Renewable and Sustainable Energy Reviews 53 (2016) 500–514

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

Layer pattern thermal design and optimization for multistream

plate-fin heat exchangers—A review
Zhe Wang a, Yanzhong Li a,b,n
a
School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an 710049, China
b
State Key Laboratory of Multiphase Flow in Power Engineering, Xi'an 710049, China

art ic l e i nf o a b s t r a c t

Article history: Cryogenic processes involve air separation and liquefaction. To reduce energy consumption in these
Received 9 June 2015 processes, many compact and efficient equipments that can handle fluid heat transfer have been
Received in revised form developed. Of these, a multistream plate-fin heat exchanger is one of the best solutions. Studies of plate-
12 August 2015
fin heat exchangers are currently focusing on four areas: heat transfer calculation, surface analysis, flow
Accepted 2 September 2015
resistance and design optimization. It is also important to optimize the layer patterns design of multi-
stream plate-fin heat exchangers. Several techniques have been proposed for this purpose thus far;
Keywords: however, most of these are based on qualitative or trial-and-error approaches. Therefore, no universally
Layer pattern accepted methodology exists designing for the layer pattern of multistream plate-fin heat exchangers.
Multistream plate-fin heat exchangers
This article starts by reviewing traditional design approaches for multistream plate-fin heat exchangers
Design criterion and evaluation
and then focuses on the development of layer pattern design methods. It highlights three types of
Optimization design
Intelligent algorithm thermal design and evaluation criteria. It then discusses some suggestions and new methods for the
optimization of the layer pattern design that have emerged in recent years. Further, newly emerging
intelligent heuristic algorithms for optimizing the layer pattern thermal design are discussed. In addition
to these basic design methodologies, the “layer pattern ring model” and “dual objective function”
optimization methods developed by the author’s research team are discussed. Finally, the status of this
research area is summarized, and emerging trends are noted.
& 2015 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
2. Brief overview of MPFHE design methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
2.1. Heat transfer design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 501
2.2. Fin surface analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502
2.3. Flow resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
3. Optimization of layer pattern design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
3.1. Early efforts toward layer pattern design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
3.2. New method for optimizing layer pattern design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
3.3. Layer pattern design criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 506
3.3.1. Local heat load balance criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
3.3.2. Equal temperature difference criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
3.3.3. Equal number of heat transfer units criterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
3.4. Layer pattern design evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
3.4.1. Cumulative zigzag curve deviation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
3.4.2. Uniformity optimization factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
3.4.3. Some commercial software model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507

n
Corresponding author at: School of Energy and Power Engineering, Xi’an Jiaotong University, Xi’an 710049, China. Tel.:þ 86 29 82668725; Fax: þ 86 29 82668789.
E-mail address: yzli-epe@mail.xjtu.edu.cn (Y. Li).

http://dx.doi.org/10.1016/j.rser.2015.09.003
1364-0321/& 2015 Elsevier Ltd. All rights reserved.
 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514 501

3.5. Intelligent algorithm for optimization of layer pattern design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 508

4. Conclusion and future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513

1. Introduction different layer pattern simulations. However, the number of

simulations required to obtain optimized results increases with
Cryogenic processes usually involve air liquefaction and the number of layers. The traditional gradient-based optimization
separation. A large amount of energy is required to achieve cooling method is also not effective because direct contact cannot be
below 120 K in these processes [1,2]. Therefore, it is necessary to established between different layer patterns and random layers
improve the thermal efficiency of cryogenic processes to avoid [4]. This produces the nonuniform distribution phenomenon,
unnecessary energy consumption. Toward this end, improved making it impossible to avoid local optimal solution trap in the
technology and high-performance equipment are required. A optimization process. Thus far, the layer pattern problem has
multi-stream plate fin heat exchanger (MPFHE) is widely used in remained the key difficulty to overcome in the design and opti-
cryogenic processes [3,4], including those in the aerospace, pet- mization of an MPFHE.
rochemical and industrial gas production industries [5,6], owing to To highlight the importance of layer pattern design, the general
its compact structure, high efficiency, low cost, and ease of design flow of the MPFHE is briefly introduced. The present review
handling multiple streams [7–9]. In air separation units and nat- mainly focuses on studies conducted on the means and methods
ural gas liquefaction plants, they represent 20–30% of the invest- for layer pattern design and optimization. The novel genetic-
ment costs [10,11]. An MPFHE serves as an important hub con- algorithm (GA)-based modeling and optimization approaches
nection between cryogenic processes and distillation compart- developed by the author's research group are also highlighted.
ments. Therefore, its performance directly affects the proper
operation and performance of the entire cryogenic process. As a
result, designers are strongly focusing on exploiting the advan-
2. Brief overview of MPFHE design methodologies
tages of MPFHE. Conventionally, a two-stream heat exchanger
design can be divided into two categories: rating design and sizing
An MPFHE can simultaneously handle many different streams
design. In the rating simulation, the exit conditions of the fluid
for heat exchange owing to its structure. As shown in Fig. 1, the
streams are estimated for a given heat exchanger geometry and
core of the MPFHE includes fins, headers and sheets. The interlayer
known inlet conditions. On the other hand, in the sizing calcula-
between two adjacent parting sheets consists of heat transfer fins,
tion, the geometrical details of the heat exchanger are determined
distribution fins and a side bar. The manufacturer suitably stacks
for specified inlet and outlet stream temperatures. For multi-
an arrangement of such layers based on different flow modes for
stream heat exchangers, mainly rating simulations have been tried
each heat transfer stream and conducts vacuum brazing of the
thus far. The sizing design of multistream heat exchangers can be
necessary headers, nozzles and structural support.
performed only under a limited scope. Obviously, for a multi-
Cryogenic heat exchangers have traditionally been designed
stream plate-fin heat exchanger, if the inlet and outlet tempera-
and rated using three models: lumped parameter, distributed
tures of different fluid streams are specified, fixing the core
parameter and stream evolution. Lumped parameter models
dimensions and fin geometry becomes very difficult, and its suc-
represent the basic design theory for heat exchangers. They are
cess often depends on the experience and intuition of the
based on the following energy balances for two single-phase
designer. Compared with a two-stream heat exchanger, the com-
streams and include the logarithmic mean temperature differ-
bination of multiple streams in an MPFHE is similar to that in a
ence (LMTD) method, effectiveness-number of transfer units
heat exchanger network; furthermore, its structure is more com-
(ε-NTU) method, and so on [13]. Distributed parameter models are
pact and process is more complex. Many problems faced in a two-
based on dividing the heat exchangers into elements of variable
stream heat exchanger have not yet been explored, and many new
size and applying a lumped parameter model to each of them;
features in the design optimization of an MPFHE are presented. For
they are widely used for applications with cryogenic heat
example, the quality of the layer pattern directly influences the
exchangers, such as complex geometries and two-phase stream
overall MPFHE performance. When the layer arrangement deviates
from the ideal layout, local heat load causes a large imbalance. The systems [14]. Stream evolution models are based on steady-state
resulting temperature crossover and internal heat loss reduce the one-dimensional mass, momentum and energy balance equations
efficiency of the MPFHE. In severe cases, we cannot use sufficient for each individual stream. These models are usually implemented
backup heat transfer areas to compensate for the loss of efficiency in proprietary software [15], and their key features are related to
[3]; in addition, local heat stress affects the structural strength and the correlations used for fluid properties and heat transfer and
span life. pressure drop characteristics.
The layer pattern problem has proved difficult to solve because
the many types of hot and cold fluids involved in the heat 2.1. Heat transfer design
exchange lead to combinatorial explosion. It is impractical for a
designer to analyze all permutations exhaustively. Moreover, the The heat transfer design of a plate-fin heat exchanger mainly
nonuniformity of the temperature field and the pressure field includes thermal property matching, determining the heat trans-
between the layers is coupled to each other, making it difficult to fer temperature difference, heat load checking, reasonable recon-
derive a theoretical general solution formula to design the layer struction of temperature field in heat exchanger and model engi-
pattern. In fact, designing the layer pattern of an MPFHE is an neering considerations. Conventionally, the heat exchanger design
integer nonlinear optimization problem, and the conventional can be of two types [13]: rating design and sizing design using
continuous variable optimization method cannot be employed. To LMTD and ε-NTU. Fig. 2 shows the common heat transfer design
optimize noncontinuous variables, it is necessary to perform many flow of heat exchangers. However, as the multistream thermal
502 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514

Nomenclature P pressure (kPa)

Q heat duty (kW)
Cpi specific heat of stream i, kJ g  1 K  1 Re Reynolds number
Ei excess heat load of each hot/cold layer bunch in layer T temperature (K)
pattern ring
cj cold stream i of MPFHE, j ¼ 1; 2; :::; nc Greek symbols
hj hot stream i of MPFHE, i ¼ 1; 2; :::; nh
m_i mass flow rate of stream i, kg=h or kg=s ε effectiveness
nc number of different cold streams in MPFHE η layer arrangement thermal efficiency of the heat
nh number of different hot streams in MPFHE exchanger
Ni number of layers of stream i σ The zigzag pattern weighted deviation of the cumu-
qi heat load per layer of stream i, kW lative heat load
Q_ i;j heat transfer rate between fluid streams i and j, kW ΔT temperature difference (K)
f Fanning friction factor ΔP pressure drop (kPa)
f(x) objective function
j Colburn factor Subscripts
Sn cumulative heat load for each layer, kW
Smid midpoint of Sn
n
mid c cold fluid
Sn arithmetic-mean of Smid
n h hot fluid
wi weighting
max maximum
ðUAÞi;j heat conductance between fluid streams i and j, WK  1
min minimum
NTU number of transfer units
i, j variable number
LMTDideal ideal logarithmic mean temperature difference

properties are not easy to match, the average temperature differ- ferent working conditions and heat transfer conditions, different
ence is difficult to calculate. Early thermal designs therefore con- fin geometries can be used. Fig. 3 shows some of the common fin
sidered the use of a simplified synthesis method that assumed that geometries used in plate fin heat exchangers (Plain, Serrated,
all cold and hot fluids were equivalent to two streams: one Perforated, Wavy, etc.) [20]. At present, a fin type structure is
endothermic stream and one exothermic stream [16–18]. Cur- usually chosen based on the fluid property and the designers'
rently, most manufacturers mainly use commercial design soft- experience [21], as shown in Fig. 4. Studies have proposed various
ware (ASPEN HYSIS, HTRI, Honeywell UniSim, etc.) for preproces- approaches for selecting fins, such as half-fin idealization [22], Z–Y
sing the heat transfer calculation. The heat transfer mechanism in graph method [23], and volume performance index (VPI) [24,25].
this software is actually an extension of the endothermic and However, these methods are only semi-qualitative and cannot
exothermic synthesis methods, i.e., the pinch method of the heat provide accurate results.
exchanger network [19]. Kao [26] first studied the performance of a fin surface. Subse-
quently Kays and London [27] conducted extensive research to
2.2. Fin surface analysis obtain a wide and reliable dataset. In recent years, many scholars
have conducted simulations and experimental verifications of fin
A fin is an important internal component for enhancing heat heat transfer, friction and efficiency. The results of these studies
transfer in an MPFHE. It is responsible for expanding the heat mainly pertained to fin structure changes and precise interpola-
transfer area, improving the heat exchange efficiency, and tion and fitting related to the surface performance. Currently, the
increasing the pressure bearing capacity. Depending on the dif- fin performance for the Colburn j-factor and Fanning friction factor

1. Block 9. Parting sheet

2. Header 10. Heat transfer fin
3. Nozzle 11. Distribution fin
4. Width 12. Side bar
5. Stacking height 13. End bar
6. Length
7. Passage outlet
8. Cover sheet
Fig. 1. Schematic drawing of MPFHE [12].
 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514 503

are mostly derived from designers' databases (Chart Industries, the manufacturing level of an enterprise [28]. As shown in Fig. 5,
FIVES, KOBELCO, Sumitomo Industries, and LINDE Industrial the heat transfer coefficient of the entire heat exchanger is
Gases). Therefore, the j-factor and friction factor of the fin reflect obtained based on the different fin structures and surface
performance.

2.3. Flow resistance

Flow resistance in the MPFHE refers to the difference between

the static pressure of the inlet pipe and that of the outlet pipe.
Usually, the friction resistance of the inlet and outlet pipes is
mainly considered at the header and nozzle structure, which is
less than the resistance of the core part; therefore, it can be
neglected in the design and calculation. The core resistance con-
sists of the shape resistance and friction resistance of the heat
exchange surface; however, it is not separated. It can be con-
sidered as the equivalent shear force acting on the total friction
area [29,30]. Therefore, a plate-fin heat exchanger core can be
considered an equicircular tube with equivalent diameter, and it
can be expressed as
L G2
ΔP ¼ 4 f ð1Þ
De 2ρ
f denotes the Fanning friction factor; De, the equivalent dia-
meter of the fin; L, the heat exchanger length and G, the fluid mass
flow rate. However, the core resistance of the MPFHE can be
changed along with the number of layers; therefore, the number of
layers should be specified before calculating the flow resistance
[31].
Fig. 6 shows the general design step of the MPFHE. The above
analysis shows that the layer pattern and design of each stage are
all closely related. At present, various analyses and calculation
models (one-dimensional model [32], two-dimensional model
[33] and heat transfer network model [34]) that can well predict
the heat transfer performance and flow resistance under a given
layer pattern are used. However, how to design and optimize the
Fig. 2. Common heat transfer design flow of heat exchangers. layer pattern according to the given heat transfer design

Plain Serrated

Perforated Wavy
Fig. 3. Some common fin geometries used in plate-fin heat exchangers [20].
504 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514

conditions remains unclear [35]. Therefore, the review and ana- exchangers were optimized by the passage-segregated method,
lysis below focus on the problem of the heat transfer mechanism and transverse heat conduction was found to be a common char-
of the layer pattern; furthermore, previous design methods are acteristic in the MPFHE's heat transfer mechanism; Weimer called
summarized and potential problems for future research are this the bypass effect of the fin. In the optimization result, the
identified. smaller the fin bypass effect, the lesser is the effect of the layer

3. Optimization of layer pattern design

The optimization of the layer pattern design of the MPFHE is

essentially a structural optimization problem. It is not restricted to
only the heat and mass transfer but to the entire heat exchanger
[36]; however, it must first be determined prior to the thermal-
hydraulic calculation. This contradiction and restriction introduces
some difficulty in the design. In addition, the corresponding pos-
sible layer patterns could be numerous and may not be possible to
check one-by-one. Some software (Aspen Platefin, HTRI Xpfe and
UniSim) can effectively evaluate the quality of a user-specified
layer pattern specified; however, they are unsuitable for the
optimization of the layer pattern design. At present, the design is
mainly based on experience and the trial-and-error method, and
the design principle is simply to avoid temperature crossover and
internal heat loss. Table 1 shows some common layer pattern units
and explanations.

3.1. Early efforts toward layer pattern design

Fan [37] proposed the concept of passage arrangement in the

design of a plate fin heat exchanger; this is the earliest report on
layer patterns. This approach has greatly influenced the design and
optimization of an MPFHE. Fan's passage-segregated method
present is innovative in that (1) the cold and hot layers are
arranged alternately and (2) the passages of the same stream are
arranged relatively together. In this method, the basic principle is
that by strengthening the degree of isolation in the layer pattern,
the number of hot layers that can reach the lowest outlet tem-
perature is increased. Compared with the passage-mixed method,
the passage-segregated method is beneficial in that it reduces the
internal heat loss and avoids temperature crossover. This design
principle can only be used when the total number of hot and cold
layers is basically equal; however, this is not always the case in
practical applications. If the number of hot or cold layers is
increased forcibly to satisfy the requirement of the passage-
segregated method, the pressure drop in the heat exchanger will
change; this, in turn, will affect the heat transfer coefficient.
Weimer and Hartzog [38] analyzed the impact of uneven dis-
tribution of streams and improperly arranged layer patterns on the
heat exchanger performance. Two streams with 12-layer heat Fig. 5. Design flow of heat transfer calculation for fin surface.

Fig. 4. Selection criteria for height and number of fins.

 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514 505

Fig. 6. General design step of MPFHE.

Table 1
Some common layer pattern units and explanations.
1000

Excess cold load

Cumulative heat load (kW)

Excess heat load

-1000

-2000

-3000
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Layer
Fig. 7. Zigzag curve method of a MPFHE.

heat load balance units, which means that heat exchange only
occurs inside the basic unit instead of among different units. From
the viewpoint of thermodynamics, the local heat load balance
method is reasonable because the distribution of heat load
between the layers can reflect, to a certain extent, the temperature
field variation in the same cross section of the MPFHE. To evaluate
a given layer pattern, they provide a zigzag graphical method, i.e.,
excess hot/cold accumulated in some layers, as shown in Fig. 7.
Ideally, the zigzag curve of the layer pattern uniformly oscillates at
the zero line. By checking the deviation of the zigzag curve from
the ideal, a layer pattern can be evaluated roughly. However,
owing to the numerous combinations of layer patterns, the
pattern on the heat exchanger performance. Sucessmann and
detection of all zigzag curves is unrealistic.
Mansour [39] found that an inappropriate layer pattern strength-
Chen [40] and Haseler [41] analyzed and evaluated the design
ens the heat load imbalance between layers, in turn leading to a
of the above layer pattern method and found that the influence of
decline in the efficiency of heat transfer. Accordingly, they pro- neighboring layers in local heat load imbalance cannot be ignored
posed the balance of local heat load method, assuming that (1) the in a single-layer arrangement. The heat transfer calculation is
cross section of the heat exchanger has the same wall temperature simplified by the local heat load balance principle for an air-
and (2) the cross section can be divided into a number of basic separation MPFHE, and the iteration convergence is accelerated.
506 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514

Prasad [42–45] determined the influence of the layer pattern on exchanger network theory and obtained a one-dimensional gen-
the MPFHE performance, surface temperature field, fin efficiency, eral solution by dynamic analysis. Cui et al. [58] considered the
and flow distribution from different aspects. The heat exchanger commutative coordination caused by the arrangement effect
performance was quantitatively evaluated using finite difference among the temperature field, velocity field, and temperature dif-
calculations and the constant wall temperature method. The ideal ference field in the MPFHE. An optimization method for the flow-
half-fin hypothesis was improved, and a secondary heat transfer field organization harmonizing temperature difference field was
surface model for MPFHE was constructed. A simulation and proposed at the altitude of field synergism. Lv et al. [59] proposed
experiment of a specific air-separation MPFHE was conducted, and the concept of the dimensionless temperature difference uni-
transverse heat conduction was found in some specific layers. The formity optimization factor. The result showed that the factor was
spacing layers between hot and cold streams should be between an effective evaluation criterion for the layer arrangement of the
1 and 2. Prasad et al. [46,47] presented a comprehensive MPFHE MPFHE. The factors are closer to 1 and the temperature difference
design study; however, they did not focus on layer pattern opti- uniformity is better, which indicated that the MPFHE has good
mization. Many studies [48,49] of heat exchangers considered heat transfer performance.
layer pattern optimization design problems as a future research Recently, the layer pattern of the MPFHE has been studied from
area. Yuan et al. [50] extended the balance of local heat load different aspects. Zhao et al. [60] studied the flow channel layout
method with an additional step of rearranging the layers according and optimized the design based on a comprehensive consideration
to the results of the differential computation of the temperature of factors such as the volume, weight and resistance. In order to
distribution. This method provides an optimal layer order with evaluate the performance comprehensively, a dimensionless ana-
limited adjustments; however, it is essentially still based on the lytical method was applied to define a synthesis factor and
initial layer pattern being given a priori and then gradually being synthesis line for the MPFHE. A comparison of the analysis results
improved based on the temperature field by trial-and-error using with design ones indicates that the discussed method is suitable
the semi-empirical method. for comprehensively optimizing the design of and improving the
In the 1990s, owing to the relative stability of energy prices and performance of multistream heat exchangers. Peng et al. [61]
the common use of heat transfer enhancement technology, few investigated the optimal design with the differential model for
studies focused on improving the efficiency of an MPFHE. Most MPFHE. An effective method for designing the layer pattern was
manufacturers used inherent principles and their own experience presented, and many conditions with different initial lengths and
to guide the design of the layer pattern. They aimed at the fol- slice thicknesses were calculated. From the calculation results, the
lowing: (1) achieve local heat load balance to the best possible initial length had almost no effect on the final divergent length;
extent and reduce excess heat load and conduction distance, however, the slice thickness greatly influenced the final result.
(2) avoid the cross temperature and reduce the internal heat loss, Guan et al. [62] introduced four connection matrices, described
(3) the calculation length of each layer should basically be similar, the connection of multilayer and stream splitting and bypass, and
(4) the pressure drop of the same stream for each layer is basically obtained the general analytical solution of the dynamic response
the same, and the pressure drop of the total layer is lower than the for MPFHE and their networks. The comparison between the cal-
control value and (5) the layer should be arranged as symme- culation and the experiment demonstrates that the general solu-
trically as possible for ease of manufacturing and assembly and to tion can be used in most cases. Jia et al. [63] presented the
reduce heat stress. structure and performance continuity principle of MPFHE and
observed that small variations in the layer pattern have an insig-
3.2. New method for optimizing layer pattern design nificant effect on the heat exchanger performance. However, this
layer arrangement optimization method based on simulated
Since 2000, with the development of energy strategies and annealing is entirely dependent on the quality of the continuity
technological innovations, studies are increasingly focusing on sequence. Shah et al. [64] analyzed the layer arrangement of the
heat exchangers; specifically, they have investigated design opti- MPFHE from the perspective of entropy production. The smaller
mization theories, experimental studies and manufacturing pro- the entropy production obtained from the layer pattern, the better
cesses. Many researchers have aimed at enhancing the heat is the layer arrangement. Guo [65] and Chen et al. [66] defined the
transfer and optimizing the efficiency of MPFHE. entransy-dissipation-based thermal resistance of the MPFHE and
Guo [51] proposed a field synergy theory and introduced the divided it into two parts: one from heat transfer between the
temperature difference uniformity factor for two flow heat layers and the other from stream mixing at the layer outlet. They
exchangers, which reflected the degree of the uneven temperature concluded that different layer arrangements correspond to differ-
difference field of the stream in the heat exchanger. The more ent thermal resistances, and the smaller the thermal resistance,
uniform the temperature difference field in the heat exchanger, the larger is the heat transfer rate.
the higher is the efficiency, and simultaneously, the smaller is the
entropy production, i.e., temperature difference field uniformity 3.3. Layer pattern design criterion
optimization principle. Reneaume et al. [52,53] used Mixed Integer
Non Linear Programming (MINLP) to investigate the manufactur- From the above literature review and research trends, it can be
ing cost, physical volume, and operating constraints as objective inferred that the goal of layer pattern design is to achieve a con-
functions and the geometrical parameters of the exchanger as stant wall temperature for the MPFHE. The heat transfer behavior
optimization variables. Sundén [54] and Picon [55] both created of all cold/hot streams and the temperature difference field of the
similar new models assuming that the MPFHE is analogous to a entire heat exchanger is uniform. Then, the MPFHE is equivalent to
two-stream heat exchange network. They introduced the pinch an ideal two-stream heat exchanger. To achieve these objectives,
analysis method to match the stream parameters and calculated Mansour and Sucessmann [39] proposed a local heat load balance
the minimum temperature driving force of the MPFHE. The results criterion with the heat load of the layer as the control object.
of the optimization examples showed that the new method is However, researchers found that achieving the ideal heat transfer
more practical and accurate compared with the traditional ε-NTU operating condition, heat transfer temperature difference [67], and
model. Luo et al. [56,57] extended the multi-stream heat thermal conduction [35] of each layer could also be used as the
 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514 507

control objects. Then, the layer pattern arrangement design cri- the maximum. The model on the above basis can be represented
teria of the MPFHE are summarized as follows: as
NTU i ðUAÞi;j
3.3.1. Local heat load balance criterion ¼ =NTU h;ideal ð7Þ
_ CpÞi
NTU h;ideal 0:5ðm
For the local heat load balancing arrangement mode, the heat
exchanger is divided into a number of transverse heat load bal- NTU j ðUAÞi;j
ancing units that should be as small as possible and that should  ¼ =NTU c;ideal ð8Þ
NTU c;ideal _ CpÞj
0:5ðm
follow the layer pattern criterion. Then, the layer heat load accu-
mulated is put into a zigzag diagram; the deviation of the zigzag Similarly, NTU calculated for the hot load layers is positive and
curve from the zero line level is studied and the layer pattern that for the cold layers, negative. The sum of these two equations
quality is evaluated. The local heat load balance criterion model in the ideal state should be equal to zero. This criterion provides a
can be expressed as reasonable layer pattern to ensure that the calculated value of the
_ hi Cphi ðt hi  T hi Þ sum of NTUs for all cold and hot layers is close to the zero line.
m
qhi ¼ ; i ¼ 1; 2; :::; nh ð2Þ
N hi
3.4. Layer pattern design evaluation
_ cj Cpcj ðt cj T cj Þ
m
qc j ¼ ; j ¼ 1; 2; :::; nc ð3Þ Just as designers can only use the finite optimization method to
N cj
achieve the optimal value to the best extent possible, thus far no
where qhi is the heat energy transferred from hot stream hi specific methods have obtained the optimum layer pattern.
whose sign is “-,” whereas qcj is the heat energy gained by cold Therefore, it is impossible to make a direct comparison between
stream cj whose sign is “ þ.” The heat load per layer of each stream the proposed approach and other established approaches, because
is calculated using Eq. (4). of which indirect evaluations are employed. Three representative
Si ¼ Si þ qhi ð4Þ indirect evaluations for the layer pattern are described below.
or cj

where Si is the zigzag curve of the longitudinal coordinate for 3.4.1. Cumulative zigzag curve deviation
each layer i or j ¼ 1; 2; :::; n. The term n represents the total The first evaluation method is the cumulative heat load zigzag
number of hot and cold layers. curve and its deviation σ [35], which accumulates the heat load of
each layer from a zigzag curve. Ideally, the zigzag curve oscillates
3.3.2. Equal temperature difference criterion regularly about zero and its deviation is 0; therefore, by checking
The ideal heat transfer state of the MPFHE is a constant tem- the deviation between the actual and the ideal, a layer arrange-
perature difference of heat transfer between the hot and the cold ment can be evaluated roughly. The related evaluation formula is
fluids for the entire heat exchanger, indicating that all the hot and as shown in Eq. (9).
cold fluids have the same flow and heat transfer performance. For vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
uP
this condition, the MPFHE can be considered a two-stream heat u n The actual zigzag  The ideal zigzag2
u
exchanger. In this case, the MPFHE can be divided into several t1 The ideal zigzag
σ¼ ð9Þ
segments for ease of analysis. Hot and cold fluid temperatures, n1
enthalpy, and slope of the enthalpy curve in each section remain
unchanged. From the outlet and inlet temperatures, which follow 3.4.2. Uniformity optimization factor
the ideal logarithmic temperature difference between the hot and The second evaluation method is the dimensionless tempera-
the cold fluids, the heat transfer for this section can be deter- ture difference uniformity optimization factor, which is based on
mined. Therefore, this model can be shown as the field synergy principle [68]. It divides the layer of a multi-
T i T j 0:5 U Q_ i stream heat exchanger into k  k  n sub-heat transfer units. This
¼ =LMTDideal ð5Þ factor can be expressed as
LMTDideal ðUAÞi;j
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
nP1 P k
Ti Tj 0:5 U Q_ j ðn  1Þ  k ½Θði þ 1; jÞ  Θði; jÞ2
 ¼  =LMTDideal ð6Þ i¼1j¼1
LMTDideal ðUAÞi;j ϕmulti ¼
nP
1 k 
P Θði þ 1; jÞ  Θði; jÞj
where Q_ i and Q_ j are the design heat loads of the two adjacent i¼1j¼1
layers. The temperature difference for the hot load calculation
Tði; kÞ T min ði; iþ 1Þ
layers is positive and for the cold layers, negative. A coefficient of Θði; jÞ ¼ ð10Þ
ΔT max ði; i þ 1Þ
0.5 is used to represent half of the layer heat load. The axis of the
layer is considered an adiabatic surface, which is an idealized where ΔT max ði; i þ 1Þ is the maximum temperature difference
assumption. To compare the heat load of each heat exchanger between the i and iþ1 layers, and T min ði; i þ1Þ is the minimum
layer, the actual temperature difference is compared with the ideal temperature of the i and iþ 1 layers. The term k denotes the sub-
logarithmic mean temperature difference. Ideally, the temperature heat transfer unit of each layer, and n is the total layer quantity. Eq.
difference between these two should be zero. This equal tem- (10) indicates that the temperature difference in the heat
perature difference criterion is a reasonable layer pattern to ensure exchanger becomes uniform as ϕmulti becomes closer to 1.
that the calculated maximum temperature difference for all layers
remains close to the zero line. 3.4.3. Some commercial software model
The evaluation method involves exchanging the positions of a
3.3.3. Equal number of heat transfer units criterion pair of hot or cold streams to verify the robustness of the opti-
The equal number of heat transfer units (NTU) criterion is used mization results and then using simulation software (Aspen Pla-
for the analysis of the MPFHEs based on the heat transfer perfor- tefin [69], HTRI Xpfe [70] and UniSim [71]) to verify the MPFHE
mance. When the NTUs of the hot and cold fluids are equal and performance. The software for layer calculations uses a distributed
also respectively equal to the fluid heat transfer unit ideal num- parameter model (DPM) based on the assumption of common wall
bers NTU h;ideal and NTU c;ideal , the efficiency of the heat exchanger is temperature. Based on the given inlet conditions of each stream
508 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514

Fig. 8. Area splitting method and flowchart of proposed methodology [81].

and exchanger geometry (layer pattern, fin structure and heat apply intelligent algorithms to optimize the layer pattern of
load), the outlet conditions for each stream and the efficiency of MPFHE. Ghosh et al. [80,81] described an area splitting method
the MPFHE can be calculated by the software. and a successive partitioning technique that can obtain the tem-
perature field of the MPFHE easily. They first applied a GA to solve
3.5. Intelligent algorithm for optimization of layer pattern design the MPFHE layer pattern problem and established that this model
is suitable for optimizing the arrangement for 3–8 different stream
For optimizing the design of the heat exchanger structure, layers. Fig. 8 shows the optimization method and process. How-
intelligent algorithms (GA, particle swarm optimization and har- ever, with an increase in the number of layers, the optimization of
mony search) have been widely used in recent years [72,73] the layer pattern becomes highly complicated. In this situation, the
because they do not need any assumptions about the objective model has no guidance to the optimization process and no
functions and are independent of the mathematical model to be restrictions to adverse results, and therefore, it cannot guarantee
optimized and the initial values. As early as 2004, Mishra et al. the final optimization results.
[74,75] used a GA to perform heat exchanger optimization with Hu et al. [82] established a linear programming model that took
minimum cost and minimum entropy production. Rao [76] and the mean-square deviation of the layer's accumulated total heat
Peng [77] both improved the particle swarm optimization algo- load as the objective function and proposed a GA for the flexible
rithm to minimize the volume of the heat exchanger. Joda [78] and optimization design of the layer pattern. The uncertain factors in
Tahouniet et al. [79], based on the design method of a heat an MPFHE, such as the actual operation parameters and external
exchange network, studied the trade-off design between the condition, were considered. The optimal passage arrangements of
pressure drop and the length and considered the variable physical some main working points were obtained using a GA and were
properties of the MPFHE using a GA. However, researchers seldom then integrated into one passage arrangement. The final layer
 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514 509

pattern is obtained after a fine-tuning operator is applied to the design of a 24-stream main heat exchanger in an air separation
passage arrangement. Fig. 9 shows the flowchart of layer pattern unit, and Fig. 10 shows the design process.
design optimization using a GA. Tian et al. [84] minimized the cumulative heat load (MCHL) to
Peng et al. [83] extended the integral-mean temperature dif- make the layer arrangement more convenient and to improve the
ference model (IMTD) from a three-stream [67] to a MPFHE and heat transfer performance, and they built a distributed-parameter
then obtained the heat exchange efficiency, layer pattern and fin model to obtain the temperature field of the MPFHE for further
improvement; Fig. 11 shows the design process.
structure based on the IMTD. The design of the layer pattern under
Skaugen et al [85,86] described and demonstrated a novel
multiple operating conditions is optimized using a hybrid PSO. The
multistream heat exchanger modeling framework to design an
effectiveness of these models is verified by the optimization
optimized structure. This framework addresses the need of models
that capture important physical mechanisms and geometrical
features. Fig. 12 shows the layer-by-layer model diagram. The
complete heat exchanger model can be generated by connecting
each individual layer to their own surface or by lumping all layers
in the layer together and connecting them to one common surface.
This new layer model framework is called a lumped layer-by-layer
model. This model could be extended through further research
and may find further applicability.
The above review suggests that studies are gradually moving
from the trial-and-error and enumeration methods to the use of
effective intelligent algorithms to search for the optimal design
scheme. However, owing to the lack of accurate mathematical
models of the temperature field of the MPFHE, researchers first
need to obtain the temperature field by modeling or using other
valid models to study layer pattern optimization. The accuracy of
the temperature field calculation of the MPFHE can directly affect
the arrangement of the layers. Moreover, the convergence speed,
accuracy and robustness of the layer pattern optimization process
are all demanding requirements for the optimization model. In
this light, the present research group [87] developed a general ring
model for layer pattern optimization based on a layer pattern
design criterion and the optimization process of nested MUSE
software. This new method can not only optimize the general
model using intelligent algorithms and a layer pattern design
criterion but also make use of sophisticated heat temperature field
calculation software for performance evaluation. As shown in
Fig. 13, the head–tail binary string x ¼ ½x1 ; x2 ; :::; xl  is adopted to
represent the layer pattern. xi ¼ 1, i ¼ 1; 2; :::; l, indicates that the
stream is selected, where xi ¼ 0 indicates that the stream is
eliminated. If the total number of hot layers is greater than the
total number of cold layers, the substring formed indicates that the
Fig. 9. Layer pattern design optimization flowchart using GA [82]. hot layers surround the cold layers, and then, the substring repeats

Fig. 10. Flowchart of optimization design for layer pattern based on IMTD [83].
510 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514

Fig. 11. Flowchart of layer pattern optimization based on the MCHL and DPM [84].

for the same number of times as the total number of cold layers to segment relative to the root mean square deviation from the zero
obtain the entire layer pattern model expressed. position is considered the optimization object.
After the establishment of the layer pattern optimization ring
model, the extra layers of aggravating local heat load and excess 1
wi ¼ Li ð11Þ
heat load were eliminated based on dual fitness function alter- P
n
Li
nating optimization. i¼1
The objective of the main fitness function is to minimize the
cumulative heat load. The main fitness function considers the zigzag mid 1 X n

segment length weighting wi for each layer, and the midpoint of the S ¼ Smid ð12Þ
n1 i ¼ 1 i
 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514 511

Fig. 12. A heat exchanger model and a lumped layer-by-layer element [85].

Fig. 13. A general layer pattern ring model [87].

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Some constraints are added to accelerate the convergence.

u
u 1 X n
min:f main ðxÞ ¼ t ðSmid  Smid Þ2 U wi ð13Þ First, the representation of the layer number for the encoding
x n1 i ¼ 1 i x ¼ ½x1 ; x2 ; :::; xl  must meet the requirement of the layer number
for each stream in the MPFHE. Second, the spacing layers between
where Li and Smidi are each line segment length and the mid- hot and cold streams should have a value between 1 and 2. The
point of the longitudinal coordinate in the layer pattern zigzag premise of this constraint is that the ratio of the total number of
curve, respectively. hot and cold layers is not less than 1:2 or not more than 2:1. In
In fact, excess heat load is always present. The objective of the addition, the layer numbers on both ends of MPFHE are required to
upgrade fitness function is to minimize the excess heat load of be 1. Fig. 14 shows the layer pattern optimization flow. The relia-
each hot/cold layer bunch with half of its two adjacent cold or hot bility of this method was provided by an experimental study of
layers. layer pattern optimization.
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u  2
u
umin P Nc ; P Nh ÞðEi
nc nh

u j j
u j ¼ 1
X j ¼ 1
4. Conclusion and future work
min :f ðxÞ ¼ t
upgrade Uw ð14Þ
i
x
i¼1
In this paper, the layer pattern thermal design and optimization
where Ei is the excess heat load of each hot/cold layer together of the MPFHE are reviewed from the viewpoints of general design
with half of its two adjacent cold or hot layers, i ¼ 1; 2; :::; methodology, design theory of layer pattern and intelligent algo-
P
nc P
nh
rithm optimization. Reported techniques based on heuristics,
minð N cj ; N hj Þ, and wi is the weighting.
j¼1 j¼1 extension of the analysis applicable for layer arrangement design
512 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514

Fig. 14. Layer pattern optimization process based on ring model and dual fitness function [35].

experience, heat transfer design criterion, and evaluation criteria will be increasingly adopted in the future to optimize the layer
are also briefly reviewed, and the advantages and limitations of pattern. Furthermore, greater emphasis will likely be given to
these methods are critically judged. In addition, a summarization multiobjective optimization including other optimization para-
of the analysis and methods for using intelligent algorithms to meters (pressure drop, fin structure, exergy efficiency, and
optimize the layer pattern design of the MPFHE is conducted. economy).
Furthermore, the ring model method and dual fitness function The present review reveals that the achievements made thus
alternating optimization proposed by the present research group far in formalizing the optimization of the layer pattern design are
are elaborated. modest but focused. At the same time, one can note a number of
In recent times, many efforts have been made to optimize the aspects that have either been ignored by researchers or have not
thermal design of a layer pattern. Intelligent optimization methods received due attention. These are listed below.
such as GA or particle swarm optimization have been often
adopted for this purpose. From the reported investigations, a dis- 1. Thermal design model and layer pattern optimization of an
tinct trend has emerged. Classical optimization algorithms are not MPFHE based on multifield synergy theory (including tem-
appropriate in this case owing to the large number and variety of perature field, velocity field, and temperature different field)
design variables involved. It is likely that evolutionary algorithms should be revived.
 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514 513

2. Multiobjective optimization, including the trade-off of multiple [23] Frank LMP, Zhu XX. Integrated heat exchanger network and equipment design
objectives, coupling of multiple parameters, and restriction of using compact heat exchangers. Heat Transf Eng 2002;23(6):18–35.
[24] Picon-Nunez M, Polley GT, Torres-Reyes E, et al. Surface selection and design
multiple constraints should be discussed for MPFHE optimization. of plate–fin heat exchangers. Appl Therm Eng 1999;19(9):917–31.
3. The overall design optimization framework of MPFHE with the [25] Picón-Núñez M, Robles JLL. Flow passage arrangement and surface selection in
comprehensive consideration of the layer pattern and fin multistream plate-fin heat exchangers. Heat Transf Eng 2005;26(9):5–14.
[26] Kao S. A systematic design approach for a multistream exchanger with
structure should be emphasized. interconnected wall. ASME Paper 1961;61:255.
4. The design criterion of the layer pattern in phase change heat [27] Kays WM, London AL. Compact Heat Exchangers. (3rd Ed.). New York:
transfer and multiple operating conditions requires further study. McGraw-Hill; 1984.
[28] Ismail LS, Velraj R, Ranganayakulu C. Studies on pumping power in terms of
5. The evaluation index of the layer pattern and the experimental
pressure drop and heat transfer characteristics of compact plate-fin heat
verification of the optimization model should be investigated exchangers-A review. Renew Sustain Energy Rev 2010;14(1):478–85.
further. [29] Kuppan T. Heat Exchanger Design Handbook. New York: Marcel Dekker; 2000.
[30] Shah RK, Sekulic DP. Fundamentals of Heat Exchanger Design. John Wiley &
Sons; 2003.
[31] Taylor MA. Plate-fin Heat Exchangers-guide to their Specification and Use,
Acknowledgments HTFS. England: Oxfordshire; 1987.
[32] Bentwich M. Multistream countercurrent heat exchangers. J Heat Transf
1973;95(4):458–63.
This research was supported by the Funds of National Key [33] Goyal M, Chakravarty A, Atrey MD. Two dimensional model for multistream
Technology R&D Program of China (Grant no. 2012BAA08B03), plate fin heat exchangers. Cryogenics 2014;61:70–8.
[34] Pattison RC, Baldea M. Multistream heat exchangers: equation-oriented
National Natural Science Foundation of China (Grant no. modeling and flowsheet optimization. AIChE J 2015;61(6):1856–66.
51376142), and the Specialized Research Fund for the Doctoral [35] Wang Z, Li Y, Zhao M. Experimental investigation on the thermal performance
Program of University by MOE China (Grant no. 20130201110069). of multi-stream plate-fin heat exchanger based on genetic algorithm layer
pattern design. Int J Heat Mass Transf 2015;82:510–20.
[36] Singh SK, Mishra M, Jha PK. Nonuniformities in compact heat exchangers—
scope for better energy utilization: a review. Renew Sustain Energy Rev
2014;40:583–96.
[37] Fan YN. How to design plate fin heat exchangers. Hydrocarb Process
References 1966;45(11):211–7.
[38] Weimer RF, Hartzog DG. Effects of Maldistribution on the Performance of
[1] Castle WF. Air separation and liquefaction: recent developments and prospects for Multistream, Multipassage Heat Exchangers. Adv Cryogen Eng. 1973. p. 52–64
the beginning of the new millennium. Int J Refrig 2002;25(1):158–72. Springer US.
[2] Li Q, Flamant G, Yuan X, et al. Compact heat exchangers: a review and future [39] W. Suessmann A. Mansour. Passage arrangement in plate-fin heat exchangers.
applications for a new generation of high temperature solar receivers. Renew In: Proceedings of the XV international congress of refrigeration. 1 1979 p.
Sustain Energy Rev 2011;15(9):4855–75. 421–429.
[3] Manjunath K, Kaushik SC. Second law thermodynamic study of heat exchan- [40] Chen C, Shen Y. Cryogenic heat exchanger. Peking: China Machine Press; 1987.
gers: a review. Renew Sustain Energy Rev 2014;40:348–74. [41] L.E. Haseler, Performance calculation methods for multi-stream plate-fin heat
[4] Das PK, Ghosh I. Thermal design of multistream plate fin heat exchangers—a exchangers[C], Heat exchangers―theory and practice. International centre for
state-of-the-art review. Heat Transf Eng 2012;33(4–5):284–300. heat and mass transfer. Symposium. 14. 1983: 495-506.
[5] Liang P, Shanxiu S, Yanzhong L, et al. Calculation and thermodynamic analysis [42] Prasad BSV, Gurukul S. Differential method for sizing multistream plate fin
on liquefaction processes of natural gas with expanders. J-XIAN JIAOTONG heat exchangers. Cryogenics 1987;27(5):257–62.
UNIV 2007;41(9):1115. [43] Prasad BSV. The performance prediction of multistream plate-fin heat
[6] Gómez MR, Garcia RF, Gómez JR, et al. Review of thermal cycles exploiting the exchangers based on stacking pattern. Heat Transf Eng 1991;12(4):58–70.
exergy of liquefied natural gas in the regasification process. Renew Sustain [44] Gurukul S, Prasad BSV. Differential methods for the performance prediction of
Energy Rev 2014;38:781–95. iultistream plate-fin heat exchangers. J Heat Transf 1992;114:41.
[7] Remeljej CW, Hoadley AFA. An exergy analysis of small-scale liquefied natural [45] Prasad BSV. The sizing and passage arrangement of multistream plate-fin heat
gas (LNG) liquefaction processes. Energy 2006;31(12):2005–19. exchangers. Heat Transf Eng 1996;17(3):35–43.
[8] Moein P, Sarmad M, Ebrahimi H, et al. APCI-LNG single mixed refrigerant [46] Prasad BSV. Fin efficiency and mechanisms of heat exchange through fins in
process for natural gas liquefaction cycle: analysis and optimization. J Nat Gas multi-stream plate-fin heat exchangers: formulation. Int J Heat Mass Transf
Sci Eng 2015;26:470–9. 1996;39(2):419–28.
[9] Na Y, Yanzhong L, Hanfei T. Novel liquid product air separation system based on [47] Prasad BSV. Fin efficiency and mechanisms of heat exchange through fins in
cold energy of liquefied natural gas. J.-XIAN JIAOTONG UNIV. 2007;41(1):122. multi-stream plate-fin heat exchangers: development and application of a
[10] Cornelissen RL, Hirs G. Exergy analysis of cryogenic air separation. Energy rating algorithm. Int J Heat Mass Transf 1997;40(18):4279–88.
Conver Manag 1998;39(16):1821–6. [48] Pingaud H, Le Lann JM, Koehret B, et al. Steady-state and dynamic simulation
[11] Zhu Y, Legg S, Laird CD. Optimal design of cryogenic air separation columns of plate fin heat exchangers. Comput Chem Eng 1989;13(4):577–85.
under uncertainty. Comput Chem Eng 2010;34(9):1377–84. [49] Reay DA. Compact heat exchangers: a review of current equipment and R&D in
[12] The schematic drawing of multi-stream plate fin heat exchangers. Available the field. Heat Recovery Syst CHP 1994;14(5):459–74.
from: 〈http://www.hts.org.uk/downloads/Linde_LNG_HEX_09Dec2008_ [50] Yuan ZX, Ma YL, Xia GP, et al. Computer implementation of the passage
Extract.pdf〉; 2008 [accessed 8.05.15]. arrangement for plate-fin heat exchangers according to local balance princi-
[13] Kakac S, Liu H, Pramuanjaroenkij A. Heat Exchangers: Selection, Rating and ple. J Therm Sci 1997;6(3):190–6.
Thermal Design. CRC press; 2012. [51] Zeng-Yuan G, Shu W, Xin-Guang C. Field synergy principle for heat exchanger
[14] Kuppan T. Heat Exchanger Design Handbook. Mechanical Engineering. New enhancement. Chin Sci Bull 2003;48(22):P56–59.
York: Marcel Dekker Inc: CRC press; 2000. [52] Reneaume JM, Niclout N. Plate fin heat exchanger design using simulated
[15] Pacio JC, Dorao CA. A review on heat exchanger thermal hydraulic models for annealing. Comput Aided Chem Eng 2001;9:481–6.
cryogenic applications. Cryogenics 2011;51(7):366–79. [53] Reneaume JM, Niclout N. MINLP optimization of plate fin heat exchangers.
[16] Yee TF, Grossmann IE, Kravanja Z. Simultaneous optimization models for heat Chem Biochem Eng Q 2003;17(1):65–76.
integration—I. Area and energy targeting and modeling of multi-stream [54] Wang L, Sundén B. Design methodology for multistream plate-fin heat
exchangers. Comput Chem Eng 1990;14(10):1151–64. exchangers in heat exchanger networks. Heat Transf Eng 2001;22:3–11.
[17] Yee TF, Grossmann IE, Kravanja Z. Simultaneous optimization models for heat [55] Picon-Nunez M, Polley GT, Medina-Flores M. Thermal design of multi-stream
integration—II. Heat exchanger network synthesis. Comput Chem Eng heat exchangers. Appl Therm Eng 2002;22(14):1643–60.
1990;14(10):1165–84. [56] Luo X, Li M, Roetzel W. A general solution for one-dimensional multistream
[18] Yee TF, Grossmann IE, Kravanja Z. Simultaneous optimization models for heat heat exchangers and their networks. Int J Heat Mass Transf 2002;45
integration—III. Process and heat exchanger network optimization. Comput (13):2695–705.
Chem Eng 1990;14(11):1185–200. [57] Luo X, Guan X, Li M, et al. Dynamic behaviour of one-dimensional flow
[19] Gorji-Bandpy M, Yahyazadeh-Jelodar H, Khalili M. Optimization of heat multistream heat exchangers and their networks. Int J Heat Mass Transf
exchanger network. Appl Therm Eng 2011;31(5):779–84. 2003;46(4):705–15.
[20] Some of the common fin geometries in plate fin heat exchangers. Available [58] G. Cui H. Lu M. Lv Application of even temperature-difference field optimizing
from: 〈https://www.htri.net/htri-xchanger-suite.aspx/downloads/fins.pdf〉; principle on fluid organization of multi-stream heat exchanger. In: Proceed-
2012 [accessed 8.05.15]. ings of the eighth international conference on advanced computational
[21] Guo K, Zhang N, Smith R. Optimisation of fin selection and thermal design of methods in heat transfer. 2004 p. 215–222.
counter-current plate-fin heat exchangers. Appl Therm Eng 2015;78(5):491–9. [59] Lv Y, Cui G, Guo J, et al. Application of temperature-difference uniformity
[22] Chato JC, Laverman RJ, Shah JM. Analyses of parallel flow, multi-stream heat optimization principle to path arrangement of multi-stream heat exchangers. J
exchangers. Int J Heat Mass Transf 1971;14(10):1691–703. Chem Indus Eng, 2007;10:010.
514 Z. Wang, Y. Li / Renewable and Sustainable Energy Reviews 53 (2016) 500–514

[60] Zhao Y, Cui G, Lu H, et al. Comprehensive-performance optimization design [75] Mishra M, Das PK, Sarangi S. Second law based optimisation of crossflow
method for multi-stream heat exchangers. J Eng Therm Energy Power, plate-fin heat exchanger design using genetic algorithm. Appl Therm Eng
2004;19(2):171–4. 2009;29(14):2983–9.
[61] Peng H, Luo L, Wang Q, et al. Numerical study of differential and optimal [76] Rao RV, Patel VK. Thermodynamic optimization of cross flow plate-fin heat
design for multi-stream plate-fin heat exchanger. J Chem Indus Eng 2004;55 exchanger using a particle swarm optimization algorithm. Int J Therm Sci
(6):876–81. 2010;49(9):1712–21.
[62] Guan X, Luo X, Li M, et al. Analytical solution of dynamic behavior for plate- [77] Peng H, Ling X, Wu E. An improved particle swarm algorithm for optimal
fin heat exchangers. J Eng Thermophys 2003;24(4):688–90. design of plate-fin heat exchangers. Indus Eng Chem Res 2010;49(13):6144–9.
[63] Guo J, Cui G, Lv Y, et al. Passage arrangement optimization of multi-stream [78] Joda F, Tahouni N, Panjeshahi MH. Application of genetic algorithms in design
heat exchanger based on structure performance continuity principle. J Eng and optimisation of multi‐stream plate–fin heat exchangers. Can J Chem Eng
Thermophys 2009;8(30):1379–82. 2013;91(5):870–81.
[64] Shah RK, Skiepko T. Entropy generation extrema and their relationship with [79] Tahouni N, Miryahyaie S, Joda F, et al. Pressure drop optimisation in design of
heat exchanger effectiveness—number of transfer unit behavior for complex multi-stream plate-fin heat exchangers, considering variable physical prop-
flow arrangements. J Heat Transf 2004;126(6):994–1002. erties. Cann J Chem Eng 2013;91(10):1650–9.
[65] Guo ZY, Liu XB, Tao WQ, et al. Effectiveness-thermal resistance method for heat [80] Ghosh I, Sarangi SK, Das PK. An alternate algorithm for the analysis of mul-
exchanger design and analysis. Int J Heat Mass Transf 2010;53(13):2877–84. tistream plate fin heat exchangers. Int J Heat Mass Transf 2006;49(17):2889–
[66] Cheng X, Zhang Q, Liang X. Analyses of entransy dissipation, entropy gen- 902.
eration and entransy–dissipation-based thermal resistance on heat exchanger [81] Ghosh S, Ghosh I, Pratihar DK, et al. Optimum stacking pattern for multi-
optimization. Appl Therm Eng 2012;38:31–9. stream plate-fin heat exchanger through a genetic algorithm. Int J Therm Sci
[67] Zhao M, Li Y. New integral-mean temperature difference model for thermal 2011;50(2):214–24.
design and simulation of parallel three-fluid heat exchanger. Int J Therm Sci [82] Yunyun H, Wu X, Gaohong H. Optimum passage arrangement for multi-
2012;59:203–13. stream plate-fin heat exchanger in various conditions through genetic algo-
[68] Wu LB, Li Z, Song YZ. Field synergy principle of heat and mass transfer. Chin rithm. Comput Appl Chem 2012;29(1):10–4.
Sci Bull 2009;54(24):4604–9. [83] Peng X, Liu Z, Qiu C, et al. Passage arrangement design for multi-stream plate-
[69] Aspen Plate-fin Exchanger. Available from: 〈http://www.aspentech.com/pro fin heat exchanger under multiple operating conditions. Int J Heat Mass Transf
ducts/aspen-plate-fin.aspx〉; 2015 [accessed 8.05.15]. 2014;77:1055–62.
[70] Htri: Design and simulate plate-fin exchangers. Available from: 〈https://www. [84] Tian Q, He G, Zhao L, et al. Passage arrangement optimization of multi-stream
htri.net/xpfe.aspx〉; 2015 [accessed 8.05.15]. plate-fin heat exchangers. Appl Therm Eng 2014;73(1):963–74.
[71] UniSim Plate-Fin Exchanger Modeler. Available from: 〈https://www.honey [85] Skaugen G, Kolsaker K, Walnum HT, et al. A flexible and robust modelling fra-
wellprocess.com/en-US/explore/products/advanced-applications/unisim/ mework for multi-stream heat exchangers. Comput Chem Eng 2013;49:95–104.
Pages/unisim-design-suite.aspx〉; 2015 [accessed 8.05.15]. [86] Skaugen G, Hammer M, Gjennestad MA. Investigation of non-ideal behavior of
[72] Gosselin L, Tye-Gingras M, Mathieu-Potvin F. Review of utilization of genetic plate-fin heat exchangers in lng services. Energy Proc 2015;64:13–22.
algorithms in heat transfer problems. Int J Heat Mass Transf 2009;52(9):2169–88. [87] Zhao M, Li Y. An effective layer pattern optimization model for multi-stream
[73] Khosravi R, Khosravi A, Nahavandi S, et al. Effectiveness of evolutionary plate-fin heat exchanger using genetic algorithm. Int J Heat Mass Transf
algorithms for optimization of heat exchangers. Energy Convers Manag 2013;60:480–9.
2015;89:281–8.
[74] Mishra M, Das PK, Sarangi SK. Optimum design of crossflow plate-fin heat
exchangers through genetic algorithm. Int J Heat Exch 2004;5(2):379–401.