![](bg1.png)
Series-Parallel Optimization Model for Heat Exchanger Network
Bin Yang
1,2
, Shiqi Liu
1*
and Zhouli Zhao
2
1
School of Mechanical Engineering and Automation, Northeastern University, Shenyang 110819, China
2
Shanghai Baosteel Company, Shanghai 201900,China
Keywords: Series-Parallel Network, Minimum Flow Rate, Mixed-Integer Nonlinear Model, GAMS
Abstract: Based on the parallel heat exchanger network, a series-parallel optimization model of the heat exchanger
network for industrial circulating water system is established. The flow rate and temperature of the heat
exchangers in the network can be calculated automatically. The mathematical formulation exhibit a
mix-integer nonlinear programming (MINLP) structure which can be solved by GAMS. A case study is
done to compare the differences between the series-parallel model and the parallel model. The result shows
that this series-parallel model can reduce fresh water and increase the outlet return temperature obviously.
1 INTRODUCTION
Cooling water system is widely used in factories to
transform waste heat from industrial equipment.
Conventional cooling water systems often use a
parallel heat exchanger network. This will bring a
huge amount of water flow rate and the return
temperature to the cooling tower might be low if the
network is arranged in parallel, which will cause a
poor cooling performance according to the theory of
Kim and Smith (Kim and Smith, 2001). In their
work, the traditional parallel exchanger networks are
changed into series types by applying water pinch
technology so that the bottleneck problems were
solved. However, if the exchangers are changed, the
cooling network might be redesigned, thus the
network will be a lack of flexibility. Therefore,
several authors (Kim et al., 2001; Kim and Smith,
2003; Kuo and Smith, 1998) built mathematical
optimization models to solve the network problem
automatically. Xiao Feng et al. (Xiao et al., 2005)
put forward an intermediate temperature in the water
network design. In this way, the recirculating
cooling water into or out of each cooler would be
from or going to one of the three mains so that the
water flow rate could be reduced and the return
temperature could be increased. Ponce-Ortega et al.
(Ponce-Ortega et al., 2007; Ponce-Ortega et al.,
2010) put forward a mixed-integer nonlinear
programming algorithm for the synthesis of cooling
networks. This work was a development of the
intermediate main which contained several stages
and the capital and utility cost was minimized.
The above papers mainly focused on reducing
the system flow rate and changing the heat
exchanger network structure. This paper proposes a
new series-parallel method to solve the exchanger
network problems. In this method, the water can be
reused so that the total water flow rate will be
reduced.And the outlet temperature can be
increased, which will improve the cooling tower
performance. The mathematical formulation exhibit
a mix-integer nonlinear programming (MINLP)
structure which can be solved by GAMS. A case
study is done to compare flow rate and outlet
temperature between the series-parallel model and
the parallel model.
2 MATHEMATICAL MODEL
In the model formulation, suppose the maximum
amount of heat exchangers is n. HE i and HE j
represents for exchanger i and exchanger j
respectively. In Figure 1, the circles represent the
mixing point and the squares represent for splitting
point. At the mixing point of each exchanger, water
mixes by part of fresh water and part of reusing
water from other exchangers. So the inlet mass flow
rate for heat exchanger i can be shown as Eq. (1).
njiiFjiFiF
n
ijj
,,1,)(),()(
,1
inin
(1)
Yang, B., Liu, S. and Zhao, Z.
Series-Parallel Optimization Model for Heat Exchanger Network.
DOI: 10.5220/0008189703270330
In The Second International Conference on Materials Chemistry and Environmental Protection (MEEP 2018), pages 327-330
ISBN: 978-989-758-360-5
Copyright
c
2019 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
327