Plate heat exchanger design software for industrial and educational

applications
Nikola R. Zlatković, Divna M. Majstorović, Mirjana Lj. Kijevčanin, Emila M. Živković
Faculty of Technology and Metallurgy, University of Belgrade, Belgrade, Serbia

Abstract
Plate heat exchanger is a type of heat exchanger that uses corrugated metal plates to TECHNICAL PAPER
transfer heat between two fluids. The plate corrugations are designed to achieve turbul-
ence across the entire heat transfer area thus producing the highest possible heat transfer UDC 004.4:620.1:66:536.2
coefficients while allowing close temperature approaches. Subsequently, this leads to a
smaller heat transfer area, smaller units and in some cases, fewer heat exchangers. In this
work, an application for thermal and hydraulic computations of plate heat exchangers had Hem. Ind. 71 (5) 439–449 (2017)
been developed using Sharp Develop, an open source programming platform. During the
development process, several literature methods and correlations for calculation of heat
transfer coefficient and pressure drop in a plate heat exchanger have been tested and the
selected four methods: Martin, VDI, Kumar and Coulson and Richardson have been
incorporated into the software. The structure of the software is visually presented through
several windows: a window for inserting input data, windows for showing the results of
computation by each of the methods, a window for showing comparative analysis of the
most important computation results obtained by all of the used methods and a help win-
dow for demonstrating the working principle of plate heat exchanger.
Keywords: plate heat exchanger, heat transfer, pressure drop, software.
Available online at the Journal website: http://www.ache.org.rs/HI/

The transfer of heat to and from process fluids is an and outlet ports, and seals designed so as to direct the
essential part of most chemical processes. Heat fluids in alternate flow passages. The flow passages are
exchanger is a heat transfer device that transfers heat formed by adjacent plates so that the two streams
between two or more process fluids. One of the most exchange heat while passing through alternate chan-
commonly used types of heat exchangers is a plate nels. When a package of plates is assembled, the holes
heat exchanger (PHE). at the corners form continuous tunnels leading the
Plate heat exchangers are widely used in dairy and fluids from the inlet into the plate package, where they
food processing plants, chemical industries, power are distributed into narrow passages between the
plants and central cooling systems. They exhibit excel- plates, and then collecting them before the outlet.
lent heat transfer characteristics, which allows a very Fluids are separated by a thin metal wall – corrugated
compact design, and can easily be disassembled for plates. The shape of corrugations is a characteristic of
maintenance, cleaning or for modifying the heat trans- each plate model and is carefully studied by the manu-
fer area by adding or removing plates. Also, low tempe- facturers. The purpose of corrugations is to provide
rature approaches can be used, as low as 1 °C, com- turbulence in order to increase the heat transfer coef-
pared to 5 to 10 °C for shell and tube exchangers (SHE). ficients and, at the same time, to increase the struc-
The mean temperature correction factor, F, will norm- tural strength of the assembly. The most common cor-
ally be higher for PHE than for SHE, as the flow is closer rugation pattern is the herringbone (chevron) pattern,
to true counter-current flow. Fouling tends to be sig- also studied in this work. Selection of the plates for any
nificantly lower in plate heat exchangers compared to application depends on the process requirements in
shell and tube heat exchangers due to the higher velo- terms of heat transfer coefficients and allowable heat
cities and the absence of dead angles. exchanger pressure drop. Patterns that provide higher
Typically, a plate heat exchanger consists of a stack heat transfer coefficients for a given flow rate also
of corrugated or embossed metal plates in mutual con- produce higher pressure drops. Thus, for any applic-
tact, each plate having four openings serving as inlet ation, the designer must choose the plate type offering
the best balance of the two effects.
The aim of this work was to develop software that
Correspondence: D.M. Majstorović, Faculty of Technology and Metal-
lurgy, University of Belgrade, Karnegijeva 4, 11120 Belgrade, Serbia.
could be used for thermal and hydraulic calculations of
E-mail: dbajic@tmf.bg.ac.rs plate heat exchangers. Due to the great variety of plate
Paper received: 21 October, 2016 dimensions and especially corrugation patterns, every
Paper accepted: 28 April, 2017 attempt to establish a general calculation method that
https://doi.org/10.2298/HEMIND161021007Z

439
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

can be applied to different types of plate heat cribed in our previous paper [2]. Experimental data
exchangers has been unsuccessful so far. Very often have been correlated with the equation:
correlations established by the producers of plate heat
exchangers are used for estimating heat transfer coef- Nu = 0.39515Re0.6244 Pr1/3 (1)
ficients and pressure drops in certain type of plates. in which Nu, Re and Pr are Nusselt, Reynolds and
However, several more general methods applicable to Prandtl numbers defined by the equations:
the herringbone plate type have been suggested in
literature. Some of these methods have been tested α De
Nu = (2)
and the results for heat transfer coefficient and pres- λ
sure drop have been compared with the values obtained
in the experimental setup already described in our ρwDe
Re = (3)
previous papers [1–3]. Finally, four methods: Martin μ
[4], VDI [5], Kumar [6] and Coulson and Richardson [7],
μcp
have been selected and, in this work, incorporated into Pr = (4)
an application called PHeatEx Designer 1.0, for thermal λ
and hydraulic design of plate heat exchangers with where w is the fluid velocity, De is the equivalent
herringbone type of plates, developed using the Sharp diameter for the fluid flow channel and ρ, λ, cp and μ
Develop 5.0 Beta 4, the open source development are fluid density, thermal conductivity, specific heat
environment. After its validation the application was and dynamic viscosity, respectively.
used for rating of two industrial heat exchangers. From the Eq. (1) and the definition of Nusselt num-
With the development of this software an attempt ber (Eq. (2)) the values of heat transfer coefficient
was made to collect in one place some of the available could be easily obtained:
literature methods for herringbone type of plates, to
compare the results obtained by using different calcul- Nu λ
α= (5)
ation methods (i.e., values of heat transfer coefficients, De
required area for heat transfer or pressure drops) and,
if possible, to give some recommendation for their use. Experimental values of pressure drop, determined
as the difference between the measured inlet and
CALCULATION PROCEDURE outlet pressures for the single phase flow [2], have also
been compared with results obtained by using lite-
In the course of investigations several literature rature methods [4–8]. The experimental measurements
methods for calculation of heat transfer coefficient in for single phase flow were conducted with the aqueous
plate heat exchangers with a herringbone type of cor- 30.6 mass% ethylene glycol solution. The same fluid,
rugation have been tested: Martin [4], VDI [5], Kumar the same process parameters and the same geometry
[6], Coulson and Richardson [7] and generalized Muley of the plates [1–3] were used for literature correlations
and Manglik [8]. All the equations have been incor- during the selection process. Thermophysical proper-
porated in an Excel sheet and the results have been ties of ethylene glycol and water binary mixture were
compared with the values calculated from the experi- taken from literature sources [9,10]. The obtained
mental correlation obtained from a procedure des- results for several experimental cases are given in Table
1, in the Results and Discussion section of this work.

Table 1. Process parameters for the experimentally investigated single-phase flow of ethylene glycol and water binary mixture

Coulson and Muley and

Parameter Kumar Martin VDI Exp.
Richardson Manglik
Mass flow rate Outlet tempe- Thermal con-
Mass flow rate Inlet tempe- Density Viscosity Specific heat
per channel rature ductivity
m / kg s–1 rature, ti / °C ρ / kg m
–3
µ / mPa s cp / kJ kg–1 K–1
mch / kg s–1 to / °C λ / W m–1 K–1
2.24 0.448 19.58 14.29 1040.9 2.429 3.790 0.4730
Heat exchanger duty, Q / kW 44.91
Velocity, w / m s–1 0.277
Reynolds number 654.3 654.3 654.3 759.0 654.3 654.3
Heat transfer coefficient 5847.9 4883.6 7042.2 4693.2 1372.2 5220.9
–2 –1
α/Wm K
Friction pressure drop, Δpk / Pa 30795 17234 17234 3566 – –
Total pressure drop, Δptot / Pa 39750 26189 26189 12521 – 43700

440
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

Table 1. Continued

Mass flow rate Outlet tempe- Thermal con-

Mass flow rate, Inlet tempe- Density Viscosity Specific heat
–1 per channel, rature –3 ductivity
m / kg s rature, ti / °C ρ / kg m µ / mPa s cp / kJ kg–1 K–1
mch / kg s–1 to / °C λ / W m–1 K–1
1.015 0.203 16.02 6.583 1045.0 2.828 3.780 0.4668
Heat exchanger duty, Q / kW 36.21
Velocity, w / m s–1 0.125
Reynolds number 254.7 254.7 254.7 295.4 254.7 254.7
Heat transfer coefficient, 3257.7 2786.0 4336.9 2677.0 498.1 3017.9
–2 –1
α/Wm K
Friction pressure drop, Δpk / Pa 7174 4804 4804 967 – –
Total pressure drop, Δptot / Pa 16123 13753 13753 9916 – 18100
Mass flow rate Outlet tempe- Thermal con-
Mass flow rate, Inlet tempe- Density Viscosity Specific heat
per channel, rature ductivity
m / kg s–1 –1 rature, ti / °C ρ / kg m–3 µ / mPa s cp / kJ kg–1 K–1 –1 –1
mch / kg s to / °C λ/Wm K
1.557 0.316 16.86 9.95 1043.5 2.669 3.784 0.4692
Heat exchanger duty, Q / kW 41.24
Velocity, w / m s–1 0.194
Reynolds number 419.4 419.4 419.4 486.5 419.4 419.4
Heat transfer coefficient, 4465.2 3713.2 5593.0 3692.8 855.9 4056.8
α / W m–2 K–1
Friction pressure drop, Δpk / Pa 16199 9632.3 9632.3 2015 – –
Total pressure drop, Δptot / Pa 25151 18538 18538 10966 – 28300
Mass flow rate Outlet tempe- Thermal con-
Mass flow rate, Inlet tempe- Density Viscosity Specific heat
per channel, rature ductivity
m / kg s–1 –3
–1 rature, ti / °C ρ / kg m µ / mPa s cp / kJ kg–1 K–1 –1 –1
mch / kg s to / °C λ/Wm K
1.190 0.238 17.34 8.69 1043.8 2.698 3.783 0.4687
Heat exchanger duty, Q / kW 38.94
Velocity, w / m s–1 0.147
Reynolds number 312.9 312.9 312.9 363.0 312.9 312.9
Heat transfer coefficient, 3688.1 3109.4 4776.2 3012.0 425.9 3389.1
α / W m–2 K–1
Friction pressure drop, Δpk / Pa 9595 6075 6075 1252 – –
Total pressure drop, Δptot / Pa 18539 15018 15018 10195 – 21800
Mass flow rate Outlet tempe- Thermal con-
Mass flow rate, Inlet tempe- Density Viscosity Specific heat
per channel, rature ductivity
m / kg s–1 rature, ti / °C ρ / kg m
–3
µ / mPa s cp / kJ kg–1 K–1
mch / kg s–1 to / °C λ / W m–1 K–1
1.730 0.346 19.59 12.98 1041.4 2.468 3.789 0.4723
Heat exchanger duty, Q / kW 43.38
Velocity, w / m s–1 0.214
Reynolds number 497.9 497.9 497.9 577.6 497.9 497.9
Heat transfer coefficient, 4899.7 4054.1 6034.4 3951.0 1016.3 4421.0
–2 –1
α/Wm K
Friction pressure drop, Δpk / Pa 19110 11048 11048 2312 – –
Total pressure drop, Δptot / Pa 28049 19987 19987 11251 – 30900
Mass flow rate Outlet tempe- Thermal con-
Mass flow rate, Inlet tempe- Density Viscosity Specific heat
per channel, rature ductivity
m / kg s–1 –1 rature, ti / °C ρ / kg m–3 µ / mPa s cp / kJ kg–1 K–1 –1 –1
mch / kg s to / °C λ/Wm K
1.360 0.272 17.96 10.02 1043.1 2.626 3.785 0.4698
Heat exchanger duty, Q / kW 40.92
Velocity, w / m s–1 0.168
Reynolds number 367.9 367.9 367.9 426.7 367.9 367.9

441
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

Table 1. Continued

Mass flow rate Outlet tempe- Thermal con-

Mass flow rate, Inlet tempe- Density Viscosity Specific heat
–1 per channel, rature –3 ductivity
m / kg s rature, ti / °C ρ / kg m µ / mPa s cp / kJ kg–1 K–1
mch / kg s–1 to / °C λ / W m–1 K–1
1.360 0.272 17.96 10.02 1043.1 2.626 3.785 0.4698
Heat transfer coefficient, 4075.8 3407.0 5177.5 3314.0 736.7 3721.7
α / W m–2 K–1
Friction pressure drop, Δpk / Pa 12292 7499 7499 1562 – –
Total pressure drop, Δptot / Pa 21234 16441 16441 10504 – 22800
Mass flow rate Outlet tempe- Thermal con-
Mass flow rate, Inlet tempe- Density Viscosity Specific heat
per channel, rature ductivity
m / kg s–1 –1 rature, ti / °C ρ / kg m–3 µ / mPa s cp / kJ kg–1 K–1 –1 –1
mch / kg s to / °C λ/Wm K
1.910 0.382 17.51 11.74 1042.5 2.576 3.786 0.4705
Heat exchanger duty, Q / kW 41.72
Velocity, w / m s–1 0.236
Reynolds number 526.1 526.1 526.1 610.3 526.1 526.1
Heat transfer coefficient, 5139.7 4244.7 6293.8 4151.2 1091.8 4628.4
α / W m–2 K–1
Friction pressure drop, Δpk / Pa 23039 13217 13217 2762 – –
Total pressure drop, Δptot / Pa 31994 22172 22172 11717 – 35900
PD (α) / % 10.03 7.99 38.43 10.70 79.60 –
PD (Δptot) / % 10.43 33.02 33.02 59.42 – –

Based on the obtained results, four literature Eq. (8), the effective plate length Lm, if known, could be
methods were selected and incorporated in the soft- used instead of the plate length Lp.
ware for thermal and hydraulic design of plate heat Required area for heat transfer Ap can be calculated
exchangers. These methods are: Martin [4], VDI [5], from the equation:
Kumar [6] and Coulson and Richardson [7]. The sel-
Q
ection of methods was mostly based on their general Ap = (9)
applicability and, in some cases, on good agreements KF Δtln
with values for heat transfer coefficient obtained from
where K is the overall heat transfer coefficient, Δtln is
the experimental correlation (1) and Eq. (2), reported in
the logarithmic mean temperature difference and F is
Table 1. All of the selected methods consist of three
the correction factor for Δtln . For plate heat exchangers
phases of calculation: 1) calculation of basic geomet-
the correction factor is often close to 1.0 since the flow
rical characteristics of the exchanger and other general
configuration is closer to true counter-current flow
characteristics of the process, 2) heat transfer calcul-
than in the shell and tube heat exchangers. Overall
ation and 3) pressure drop calculation.
heat transfer coefficient K (W m–2 K–1) is determined
Heat exchanger duty is determined from one of the
from the equation:
following equations:
1 1 1 δpl
Q = mh cp ,h ( thi − tho ) (6) = + + + Rp,h + Rp,c (10)
K αh α c λpl
Q = mc cp ,c ( tco − tci ) (7)
where ߙh and ߙc are heat transfer coefficients for hot
where m is the mass fluid flow rate, cp is the fluid and cold fluid, respectively, Rp,h and Rp,c are corres-
specific heat, and ti and to are inlet (i) and outlet (o) ponding fouling factors, ߜpl is the plate thickness and ߣpl
temperatures, while subscripts h and c designate hot is the thermal conductivity of the plates.
and cold fluid, respectively. Reserve in heat transfer area ΔA (%) is calculated
Available area for heat transfer, As, is defined by the from the equation:
equation: As − Ap
ΔA = 100 (11)
(
As = LpW Npl − 1 ) (8) As

in which Lp is the plate length, W is the plate width and Cross sectional area of the channel S is defined by
Npl is the number of plates in the heat exchanger. In the the equation:

442
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

S = BW (12) operating conditions accounted for by empirical coef-

ficients and constants. More details on calculation pro-
where B is the average distance between adjacent
cedures could be found elsewhere [4,5]. The charac-
plates determined as:
teristic length L (m) could be determined as a function
B = 2a (13) of the corrugation angle (αpl) and wave length (Λ), from
the equation:
in Martin [4] and VDI [5] correlations. In Kumar [6] and
Coulson and Richardson [7] methods parameter B is Λ
L= (21)
calculated from the expression: sin(2αpl )
B = pp − δpl (14)
The heat transfer coefficient in Kumar method [6] is
in which pp (m) is the plate profile pitch and δpl (m) is calculated using the expression:
the plate thickness. 0.17
 μ 
Equivalent diameter De for the fluid flow channel in jhλ Pr 0.33  
Coulson and Richardson [7] method is determined as:
α=  μw  (22)
De
De = 2B (15)
Heat transfer factor (jh) is determined as a function
In Martin [4], VDI [5] and Kumar[6] methods the
of Reynolds number (Re) and plate profile angle
following equation for equivalent diameter is used:
( β = 90 − αpl ) according to the described procedure
2B [6].
De = (16)
μp Coulson and Richardson [7] method uses the fol-
lowing expression for calculation of the heat transfer
where μp is the plate enhancement factor due to cor- coefficient:
rugation. If not given by the plate manufacturer, it
α De
= 0.26Re0.65 Pr 0.4 ( μ / μw )
0.14
could be calculated as a function of average distance (23)
between adjacent plates (B) and wave length of the λ
plate (Λ) according to the described procedure [4,5]: The total fluid pressure drop Δptot is the sum of

( )
1 three values: a) pressure drop in exchanger channels
μp = 1 + 1 + X2 + 4 1 + X2 / 2 (17) (friction pressure drop), b) pressure drop in nozzles,
6
and c) pressure drop due to elevation. The pressure
where X is calculated as: drop in exchanger channels according to Martin [4] and
VDI [5] methods is calculated:
Bπ
X= (18)
Λ ξρw 2Lp
Δpk = N (24)
Heat transfer coefficient in Martin [4] and VDI [5] 2De
methods is determined from the Eq. (5). The difference
in which N represents the number of fluid passes
between these two methods is in the equations used
through the exchanger while ξ is the fluid friction
for calculating the Nusselt number. In Martin [4]
coefficient, already mentioned in the Eqs. (19) and (20).
method, the following equation is used:
In the Kumar method [6], friction coefficient ξ is rep-
( )
Nu = 0.122Pr1/3 (μ / μw )1/6 ζ sin(2αpl Re2 )0.374 (19) laced by the Fanning – type friction factor f, and a cor-
rection factor due to the wall temperature influence is
while in VDI correlation [5], the Nusselt number is cal- taken into account:
culated from the equation:
4 f ρw 2Lp
Δpk = N (25)
Nu = 1.615[(ξ Re/ 64)RePr De / L] 1/3
(20)  μ 
0.17
2De  
In both equations Re and Pr are Reynolds and  μe 
Prandtl numbers of the fluid, μ (Pa s) and μw (Pa s) are
viscosities of the fluid at the mean fluid temperature The friction factor, similar to the heat transfer
and wall temperature, respectively, while ξ represents factor (jh), is determined as a function of the Reynolds
the fluid friction coefficient. The friction coefficient ξ number (Re) and plate profile angle ( β = 90 − αpl )
depends on three friction components (ξ0, ξ1,0, ξ1) and according to the procedure described in [6].
plate corrugation angle αpl. Friction components (ξ0, According to Coulson and Richardson [7], pressure
ξ1,0, ξ1) are functions of the Reynolds number and drop in exchanger channels is calculated:

443
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

8 jf ρw 2Lp
Δpk = N (26)
2De

while the friction factor jf can be determined from the

expression: (a)
−0.3
jf = 0.6Re (27)

Pressure drop in nozzles, in all four selected

methods, is calculated from the equation:

1.3wp2 ρ
Δpp = N (28)
2
where wp represents the fluid velocity in nozzles.
Pressure drop due to elevation has to be taken into
account in the case of uneven number of fluid passes,
and can be determined from the equation:
Δpel = ρ gLp (29) (b)

Figure 1. Application shortcut (a) and login form window (b).

Software description
The software PHeatEx Designer 1.0 for thermal and After choosing one of the calculation methods from
hydraulic calculation of plate heat exchangers has been Methods of Calculation drop-down menu, a new form
developed in Sharp Develop 5.0 Beta 4, the open which depends on the chosen method will open. If any
source development environment. The software is acti- of the input parameters necessary for calculation are
vated by clicking the shortcut shown in Figure 1a. After missing, a message, in a form of Message Box as shown
running the application, the Login Form window, in Figure 4, will appear.
shown in Figure 1b, will appear. When one of the methods of calculation is chosen,
After entering the correct password Application a form which describes that method will open as can be
Loading message will appear and the main window, seen in Figure 5 for Kumar method. In the upper part of
shown in Figure 2, will open. the window four commands are offered: Back – for
From the menu in the upper part of the window, returning to the main window, Start Calculating, Save
the following sections could be chosen: Insert Input Results–for saving the calculation results in a form of a
Data, Methods of Calculation, Working Principle and Notepad Document, Microsoft Word Document or
Analysis of Results. After choosing Insert Input Data, a Microsoft Excel Document and Delete Results.
new form, Insert Input Data window, shown in Figure
3, will open.

Figure 2. Application main window.

444
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

Figure 3. Insert input data window.

our previous work [2]. A survey of the obtained results

for seven experimental cases is given in Table 1. Lite-
rature methods were assessed by calculating percent-
age deviation for heat transfer coefficient and pressure
drop from the equation:
m
Yexp − Ycal

100
PD(Y ) = (30)
Figure 4. Missing parameter warning window.
m i =1
Yexp
i

After running the calculation, all output parameters where Yexp and Ycal denote experimental and calculated
will be shown in a separate window and the most values of the heat transfer coefficient and pressure
important results (heat transfer coefficients of both drop, while m is the number of experimental data points.
fluids, overall heat transfer coefficient, available sur- From the Table 1 it can be concluded that the best
face area, required surface area, total pressure drops of agreement with experimental values for heat transfer
both fluids, etc.) will be also shown in a Message Box. coefficient, with around 8% deviation, was obtained by
The Working Principle window, shown in Figure 6, using the Martin method [4]. This result is considered
contains two video clips which show working principle as very good since the experimental uncertainty in
and components of a plate exchanger. determination of heat transfer coefficient is estimated
The Analysis of Results window is a part of the between 10% and 15% [2]. Satisfactory results, below
software, which allows comparison of the results the value of experimental uncertainty, were also
obtained by different methods. If so desired, some of obtained with Kumar [6] and Coulson and Richardson
the methods could be excluded from the comparative [7] methods, with the difference that the Kumar
analysis, as shown in Figure 7. method [6] gave higher and Coulson and Richardson
method [7] lower values of the heat transfer coefficient
than the experimental ones. VDI method [5], which is
RESULTS AND DISCUSSION
based on similar calculation procedure as the Martin
The process of selection of literature methods to be method [4], unexpectedly gave much higher deviation,
used for the development of the application PHeatEx around 38%. The Muley and Manglik method [8] has
Designer 1.0 for thermal and hydraulic calculations of been known for giving considerably lower values for
plate heat exchangers, was based on comparison of the Nusselt number and heat transfer coefficient than the
values determined using literature methods [4–9] with other literature methods, and this was confirmed by
experimental data measured on the setup described in our results, given in Table 1. Consequently, this method

445
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

Figure 5. Kumar method window.

was not included in the software for thermal and hyd- calculated in an Excel sheet for each of the selected
raulic design of plate heat exchangers. methods. After its validation the software was applied
The best agreement with experimental values for on two industrial plate heat exchangers. Process para-
pressure drop, below estimated experimental uncert- meters, geometry of the plates and thermophysical
ainty of 15–17% [2] was obtained by using the Kumar properties of working fluids were taken from the pro-
method [6]. Martin [4] and VDI [5] methods gave 33% ject documentation.
lower values than the measured pressure drop. The The most important geometrical data and process
highest deviation from experimental data was obtained parameters for the first tested plate heat exchanger are
by using the Coulson & Richardson method [7], due to given in Table 2. Process fluid, light vacuum gas oil
very low values of friction pressure drop calculated by (LVGO), was going through the temperature change
the equation (26). from 184 to 104 °C. As a cooling medium, water with
inlet temperature of 29 °C was used.
Summary of the software results calculated by four
literature methods, Martin [4], VDI [5], Kumar [6] and
Coulson and Richardson [7], as well as the values
obtained from the Excel sheet by using the experi-
mental correlation (Eq. (1)) for calculation of the heat
transfer coefficient, is given in Table 3. The experimen-
tal correlation was not included in the software since
its applicability was never properly tested under con-
ditions different from the ones under which it was ori-
ginally established.
From Table 3 it can be concluded that in the first
investigated case, the greatest disagreement is
between the Martin [4] and VDI [5] methods, which
were, interestingly, both established by the same
author. These two methods give the upper and lower
limits of possible values for the heat transfer coefficient
and consequently the area required for heat transfer.
Figure 6. PHE working principle window. As expected, values of the heat transfer coefficient for
hot fluid are much lower than for the cold fluid due to
Validation of the software was performed by the difference in thermophysical properties (especially
comparison of the software results with the values viscosity) between LVGO and water. Also, the velocity

446
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

Figure 7. Analysis of results window.

Table 2. Process parameters and geometry of the plates for the first tested plate heat exchanger

Geometrical data Value Process parameter Value

Length, Lp / m 0.872 Hot fluid LVGO
Width, W / m 0.486 Cold fluid water
Corrugation angle, αpl / ° 60 Hot fluid mass flow rate mh / kg s–1 21
Plate thickness, δpl / m 0.0006 Hot fluid inlet temperature, thi / °C 184
Plate enhancement factor, μp 1.16 Hot fluid outlet temperature, tho / °C 104
Amplitude of corrugation, 2a 0.0032 Cold fluid inlet temperature, tci / °C 29
Flow configuration 1–1 Cold fluid outlet temperature, tco / °C 40
Area of heat exchanger, As / m2 67.4 Hot fluid density, ρh / kg m–3 825
(available for heat transfer) Hot fluid viscosity, μh / Pa s 0.00138
Hot fluid thermal conductivity, λh / W m–1 K–1 0.106
Hot fluid specific heat, Cp,h / kJ kg–1 K–1 2.36
Hot fluid fouling factor, Rph / m2 K W–1 0.0005
Cold fluid fouling factor, Rpc / m2 K W–1 0.00017

Table 3. Calculation results for the first tested plate heat exchanger; Exp. values calculated from the Excel sheet

Parameter Martin VDI Kumar Coulson and Richardson Exp.

Heat exchanger duty, Q (kW) 3964.8 3964.8 3964.8 3964.8 3964.8
Hot fluid heat transfer coefficient, αh / W m–2 K–1 1005.9 1878.3 1212.0 1049.7 1318.9a
Cold fluid heat transfer coefficient, αc / W m–2 K–1 17422.7 21695.0 19370.4 13754.7 15032.0a
Overall heat transfer coefficient, K / W m–2 K–1 567.7 776.1 630.2 576.3 651.6
2
Required area for heat transfer, Ap / m 66.0 48.3 59.5 65.0 57.5
Reserve in heat transfer area, ΔA / % 2.0 28.3 11.7 3.6 14.7
Hot fluid friction pressure drop, Δpk,h / Pa 7488 7488 17317 1546 –
Cold fluid friction pressure drop, Δpk,c / Pa 82136 82136 138443 11456 –
a
Values calculated from the Excel sheet using the empirical correlation (1) and Eq. (2)

of hot fluid through one channel is more than three heat transfer area are much higher than the available
times lower than the velocity of cold fluid (0.21 in area for heat transfer, while all the other literature
comparison to 0.69 m s–1), which results in much higher correlations showed the existence of some reserve in
pressure drop on the cold fluid side. Values of the fric- the heat transfer area, although the numbers varied
tion pressure drop, calculated by using the investigated from 2.0 to 28.3%.
methods, are very different, the upper and the lower As an illustration, another example of the results
limits being defined by Kumar [6] and Coulson & Rich- obtained by the plate heat exchanger design software
ardson [7] methods. is shown here. In this case hot and cold fluids had simi-
The Muley and Manglik [8] method, although not lar thermophysical properties: demineralized water
included in the software, was also tested by means of was heated from 40 to 60 °C while process water was
the Excel sheet. According to this method, demands for used as a heating medium. The most important geo-

447
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

metrical data and process parameters are presented in ation, and the lower limit is, in this case, predicted by
Table 4 and calculation results are compared in Table 5. the Coulson and Richardson [7] method. Values of the
From the results presented in Table 5 it can be con- friction pressure drop are, as in the previous case, very
cluded that the upper limit for heat transfer coefficient different, with the upper and lower limits obtained by
and consequently the largest reserve in heat transfer Kumar [6] and Coulson and Richardson [7] methods.
area was again obtained by using the VDI [5] correl-

Table 4. Process parameters and geometry of the plates for the second tested plate heat exchanger

Geometrical data Value Process parameter Value

Length, Lp / m 0.472 Hot fluid water
Width, W / m 0.286 Cold fluid Demineralized water
Corrugation angle, αpl / ° 60 Cold fluid mass flow rate mh / kg s–1 10
Plate thickness, δpl / m 1-1 Hot fluid inlet temperature, thi / °C 100
Plate enhancement factor, μp 6.61 Hot fluid outlet temperature, tho / °C 90
Amplitude of corrugation, 2a 0.472 Cold fluid inlet temperature, tci / °C 40
Flow configuration 0.286 Cold fluid outlet temperature, tco / °C 60
Area of heat exchanger, As / m2 60 Hot fluid fouling factor, Rph / m2 K W–1 0.00017
(available for heat transfer) Cold fluid fouling factor, Rpc / m2 K W–1 0.00017

Table 5. Calculation results for the second tested plate heat exchanger; Exp. values calculated from the Excel sheet

Parameter Martin VDI Kumar Coulson and Richardson Exp.

Heat exchanger duty, Q / kW 839.4 839.4 839.4 839.4 839.4
Overall heat transfer coefficient, K / W m–2 K–1 3212.0 3411.3 3302.3 2888.9 3020.1a
Required area for heat transfer, Ap / m2 5.83 5.49 5.67 6.48 6.20
Reserve in heat transfer area, ΔA / % 11.8 17.0 14.2 1.9 6.2
Hot fluid friction pressure drop, Δpk,h / Pa 70790 70790 111453 7752 –
Cold fluid friction pressure drop, Δpk,c / Pa 19608 19608 32688 2768 –
a
Values calculated from the Excel sheet using the empirical correlation (1) and Eqs. (2) and (10)

CONCLUSION Acknowledgements
The authors gratefully acknowledge the financial
Software PHeatEx Designer 1.0 for the thermal and
support received from the Research Fund of Ministry of
hydraulic design of plate heat exchangers based on
Education, Science and Technological Development,
four literature methods: Kumar, Martin, VDI and Coul-
Serbia and the Faculty of Technology and Metallurgy,
son and Richardson, has been developed in the Sharp
University of Belgrade, Serbia (project No. 172063).
Develop programming environment. The selection of
methods to be incorporated into the application was
REFERENCES
based on comparison of the results obtained by lite-
rature methods with experimental values. The applic- [1] E. Djordjević, S. Kabelac, Flow boiling of R-134a and
ation offers to the user a number of options regarding ammonia in a plate heat exchanger, Int. J. Heat Mass
the calculation method, saving results, review and rep- Tran. 51 (2008) 6235–6242.
resentation of the results, comparative analysis of the [2] E. Djordjević, S. Kabelac, S. Šerbanović, Mean heat
results obtained by different calculation methods as transfer coefficients during evaporation of 1,1,1,2-tetra-
well as the explanation of plate heat exchanger compo- fluoroethane (R-134a) in a plate heat exchanger, J. Serb.
nents and working principle in two video clips. After its Chem. Soc. 72 (2007) 833–846.
validation, the application has been tested on heat [3] E. Živković, S. Kabelac, S. Šerbanović, Local heat transfer
exchangers from industrial practice and the results coefficients during evaporation of 1,1,1,2-tetrafluoro-
have been compared and analyzed. ethane (R-134a) in a plate heat exchanger, J. Serb.
It is hoped that this software will help to bridge the Chem. Soc. 74 (2009) 427–440.
gap between teaching the fundamentals and industrial [4] H. Martin, A theoretical approach to predict the per-
formance of chevron-type plate heat exchangers, Chem.
design and application of plate heat exchangers.
Eng. Process. 35 (1996) 301–310.

448
N.R. ZLATKOVIĆ et al.: PLATE HEAT EXCHANGER DESIGN SOFTWARE Hem. ind. 71 (5) 439–449 (2017)

[5] H. Martin, Druckverlust und Wärmeübergang in Platten- [8] S. Kakaç, H. Liu, Heat Exchangers: selection, rating and
wärmeübertragern, in: VDI – Wärmeatlas, Springer thermal design, 2nd ed., CRC Press, Boca Raton, FL,
Verlag, Heidelberg, 2002, pp. Mm1–Mm7. 2002.
[6] H. Kumar, The plate heat exchanger: construction and [9] T. Sun, A.S. Teja, Density, Viscosity, and Thermal Con-
design, in Proceedings of First UK National Conference ductivity of Aqueous Ethylene, Diethylene, and Triethyl-
on Heat Transfer, University of Leeds, Inst. Chem. Eng. ene Glycol Mixtures between 290 K and 450 K, J. Chem.
Symp. Series No. 86, Leeds, UK, 1984, pp. 1275–1288. Eng. Data 48 (2003) 198–202.
[7] R.K. Sinnott, Chemical Engineering Design, Coulson & [10] C. Yang, P. Ma, F. Jing, D. Tang, Excess Molar Volumes,
Richardson’s Chemical Engineering, Vol. 6, 4th ed., Else- Viscosities, and Heat Capacities for the Mixtures of
vier Butterworth-Heinemann, Oxford, 2005. Ethylene Glycol + Water from 273.15 K to 353.15 K, J.
Chem. Eng. Data 48 (2003) 836–840.

IZVOD

PROGRAMSKI PAKET ZA PROJEKTOVANJE PLOČASTOG RAZMENJIVAČA TOPLOTE ZA INDUSTRIJSKU I EDUKATIVNU

PRIMENU
Nikola R. Zlatković, Divna M. Majstorović, Mirjana Lj. Kijevčanin, Emila M. Živković
Tehnološko-metalurški fakultet, Univerzitet u Beogradu, Karnegijeva 4, 11120 Beograd, Srbija
(Stručni rad)
Pločasti razmenjivači toplote su tip aparata koji danas nalazi veoma široku Ključne reči: Pločasti razmenjivač toplote
oblast primene, od prehrambene i hemijske industriji do elektroenergetskih pos- • Prenos toplote • Pad pritiska • Pro-
trojenja. Osnovni konstruktivni element ovih aparata je slog orebrenih metalnih gramski paket
ploča. Orebrenja na pločama su tako dizajnirana da omogućavaju postizanje tur-
bulentnog strujanja duž cele površine za toplotnu razmenu, čime se postižu najviši
mogući koeficijenti prenosa toplote pri relativno maloj razlici temperatura fluida.
Ovo dovodi do manje potrebne površine za toplotnu razmenu, manjih dimenzija
aparata i, u nekim slučajevima, manjeg broja razmenjivača toplote. Pored odličnih
karakteristika u pogledu prenosa toplote, imaju i brojne druge prednosti u odnosu
na najčešće korišćene razmenjivače toplote sa cevnim snopom i omotačem: lako
se sastavljaju i rastavljaju radi čišćenja i popravki, površina za toplotnu razmenu se
lako može menjati promenom broja ploča u slogu, mogu raditi pri vrlo malim
razlikama temperature fluida koji učestvuju u toplotnoj razmeni, a stvaranje one-
čišćenja je manje izraženo zbog većih brzina strujanja i bolje distribucije toka
fluida. Osnovni problem pri termohidrauličkom proračunu pločastih razmenjivača
toplote je ograničen broj literaturnih korelacija, posebno onih koje bi imale širi
opseg primenljivosti, što je i razumljivo obzirom na različite geometrije i tipove
profila ploča koji se javljaju kod ovih aparata. U ovom radu je testirano nekoliko
literaturnih metoda za proračun koeficijenta prelaza toplote i pada pritiska u plo-
častom razmenjivaču toplote sa orebrenjem tipa “riblja kost” i dobijeni rezultati
su upoređeni sa eksperimentalno određenim vrednostima za koeficijent prelaza
toplote i pad pritiska. Izabrane su četiri metode koje su ugrađene u programski
paket za termički i hidraulički proračun pločastog razmenjivača toplote razvijenog
u okruženju Sharp Develop. Ovaj programski paket za termički i hidraulički pro-
račun pločastog razmenjivača toplote koristi sledeće četiri metode proračuna:
Martinov metod, VDI metod, Kumarov metod i metod Kulsona i Ričardsona
(Coulson–Richardson). Razvijeni softver je vizuelno prikazan preko prozora za
unos ulaznih podataka, prozora za prikaz rezultata proračuna dobijenih svakom od
metoda, prozora na kome je prikazano poređenje najbitnijih rezultata proračuna
za sve korišćene metode i pomoćnog prozora za prikaz principa rada pločastog
razmenjivača toplote. Nakon izbora metoda, razvoja i verifikacije softvera pristu-
pilo se njegovoj primeni na dva industrijska razmenjivača toplote.

449