Case Studies in Thermal Engineering 12 (2018) 497–509

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

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

Numerical investigation of heat transfers in the water jacket of

T
heavy duty diesel engine by considering boiling phenomenon
⁎
M. Gholiniaa, M. Pourfallaha, , H.R. Chamanib
a
Department of Mechanical Engineering, Mazandaran University of Science and Technology, Iran
b
Department of Mechanical Engineering, Iran University of Science and Technology, Iran

A R T IC LE I N F O ABS TRA CT

Keywords: The flow of the coolant fluid and its heat transfer directly affect the cooling performance, heat
Heavy duty diesel engine load of the hot parts and also the thermal efficiency of the diesel engine. The proper estimation of
Cooling jacket heat transfer and temperature distribution in a diesel engine is essential for investigating the
Heat transfer thermal stresses and calculating its performance, which requires a precise simulation of the
Subcooled nucleate boiling
cooling water jacket. An efficient approach to study the cooling system is to simulate using
Numerical Simulation
Computational Fluid Dynamics (CFD) as a three-dimensional model by simultaneously solving
the structure and fluid, which leads to accurate prediction of wall temperature and heat flux. In
the present paper, the distribution of heat transfer coefficients (HTC) in the cooling jacket of a 16-
cylinder heavy-duty diesel engine has been calculated using ANSYS/Fluent based on 3D-CFD
method. Also, equations of subcooled boiling phenomenon have been solved based on methods of
Chen and BDL, and the effects of fluid pressure, velocity, and temperature (At the time of the
phenomenon of boiling) on the heat transfer of cooling jacket wall have been studied. The results
show that the sensitive thermal region that is at risk is the region between the exhaust valve and
around the glow-plug. This region if not properly cooled, will result in gas leakage from the
combustion chamber, which will result in a decrease in engine power and torque.

1. Introduction

Heavy duty diesel engines (CI) are used in a wide range of applications such as power generators, navy propulsion and rail
traction applications, each of which uses different design of cooling systems. For this reason, a fundamental understanding of the
processes considered in designing a cooling system for diesel engines is a necessary prerequisite for optimal cooling based on di-
mensions, type of components and thermal capacity. One of the most important parts in the cooling system of diesel engines is the
water jacket installed around the liner and inside the cylinder head for the coolant fluid to flow inside it. The ability to predict the
amount of heat transfer between the fluid and the walls of the suction system, the exhaust system and the cooling system is very
momentous for the designer engineer, because the heat transfer of CI engines is important in many ways:

• Protecting the materials used in sensitive parts of the engine against melting or deformation
• Increasing the engine's useful power by reducing energy leakage inside the combustion chamber
• Reducing contamination by reducing engine warm-up time
• Improving lubrication performance and reducing the hammering phenomenon.
⁎
Corresponding author.
E-mail address: m.pourfallah@ustmb.ac.ir (M. Pourfallah).

https://doi.org/10.1016/j.csite.2018.07.003
Received 5 May 2018; Received in revised form 17 June 2018; Accepted 8 July 2018
Available online 10 July 2018
2214-157X/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/BY/4.0/).
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

Heat transfer from the wall of the cylinder head to the cooling fluid is a combination of convective heat transfer and nucleate
boiling. Due to the passage of the operating fluid from the cooling water jacket, the bubbles formed from the hot surface of the wall
are separated and enter a lower temperature region, and thus a large amount of energy on the hot wall of the cylinder head is
transferred. By increasing the wall temperature, the number of bubbles formed at the surface also increases. The bubble formation
regime gradually reaches a stage where the number of bubbles separated from the hot surface of the cylinder head is more than the
number of bubbles formed. The time from the formation of the first bubble until this stage is called nucleate boiling. Nucleate boiling
is beneficial in the hot regions of the cylinder head, i.e. the region between the exhaust valves and around the glow-plugs, since it
transmits a high thermal flux for a low temperature change. It is worth noting that nucleate boiling consists of two regions; saturated
boiling and subcooled boiling. Since the type of boiling phenomenon that occurs in the engine is of a kind of subcooled flow boiling,
therefore the stated topics relate to the subcooled nucleate boiling. Many studies have been conducted to optimize the design of
cooling water jacket of CI engines. One of the first activities in the field of boiling heat transfer is the work done by Norris et al. [1]. In
a study, they tried to find pure convection regions and nucleate boiling in a 6-cylinder diesel engine by using the measurement of
cylinder head temperature and finite element analysis. Another very effective research on the field of nucleate boiling for use in
internal combustion engines is the work done in 2003 by Robinson et al. [2]. The results of Robinson et al. show that the effect of
convective heat transfer is so high at high speeds that boiling does not actually occur. In the context of subcooled boiling modeling,
Lee and O’ Nill [3] calculated the temperature at the cylinder head of an internal combustion engine experimentally by implanting a
thermocouple. Lee has done similar works in the same field as can be seen in references [4,5]. Furthermore, computation fluid
dynamics (CFD) was investigated by You-chang and Xiao-hong [6] for studying the cooling system in a diesel engine in three
dimensions for the water jacket and in one-dimensional simulation for the cooling system. Their results show that the cooling system's
performance can be optimized by changing the speed, pressure and distribution of the heat transfer coefficient in the cooling jacket.
Jafarabadi et al. [7] investigated the two-phase mixing of passing through the cooling water jacket of a 12-cylinder heavy duty diesel
engine by simultaneous three-dimensional analysis of the structure and fluid. They showed that the boiling phenomenon occurs at hot
points with a low flow of coolant, which results in an increase in the convective heat transfer coefficient. In the continuation of their
studies in the field of cooling systems, Jafarabadi et al. [8] presented four designs for cooling passages and selected the proposed
design based on the advantages and disadvantages of the presented designs. In addition, Punekar and Das [9] of ANSYS Company
investigated the boiling phenomenon in the engine's cooling jacket by presenting a computational fluid dynamic (CFD) model. Their
study was carried out in order to investigate the temperature distribution in two cases, taking into account the boiling phenomenon
and without it. The results show a temperature decrease at critical points with consideration of the boiling phenomenon. Another
very effective research on the diesel engine's cooling is the work by Romanov and Khozeniuk [10] in 2016. Their goal was to achieve
a high quality engine cooling design and to obtain uniform cylinders at the same level of thermal stresses. The results of Romanov for
the considered diesel engine showed that it is possible to control the cooling in a uniform path without changing the design of the
chamber, crankshaft and cylinder head.
All these activities were performed with the aim to develop predictive tools to help design high-precision cooling systems. The
aim of this paper is the numerical simulation of the subcooled boiling phenomenon in the cooling jacket of a 16-cylinder heavy duty
diesel engine based on the two commonly used models of Chen and Boiling departure lift-off (BDL). In addition, the investigation of
the indices of velocity, pressure and heat transfer coefficient (HTC) in the engine's cooling jacket has been done using a three-
dimensional fluid simulation using Fluent software with proper mesh density, which guarantees the accuracy of the results and a
logical time of the calculation. A noteworthy point in the present simulation is the simultaneous solution of the structure (cylinder
head) - fluid.

2. Numerical simulation of boiling

For modeling the boiling phenomenon, all the mechanisms governing this phenomenon should be considered. However, finding
the exact relations governing all these mechanisms, such as the volume of the formed bubble, the frequency of bubble separation, the
diameter of the bubble separation and so on is very difficult. Therefore, a series of simpler assumptions is used to simulate the heat
transfer of the boiling phenomenon. One of the patterns used to simulate boiling is the following pattern proposed on the basis of the
assumptions of Rohsenow [11] on the sum of the values of forced convective heat transfer and nucleate boiling thermal flux.

″ = qconv
qtotal ″ + qboil
″ = hconv (Tw − Tf ) + hboil (Tw − Ts ) (1)

In the above equation, hconv is the forced convective heat transfer coefficient, and hboil is the nucleate boiling heat transfer
coefficient. The most known relationships proposed for the forced convective heat transfer coefficient and the nucleate boiling heat
transfer coefficient are the relationships proposed by Ditus Bulter and Foster-Zuber [12].

hconv = k / Dh 0.023 Re0.8 . Pr 0.4 (2)

0.79 0.45 0.49

⎡ k1 Cp, l ρl ⎤
hconv = 0.00122 ⎢ 0.5 0.29 × (ΔTs )1.24 (Δps )0.75
σ μ1 (hρl, g )0.24 ⎥ (3)
⎣ ⎦

Several researchers used this model in a variety of ways, among which Chen's method and the BDL method were more accepted.

498
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

2.1. Chen's method

Following the equations proposed for various types of boiling, Chen [13] corrected the equation of the subcooled flow boiling as
follows:
″ = qconv
qtotal ″ + qboil
″ × Schen = hconv (Tw − Tl ) + hboil (Tw − Ts ) × Schen (4)
Following the occurrence of boiling when flowing, the effect of convective heat transfer is reduced in comparison with the
nucleate boiling. Therefore, Chen introduced a corrective factor in Eq. (1) to resolve the effect of this heat transfer reduction. The
factor introduced in the Rohsenow Equation [11] by Chen [13] is known as the suppression factor (S), and its value is obtained from
this equation:
1
Schen =
1 + 0.12(Re×10−4)1.17 (5)
Later, Butterworth [14] presented another equation for the suppression factor as follows.
1
Schen, corrected =
1 + 0.12 × 10−6 (Re)1.17 (6)
With respect to these topics, Eq. (7) is presented for the subcooled flow boiling according to Chen's method:

1
″ = qconv
qtotal ″ × Schen = hconv ⎜⎛Tw − Tf ⎟⎞ + hboil ⎜⎛Tw − Ts ⎞⎟ ×
″ + qboil .
⎝ ⎠ ⎝ ⎠ 1 + 0.12 × 10−6 (Re)1.17 (7)

2.2. The BDL method

The general heat transfers in the BDL method, as in Chen's method, is obtained from Eq. (1). The only difference of the BDL
method with Chen's method is in the value of the suppression factor (S). This method, first proposed by Steiner [15] with the use of
Zeng's work [16], attempts to determine the suppression factor using the steam dynamics and the flow field in the neighborhood of
the hot surface.
SBDL = Sflow × Ssub (8)

Sflow = (dD/ dL)n (9)

Ssub = ((Tw − Ts )/(Tw − Tb)) (10)

In Eq. (8), Sflow is called the suppression factor caused by flow and Ssub is called the suppression factor caused by subcooling.
According to Zeng's theory, bubble separation from the hot surface consists of three stages as shown in Fig. 1(a). The bubble in the
first stage at the point of the bubble nucleus inclines to size (θ) due to the hydrodynamic force. Then in the second stage, the bubble
adhering to the surface grows so that it reaches a size making it slip on the hot surface. The radius of bubble in this stage is
represented by rD. The bubble growth process continues until the bubble reaches the rL radius. At this stage, the floating force causes
the bubble to separate from the hot surface and enter the fluid mass having a lower temperature. In order to obtain the separation
radius and the climb radius, the momentum equations in the direction of flow and in the direction perpendicular to the flow must be
calculated in accordance with Fig. 1(b).
Fd + Fdu sin θ = 0, (11)

Fbcy + Fdu cos θ + Fsl = 0. (12)

In Eqs. (11) and (12), Fd, Fdu, Fbcy and FsL are the drag, bubble growth, buoyancy, and shear forces respectively. These forces are

(a) Inclination (b)

U(y) Angle Lift-off
Center of
Mass Bubble

Heating
Surface
Sliding
Levels Bubble

Fig. 1. Bubble separation from the hot surface (a) and force balance for a bubble (b).

499
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

obtained as follows:
Fd = 6μl πur ((2/3) + [(12/Reb) + 0.796n3]) , n3 = 0.65 (13)

Fsl = (1.938) ρl u2πr 2Gs1/2 × ((1/Re2b) + 0.014Gs2)1/4 (14)

Fbcy = (4/3) πr 3g (ρl − ρg ) (15)

Fdu = ρl πr 2 (3/2. Cs r ′2 − rr ″) (16)

Reb = 2ρl ur / μl (17)

du r
Gs = .
dy y=r
u (18)
The value of the parameters (16) and (17) is proposed by reference [16] to be 20.3. Furthermore, in the above equations, Reb is
called the bubbles’ Reynolds and Gs is called dimensionless shear stress. In Eq. (16), the bubble radius is obtained by Zuber's equation
[17]:
r (t ) = 2b/ π . Ja. αl t (19)

Ja = ρl Cp, l (Tw − Ts )/ ρg hlg (20)

αl = kl/ ρl Cp, l (21)

In Eq. (19), Ja is the Jaccobian number, b is an experimental constant and α is the heat distribution coefficient in liquid state,
which is proposed to be 0.21 for horizontal flows [15]. In Eqs. (17)–(18) the parameter u is the flow velocity at the height of the
bubble center inside the boundary layer. Before solving the momentum equations, the flow profile must be determined. The velocity
profile inside the boundary layers is obtained using the following equations [18].

u 1 ⎛ ⎞ ⎡ y+ y+ y+ ⎤
u+ = = ln ⎜1 + ky+ ⎟ + C ⎢1 − exp ⎛⎜− ⎞⎟ − exp ⎛⎜− ⎞⎟ ⎥.
uτ k ⎝ χ⎠ χ ⎝ 3 ⎠⎦ (22)
⎝ ⎠ ⎣

y+ = ρl u τ y / μl (23)

u τ = (Tw / ρl )1/2 (24)

+
In the above equation, y is the dimensionless normal distance and uτ is the friction velocity of the wall. Furthermore, the values
accepted for the constants are determined at κ = 0.41, χ = 11 and C= 7.4 depending on the surface roughness. There is no slip
between the bubble and the surrounding liquid at the moment of the bubble rise. Therefore, the drag and shear forces and the angle of
bubble deviation (θ) are zero relative to the vertical axis. Hence, the momentum equation will be converted to the following equation
for determining rL.
Fdu + Fbcy = 0 (25)
After formulating Eqs. (8)–(9) and (25) and obtaining the values of rD and rL, the suppression factor caused by flow Sflow is
obtained.

3. Geometry production and computational mesh preparation

The targeted engine in this study is a heavy duty 16-cylinder turbocharged diesel engine equipped with median cooler. The
diameter-to-length ratio of the pathway is 165 mm/195 mm and the ratio of the nominal power/nominal speed is 2000 kW/1800 rpm
[19]. The cylinder head of this engine is ductile cast iron. The three-dimensional model of the cooling pathway around the liner and
inside the cylinder head can be seen in Fig. 2(a-c). Each row of the engine consists of eight cylinders cooled in series. Coolant fluid is
divided into two parts after circulating around each liner. Part of it goes to the next liner, and the other part enters the cooling
pathway inside the cylinder head. To carry out the thermal analysis, the solid and liquid parts of one cylinder (including the cooling
pathway around the liner and the cooling pathway inside the cylinder head) have been simultaneously entered to the ANSYS Meshing
software and meshed using the unstructured meshing method. Furthermore, in order to reduce the number of meshes and meshing
errors, geometry correction and removal of additional lines have been done. The number of the generated meshes in the software is
about 2,500,000, mainly consisting of hexagonal cells, as well as a small number of pyramidal cells and prismatic cells. The meshed
model of this geometry is shown in Fig. 2(d-e). In addition, it is worth noting that in order to increase the quality of the computational
mesh, the mesh dimensions in a range of geometries such as regions with extreme temperature gradients, surfaces in contact with
each other, and surfaces that have a common interface with fluid analysis have been fined, and thus, the precision of the calculations
has increased. On the other hand, by expanding mesh dimensions in regions that do not have much effect on the results of the
analysis, the computational time can be reduced.

500
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

Fig. 2. Overall view of the water jacket and cylinder head (a-c); meshing of the Geometry(d-e).

3.1. Boundary conditions

Boundary conditions are divided into two categories; fluid flow boundary conditions and thermal boundary conditions. To
analyze fluid flow, the flow characteristics must be defined in the inlet and outlet channels of the water jacket. In the inlet channel,
the mass flow boundary condition is used, and the outlet pressure boundary condition is used in the outlet flow. The static pressure of
the fluid in the outlet is 0.4 bar. Furthermore, the cooling fluid is considered to be in a general combination of water and ethylene
glycol with a volume ratio of 50% water and 50% ethylene glycol, where the physical and thermal properties of this combination are
considered at the engine operating temperature (i.e. 363.8 K) [20]. (See Table 1). It should be noted that given the critical condition
of the cylinder head operation at maximum engine power, all values have been calculated at engine speed of 1800 (rpm) and full-load
operating conditions. Table 2 shows the boundary conditions of the wall of the cylinder head. In addition, for a better understanding,
a view of the points defined in geometry as boundary conditions is shown in Fig. 2(b).
In this research, after applying the boundary conditions and choosing the K-ε flow turbulence pattern, a couple solver method has

501
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

Table 1
Thermophysical properties of 50% water and 50% ethylene glycol
at the engine operating temperature (i.e. 363.8 K) [20].
Physical properties

Specific heat capacity 3470 ( j/kg K)

Density 1030 (kg/m3)
Heat conduction coefficient 0.44 (W/m K)

Table 2
Temperature and heat transfer coefficient in cylinder head walls [19,24].
Region Temperature (K) Heat transfer coefficient (W/m2 k)

Flam face 1050 600–650

Smoke port location 880 300–350
Air port location 380 100–110

been used to solve the flow heat. This means that all the equations of fluid and energy are solved simultaneously in this method.
Furthermore, in this simulation for determining the relationship between momentum, continuity and pressure equations, the solution
starts with a simple method, and the repetition continues until the variations in the transient and steady solution reach the lowest
possible value. At the end, the values of temperature distribution and heat flux in the cylinder head, velocity distribution, pressure
distribution, and convection heat transfer coefficient in the water jacket are obtained.

4. Results and discussions

4.1. validation for boiling phenomenon

In order to verify the precision of the present research, we compare some of our solutions with the outcomes of the formerly
published studies of Ref (Finaly et al. [21]) for the sub-cooled and film boiling of water and ethylene glycol within an aluminum
cylindrical tube. (See Fig. 3a-e). BDL and Chen correlations are used for validation. This research shows that BDL and Chen are great
models for viewing the boiling phenomenon, such as the problem raised in this research.

4.2. Simulation of the boiling phenomenon

In order to investigate the precision and accuracy of the results of the boiling phenomenon, simulation was carried out using two
different methods based on Foster-Zuber patterns [12]. Fig. 4(a-b) shows the comparison of the proposed methods. The difference in
the graphs in this figure does not represent the superiority of a method on the other, because in some studies, Chen's simulation
system has an acceptable agreement with experimental results according to the problem type and operating conditions [15], while
the BDL results are in better agreement with experimental results in some others [7]. The purpose of this paper is to show the boiling
phenomenon by changing the active indices. Therefore, validation was performed in seven groups of calculations according to
Table 3 based on the BDL method to study the effect of pressure, velocity and temperature on wall temperature and thermal flux. The
fluid is the volume composition (50−50) of water and ethylene glycol. To calculate the thermal flux values, the equations have been
analyzed using Maple-18 software with a simple algorithm. As shown in Figs. (4) to (5), the variations thermal flux change linearly
with the wall temperature before the boiling point (near to saturation temperature at the corresponding pressure). This indicates
forced heat transfer, and until this stage, the thermal flux for both simulations is approximately the same (according to Fig. 4(a)-b).
With the start of the nucleate boiling, the heat transfer coefficient increases, and as a result, the slope of the thermal flux graph
suddenly enhances according to the wall temperature. This suggests that for a constant surface temperature, more heat can be
eliminated if nucleate boiling phenomenon occurs. In fact, this topic represents the potential for heat transfer increase by the op-
erating fluid in the nucleate boiling region. The graph of the effects of the coolant fluid velocity on the thermal flux as a function of
the temperature of the cylinder head wall shows that the effect of nucleate boiling is observable at speeds lower than 1.5 m/s.
However, almost no boiling phenomenon occurs at speeds higher than 2.5 m/s, which means that the dominant regime is the
convective heat transfer at higher speeds. In other words, by decreasing the velocity of the operating fluid, the starting temperature of
the nucleate boiling will decrease. However, it should not be forgotten that the sensitivity to the film boiling start point increases by
decreasing the inlet velocity. (See Fig. 5(a)-b). The effects of the inlet temperature of the coolant fluid on the thermal flux as a
function of the temperature of the cylinder head wall are shown in Fig. 5(c-e). As it is shown, by decreasing the inlet temperature of
the operating fluid, the temperature point at which the boiling phenomenon occurs is at lower temperature of the wall. In fact, this
means that by decreasing the temperature of the inlet fluid, the effects of boiling heat transfer will gradually appear, such that the
lower the inlet temperature, the sensitivity to the starting point of the film boiling will increase. However, the excessive increase in
the inlet temperature will also make the starting point of the nucleate boiling occur earlier, which makes the film boiling phe-
nomenon occur earlier in time. According to the presented topics, it should be admitted that the inlet temperature is one of the

502
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

Pressur e Tappings
(a)

Thermocouple
T
P
Position

(c) 8 (d) 8
Velocity=0.5m/s Chen model

Experiment [21]
6
Heat Flux (W/M2). 10-5

6
Heat Flux (W/M2). 10-5

BDL model

4 4

Chen model

2 BDL model 2
Velocity=1m/s
Experiment [21]

0 0
80 100 120 140 160 180 200 220
SurfaceTemperature ( 0C) SurfaceTemperature ( 0C)

(e) 12
Velocity=3m/s
10
Heat Flux (W/M2). 10-5

Experiment [21]
4
Chen model
2
BDL model
0

SurfaceTemperature ( 0C)
Fig. 3. Water-jacket considered in the experiment of Finlay et al. [21] that is used for validation (a-b); The results obtained from Chen and BDL
models against experimental ones for various inlet velocities (c-e).

important factors in the development of the boiling phenomenon, because decreasing or increasing the inlet temperature has a
significant effect on the time and temperature of the occurrence of the boiling phenomenon. If the heat transfer is of the convective
type (i.e. before the start of the boiling phenomenon), pressure variations cannot have much effect on the heat transfer rate, but when
the heat transfer occurs through boiling, pressure can have a significant effect on the heat transfer rate. In fact, at the low speeds
where boiling has more importance, the pressure effect should be taken into account. In this way, high pressure can delay the start of

503
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

(a) ×10
6
(b) 2 ×10
6
2.5
S Chen , P=1.0 (bar), V= 0.2 (m/s) S Chen , P=2.0 bar, V=1.2 m/s.

2 S BDL , P=2.0 bar, V= 1.2 m/s.

Heat Flux (W/m2) S BDL, P=1.0 (bar), V= 0.2 (m/s) 1.5
S Chen , P=2.0 bar, V= 4.0 m/s.

Heat Flux (W/ m2)

S Chen , P=1.4 (bar), V= 0.6 (m/s)
1.5
S BDL , P=2.0 bar, V= 4.0 m/s.
S BDL , P=1.4 (bar), V=0.6 (m/s) 1

0.5
0.5
Temp inlet=100 c

0 0
105 120 135 150 165
T (˚C) T( 0C)
Fig. 4. (a-b). Comparison of simulation results for BDL and Chen methods.

Table 3
The condition of seven groups of simulations.
Numbers Temp inlet (°C) Velocity inlet (M/s) Pressure inlet (Bar)

1 80 0.3–0.9–1.5–3 1
2 90 0.25 1–1.4–2
3 75–95–105 0.25 1
4 80 0.05 1–1.4–2
5 90 0.5–1.2–4. 2
6 75–95–105 1 1
7 75–95–105 1.2 2

film boiling and decrease the sensitivity of entering this area. (See Fig. 5f-g).

4.3. CFD simulation validation

One of the important issues in turbulent flow simulations is to ensure the presence of sufficient number of boundary layers to
observe the gradient of velocity. Fluid dynamics mechanisms in the near-wall region are investigated with quasi-experimental
equations in dependent variables. These equations are based on functions of a dimensionless distance perpendicular to the wall [22].
In addition, considering the dependence between the determined thickness of the boundary layer and the value of y+ is important.
The value of y+ depends on the geometry and the fluid flow variables, whose actual value is determined after the simulation. Fig. 6(a)
shows the value of y+ within the acceptable range [23]. Furthermore, in order to validate the simulation results, the available
experimental data for heavy duty diesel engine with nominal power and speed in the range from 175 kw/450 rpm to 1030 kw/
1000 rpm were compared with the results of the theoretical method for the same engine [7,24]. The comparison of the results of CFD
analysis with experimental data indicates the precision and accuracy of the present simulation.

4.4. Pressure distribution in the cylinder head water jacket

As noted above, the pressure variations in the convective heat transfer do not have much effect on the heat transfer rate. However,
when the heat transfer of the boiling type occurs, fluid pressure can have a considerable effect on the heat transfer rate. Therefore, the
effect of pressure should be considered at low velocities where boiling has more importance. Pressure distribution in the engine's
cooling water jacket in the present problem with a mass flow of 0.65 (kg/s2) is shown in Fig. 6(b). In the first region, where the fluid
is moving around the liner, the pressure drop is almost negligible due to the relatively large hydraulic diameter of the cooling path
around the liner. In contrast, in the second region, where the coolant fluid is moving in the cylinder head, a relatively large pressure
drop occurs in these regions due to the smaller hydraulic diameter of the cooling paths and the more complex fluid flow path. A
notable point in the design of the present geometry is the relatively high pressure of the coolant fluid in the regions of exhaust valves
and glow-plugs, which delays the film boiling and makes it stay in the nucleate boiling area. Fig. 2(c) shows the geometry of fluid
flow in the regions of exhaust valves and glow-plugs.

4.5. Velocity distribution in the cylinder head water jacket

The distribution of the velocity field and the flow lines in the cooling water jacket around the liner and inside the cylinder head is
shown in Fig. 7(a-b). The highest velocity of the coolant fluid is 8.43 m/s, which occurs in the region around the valves and around

504
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

(a) ×10
6 (b) ×10
6
2.1 1.8
Velocity=0.3 (m/s) Velocity=0.5 m/s
1.8
Velocity=0.9 (m/s) Velocity=1.2 m/s
1.5
Heat Flux (W/m2)

Heat Flux (W/ m2)

1.2
Velocity=1.5 (m/s)
Velocity=4 m/s
1.2
Velocity=3.0 (m/s)
0.9
0.6
0.6

Pressure inlet=1 (bar)

Pressure=2 bar
0.3 Temp =90 0C
Temp inlet=80 ( ˚C)
0
0
85 98 111 124 137 150 163 90 115 130 145 160
T (˚C) T( 0C)
(c) 2.1 ×106 (d) ×10
6
2.4
Temp=75 C Temp=75 (˚C)
2
Temp=95 C Temp=95 (˚C)
Heat Flux (W/ m2)

Heat Flux (W/m2)

1.4 Temp=110 C 1.6
Temp=110 (˚C)

1.2

0.7 0.8
Pressure inlet=1 (bar)
Velocity inlet=0.25 ( m/ s)
0.4
Velocity inlet=1 m/s
Pressure inlet=1bar
0 0
75 95 115 135 155 175 75 95 115 135 155 175

T(0C) T (˚C)
5
(e) 20 ×10 Velocity=1.2 m/s pressure=2bar
6
(f) ×10
2
Temp=75 C

Temp=95 C Pressure=1
15
Heat Flux (W/ m2)

Temp=110 C 1.5
5 Pressure=1.4
Heat Flux (W/ m 2)

Pressure=2
10
1

5
0.5
5 Velocity= 0.05
0 m/s
Temp inlett= 800C

0 0
80 110 140 170 80 120
100 120
13 140 160
165
T( C) 0
T( 0 C)

(g) ×10
6

1.8
Pressure=1.0
r (bar)
1.5
Pressure=1.4
r (bar)
Heat Flux (W/m 2)

1.2
Pressure=2
r (bar)

0.9

0.6

Velocity inlet=
=1 (bar)
0.3
Temp inlet=909 ( ˚C)

0
90 105 120 135 150 16
65
T (˚C)

Fig. 5. Effect of velocity inlet (a-b), temperature inlet (c-e) and pressure inlet (f-g) on the heat flux graph in terms of wall temperature.

505
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

Fig. 6. View of the contour y + (a) and distribution of pressure over the cooling jacket (b).

the glow-plugs, and the lowest fluid velocity is about 0.3 m/s near the exhaust gas outlet. The high velocity of the coolant fluid in
these sensitive thermal regions, where there is a risk of the occurrence of film boiling phenomenon, causes these regions to be
protected from this phenomenon. In addition, the low velocity of the coolant fluid around the seat of the exhaust valve and around
the exhaust pathway increases the sensitivity to the starting point of the film boiling due to the high thermal flux, so it is necessary to
modify the design to improve the fluid motion. plus, in using the boiling phenomenon to cool the narrow pathways such as the
regions around the valves (See Fig. 2(c)), it should be noted that due to increased bubble production, there is a probability of closing
these narrow paths by bubbles. Therefore, an acceptable heat transfer rate can be achieved in this region by monitoring the velocity.

4.6. Temperature distribution is the cylinder head water jacket

Figs. 7(c-e) to 8(a-b) show the results of the distribution of temperature and thermal flux in the cylinder head and the cooling
water jacket. The results indicate that the cylinder head maximum temperature in the bridge area occurs between the exhaust valves.
An increase in the temperature of the coolant fluid at the outlet of the exhaust pathway can also be seen in Fig. 7(c-e). This is due to
the low velocity of the coolant fluid and the lack of design of the vortex path behind the exhaust valves. In addition, according to the
simulation results, it is predicted that the coolant fluid temperature will increase by 12.5 degrees in the pathway from the cylinder
inlet to the outlet of the cylinder head. The heat flux passing through the wall of the cylinder head without considering the boiling
phenomenon is also depicted in Fig. 8(a-b). The comparison of this figure with the numerical simulation diagrams of the boiling
phenomenon (Figs. 4 to 5) shows that the use of boiling phenomenon can significantly increase the heat flux. As shown in Fig. 8(a),
the heat flux passing through the valves is larger than the other parts. Therefore, the lack of proper cooling in these regions can cause
heat cycle fatigue at the level of the cylinder head flame. Therefore, this necessitates more precision in designing cooling water jacket
and better use of the boiling phenomenon around this region. On the other hand, the thermal flux in the seat of the exhaust valve is
significantly low, which indicates low heat transfer rate in these regions. The reason for this according to the above explanations is
the lack of design of the vortex path for cooling in these sensitive thermal regions, which reduces the thermal flux and increases
temperature at these surfaces (Fig. 8b). The strength point of this research is the integrated or simultaneous solution for the thermal
analysis of the coolant fluid flow. That is, the thermal analysis of the structure of the cylinder head and the coolant fluid inside it has
been carried out simultaneously.

4.7. Distribution of heat transfer coefficient in the water jacket

The results of the heat transfer coefficient (HTC) for the cooling water jacket are shown in Fig. 8(c-d). The heat transfer coefficient
has increased considerably in the points of the water jacket where the temperature is higher. In the present simulation, the maximum
HTC is about 49,000 W/m2k, whose maximum values occur in the vicinity of the glow-plugs according to the temperature dis-
tribution and velocity distribution. Also, HTC is greatly reduced in the vicinity of the exhaust path due to the high temperature of the
exhaust gases and the low velocity of the coolant fluid around the wall, which causes the accumulation of heat in these regions.

5. Conclusion

In this paper, the distribution of the velocity field, pressure, and heat transfer coefficient (HTC) in the cooling water jacket of a
diesel engine has been investigated through three-dimensional CFD method using ANSYS/Fluent software. Moreover, by presenting
mathematical models based on methods of Chen and BDL, the numerical simulation of subcooled boiling in the water jacket has been
considered in this paper. The significant point is the simultaneous solution of the structure and fluid in this analysis. The results show
that:

506
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

Fig. 7. View of the velocity (a-b) and temperature contours (c-e).

• Using nucleate boiling flow regime can increase the heat transfer coefficient, which results in dissipating more heat from the
surface and provides a more uniform temperature distribution for the engine.
• The temperature at which the nucleate boiling will start increases by increasing water flow velocity.
• The effect of pressure on decreasing or increasing the heat transfer rate is very important when the heat transfer occurs through
boiling. In this way, high pressure can delay the start of film boiling.
• The low velocity of the coolant fluid and the lack of design of the vortex path behind the exhaust valves lead to an increase in the
fluid temperature in the seat of the exhaust valve and around it, especially the narrow bridge between the two valves.
• The comparison of the thermal flux passing through the wall of the cylinder head with and without considering the boiling
phenomenon shows that the occurrence of the boiling phenomenon can significantly increase the heat flux.
• In the thermal analysis of the water jacket, it was determined that the maximum value of the convective heat transfer coefficient is
about 49,000 W/m2k, that its maximum value occurs in the vicinity of the glow-plug according to the temperature distribution

507
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

Fig. 8. View of the heat flux (a-b) and heat transfer coefficient (c-d).

and velocity distribution.

References

[1] P.M. Norris, W. Wepfer, K.L. Hoag, D. Courtine-White, Experimental and analytical studies of cylinder head cooling, SAE Technical, Paper 931122, 1993.
[2] K. Robinson, J. Hawley, N. Campbell, Experimental and modelling aspects of flow boiling heat transfer for application to internal combustion engines, Proc. Inst.
Mech. Eng. Part D: J. Automob. Eng. 217 (2003) 877–889.
[3] H.S. Lee, A.T. O'Neill, Comparison of boiling curves between a standard SI engine and a flow loop for a mixture of ethylene glycol and water, SAE Technical,
Paper 2006-01-1231, 2006.
[4] H. Lee, Heat Transfer Predictions using the Chen Correlation on Subcooled Flow Boiling in a Standard IC Engine, SAE Technical Paper 2009-01-1530, 2009.
[5] H. Lee, A. O'Neill, Forced convection and nucleate boiling on a small flat heater in a rectangular duct: experiments with two working fluids, a 50–50 ethylene
glycol—water mixture, and water, Proc. Inst. Mech. Eng., Part D: J. Automob. Eng. 223 (2009) 203–219.
[6] L. You-chang, G. Xiao-hong, C. Dan, Research on cooling system for 4-cylinder diesel engine, SAE Technical 2007-01-2064, 2007.
[7] M. Jafarabadi, H. Chamani, A. Malakizadi, S.A. Jazayeri, a fast coupled CFD-Thermal analysis of a heavy duty diesel engine water cooling system, In ASME 2008
International Mechanical Engineering Congress and Exposition, pp. 663–670, 2008.
[8] M. Jafarabadi, H. Chamani, S.A. Jazayeri, Improvement of a Diesel Engine Water Cooling Performance Through Implementation of Different Cooling Designs, In
ASME 2013 Internal Combustion Engine Division Fall Technical Conference, pp. V002T07A013, 2013.
[9] H. Punekar, S. Das, Numerical Simulation of Subcooled Nucleate Boiling in Cooling Jacket of IC Engine, SAE Technical Paper 2013-01-1651, 2013.
[10] V.A. Romanov, N.A. Khozeniuk, Experience of the Diesel Engine Cooling System Simulation, International Conference on Industrial Engineering, Procedia
Engineering 150, 490–496, 2016.
[11] W.M. Rohsenow, Heat Transfer, A Symposium, Engineering Research Institute, University of Michigan, Michigan, 1952.
[12] H. Forster, N. Zuber, Dynamics of vapor bubbles and boiling heat transfer, AIChE J. 1 (1955) 531–535.
[13] J.C. Chen, Correlation for boiling heat transfer to saturated fluids in convective flow, Ind. Eng. Chem. Process Des. Dev. 5 (1966) 322–329.
[14] D. Butterworth, The Correlation of Cross Flow Pressure Drop Data by Means of Permeability Concept, (1979) (UKAEA Report, No. AERE-R9435).
[15] H. Steiner, A. Kobor, L. Gebhard, A wall heat transfer model for subcooled boiling flow, Int. J. Heat Mass Transf. 48 (2005) 4161–4173.
[16] L.Z. Zeng, J.F. Klausner, D. Bernhard, R. Mei, A unified model for the prediction of bubble detachment diameters in boiling systems, Int. J. Heat Mass Transf. 36
(1993) 2271–2279.
[17] N. Zuber, The dynamics of vapor bubbles in nonuniform temperature fields, Int. J. Heat Mass Transf. 2 (1961) 83–98.
[18] J.O. Hinze, Turbulence, McGraw-Hill, New York, 1975.

508
M. Gholinia et al. Case Studies in Thermal Engineering 12 (2018) 497–509

[19] M. Pourfallah, CFD Simulation of Heavy Duty Diesel Engine Combustion with 2000KW, 1800 RPM and Marine Application (M.Sc. Thesis), (2009).
[20] M.G. Sitkei, Heat transfer and thermal loading in internal combustion engines, Akademiai Kiado, 1974.
[21] I.C. Finlay, R.J. Boyle, J.P. Pirault, T. Biddulph, Nucleate and film boiling of engine coolants flowing in a uniformly heated duct of small cross section, SAE
Technical Paper 870032, 1987.
[22] B.E. Launder, D.B. Spalding, The Numerical computation of turbulent flows, Comp. Methods Appl. Mech. Eng. 3 (1974) 269–289.
[23] T. Wintterle, L. Eckart, Y. Egorov, F. Menter, On the Customizing of a Subcooled Boiling Model for Industrial CFD Simulation, 11th International Topical Meeting
on Nuclear Reactor Thermal-Hydraulics (NURETH-11), Avignon, France, October, p. 186, 2005.
[24] Engine Acceptance Test Report, DESA Co. Test Running Department, Issued March, 2007.

509