations in external supply, demand, and fluctuations in plant throughput. Uncertainties should be taken into account when designing an integrated site in order to achieve long-term operational goals such as availability or probability of feasible operation. To this end the optimal design of integrated sites should consider as decision variables the selection of alternate processes from a superstructure, location of process redundancies, expansion of process capacities over a given time horizon and location and size of intermediate storage.

Past research efforts aimed at synthesizing the optimal design of a given processing network with discrete and continuous uncertainties, include the following. Straub & Grossmann (1,2) propose a method for optimizing the expected stochastic flexibility E(SF) as measure of long-term probability of feasible operation of the process network in the face of continuous and discrete uncertainties. Pistikopoulos et al. (3,4) and Bansal et al. (5,6) solved the same problem through the use of a flexibility reliability index FRI. These works however, do not include the use of intermediate storage as a design alternative for hedging against the uncertainties. The effect of intermediate storage on plant reliability/availability and related measures has only been studied for fixed plant configurations. Some examples of their works are Henley & Hoshino (7), Orban-Mihalyko & Lakatos (8) and Davies & Swartz (9).

This work makes three contributions to the related fields. First, it proposes a novel and simple Mixed Integer Linear Formulation for the optimization of expected stochastic flexibility E(SF) that can exploit existing algorithmic techniques proposed by Straub and Grossmann (1,2) for handling large process networks. The basic idea relies on considering fixed quadrature points and use of auxiliary 0-1 variables to determine feasibility/infeasibility of operation. Second, it integrates the effect of intermediate storage to the framework for evaluating and optimizing expected stochastic flexibility E(SF); a nontrivial task given that no explicit timing considerations are included in existing formulations for optimizing E(SF) or FRI. The basic idea here relies on the use of Markov chains and random walks to capture the discrete-event nature of the problem. Finally, our work also proposes a novel approach for integrating superstructure optimization in the determination of optimal E(SF) in the design of process networks, as opposed to considering a fixed network structure as in past research works.

We illustrate our approach with simple existing examples from the literature and with industrial case studies. As a result of applying our methodology to these examples we provide a trade-off curve of optimal expected stochastic flexibility vs. capital investment.


(1) Straub D.A. and I. E. Grossmann,. (1990). Integrated stochastic metric of flexibility for systems with discrete state and continuous parameter uncertainties. Computers and Chemical Engineering, 14(9), 967 ? 985.

(2) Straub D.A. and I. E. Grossmann,. (1993). Design Optimization of Stochastic Flexibility. Computers and Chemical Engineering, 17(4), 339 ? 354.

(3) Pistikopoulos E. N., T. V. Thomaidis, A. Melin, and M.G. Ierapetritou (1996). Flexibility, reliability and maintenance considerations in batch plant design under uncertainty. Computers and Chemical Engineering, 20, S1209 ? S1214.

(4) Pistikopoulos E. N., C. G. Vassiliadis, J. Arvela and L.G. Papageorgiou, (2001). Interaction of maintenance and production planning for multipurpose process plants ? a system effectiveness approach. Industrial and Engineering Chemistry Research, 40, 3195 ? 3207.

(5) Bansal V., D. P. Perkins and E. N. Pistikopoulos, (2000). Flexibility analysis and design of linear systems by parametric programming. AIChE Journal, 46(2) 335 ? 354.

(6) Bansal V., D. P. Perkins and E. N. Pistikopoulos, (2002). Flexibility analysis and design using a parametric programming framework. AIChE Journal, 48(12) 2851 ? 2868.

(7) Henley E.J. and H. Hoshino, (1977). Effect of Storage Tanks on Plant Availability. Industrial and Engineering Chemistry Fundamentals, 16(4), 439 ? 443.

(8) Orban-Mihalyko E. and B. G. Lakatos, (2004). Intermediate storage in batch/continuous processing systems under stochastic operation conditions. Computers and Chemical Engineering, 28, 2493 ? 2508.

(9) Davies K. M. and C. L. E. Swartz. MILP Formulations for Optimal Steady-State Buffer Level and Flexible Maintenance Scheduling. MASs Thesis, McMaster University.