IDEAS approach to process network synthesis: minimum utility cost for complex distillation networks☆
Introduction
Distillation is an inefficient, energy intensive separation process in which up to 95% of the heat entering the reboiler is removed by the condenser (Mix, Dweck, Weinberg, & Armstrong, 1978). This inefficiency, coupled with its widespread use, makes distillation a heavy energy user, accounting for three percent of US energy usage; a number which has remained constant for 20 years (Mix et al., 1978, Department of Energy, 1993; Department of Commerce, 1997).
Several methods have been developed to increase the energy efficiency of distillation processes. Heat integration of distillation columns, was used as early as 1910 by Karl von Linde in cryogenic distillation (Linde, 1921). Petlyuk, Platonov, and Slavinskii (1965), and Doukas and Luyben (1978) showed significant reduction in energy use through the use of complex column configurations. Andrecovich and Westerberg (1985) proposed the column stacking method as a means of designing more efficient distillation networks.
Column sequencing and heat integration may be combined in the same design procedure through the use of optimization, such as mixed integer nonlinear programming (MINLP) (Yeomans & Grossmann, 1999). Such a method requires a “superstructure” containing the preconceived sequences from which the optimal solution will be chosen (Hohmann, Sanders, & Dunford, 1982). As a result, nonintuitive and counterintuitive designs such as those identified by Bagajewicz and Manousiouthakis (1992) and Kovacs, Friedler, and Fan (1993) may not be included. Furthermore, the nonconvex nature of the resulting mathematical programs, provides no guarantee that the identified solutions are globally optimal.
In this paper, we introduce the Infinite DimEnsionAl State-space (IDEAS) process representation as a new paradigm for the design of chemical process networks, and apply this conceptual breakthrough to the synthesis of complex distillation networks. A case study on the distillation of a nitrogen/oxygen mixture is employed to illustrate the proposed method. IDEAS designs are compared with designs developed using conventional methods.
Section snippets
Infinite DimEnsionAl State-space (IDEAS) process representation
The IDEAS process representation is a conceptual advance in the development of process networks, which has two advantages over all other process network design methods. First, it allows consideration of all process networks employing a given set of unit operations. Second, IDEAS naturally results in convex (linear) optimization formulations, thus guaranteeing the global optimality of the obtained solution.
The IDEAS representation of a complex distillation networks consists of two subnetworks (
Case study, nitrogen/oxygen separation
A two part case study is used to demonstrate the power and flexibility of the IDEAS process design representation. Nitrogen/oxygen separation was chosen because of its industrial importance. Nitrogen and oxygen rank second and third, after sulphuric acid, in US production, with 1999 rates of 3.51×1010 and , respectively (Chemical Engineering News, 2000).
We consider that a hot utility is available at 92K at a cost of $105/MJ, and a cold utility is available at 76K at a cost of
Conclusions
The Infinite DimEnsionAl State-space representation has been introduced to determine the minimum utility cost for complex distillation networks. Two examples in the separation of a nitrogen/oxygen mixture are used to illustrate the method. In the first example, with feed and product streams specified as subcooled liquids, the separation efficiency of the IDEAS design with reverse exchangers was 30%; twice that of the conventional McCabe–Thiele minimum utility design. In the second example, with
Physical parameters
Antoine equation constants for oxygen:
Relative volatilities:
Reference temperature:
Latent heat:
Heat capacities:
Notation
Sets of infinite sequences norm of any element finite; absolute sum of elements finite: 2-tuples with nonnegative elements and finite sum . nonnegative elements and finite sum . Other sets N-dimensional vectors with real, rational elements, N-dimensional vectors with real elements. Other notation Greek letter finite number of positive elements
Acknowledgements
The support of the National Science Foundation is gratefully acknowledged under grant numbers CTS 9528553, GER 9554570, and CTS9876489.
References (20)
- et al.
A structural optimization approach in process synthesis—II heat recovery systems
Computers in Chemical Engineering
(1983) - et al.
A systematic modeling framework for superstructure optimization in process synthesis
Computers & Chemical Engineering
(1999) - et al.
A simple synthesis method based on utility bounding for heat-integrated distillation sequences
AIChE Journal
(1985) - et al.
Mass/heat exchange network representation of distillation networks
AIChE Journal
(1992) - Chemical Engineering News, (2000). Production: Gain best losses, vol. 78...
Statistical abstract of the United States
(1997)- Department of Energy, (1993). New energy-efficient cocurrent distillation technology is a major step forward for...
- et al.
Economics of alternative distillation configurations for the separation of ternary mixtures
Industrial and Engineering Chemistry Process Design and Development
(1978) - Drake, J.E. (2000). Globally optimum process network synthesis using the infinite dimensional state-space (IDEAS)...
- et al.
Synthesis of mass exchange networks
AIChE Journal
(1989)
Cited by (0)
- ☆
Part of this work was first presented in Session 237 at the 1998 A.I.Ch.E. Annual Meeting, Paper 237i.