CHEMICAL ENGINEERING TRANSACTIONS
VOL. 45, 2015
A publication of
The Italian Association
of Chemical Engineering
www.aidic.it/cet
Guest Editors: Petar Sabev Varbanov, Jiří Jaromír Klemeš, Sharifah Rafidah Wan Alwi, Jun Yow Yong, Xia Liu
Copyright © 2015, AIDIC Servizi S.r.l.,
ISBN 978-88-95608-36-5; ISSN 2283-9216 DOI: 10.3303/CET1545309
Please cite this article as: Short M., Isafiade A.J., Fraser D.M., Kravanja Z., 2015, Heat exchanger network synthesis
including detailed exchanger designs using mathematical programming and heuristics, Chemical Engineering Transactions,
45, 1849-1854 DOI:10.3303/CET1545309
1849
Heat Exchanger Network Synthesis Including
Detailed Exchanger Designs Using Mathematical
Programming and Heuristics
Michael Short
a
, Adeniyi J. Isafiade*
a
, Duncan M. Fraser
a
, Zdravko Kravanja
b
a
Department of Chemical Engineering, University of Cape Town, Private Bag, Rondebosch 7701, South Africa
b
Faculty of Chemistry and Chemical Engineering, University of Maribor, Slovenia
AJ.Isafiade@uct.ac.za
The synthesis of heat exchanger networks (HENs) has mainly been done through the use of approximate
models for each of the individual heat exchangers that comprises the network. These approximate models
do not adequately take into account key parameters such as the overall heat transfer co-efficient, TEMA
standards, pressure drops, FT correction factors, and multiple shells. These factors can significantly alter
the cost of the network. This paper presents a new methodology for the synthesis of heat exchanger
networks using detailed heat exchanger design models that takes into account the aforementioned design
parameters. The newly developed method involves the following steps. First, a SYNHEAT (Yee and
Grossmann, 1990) MINLP model is solved. The individual exchangers for the resulting network are then
designed using heuristics, TEMA standards and the Bell-Delaware method. From the designs obtained for
these individual exchangers, correction factors are inserted into the SYNHEAT model that account for
changes in overall heat transfer coefficient, TEMA choices, pressure drops, Ft correction factors and the
effect of multiple shell passes. The SYNEAT model is then re-run and individual exchangers re-designed
and the procedure repeated until convergence is achieved. For each iteration the change in each
correction factor is limited to avoid the omission of certain solutions. While the methodology cannot
guarantee global optimality it can ensure that the synthesised processes are physically achievable and has
also been shown to converge on physically meaningful parameters without the explicit formulation of
complicated non-linear equations in the MINLP formulation.
1. Introduction
Heat Exchanger Network Synthesis (HENS) is one of the most well-known subjects in Process Integration
as it can be used to decrease energy costs and environmental impact of a process through the utilisation
of process heat rather than through the use of utilities. HENS has been attempted using a variety of
methods, with most methods falling under either sequential approaches, like Pinch Technology, or
simultaneous approaches. The simultaneous mathematical programming approach has received the most
attention in recent years due to advances in solver capabilities and the ability to simultaneously consider
multiple factors relating to the overall cost of the network using mixed-integer non-linear programming
(MINLP).
The majority of mathematical programming approaches to HENS use a stage-wise superstructure-based
model (SYNHEAT) first proposed by Yee and Grossmann (1990) that embeds a large number of possible
stream matches into a superstructure that allows for stream splitting and isothermal mixing. This method is
very good for considering potential networks, however the NP hard formulation makes it difficult to solve
to global optimality with current solvers (Furman and Sahinidis, 2002). The formulation fails to consider
details involved in heat exchanger design, such as changing heat transfer coefficients, pumping costs,
number of baffles, tube passes, and number of shells, and cannot be extended to include these as the
combinatorial nature of the problem combined with the increased non-linearity will result in non-optimal
solutions. While the use of constant heat transfer coefficients and the simplifications of ignoring design