i An update to this article is included at the end

Energy Conversion and Management 175 (2018) 86–98

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Experimental investigation on thermal performance and economic analysis T

of cosine wave tube structure in a shell and tube heat exchanger
⁎
Kamel Milani Shirvan, Mojtaba Mamourian , Javad Abolfazli Esfahani
Department of Mechanical Engineering, Ferdowsi University of Mashhad, P.O.B. 91775–1111, Mashhad, Iran

A R T I C LE I N FO A B S T R A C T

Keywords: A novel design of tube structure with cosine wave is proposed in this study to investigate effects of wavy surface
Wavy tubes characteristics on enhancement of heat transfer in a shell and tube heat exchanger, experimentally. All ex-
Thermal performance factor periments have been designed and carried out on water/water loops. Response Surface Methodology is used to
Shell and tube heat exchanger investigate the optimization procedure. Four effective parameters of wavy starting length
Economic analysis
(0 mm ≤ b ≤ 120 mm), hot water flow rates (11 l/min ≤ Qc ≤ 19 l/min), cold water flow rates (7 l/
min ≤ Qh ≤ 11 l/min) and wavy wavelength (0 mm ≤ λ ≤ 80 mm) in five levels was selected to obtain the
maximum thermal performance factor. The results showed that the thermal performance factor reduces by in-
creasing of the hot water flow rate. The thermal performance factor of wave tubes is found to be larger than that
of smooth tubs. Optimizing of the procedure indicated that to maximize the thermal performance factor, the
lower values of wave starting length and hot water flow rate, and higher values of cold water flow rate and a
wavelength of 33.13 mm, must be selected. Also results show that the corrugating process is useful from eco-
nomic analysis point of view.

1. Introduction
Shell and tube heat exchangers are the most popular types of heat
exchangers in industries, especially in petrochemical industry, oil and
gas refinery, food industries and etc. [1,2]. This type of heat exchangers
is usable for high temperature and high pressure conditions. The
structure of the shell and tube heat exchangers is consisting of a number
of tubes that are placed inside a cylinder; and the two fluids of cold and
hot, exchange their heat through metal wall of the pipes without in-
teracting directly. In other words, one of the two fluids flows inside the
tubes, and the other enters around them within a shell. One of the
characteristics of these types of heat exchangers is that the two fluids
can be moved in two phases and in counter flow directions. The shell
and tube heat exchangers have vast of advantages, such as high contact
levels in low volume, good mechanical design and uniform distribution
of pressure and easy cleaning procedure for tubes [3].
In recent years, a significant attention to the shell and tube heat
exchangers has been observed and according to their importance and
application in industries, many studies have been addressed by the
scientific communities and academic researchers in this field. One of
the most popular issues, which researchers in the field of heat ex-
changers are interested in is baffle. They presented numerical and ex-
perimental studies on baffle type and changed its structure or the space
between baffles or shell to investigate their effects on heat transfer and
flow characteristics, such as: Yang and liu [4] have designed and per-
formed a numerical and experimental research on a novel shell-and-
tube heat exchanger with new plate baffles. Their results showed that
the novel design in plate baffle have higher performance than those
with rod baffles. Maakoul et al. [5] have carried out a 3-D numerical
simulation on low shell side flow rates conditions to evaluate effects of
novel trefoil-hole, helical baffles and the conventional segmental baffles
on the shell-side flow behavior, heat transfer characteristics and the
pressure drop. The results showed that using of helical baffles and
trefoil-hole baffles may be more effective than segmental baffles. The
shell-side thermo-hydraulic characteristics have been studied in a shell
and tube heat exchanger with trefoil-hole baffles experimentally by You
et al. [6]. They have investigated their study in turbulent conditions
and found correlations to achieve the Nusselt number and pressure loss.
In their correlations, studied parameters were summarized base on the
Reynolds number. Wang et al. [7] have done an experimental and nu-
merical investigation to evaluate the flow characteristics and heat
transfer performance within shell side of a shell and tube heat ex-
changer with fold baffles. They observed an enhancement in heat
transfer performance. The heat transfer characteristics and the pressure
drop in a small shell and tube heat exchanger with defining conditions
are studied numerically by Ozden and Tari [8]. They investigated the

⁎
Corresponding author.
E-mail address: mamourian@um.ac.ir (M. Mamourian).

https://doi.org/10.1016/j.enconman.2018.08.103
Received 7 June 2018; Received in revised form 29 August 2018; Accepted 30 August 2018
Available online 04 September 2018
0196-8904/ © 2018 Elsevier Ltd. All rights reserved.
K. Milani Shirvan et al.
Nomenclature
A Area (m2)
Cp specific heat at constant pressure (J/kg K)
Di the tube inside diameter (m)
ΔP pressure drop, Pa
ft the friction factor
h convective heat transfer coefficient (W/m2 K)
k thermal conductivity (W/m K)
K the coverage factor
L eff effective tube length
ṁ mass flow rate (kg/s)
Nu Nusselt number,
Np The number of the tube passes
Nt The tube numbers
p pressure (Pa)
Pr Prandtl number,
Re Reynolds number,
T temperature (K)
effect of the baffle spacing to shell diameter ratio on the heat exchanger
performance with changing the flow rate. Lei et al. [9] have in-
vestigated a numerical study to evaluated thermo-hydraulic perfor-
mance of two novel shell and tube heat exchangers with louver baffles.
They have also compared their results with a shell and tube heat ex-
changer with segmental baffles. They found that in their novel studied
heat exchangers with louver baffles, the heat transfer coefficient per
pressure drop is higher than exchangers with segmental baffles. Wen
et al. [10] used a new design of shell and tube heat exchangers with
ladder-type fold baffle to study the flow and heat transfer character-
istics, numerically. They found that their modification leads to en-
hancement in flow and heat transfer characteristics. Yang et al. [11]
carried out an experimental study to investigate the thermal perfor-
mance in a combined serial two shell-pass shell-and-tube heat ex-
changer (CSTSP-STHX) with continuous helical baffles. They had a
comparison between CSTSP-STHX with the double shell-pass shell-and-
tube heat exchanger with segmental baffles (SG-STHX). They found that
the CSTSP-STHX has better performance compared to SG-STHX. Wang
et al. [12] have studied and optimized a shell and tube heat exchanger
with staggered baffles. They found that from the heat transfer im-
provement point of view, the shell and tube heat exchanger with seg-
mental baffles is not always the best option. Moreover, some other re-
searchers such as Yang et al. [13] and Chen et al. [14] have carried out
other studies in this field of heat exchangers.
Moreover, some researchers have presented some studies on
thermo-economic and cost analysis in heat exchangers such as:
Sadeghzadeh et al. [15] have done a numerical investigation to find
techno-economically optimum condition in a shell-and-tube heat ex-
changer. They used Delaware method to estimate the heat transfer ef-
ficiency and the shell-side pressure drop. Yang et al. [16] have used a
modified optimization method to present a theory for minimizing the
function cost in shell and tube heat exchangers. They divided a shell-
and-tube heat exchanger into several in-series heat exchangers. They
found that their method can significantly decrease the total cost. A
multi-objective optimization algorithm is used by Mirzaei et al. [17] to
optimize simultaneously cost and effectiveness in a shell and tube heat
exchanger. Their results showed an enhancement in thermal efficiency.
Hadidi et al. [18] have used the imperialist competitive algorithm to
minimize the total cost in shell and heat exchangers. Their results
showed that the used algorithm has an effective benefit for designing a
shell and tube heat exchanger with optimal design and it has higher
accuracy in less computational time. Another optimization of shell and
tube heat exchanger has been successfully done by DeVasconcelos Se-
gundoa et al. [19] to minimize the total annual cost by Differential
87
Energy Conversion and Management 175 (2018) 86–98
TPF The thermal performance factor
Re Reynolds number,
Greek symbols
ρ density (kg/m3)
μ dynamic viscosity (kg/m s)
Subscripts
c cold
h hot
i inner
m mean
s shell side
t tube side
o outer
w wall
Evolution (DE) and a novel Differential Evolution variant, denominated
Tsallis Differential Evolution (TDE). Some other researchers such as Yin
et al. [20] and Hatami et al. [21] have survived in this field of heat
exchangers.
Some other researchers surveyed on effect of tubes arrangements
such as: The effects of tubes arrangements on the flow pattern in a shell
and tube heat exchanger have been studied numerically by Labbadlia
et al. [22]. They found that the studied parameters have an important
effect on the flow pattern.
Some other studies are focused on novel methods such as injection
of air flow, insert of twisted tapes or using new structure of tubes such
as coiled tubes to increase heat transfer. For example: Ayub et al. [23]
have carried out an experimental study to evaluate the thermal per-
formance of a shell and tube heat exchanger with interstitial twisted
tapes. They used the propylene glycol/water solution as a working fluid
and compared their heat exchanger with a same sized shell and tube
exchanger, but with single segmental baffles. They found that tube heat
exchanger with interstitial twisted tapes has better thermal enhance-
ment against the one with single segmental baffles. El-Said and Alsood
[24] have done an experimental investigation to study the air injection
effects on enhancement of the thermal performance within the shell
side of the shell and tube heat exchangers. Their results showed that the
thermal performance increases by injection of air flow. Dizaji et al. [25]
have done an experimental study to investigate the influences of flow,
thermodynamic and geometrical characteristics on exergy loss in shell
and coiled tubes heat exchangers. Sadighi Dizaji et al. [26] have in-
vestigated a shell and tube with corrugated tube to analyze the exergy.
They have used various arrangements of convex and concave corru-
gated tubes. Their results showed that, heat exchanger with corrugated
tubes have more exergy loss and NTU than heat exchanger with smooth
tubes. Sepehr et al. [27] have done a numerical study to investigate the
heat transfer, pressure drop and entropy generation in shell and heli-
cally coiled tube heat exchangers. They found some correlations for
estimating of the Nusselt number and friction factor in the shell side.
Moreover, they studied about the thermal effectiveness, NTU and the
entropy generation rate relations.
According to the literature review, it can be concluded that most of
the researches have been done on baffles with less consideration on the
tubes structure. Accordingly, the researchers decided to develop their
research on the tubes structure. Therefore, the novelty of the present
study lies on this fact that the structure of tube bundle in heat ex-
changers with cosine corrugated tubes is considered for first time. On
the other hand, as much as the authors know, the thermal performance
factor (TPF) on the tube side of the shell and tube heat exchanger with
K. Milani Shirvan et al.
Fig. 1. A schematic presentation of studied corrugated tubes no initial length.
cosine corrugated tubes in tube bundle has not been investigated till
date. Moreover, the sensitivity analysis of effective parameters on the
thermal performance factor and thermal economic analysis are carried
out in this study.
2. Experimental configuration
2.1. Corrugated tubes and heat exchanger geometry
In this research, all of the smooth and corrugated tubes are copper
tubes with diameter of 16 mm (5/8″) and thickness of 0.63 mm. The
length of tubes is designed to be 1150 mm. Since, the length of the tubes
decreases during the corrugating process, the corrugated tubes are
shorter than the smooth tubes. Therefore, to compensate the length
reduction, copper tubes used in corrugating process were initially
considered longer than those used in smooth tubes bundle case. On the
other hand, with this politic, the smooth tubes and the corrugated tubes
have the same length for comparing.
Fig. 1 indicates a schematic presentation of studied corrugated
tubes. As mentioned above in introduction section and shown in Fig. 1,
in some cases, the wave in tubes starts after a starting length.
In this experiment, every tube bundle is consisting of 10 tubes.
Fig. 2 represent the layout of the tubes in tube bundles. Tubes in tube
sheet are located in triangular layout.
A graphical schematic from studied heat exchanger is indicated in
Fig. 3; which shows the studied shell and tube type heat exchanger with
shell diameter of 100 mm (4″). The length of the heat exchanger is
1150 mm. Additionally, this heat exchanger has 3 baffles with 22% cut.
The thickness of baffles plate is 3 mm. In tests, the wavy tubes are
fabricated with following equation:
y = α cos(λx ) (1)
Fig. 2. Layout of tubes in tube bundles.
88
Energy Conversion and Management 175 (2018) 86–98
According to Eq. (1), x and y are the coordinates in x-axis and y-axis
directions, λ is the wave length and α is the amplitude of the wavy wall
that is fixed at 2 mm.
2.2. Experimental set-up and procedure
The studied experimental setup is illustrated in Fig. 4 with details.
As can be seen in this figure, the experimental setup is consisting of two
main loops. It must be noted that, in this research, all of experiments
have been designed and carried out on water/water loops for the
counter current flow.
These two loops are explained as follows:
– Hot water loop:
This loop is consisting of pump, tank and electrical heater. The
heated water is produced inside of a tank with 140-liter capacity. In this
tank the hot water is kept at constant temperatures that defined by
thermostat. For this purpose, three electrical heaters with 3 kw power
(9 kw in total) have been located in the tank. Produced hot water
pumps inside the inner tube of heat exchanger by a centrifugal pump of
0.75 hp to exchange heat with cold water in shell of the heat exchanger.
To measure and control the hot water flow rate, a plug valve and a
Rotameter are used during the experiments. 3 K-type thermocouples are
used to measure the tube wall temperatures; moreover, two thermo-
meters are located on entrance and outlet port of hot water to obtain
their temperatures. It must be noted that, to remove heat loss from hot
water tank and also cold water tank to surroundings, body of tanks were
covered by the glass wool insulator.
– Cold water loop
K. Milani Shirvan et al. Energy Conversion and Management 175 (2018) 86–98

Fig. 3. A graphical schematic from studied heat exchanger: A: Hot water entrance, B: Cold water exit, C: Thermocouple, D: Cold water entrance, E: Hot water exit, F:
Thermometer, G: Baffle, H: Insulation, I: Header, J: Tubes, K: Tie rod.

Cold water pumps by a centrifugal pump with 0.75 hp power to exchanger using TEMA standard [28]. To calculate the tube-side heat
shell side from a storage tank with 70-liter capacity. It is used for ex- transfer coefficient and pressure drop, the Bell-Delaware method has
changing heat with hot water in tubes. Cooling cycle is consisting of a been used. The tube side Nusselt number and pressure drop correlations
compressor, condenser, a coil inside of cold water tank as an evaporator according to the Reynolds number, is calculated by Eqs. (2)–(7) [2,29].
and expansion valve. In the cooling cycle, R22 is used as a refrigerant. For Re < 2300, Sieder and Tate correlation is given as follows
Assame as the hot water loop, a plug valve and a Rotameter are used for [30,31]:
measuring and controlling of cold water flow rate during the experi- 1/3
ments. Two thermometers are also used for measuring of entrance and D
Nu = 1.86 ⎜⎛Re ·Pr· i ⎟⎞
outlet of cold water temperatures. ⎝ Leff ⎠ (2)
Table 1, gives more vital details about the two loops.
Leff = L−2Ls (3)

3. The heat exchanger thermal modeling and validation In Eq. (2) and (3), Di is the tube inside diameter, Leff is effective tube
length and Ls is the tube sheet thickness. Additionally, the Reynolds
In this investigation, it is tried to perform the design of a heat number is denoted by Re, which can be calculated using Eq. (4):

Fig. 4. The geometric structure of the main loops of cold and hot waters in the studied heat exchanger.

89
K. Milani Shirvan et al. Energy Conversion and Management 175 (2018) 86–98

Table 1 qh
hh =
The range and accuracy of measurement devices. |Tm−Tw | (13)
Instruments Range Accuracy In which, Tm and Tw are the mean temperatures of the fluid inside
tube and the tube wall, respectively. These temperatures are obtained
Rotameter 4–25 l/min 1 l/min
Thermometer 0–80 °C 0.1 °C by averaging process (Tin and Tout ) and threethermocouples tempera-
Pressure gauge 0–500 pa 5 pa tures, which are located on tube bundle, respectively.
hh Di
Nuh =
ρVDi k (14)
Re =
μ (4) In order to validate the study, the predicted theoretical pressure
drop and Nusselt number (determined by Eqs. (7) and (2)) are com-
Moreover, the tube side Prandtl number is denoted by Pr and can be
pared to experimental data obtained in the smooth tube and wavy cases
determined by Eq. (5):
with λ = 80 mm, b = 60 mm, Qh = 9 l/min, Qc = 15 l/min for pressure
μcp drop and Nusselt number (defined by pressure gauge and Eq. (14)),
Pr =
k (5) respectively. The results of the comparison are presented in Table 2.
According to Table 2, the presented experimental study has suffi-
where k and cp are the water thermal conductivity and specific heat
ciently accurate to use for calculating other cases.
coefficient, respectively.
To compute the flow velocity (V) in tube side, Eq. (6) can be used as
4. Uncertainty in experimental data
follows:

The measuring process of parameters such as volume flow rate and

V=
mṫ ⎛ Np ⎞
⎜ ⎟
temperature always have some errors, which leads to an uncertainty in
()
π
4
Di2 ρ ⎝ Nt ⎠ (6) experimental data obtained. The thermocouples and flow meter outputs
Here, Np is the number of the tube passes, which is defined to be 1 in are using to calculate Nu number for the heat transfer experiment. It
present research, and Nt is tube numbers that it is fixed at 10. must be noted that the experimental uncertainty has been calculated by
Pressure drop in tube side of heat exchanger is calculated by Eq. (7): Kline and McClontock [33] and Moffat [34] method. The Eqs. (15) and
(16) have been used to evaluate the uncertainty. Meanwhile, the un-
f ·Leff ·Np V 2 certainty is consists of two parts: the uncertainty of Instruments
ΔP = ⎛ t
⎜ + 4·Np ⎞ ρ· ⎟

⎝ D i ⎠ 2 (7) ((uNu, Ins ) and the uncertainty of repeatability (uNu, Rep ).

In above equation, ft is the friction factor that can be determined by UNu = K·uNu ; uNu = (uNu, Ins )2 + (uNu, Rep)2 (15)
Eq. (8) [29]:
2 2 2

⎧ (1.82log(Re )−1.64)−2 , Re ≤ 2100 uNu, Ins = ( ∂Nu

) + ( u ) + ( u ) and
u
∂hh hh
∂Nu
∂Di Di
∂Nu
∂k k

ft = 0.0054 + 0.00000023(Re−3/2), 2100 < Re < 4000 2 2 2

⎨
⎩ 0.00128 + 0.1143(Re−1/3.214 ), Re ≥ 4000 (8)
uNu, Rep = ( u ) +( u ) +( u )
∂Nu
∂hh hh
∂Nu
∂Di Di
∂Nu
∂k k (16)

The thermal performance factor (TPF) in this paper is defined by Eq. Here u is the contribution of the uncertainty in the results from
(9): parameters hh , Di and k , and K is the coverage factor, which is con-
sidered 2 in this case. As a result, this leads to an uncertainty of ef-
Nuimproved / NuSmooth fectiveness less than 10%.
TPF =
Δpimproved /ΔpSmooth (9)
5. Optimization procedure
On the other hand, TPF values has been normalized by dividing TPF
of each test by the its values of the smooth tube case.
5.1. Response surface methodology
In this paper, the experimental Nusselt number is calculated using
Eqs. (10)–(13) [32]:
The Response Surface Methodology (RSM) is a series of statistical
Qh = ṁ h cp, h (Tin−Tout )h (10) and mathematical methods that it is used to analyze empirical results
[35–37]. This method is also very practical in designing, improving and
where Qh is the hot heat transfer rate and Tin and Tout are the inlet and formulating new products [38–40]. The most widely application of the
outlet temperatures of the hot water, respectively. mentioned method is in industrial fields. The Response surface meth-
Qh odology is used in situations in which input variables have a high im-
qh = pact on yield and product quality characteristics (responses) in process
Ao (11)
[41,42]. The dependence of a response to inputs is given by Eq. (17)
In which, A o is total surface area, that it is calculated using Eq. (12): [43–45]:

Ao = Nt (πDi Leff ) (12) Y= f(X1, X2, …, Xk) + ε (17)

Table 2
The comparison between the theoretical and experimental Nu number and pressure drop obtained in the smooth tube case and wavy case with λ = 80 mm,
b = 60 mm, Qh = 9 l/min, Qc = 15 l/min.
Parameter Theoretical mount Experimental amount Theoretical mount Experimental Error in smooth Error in wavy
smooth case smooth case wavy case amount wavy case case (%) case (%)

Nu number 8.04 9.99 8.23 13.17 19.46 37.51

Pressure drop (Pa) 34.27 45 37.63 50 23.84 24.74

90
K. Milani Shirvan et al. Energy Conversion and Management 175 (2018) 86–98

In above equation, Y and X are the input variables and ε is the error Table 3
function. A complete description of a process requires it to be modeled Input variables and the levels of their values.
as a polynomial function; generally, as a polynomial of degree 2 Parameters Symbol −2 −1 0 +1 +2
(quadratic) or higher. The quadratic model is suitable for industrial
processes and has the following strengths [46–48]: Wave wavelengths (mm) λ 0 20 40 60 80
Wave starting lengths (mm) b 0 30 60 90 120
The quadratic model is highly flexible and has a proper function in
Hot water flow rates (l/min) Qh 7 8 9 10 11
correcting approximation of response. Many experiments have been Cold water flow rates (l/min) Qc 11 13 15 17 19
described the proper function of a quadratic model for describing the
response behavior.
Considering the advantages of the quadratic model, this type of summarized in Table 3.
model is used to examine the effect of input variables on the output The values of (-2), (0) and (+2) are indicators of the lower, equal to
variables in the present experiment. In fact, since operating conditions and higher values to the central point. Moreover, the thermal perfor-
may be accompanied by non-standard changes, a nonlinear quadratic mance factor is considered as a dependent variable.
model is capable of describing it. The quadratic polynomial model can
be considered as the form of Eq. (18) [44]: 6. Results and discussion
4 4 3 4
Z = β0 + ∑ βi xi + ∑ βii xi xi + ∑ ∑ βij x i x j + ei In this paper, besides defining the optimum status of the effective
i=1 i=1 i=1 j=i+1 (18) parameters to achieve the maximum of the thermal performance factor,
an economic analysis is carried out to determine if the corrugating
In this equation Z is the response, x i and x j are independent vari- process is useful from economic and saving energy point of views or its
ables, β0 is the constant number and βi, βii and βij are linear, degree 2, costs is more than its benefits in this studied heat exchanger.
and interacting coefficients, respectively. In addition, ei is considered as
an error [49]. 6.1. Statistical analysis
Using ANOVA1 analysis, the presented models for the responses
were evaluated and the regression coefficients for linear, interactions Analytical analysis in this paper is carried out by initial data that are
and degree 2 sentences were estimated. Moreover, using the coefficient presented in Table 4. According to this table, the effective parameters
of determination (R2), the fitting quality of the equation of the models and the runs of the test along with amount of dependent variable (TPF)
was expressed. The coefficient of the determination or the fit is one of are summarized. The basis of this table is on 31 randomized trials.
the used tests in the analysis of variance for identifying the superior After analyzing and evaluating the data in Table 4, the mathema-
model [43–45]. Statistically, when a model has the highest coefficient tical model of responses is determined with regression of degree 2
of determination, it is a proper model; however, it must be noted that, (quadratic). In addition, the accuracy and acceptability of regression
although this condition is necessary, but it is not enough to determine models are analyzed using variance analysis. The results of the variance
the accuracy of the model. Statistically, the significance of the models analysis of the quadratic model are presented in Table 5.
evaluates with P-value and F-value. In fact, as the F-value gets higher
values, it shows that the model has a greater validity [43–45]. The P- Table 4
value represents the meaningfulness of the model, as well as the ef- The levels of the factors values and the results of the experiments.
fectiveness of the factors in the model [44–46]. A model is meaningful
Run order Values of parameters TPF
when its P-value is less than 0.05 [36–39]. In addition, the P-value and
the F-value for each of the factors indicate their effects on the model λ b Qh Qc
[44–46,50].
1 40 60 9 15 1.80
2 40 60 9 15 1.85
5.2. Test run parameters 3 60 30 8 17 1.57
4 40 0 9 15 1.84
5 20 30 8 17 1.63
Since, the summarized aim of this paper is to obtain the optimized 6 0 60 9 15 2.01
geometry of tubes in shell and tube heat exchanger in order to max- 7 20 90 10 13 1.75
imize the thermal performance factor; four effective parameters of 8 20 90 8 13 2.08
wave starting length, hot water flow rate, cold water flow rate and 9 40 60 9 15 1.77
10 40 60 9 15 1.75
wavy wavelength are studied. As presented above, the statistical RSM
11 20 90 10 17 1.72
model is used as analytical procedure. The considered ranges of effec- 12 60 30 10 17 1.69
tive parameters are given as follows: 13 20 90 8 17 1.72
14 60 90 8 13 3.10
- Wave wavelength (λ): The variation range of λ is considered to be 15 20 30 10 17 1.64
16 40 120 9 15 1.87
0 mm, 20 mm, 40 mm, 60 mm and 80 mm. 17 40 60 9 19 1.00
- Wave starting length (b): The variation range of b is considered to be 18 60 30 10 13 1.77
0 mm, 30 mm, 60 mm, 90 mm and 120 mm. 19 80 60 9 15 1.61
- Hot water flow rates (Qh): The variation range of Qh is considered to 20 40 60 9 15 1.80
21 60 90 10 13 1.74
be 7 l/min,
22 60 90 8 17 1.57
- 8 l/min, 9 l/min, 10 l/min and 11 l/min. 23 20 30 10 13 1.82
- Cold water flow rates (Qc): The variation range of Qc is considered to 24 40 60 11 15 1.75
be 11 l/min, 13 l/min, 15 l/min, 17 l/min and 19 l/min. 25 40 60 9 19 1.32
26 20 30 8 13 1.90
27 60 30 8 13 1.89
It must be noted that, in this research five levels are defined to
28 40 60 7 15 1.53
analyze by RSM method. These parameters and their levels are 29 60 90 10 17 1.63
30 40 60 9 15 1.73
31 40 60 9 15 1.83
1
Analyze of Variance

91
K. Milani Shirvan et al. Energy Conversion and Management 175 (2018) 86–98

Table 5 Table 6
The variance analysis results of the TPF. The variance analysis results of TPF (for P-value < 0.05).
Source DOF Adj. SUM of Adj. Mean F-Value P-Value Source DOF Adj. SUM of Adj. Mean F-Value P-Value
Squares (SS) squares (MS) Squares (SS) squares (MS)

Model 14 1.67141 0.119386 1.43 0.024 Model 5 0.86382 0.172764 2.01 0.011
linear 4 0.36445 0.091113 1.09 0.039 linear 4 0.36445 0.091112 1.06 0.039
A 1 0.20907 0.209067 2.50 0.013 A 1 0.20907 0.209067 2.43 0.013
B 1 0.06615 0.066150 0.79 0.038 B 1 0.06615 0.066150 0.77 0.038
C 1 0.00042 0.000417 0.00 0.044 C 1 0.00042 0.000417 0.00 0.044
D 1 0.0882 0.088817 1.06 0.031 D 1 0.08882 0.088817 1.03 0.031
Square 4 0.61271 0.153178 1.83 0.173 Square 1 0.49937 0.499370 5.81 0.024
A*A 1 0.42911 0.429115 5.12 0.038 A*A 1 0.49937 0.499370 5.81 0.024
B*B 1 0.00018 0.000179 0.00 0.964 Error 25 2.14792 0.085917 – –
C*C 1 0.04575 0.045753 0.55 0.471 Lack-of-Fit 19 2.13692 0.112469 61.35 0.000
D*D 1 0.07511 0.075109 0.90 0.358 Pure Error 6 0.01100 0.001833 – –
2-way Interaction 6 0.69425 0.115708 1.38 0.281 Total 30 3.01174 – – –
A*B 1 0.27040 0.270400 3.23 0.091
A*C 1 0.09000 0.090000 1.07 0.315 R2 = 88.68%, R2−adj = 64.42% .
A*D 1 0.08703 0.087025 1.04 0.323
B*C 1 0.05063 0.050625 0.60 0.448
this diagram, plotted responses show how the errors are dispersed.
B*D 1 0.15210 0.152100 1. 82 0.197
C*D 1 0.04410 0.044100 0.53 0.479 Errors are the clutter between the actual values and the predicted va-
Error 16 1.34032 0.083770 – – lues of the responses by the model [51–53]. Proper and normal dis-
Lack-of-Fit 10 1.32932 0.132932 72.51 0.000 tribution of points around the right line represents the proper dis-
Pure Error 6 0.01100 0.001833 – –
tribution of errors. On the other hand, less skewed distributions in the
Total 30 3.01147 – – –
residual histograms can be observed. As a result, it can be dedicated
R2 = 75.50%, R2−adj = 56.56% . that these distributions are more similar to symmetrical distribution.
According to these diagrams, since the errors are dispersed normally,
According to above table, the coefficient of determination (R2) of the models are meaningful and the predicted responses are consistent
this model is 75.50% and F-value is 1.43, which indicates that the re- with the actual information; or it can be dedicated that a good corre-
gression model is meaningful. In this model, the sentences in which P- lation between observed and fitted values has been achieved [50–52].
value is less than 0.05 are more important in the regression model, and A good condition can be observed in Fig. 5, for the normal prob-
their model terms are more statistically significant [50,51]. As a result, ability plots of residuals. Beside above description, a general correlation
by removing the sentences with P-values higher than 0.05 from the form between the independent parameters (four effective parameters)
model, the data in Table 6 are obtained. and dependent parameters (TPF) is presented as below
As discussed in Section 4, F-value is the basis for measuring the
TPF = a + bλ + cb + dQh + eQc + fλ2 + gb2 + hQh 2 + iQc 2 + jλ·b
significance of the model. In other words, the meaningfulness of the
model for TPF is determined by F-value. The F-Value is the ratio of + kλ·Q h + lλ·Qc + mb·Q h + nb·Qc + oQ h ·Qc (19)
residual variances in a model with and without the predictor. The va- Table 7, summarizes the related coefficients to Eq. (19). It must be
lues close to 1 presenting that the residual variance is not under effect noted that this table also consists of insignificant coefficients. The basis
of the predictor; and its high values are indicators of reduction in re- of determining the significance of the coefficients is the values of P-
sidual variance, which can be resulted from the predictor. As a result, F value. In other words, parameters with P-value higher than 0.05, are
value is obtained dividing the variance of the group means by the mean ignored [50–52].
of the within group variances [52]. Using the coefficients given in the Table 7, in Eq. (17), the thermal
According to Table 5, R2 is75.50%. Therefore, it can be dedicated performance factor based on effective parameters is obtained as given
that this estimated model cannot consider 24.50% of the total changes in Eq. (20):
for the thermal performance factor in the studied heat exchanger.
As stated above, from Table 5, it can be dedicated that in addition to TPF = 1.8647−0.0933λ−0.05258b−0.0042Qh + 0.0608Qc−0.1303λ2 (20)
the linear terms of λ, b, Qh and Qc , the square term of λ2 is the significant The variations of the thermal performance factor are presented in
model term associated with the thermal performance factor for heat terms of effective test parameters of λ-b (a), λ-Qh (b), λ-Qc (c), b-Qh (d),
exchanger. On the other hand, the square terms of b2 , Qh 2 and Qc 2 and b-Qc (e) and Qc -Qh (f) in Fig. 6.
interaction terms of λ·b, λ·Qh , λ·Qc , b·Qh b·Qc and Qh·Qc are low im- According to Fig. 6, the following results can be dedicated:
portance parameters. In summary, these parameters have insignificant
effects on amount of the thermal performance factor. (a) To consider the wavy wavelength effects on the thermal perfor-
Since, the F-value is an important factor in determining the mean- mance factor, it can be observed that the maximum of the thermal
ingfulness of a parameter; it indicates the sensitivity of the parameter performance factor achieves in the vicinity of level (0), and its
on the objective function. Therefore, according to the F-value amounts minimum amount is at level (+2). Since, the corrugating process
in Table 6, to find more sensitive parameters, the classified terms from through corrugated tubes led to further surface area in comparison
high to low sensitivity values can be listed as the square term of λ2 and with the smooth tubes; thereby, an increase in amount of input heat
the linear terms of λ, Qc , b and Qh . In summary, the square term of λ2 is energy to shell from the tubes is expected. As a result, the amount of
the most significant parameter and the linear terms are the secondly heat transfer in the corrugated tubes will be greater than the
more significant on the thermal performance factor. smooth tubes. On the other hand, reduction in the number of the
waves, which means an enhancement in tube wavelength, leads to
6.2. Analysis of variance (ANOVA) weaker re-circulation zones and swirl flow intensity [54]. There-
fore, it can be concluded that the heat transfer decreases with en-
In this study, to obtain the residual plots, the data are processed by hancement of tube wavelength. Moreover, by increasing of the tube
the analysis of variance (ANOVA) in analytical software. wavelength, the value of the pressure drop decreases. But according
Fig. 5, indicates normal probability plots with the response error. In to Fig. 6, the amount of decrement in the pressure drop compared to

92
K. Milani Shirvan et al. Energy Conversion and Management 175 (2018) 86–98

Fig. 5. The residual plots of the thermal performance factor (TPF)

Table 7 increment in number of tube waves overcomes the negative effects

Estimated regression coefficients.
Term Coefficients for TPF
a 1.8647
b −0.0933
c −0.05258
d −0.0042
e 0.0608
f −0.1303
g 0.0000
h 0.0000
i 0.0000
j 0.0000
k 0.0000
l 0.0000
m 0.0000
n 0.0000
o 0.0000
the amount of enhancement in heat transfer can be ignored. It must
also be noted that considering the region in the vicinity of level 0,
the thermal performance factor of the smooth tube reduces (level
(−2)). This fact shows that the amount of increment in heat
transfer is less than enhancement in the pressure drop resulting
from increasing in the tube wavelength.
(b) The thermal performance factor decreases with the increment in
wavy starting length and hot water flow rate. It means that the
highest value of the thermal performance factor for mentioned
parameters is seen at level (+2) and its lowest amount at level
(−2). As mentioned above, increment in number of tube waves
leads to an enhancement in the re-circulation zones and swirl flow
intensity. As a result of this enhancement, the heat transfer in-
creases. On the other hand, increasing the tube waves leads to in-
crement in pressure drop. However, as can be seen in Fig. 6, the
positive effects of increasing the heat transfer value resulting from
of increasing pressure drop. Thereby, an improvement in thermal
performance factor is observed with reduction of wavy starting
length. Moreover, the fact that increasing the hot water flow rate
leads to a decrement in the thermal performance factor may be
explained through this fact that the enhancement of the hot water
flow rate, increases Re number. As Reynolds number increases, both
the pressure drop and the heat transfer are increased. However, in
this case, the positive effects of increasing the amount of heat
transfer is affected by the more dominant negative effects of in-
creasing pressure drop.
(c) By increasing the cold water flow rate, the thermal performance
factor increases. This may be explained by this fact that the en-
hancement of the cold water flow rate leads to an increase in heat
transfer rate between inside and outside of tubes in heat exchanger.
This increment is caused by enhancement of turbulence and vor-
tices of the flow in the vicinity of the tube walls, which led to a
higher heat transfer rate increment compared to the case with lower
cold water flow rate. Thereby, whenever the cold water flow rate
increases, the thermal performance factor enhances due to constant
pressure drop within tubes. For the case with λ = 20 mm,
b = 30 mm and Qh = 8 l/min, enhancement in cold water flow rate
from 11 l/min to 17 l/min, leads to 9.47% increase in the thermal
performance factor.
(d) In all cases, the thermal performance factor is larger than unity.
This fact is important form the energy saving point of view. On the
other hand, the application of wave tubes in tube bundle in a shell
and tube heat exchanger can lead to an enhancement in the thermal
performance factor and the energy saving in reasonable levels.
6.3. Optimization plots
Optimization of industrial processes is important because of their
direct impact on products quality and economic conditions of the

93
K. Milani Shirvan et al.
Fig. 6. The TPF variations as function of λ-b (a), λ-Q h (b), λ-Qc (c), b-Q h (d), b-Qc (e) and Qc -Q h (f). It must be noted that the optimization points show with ( )
index. The design points can be seen in left with ( ) index.
processes. Optimizing procedure can optimize each response in-
dividually but selecting optimal conditions for just one answer may
have a negative effect on other responses [55].
The multi-criteria methodology is relatively a common method by
researchers to evaluate of the optimization of goal functions. Whenever
several responses have to be considered at the same time, this approach
can be applied. Total numbers of responses taken into account by this
method needs to have optimal compromises between them. Considering
the optimization of analytical procedures, the most used and most
important multi-criteria methodology is the desirability function or the
Derringer function [56]. In this approach, a desirability function must
be constructed for any individual response; and as a result, a di-
mensionless desirability (di) scale must be considered instead of the
measured properties of any response. Each response must fulfill some
specifications in the measuring procedure, which will be defined by the
analyst through the individual functions. The individual desirability
function is between 0 (fully undesirable response) and 1 (completely
desired response). The results of measuring properties on different
94
Energy Conversion and Management 175 (2018) 86–98
orders of magnitude can be combined through this transformation.
Considering each of these desirabilities, the overall desirability (D) can
be obtained. The purpose of this section is to find the optimal condition
for the TPF as the goal function. For this purpose, the response surface
methodology by Minitab software is used to find optimal levels. Also,
the optimization procedure graphs are shown. To gather more in-
formation and details about optimization, referring to reference [46,47]
might be useful.
The result of optimization and optimal level in ranges of in-
dependent variables is shown in Fig. 7.
According to Fig. 7, optimization procedure performed on the basis
of experimental information by the response surface methodology
model with desirability of 0.5316. In addition, the maximum amount of
the thermal performance factor is predicted at 2.1164. The aim of this
paper is to find optimum conditions to maximize the thermal perfor-
mance factor; thereby, as summarized in Fig. 7, this goal can be found
at levels (−0.3434), (−2), (−2) and (+2) for the wavy wavelength,
wavy starting length, hot water flow rates and the cold water flow rates,
K. Milani Shirvan et al. Energy Conversion and Management 175 (2018) 86–98

Fig. 6. (continued)

Fig. 7. The optimization graphs for TPF.

95
K. Milani Shirvan et al. Energy Conversion and Management 175 (2018) 86–98

Table 8 Table 9
The total cost of studied shell and tube The normalized thermal performance/the
heat exchanger (EUR). normalized total cost.
Cases Ctot Cases (Nu/Δp)/ Ctot

1 1.55 1 1.16
2 1.55 2 1.16
3 1.55 3 1.01
4 1.55 4 1.19
5 1.55 5 1.05
6 1.55 6 1.29
7 1.57 7 1.12
8 1.57 8 1.32
9 1.55 9 1.16
10 1.55 10 1.16
11 1.55 11 1.11
12 1.55 12 1.09
13 1.55 13 1.11
14 1.57 14 1.97
15 1.55 15 1.06
16 1.56 16 1.16
17 1.00 17 1.00
18 1.57 18 1.13
19 1.55 19 1.04
20 1.55 20 1.16
21 1.57 21 1.11
22 1.55 22 1.01
23 1.57 23 1.16
24 1.55 24 1.13
25 1.55 25 0.85
26 1.57 26 1.21
27 1.57 27 1.2
28 1.56 28 0.98
29 1.55 29 1.05
30 1.55 30 1.16
31 1.55 31 1.16

respectively. It must be noted that, to estimate the amount of the wavy
wavelength from level of -0.3434, an interpolation between levels of
−1 and 0 must be done, which leads to the wavy wavelength of
33.13 mm.
7. Economic analysis and estimation costs for studied heat
exchanger
For estimating the total cost in studied shell and tube heat ex-
changer, in this paper, summation of two parameters consists of setup
cost or the investment cost of heat exchanger and the total discounted
operating cost are calculated. This equation is given as follows:
Ctot = Ci + Cod (21)
In Eq. (21), Ctot , Ci and Cod are the total annual cost, setup cost and
the total discounted, respectively [2]. For calculating the investment
cost, the following equation is used [57].
Ci = 4138.1At0.68 (22)
As can be seen in Eq. (22), this equation is function of the total tube
outside heat transfer surface area ( At ). It must be noted that in this
equation is defined for heat exchangers made with Carbon steel for shell
and Aluminum for tubes. Also, here it should be noted that the above
equation relates to smooth tubes. In this paper, in order to make the
data of the corrugated tube more realistic, according to the calculations
done on the costs, 40% is added to the values obtained from the Eq.
(22) for the corrugated tubes.
The total discounted operating cost in Eq. (21) is function of the
pressure drops in shell side and tube side of heat exchanger and it can
be determined as Eq. (23):
ny
C 8. Conclusion
Cod = ∑ (1 +o k )i
i=1 (23)
In Eq. (23), Co , ny and k are annual operating cost that it is defined
as Eq. (24), equipment life (years) that a period of 10 years is most often
taken [58–60] and the annual discount rate (%) that it is taken equal
10%, respectively.
Co = PK elctric H (24)
where K elctric and H arethe price of electrical energy, EUR/(W·h) that it
is defined 0.00012 EUR/(W·h) [61] and the hours of operation per year,
that it is defined equal to 330 days or 7920 h/year [62]. Here, P is
pumping power (W) and defined as can be seen in Eq. (25):
1 ⎛ ṁ tube ṁ
P= ⎜ Δp + shell Δpshell⎞⎟
η ⎝ ρtube tube ρshell ⎠ (25)
where η is the pump efficiency (here is defined equal to 0.7), ṁ tube ,
ṁ shell , ρtube and ρshell are the tube side flow rate (kg/s), the shell side flow
rate (kg/s), the tube side fluid density and the shell side fluid density
(kg/m3), respectively. It should be noted that the properties are read for
average fluid temperature. Also Δptube and Δpshell are the pressure drop
for tube side and shell side (Pa).
According to above mentioned discussion, the total cost of studied
shell and tube heat exchanger is summarized in Table 8.
It must be noted that in Table 8, all data's normalized by data of
smooth tube. By dividing the normalized TPF data’s into the data’s in
Table 8, the data’s for Table 9, is obtained.
By comparing the data’s in this table, it can be seen, most cases have
values higher than the unit. As a result, it can be dedicated that the
corrugating process in the tubes and using them in tube bundle in-
creases the costs, but from economic analysis point of view, it is com-
pletely economic.
A novel design of tube structure with cosine wave is proposed to

96
K. Milani Shirvan et al.
study the effects of wavy surface characteristics on the thermal per-
formance factor in a shell and tube heat exchanger experimentally. In
this paper the focus was on the optimization of the thermal perfor-
mance factor. For this aim, the Response Surface Methodology in dif-
ferent levels of effective parameters was used. The following findings
can be summarized from the results obtained:
– Considering the wavy wavelength effects on the thermal perfor-
mance factor, the maximum of the thermal performance factor is
observed in the vicinity of λ = 40 mm, and its minimum amount at
λ = 80 mm.
– The thermal performance factor decreases with increasing the wavy
starting length and hot water flow rate.
– By increasing of the cold water flow rate, the thermal performance
factor increases.
– In all cases, the thermal performance factor is larger than unity. This
fact is important form energy saving point of view.
– The aim of this paper is to find optimum conditions to maximize the
thermal performance factor. The goal is found at λ = 33.13 mm,
b = 0 mm, Q h = 7 l/min and Qc = 19 l/min for the wavy wave-
length, the wavy starting length, the hot water flow rates, and the
cold water flow rates, respectively.
– It can be dedicated that the corrugating process in the tubes and
using them in tube bundle increases the costs, but from economic
analysis point of view, it is completely economic.
– The results of this paper are of great importance for researchers to
obtain the optimal design conditions in heat exchangers.
Acknowledgement
This research is done under the direct support of the Islamic
Republic of Iran's oil ministry, National Iranian Gas Company, South
Pars Gas Complex with facilities and resources. The authors gratefully
acknowledge them.
References
[1] Mamourian M, Shirvan K Milani, Ellahi R, Rahimi AB. Optimization of mixed
convection heat transfer with entropy generation in a wavy surface square lid-
driven cavity by means of Taguchi approach. Int J Heat Mass Transf
2016;102:544–54.
[2] Asadi M, Song Y, Sunden B, Xie G. Economic optimization design of shell and tube
heat exchangers by cuckoo-search-algorithm. Appl Therm Eng 2014;73:1032–40.
[3] Fesanghary M, Damangir E, Soleimani L. Design optimization of Shell and tube heat
exchangers using global sensitivity analysis and harmony search algorithm. Appl
Therm Eng 2009;29:1026–31.
[4] Yang Jie, Liu Wei. Numerical investigation on a novel shell-and-tube heat exchanger
with plate baffles and experimental validation. Energy Convers Manage
2015;101:689–96.
[5] Maakoul AE, Laknizi A, Saadeddine S, Metoui ME, Zaite A, Meziane M, et al.
Numerical comparison of shell-side performance for shell and tube heat exchangers
with trefoil-hole, helical and segmental baffles. Appl Therm Eng 2016;109(Part
A):175–85.
[6] You Y, Fan A, Lai X, Huang S, Liu W. Experimental and numerical investigations of
shell-side thermo-hydraulic performances for shell-and-tube heat exchanger with
trefoil-hole baffles. Appl Therm Eng 2013;50(1):950–6.
[7] Wang S, Wen J, Yang H, Xue Y, Tuo H. Experimental investigation on heat transfer
enhancement of a heat exchanger with helical baffles through blockage of triangle
leakage zones. Appl Therm Eng 2014;67(1–2):122–30.
[8] Ozden E, Tari I. Shell side CFD analysis of a small shell-and-tube heat exchanger.
Energy Convers Manage 2010;51(5):1004–14.
[9] Lei Y, Li Y, Jing Sh, Song Ch, Lyu Y, Wang F. Design and performance analysis of the
novel shell-and-tube heat exchangers with louver baffles. Appl Therm Eng
2017;125:870–9.
[10] Wen J, Yang H, Wang S, Xu Sh, Xue Y, Tuo H. Numerical investigation on baffle
configuration improvement of the heat exchanger with helical baffles. Energy
Convers Manage 2015;89:438–48.
[11] Yang JF, Lin YSh, Ke HB, Zeng M, Wang QW. Investigation on combined multiple
shell-pass shell-and-tube heat exchanger with continuous helical baffles. Energy
2016;115(Part 3):1572–9.
[12] Wang X, Zheng N, Liu Zh, Liu W. Numerical analysis and optimization study on
shell-side performances of a shell and tube heat exchanger with staggered baffles.
Int J Heat Mass Transf 2018;124:247–59.
[13] Yang Sh, Chen Y, Wu J, Gu H. Performance simulation on unilateral ladder type
97
Energy Conversion and Management 175 (2018) 86–98
helical baffle heat exchanger in half cylindrical space. Energy Convers Manage
2017;150:134–47.
[14] Chen YP, Wang W, Wu JF, Dong C. Experimental investigation on performances of
trisection helical baffled heat exchangers for oil/water–water heat transfer. Energy
Convers Manage 2015;101:460–9.
[15] Sadeghzadeh H, Ehyaei MA, Rosen MA. Techno-economic optimization of a shell
and tube heat exchanger by genetic and particle swarm algorithms. Energy Convers
Manage 2015;93:84–91.
[16] Yang J, Fan A, Liu W, Jacobi AM. Optimization of shell-and-tube heat exchangers
conforming to TEMA standards with designs motivated by constructal theory.
Energy Convers Manage 2014;78:468–76.
[17] Mirzaei M, Hajabdollahi H, Fadakar H. Multi-objective optimization of shell-and-
tube heat exchanger by constructal theory. Appl Therm Eng 2017;125:9–19.
[18] Hadidi A, Hadidi M, Nazari A. A new design approach for shell-and-tube heat ex-
changers using imperialist competitive algorithm (ICA) from economic point of
view. Energy Convers Manage 2013;67:66–74.
[19] DeVasconcelos Segundoa EH, Amoroso AL, Mariania VC, dos Santos Coelho L.
Economic optimization design for shell-and-tube heat exchangers by a Tsallis dif-
ferential evolution. Appl Therm Eng 2017;111:143–51.
[20] Yin Q, Du WJ, Ji XL, Cheng L. Optimization design and economic analyses of heat
recovery exchangers on rotary kilns. Appl Energy 2016;180:743–56.
[21] Hatami M, Ganji DD, Gorji-Bandpy M. Experimental and numerical analysis of the
optimized finned-tube heat exchanger for OM314 diesel exhaust exergy recovery.
Energy Convers Manage 2015;97:26–41.
[22] Labbadlia O, Laribi B, Chetti B, Hendrick P. Numerical study of the influence of tube
arrangement on the flow distribution in the header of shell and tube heat ex-
changers. Appl Therm Eng 2017;126:315–21.
[23] Ayub ZH, Yang D, Khan TS, Al-Hajri E, Ayub AH. Performance characteristics of a
novel shell and tube heat exchanger with shell side interstitial twisted tapes for
viscous fluids application. Appl Therm Eng 2018;134:248–55.
[24] El-Said EMS, Abou Alsood MM. Experimental investigation of air injection effect on
the performance of horizontal shell and multi-tube heat exchanger with baffles.
Appl Therm Eng 2018;134:238–47.
[25] Sadighi Dizaji H, Jafarmadar S, Hashemian M. The effect of flow, thermodynamic
and geometrical characteristics on exergy loss in shell and coiled tube heat ex-
changers. Energy 2015;91:678–84.
[26] Sadighi Dizaji H, Jafarmadar S, Asaadi S. Experimental exergy analysis for shell and
tube heat exchanger made of corrugated shell and corrugated tube. Exp Therm
Fluid Sci 2017;81:475–81.
[27] Sepehr M, Hashemi SS, Rahjoo M, Farhangmehr V, Alimoradi A. Prediction of heat
transfer, pressure drop and entropy generation in shell and helically coiled finned
tube heat exchangers. Chem Eng Res Des 2018;134:277–91.
[28] TEMA. Standards of the tubular exchanger manufacturers association. ninth ed.
New York: Tubular Exchanger Manufacturers Association; 2007.
[29] Shah RK, Sekulic DP. Fundamentals of heat exchanger design. Hoboken, USA: John
Wiley & Sons Inc.; 2003.
[30] Heat Heat Exchanger Rating (Bell-Delaware Method), < https://cheguide.com/
2017/03/heat-exchanger-rating-bell-delaware-method/ > [accessed 21 April
2018].
[31] Field RW. A theoretical viscosity correction factor for heat transfer and friction in
pipe flow. Chem Eng Sci 1990;45(5):1343–7.
[32] Incropera FP, Dewitt DP. Fundamentals of heat and mass transfer. fourth ed. New
York: Wiley; 1996.
[33] Kline SJ, McClintock FA. Describing uncertainties in single-sample experiments.
Mech Eng 1953;75(1):3–8.
[34] Moffat Robert J. Describing the uncertainties in experimental results. Exp Therm
Fluid Sci 1988;1(1):3–17.
[35] Han HZ, Li BX, Wu H, Shao W. Multi-objective shape optimization of double pipe
heat exchanger with inner corrugated tube using RSM method. Int J Therm Sci
2015;90:173–86.
[36] Milani Shirvan K, Ellahi R, Mirzakhanlari S, Mamourian M. Enhancement of heat
transfer and heat exchanger effectiveness in a double pipe heat exchanger filled
with porous media: numerical simulation and sensitivity analysis of turbulent fluid
flow. Appl Therm Eng 2016;109:761–74.
[37] Milani Shirvan K, Mamourian M, Mirzakhanlari S, Ellahi R. Two phase simulation
and sensitivity analysis of effective parameters on combined heat transfer and
pressure drop in a solar heat exchanger filled with nanofluid by RSM. J Mol Liq
2016;220:888–901.
[38] Mamourian M, Milani Shirvan K, Mirzakhanlari S. Two phase simulation and sen-
sitivity analysis of effective parameters on turbulent combined heat transfer and
pressure drop in a solar heat exchanger filled with nanofluid by response surface
methodology. Energy 2016;109:49–61.
[39] Akbarzadeh M, Rashidi S, Bovand M, Ellahi R. A sensitivity analysis on thermal and
pumping power for the flow of nanofluid inside a wavy channel. J Mol Liq
2016;220:1–13.
[40] Rashidi S, Bovand M, Esfahani JA. Heat transfer enhancement and pressure drop
penalty in porous solar heat exchangers: a sensitivity analysis. Energy Convers
Manage 2015;103:726–38.
[41] Rashidi S, Bovand M, Esfahani JA. Sensitivity analysis for entropy generation in
porous solar heat exchangers by RSM. J Thermophys Heat Transf
2016;31(2):390–402.
[42] Mamourian M, Milani Shirvan K, Mirzakhanlari S, Rahimi AB. Vortex generators
position effect on heat transfer and nanofluid homogeneity: a numerical in-
vestigation and sensitivity analysis. Appl Therm Eng 2016;107:1233–47.
[43] Montgomery DC, Myers RH. Response surface methodology: process and product
optimization using designed experiments. A Wiley-Interscience Publications; 1995.
K. Milani Shirvan et al.
[Raymond H. Meyers and Douglas C. Montgomery].
[44] Wang S, Xiao J, Wang J, Jian G, Wen J, Zhang Z. Configuration optimization of
shell-and-tube heat exchangers with helical baffles using multi-objective genetic
algorithm based on fluid-structure interaction. Int Commun Heat Mass Transf
2017;85:62–9.
[45] Mirzakhanlari S, Milani Shirvan K, Mamourian M, Chamkha AJ. Increment of mixed
convection heat transfer and decrement of drag coefficient in a lid-driven nanofluid-
filled cavity with a conductive rotating circular cylinder at different horizontal
locations: a sensitivity analysis. Powder Technol 2017;305:495–508.
[46] Bezerra MA, Santelli RE, Padua Oliveira E, Villar LS, Escaleira LAl. Response surface
methodology (RSM) as a tool for optimization in analytical chemistry. Talanta
2008;76:965–77.
[47] Milani Shirvan K, Mamourian M, Mirzakhanlari S, Ellahi R, Vafai K. Numerical
investigation and sensitivity analysis of effective parameters on combined heat
transfer performance in a porous solar cavity receiver by response surface metho-
dology. Int J Heat Mass Transf 2017;105:811–25.
[48] Milani Shirvan K, Mamourian M, Mirzakhanlari S, Öztop HF, Abu-Hamdeh N.
Numerical simulation and sensitivity analysis of effective parameters on heat
transfer and homogeneity of Al2O3 nanofluid in a channel using DPM and RSM. Adv
Powder Technol 2016;27(5):1980–91.
[49] Noshadi I, Amin NAS, Parnas RS. Continuous production of biodiesel from waste
cooking oil in a reactive distillation column catalyzed by solid heteropolyacid:
optimization using response surface methodology (RSM). Fuel 2012;94:156–64.
[50] Milani Shirvan K, Mirzakhanlari S, Mamourian M, Abu-Hamdeh N. Numerical in-
vestigation and sensitivity analysis of effective parameters to obtain potential
maximum power output: a case study on Zanjan prototype solar chimney power
plant. Energy Convers Manage 2017;136:350–60.
[51] Mamourian M, Milani Shirvan K, Pop I. Sensitivity analysis for MHD effects and
inclination angles on natural convection heat transfer and entropy generation of
Al2O3-water nanofluid in square cavity by response surface methodology. Int
Commun Heat Mass Transf 2016;2016(79):46–57.
98
Energy Conversion and Management 175 (2018) 86–98
[52] Milani Shirvan K, Mirzakhanlari S, Chamkha AJ, Mamourian M. Numerical simu-
lation and sensitivity analysis of effective parameters on natural convection and
entropy generation in a wavy surface cavity filled with a nanofluid using RSM.
Numerical Heat Transf Part A: Appl 2016;70(10):1157–77.
[53] Milani Shirvan K, Mirzakhanlari S, Mamourian M, Kalogirou SA. Optimization of
effective parameters on solar updraft tower to achieve potential maximum power
output: A sensitivity analysis and numerical simulation. Appl Energy
2017;195:725–37.
[54] Yin J, Yang G, Li Y. The effects of wavy plate phase shift on flow and heat transfer
characteristics in corrugated channel. Energy Procedia 2012;14:1566–73.
[55] Šumić Z, Vakula A, Tepić A, Čakarević J, Vitas J, Pavlić B. Modeling and optimi-
zation of red currants vacuum drying process by response surface methodology
(RSM). Food Chem. 2016;203:465–75.
[56] Murphy TE, Tsui KL, Allen JK. A review of robust design methods for multiple
responses. Res. Eng. Des. 2005;16:118.
[57] Wildi-Tremblay P, Gosselin L. Minimizing shell-and-tube heat exchanger cost with
genetic algorithms and considering maintenance. Int J Energy Res 2007;31:867–85.
[58] Sanaye S, Hajabdollahi H. Multi-objective optimization of shell and tube heat ex-
changers. Appl Therm Eng 2010;30:1937–45.
[59] Marenjak S, Krstic H. Sensitivity analysis facilities life cycle costs. Tech Gazette
2010;17(4):481–7.
[60] Caputo CA, Pelagagge MP, Salini P. Heat exchanger design based on economic
ptimization. Appl Therm Eng 2008;28:1151–9.
[61] Hochsteiner de Vasconcelos Segundo E, Levati Amoroso A, Cocco Mariani V, dos
Santos Coelho L. Economic optimization design for shell-and-tube heat exchangers
by a Tsallis differential evolution. Appl Therm Eng 2016. https://doi.org/10.1016/
j.applthermaleng.2016.09.032.
[62] Slavković G, Budimir SJ, Rakonjac IM, Jarić MS, Budimir NJ. Techno-economic
analysisof heat exchangers with parallel helical tube coils. Tehnički vjesnik
2014;21(4):861–6.
 Update
Energy Conversion and Management
Volume 205, Issue , 1 February 2020, Page

DOI: https://doi.org/10.1016/j.enconman.2019.112399
 Energy Conversion and Management 205 (2020) 112399

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Corrigendum

Corrigendum to “Experimental investigation on thermal performance and T

economic analysis of cosine wave tube structure in a shell and tube heat
exchanger” [J. Energy Convers. Manage. 175 (2018) 86–98]
⁎
K. Milani Shirvan, M. Mamourian , J. Abolfazli Esfahani
Department of Mechanical Engineering, Ferdowsi University of Mashhad, P.O.B. 91775-1111, Mashhad, Iran

Unfortunately, we made error in Eq. (16) in our original paper [1]. confusion that it has caused.
The uncertainty can be calculated in Eq. (16) as shown below:
− References
n
∑1 (Nu − Nu)2
uNu, Rep =
n (n − 1) (16) [1] Milani Shirvan K, Mamourian M, Abolfazli Esfahani J. J Energy Convers Manage
−
2018;175:86–98.
here, Nu is the average of the Nu Number.
The authors would like to apologize for this error and for any

DOI of original article: https://doi.org/10.1016/j.enconman.2018.08.103

⁎
Corresponding author.
E-mail address: mamourian@um.ac.ir (M. Mamourian).

https://doi.org/10.1016/j.enconman.2019.112399
Received 9 December 2019; Accepted 10 December 2019
Available online 30 December 2019
0196-8904/ © 2019 Elsevier Ltd. All rights reserved.