Propulsion and Power Research 2017;6(3):195–205

HOSTED BY http://ppr.buaa.edu.cn/

Propulsion and Power Research

www.sciencedirect.com

ORIGINAL ARTICLE

Three-dimensional simulation of a novel

rotary-piston engine in the motoring mode
Mohammadreza Khania,⁎, Ghazaleh Esmaeelzadeb

a
Department of Biological Engineering, University of Idaho, Moscow, Idaho, USA
b
Faculty of New Sciences and Technologies, University of Tehran, Tehran, Iran

Received 27 October 2015; accepted 22 January 2016

Available online 4 September 2017

KEYWORDS Abstract In this simulation study, the flow and thermal characteristics of a novel rotary-
piston engine, which is a kind of internal combustion engines, were investigated by
Novel rotary engine;
Computational fluid computational fluid dynamics and the finite volume method. The structure of this engine is
dynamics; different to others, mainly for having 24 cylinders during the motoring mode. As a novel
Swirl ratio; engine, creation of numerical models based on Reynolds average Navier Stokes (RANS)
Flow and thermal simulation and analysis of various speed engines on the flow and thermal fields during intake
characteristics; and compression strokes are the focus of this work. The results were illustrated in term of the
Dynamic mesh motion streamline patterns, in-cylinder temperature and pressure profile, swirl ratio (SR), wall heat
flux, and turbulent velocity fluctuation. The present study indicates that, the mean pressure,
temperature trace, and heat loss from the wall increase when switching to a higher engine
speed. The temperature distribution reveals that the maximum temperature is restricted in the
center of the combustion chamber near top dead center (TDC). Also, the maximum amount of
turbulent velocity and swirl ratio are achieved at the beginning of the intake stroke and near
TDC. It is observed that the obtained numerical results are in general agreement with the
available experimental data.
& 2017 National Laboratory for Aeronautics and Astronautics. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

⁎
Corresponding author. Tel.:+1 (208) 596-078. 1. Introduction
E-mail address: [email protected]
(Mohammadreza Khani). The rotary-piston engine is very efficient in comparison
Peer review under responsibility of National Laboratory for Aeronautics to the conventional internal combustion engines. The rotor
and Astronautics, China. is driven by 24 combustion events in each revolution. Since

2212-540X & 2017 National Laboratory for Aeronautics and Astronautics. Production and hosting by Elsevier B.V. This is an open access article under the
CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

http://dx.doi.org/10.1016/j.jppr.2017.07.005
196 Mohammadreza Khani, Ghazaleh Esmaeelzade

Nomenclature u radial velocity (unit: m/s)

v tangential velocity (unit: m/s)
CFD computational fluid dynamics u' velocity fluctuation (unit: m/s)
rpm revolutions per minute h cam height (unit: mm)
TDC top dead center
aTDC after top dead center Greek letters
bTDC before top dead center (unit: kg)
UDF user defined function ρ density (unit: kg/m3)
Vp mean rotor speed (unit: m/s) θ rotor angle (unit: °)
V flow velocity (unit: m/s) ∀ cell volume (unit: m3)
k turbulent kinetic energy (unit: m2/s2) ε dissipation rate of turbulence (unit: m2/s3)
n engine speed (unit: rpm)
SR swirl ratio

the upper and lower faces of the cam and rotor are 180 The pressure of combustion forces the rotor to turn as the
degrees out of phase, the engine is always balanced and chamber volume increases. The combustion chamber con-
exhibits minimal vibration. tinues to expand, moving the rotor and creating power. At
There are numerous defining characteristics that distin- the moment the peak of the chamber passes the exhaust
guish a rotary-piston engine from a usual conventional port, the high-pressure combustion gases expel the exhaust
reciprocating engine. The rotary-piston engine has far fewer port. As the rotor continues to move, the chamber begins to
moving parts, including the rotor and the 12 vanes. On the contract, forcing the remaining gasses out of the exhaust
other hand, a similar four-stroke piston engine has at least port. Once the volume of the chamber is close to its
40 moving parts, as well as pistons, connecting rods, minimum, the peak of the chamber passes the intake port
camshaft, valves, valve springs, rockers, timing belt, timing and the whole cycle begins again.
gears and crankshaft. According to aforementioned literature review, almost all
The rotor and integral shaft in the rotary-piston engine numerical and experimental studies on performance have
spin constantly in one direction, instead of violently focused on typical internal combustion engines. It is
changing directions like the pistons in a conventional important to note that despite the RadMax Technologies
reciprocating engine do. The rotary-piston engine is entirely constructed this rotary piston engine [1], the numerical
balanced with Combustion chambers on opposite sides of simulation of this engine has never been accomplished yet.
the rotor positioned to annihilate vibrations completely. In this study, the flow and thermal characteristics of novel
The power transmission system in a rotary-piston engine rotary engines are numerically investigated by considering
is also smoother. Each combustion event lasts through 30° the RANS turbulence model. In fact, the temperature
of the rotor's revolution; therefore, 24 combustion events distribution, swirl ratio and the flow pattern inside the
power each revolution of the rotor. This means that the combustion chamber are the key parameters for the
rotary-piston engine delivers continuous power during each combustion simulation [2]. Therefore, the research focuses
rotation of the output shaft. on analyzing the cold flow characteristics and the factors
Each chamber performs the “Otto” engine four-cycle that are important for combustion process.
process: intake, compression, power and exhaust. For each The motoring mode of the conventional internal combus-
rotation of the rotor, the volumes of the 24 chambers tion engines had been investigated numerically in numerous
alternatively expand and contract to draw air into the rotary- studies [3–7]. In numerical studies characteristics of the
piston engine, compress the air, inject fuel at the defined flow motion were investigated in cylinder and piston bowl
moment to make optimum power as the gases expand, and during intake and compression strokes.
then expel the exhaust. Mao et al. [8] calculated the complete simulation of both
The intake phase begins when each chamber moves intake and compression strokes in an axisymmetric engine.
through the intake port. As the intake port is exposed to the They showed the influence of different turbulence models
chamber, the volume of that chamber is nearing its and the numerical precision of the simulations. However, the
minimum. When the rotor moves past the intake port, the numerical simulation results were not validated with experi-
volume of the chamber expands, drawing air into the mental results. Later, Aita et al. [9] investigated the effect of
chamber. By the time the chamber passes the intake port, swirl motion inside the cylinder during a complete intake and
that chamber is sealed off and compression starts. compression stroke. They showed that a strong interaction
When the rotor continues its motion around the housing, was found between the swirling motion and the shape and
the volume of the chamber gets smaller and the air gets position of the piston bowl. The obtained simulation results
compressed. Once the chamber is exposed to the injector, then were compared with experimental data.
the volume of the chamber is again nearing its minimum Based on numerical studies of Chen et al. [10], a
and spontaneous combustion begins. complete intake and compression stroke was performed
Three-dimensional simulation of a novel rotary-piston engine in the motoring mode 197

and their results were validated with experimental data.

They clarified that the calculation was not able to predict
the turbulent velocity accurately due to differences between
the experimental result errors and k-ɛ turbulence model
limitations. Also a similar numerical simulation [11] was
performed for diesel engines with one cylinder and intake
valve.

2. Geometry and numerical methodology

2.1. Geometry and dynamic mesh

Power, weight, and fuel consumption are three important

and influential parameters in the design of internal combus-
tion engines. Rotary-piston engines are a different type of
internal combustion engines in many fields of applications
from aerospace propulsion, gas turbines industry to trans-
portation area [1]. Some of the main advantages of these
engines with respect to the conventional reciprocating
internal combustion engines consist of higher power to
weight ratio, fewer dynamic parts, and lower noise level.
As shown in Figure 1(a), the rotor is partitioned into 24
chambers, 12 on each side of the rotor. Separating each
chamber are 12 axial vanes. The combination of the rotor
and vanes form combustion chambers that act like the
cylinder of a conventional reciprocating engine. The
revolution of the rotor and the axial movement of the vanes
alternatively expand and contract the volume of the
chambers. The output shaft is an integral part of the rotor,
eliminating the need for a crank shaft, which is a common
part of a conventional reciprocating engine. In the rotary-
piston engine, as the chamber follows its path on the cam,
the rotor creates torque in the shaft to drive the load.
Each cam has three ports that are accurately positioned to
facilitate the intake, compression, power, and exhaust
cycles without the use of valves or timing gears. The same
locations are provided on the opposite cam and the
sinusoidal profile is repeated on the front and back cams.
Figure 1(b) shows the fixed locations of the injector ports,
besides the intake and exhaust ports.
Compression is achieved when the vanes, which form the
sides of each combustion chamber, follow a sinusoidal pin
track in the housing as shown in Figure 1(c). The ends of
the housing are fitted with titanium front and back cams. On
each rotation of the rotor, the vanes move sinusoidally from
minimum length to maximum height of the cam (approxi-
mately 23 mm) for the sides of each combustion chamber.
The vanes always stay in contact with the cam and the
housing wall, forming 24 sealed combustion chambers as
shown in Figure 1(d).
The computational domain includes intake ports, cam,
lateral walls, rotor and the bowl as shown in Figure 2. The
number of volume cells varies during simulation from Figure 1 Schematics of investigated rotary-piston engine.
30,000 cells in TDC, to 80,000 cells in 90° bTDC. Each (a) Schematic of rotor, (b) schematic of cam, (c) schematic of
face of the combustion chamber has been meshed with rotary-piston engine assembly, and (d) schematic of assembled
specific size and scheme to reduce dynamic mesh motion rotary-piston engine without housing.
198 Mohammadreza Khani, Ghazaleh Esmaeelzade

Table 1 Computational details of the engine.

Displacement volume 1.08 lit

Cam maximum height ðhmax Þ 23 mm
Outer radius 127 mm
Inner radius 89 mm
Cylinder number 24
Bowl depth 3.5 mm
Static bumping clearance 0.5 mm
Compression ratio 1:21
Intake valve opening 178° bTDC
Intake valve closure 88° bTDC
Exhaust valve opening 92° aTDC
Exhaust valve closure 178° aTDC
Figure 2 Computational domain. Maximum rotor angle ðθmax Þ 360°
Minimum rotor angle ðθmin Þ 0°
time and achieve a high quality grid. The smaller mesh size
was created on an intake port and lateral walls with the
scheme named map split. The mesh was expanded on these
zones during the intake stroke. Tetrahedral mesh was used
for the combustion chamber volume due to the complex
pseudo-sinusoidal mesh motion of the cylinder. The opera-
tional and geometrical characteristics of this engine are
presented in Table 1. The calculation in this paper was
performed under the operational condition mainly used for
diesel engines.
The dynamic mesh motion was presented to FLUENT
with a user-defined function (UDF) during the simulation.
Node displacement along Z-axis (dz) was calculated using
Eq. (1) for the rotor, lateral walls, and intake and exhaust
ports. The Z displacement was set to zero during TDC and Figure 3 Dynamic mesh motion during one complete mechanical cycle.
90° bTDC stages due to the distinct shape of the cam, so the
cam apex was constant in these phases. Node displacement working fluid (air) is Newtonian and compressible, and
along the angular direction (ds) was calculated using Eq. thermal radiation and viscous dissipation effects are
(2). The rotational speed (ω) was constant and equal to neglected in the energy equation.
engine speed for the rotor and lateral wall nodes during the In view of the above assumptions, the governing
whole cycle. However, it was altered with Eqs. (3) and (4) equations of the problem are the continuity, momentum,
on intake and exhaust ports during the opening and closing and energy equations and the transport equations for the
phases, respectively. In addition, the cam surface was turbulent kinetic energy and dissipation added through the
deformed with a faceted method mesh generation by the turbulence model. The frequently used RNG k–ɛ model
rotation of lateral walls. At the end of the intake stroke, with the wall function was used to simulate the in-cylinder
once the intake port was closed, the intake port was merged turbulence flow field [12–14]. With this model, the turbu-
with the cam, so that only the cam surface was considered. lence is considered as isotropic; however, this hypothesis
The cylinder condition is shown in Figure 3 at different seems reasonably correct only near the TDC in the
positions during one complete mechanical cycle. compression stroke, as observed by Bopp et al. [15] and
dz ¼  2ñhmax ñ sin ð2θÞ ñωñtime ð1Þ Brandstätter et al. [16] and for quiescent types of combus-
tion chambers. In general, this study shows that in-cylinder
ds ¼ rñωñdt ð2Þ CFD predictions yield reasonably accurate results that allow
improving the knowledge of the air flow characteristics
ðθnode −θmin Þ
ω ¼ ωrotor ñ ð3Þ during the intake and compression strokes. Since the air
ðθmax −θmin Þ motion at TDC strongly influences the development of the
ðθnode −θmax Þ injection and combustion processes in the combustion
ω ¼ ωrotor ñ ð4Þ chamber of diesel engines, the analysis presented here
ðθmin −θmax Þ
represents a first step towards their understanding. In this
sense, CFD represents an efficient design tool to develop
2.2. Mathematical modeling less polluting and more efficient direct-injection diesel
engines. Although it has been known that LES model
In this study, an unsteady and turbulent three-dimen- shows better performance than RNG k-ɛ model in the
sional flow is considered. It is also assumed that the numerical simulation of internal combustion engines, the
Three-dimensional simulation of a novel rotary-piston engine in the motoring mode 199

large computer time requirement, storage and analysis of that was achieved for the first calculation (case 1) was about
the large data sets that are generated are significant practical 30,000 cells at TDC (intake stroke). The second case (case
problems for the LES turbulence model. Furthermore, the 2) was considered with a smaller mesh size on intake port
convergence in LES model is not applicable due to complex and lateral walls that were expanding during the simulation
dynamic mesh motion and coarsen mesh generation during so the number of cylinder volume mesh increased (42,000
the intake stroke when the intake port is opening and the cells). Finally, in case three, the total number of mesh on
mesh resolution increases temporally. Mass, momentum cylinder surface was doubled and the volume mesh size was
and energy conservation equations have been discretized by halved so the total number of meshes increased (approxi-
the control volume technique and FLUENT commercial mately 55,000 cells).
software [17]. Flow characteristics such as non-dimensional radial
velocity and swirl ratio are presented in Figure 4 at a point
2.3. Boundary and initial conditions located 3 mm from the rotor bowl and at a radius of
107 mm from the engine axis for different cases. The results
In order to study the in-cylinder flow field characteristics show that there are no wide differences between these cases
and validation of numerical results with the experimental during the compression stroke. Case 1 predicted adequately
data, CFD simulation was performed for the complete well the trend of swirl ratio and non-dimensional radial
mechanical cycle in the motoring mode. The pressure- velocity with maximum error of 5% achieved near TDC in
velocity coupling has been established through the PISO comparison with case 3. So, in this study, case 1 was used
scheme [18]. The PRESTO spatial discretization was used for all simulations.
for the pressure and the second-order upwind method and As there is no physical case for the novel rotary-piston
central difference method have been used for calculating the engine, the validation was performed with the experimental
convective and diffusive terms, respectively. The transient measurement obtained by Pariotis et al. [25] for a recipro-
formulation was implicit and the time stepping method was cating engine with the same compression ratio. The experi-
adaptive due to the cylinder location. The time step was mental measurements were conducted for 20 motoring
specified approximately 0.01 angle degrees per time step cycles and data recorded for every 1° CAD at each engine
during the intake stroke when the intake port was slightly speed. The mean pressure trace achieved for 3000 rpm with
open because of high skew mesh and local velocity. After a maximum error of 1% was used for the validation in this
the intake port was completely opened and during the study. Variation of the cylinder mean pressure with rotor
compression stroke, the time step enlarged (approximately angle for different engine speeds is shown in Figure 5. As
doubled) due to the lower mesh skew and local velocity. explained before in Section 2.1, the motion of each cylinder
Also at the end of the compression stroke due to the small in the rotary-piston engine is more similar to the conven-
clearance between the rotor and cam and the importance of tional reciprocating engine than the wankel rotary engine.
the squish effect, the time step reduced to comply with the
stability standard [19].
The CFD calculations started at TDC (intake stroke) with
an initially quiescent flow due to discounting the remaining
swirl from the exhaust stroke. In addition, the temperature
and pressure values, which were considered homogeneously
constant at the beginning of the intake stroke, were attained
from the experimental data [20]. An initial value of 10%
was set for turbulent intensity to consider the fully turbulent
flow and the turbulent length scale calculated from Prandtl
mixing length model [21]. The parameters obtained for this
section were developed by Inoue et al. [22] and Henriot
et al. [23]. Constant pressure inlet was set for the intake port
so the dynamic effect was not accounted in the simulation.
Constant temperature was set for thermal condition of
cylinder walls following a correlation achieved by Salavert
[20] that depends on the mean rotor speed. This correlation
follows the proposed model by Woschni [24].

3. Results and discussions

In order to investigate grid independency in the numer-
ical results, three CFD calculations of complete intake and
compression stroke were considered. The number of meshes Figure 4 Effect of grid size on swirl ratio and radial velocity.
200 Mohammadreza Khani, Ghazaleh Esmaeelzade

The back and forth motion of the combustion chamber that

was produced by the combination of rotary motion and
sinusoidal shape of the cam produced the similar flow
pattern that could be seen in the conventional reciprocating
engine. Even though the rotary motion of the cylinder has
some effect on the flow pattern, as it was concluded from
the results, the total flow pattern and turbulence character-
istic are similar to conventional reciprocating engines. With
this in mind, one typical numerical simulation was con-
ducted for this conventional reciprocating engine and the
results were validated with the presented experimental data
in Pariotis et al. [25]. Formerly, exactly the same techniques
of mesh generation, dynamic mesh and mathematical
modelling were used for numerical simulation of the
rotary-piston engine. Although the numerical simulation
predicted the peaks of the mean pressure profile with a
maximum difference of 0.2 bar in comparison with the Figure 7 Wall-heat transfer calculated for various engine speeds.
experimental results, the experimental results vary in a
wider range because of physical and geometrical differences in Figure 5, the cylinder mean pressure increased for an
between the reciprocating and the rotary engine. As shown augmentation in the rotor speed.
The cylinder mean temperature was calculated during the
compression stroke for four engine speeds as shown in
Figure 6. With increasing the engine speed, cylinder mean
temperature increases. The maximum difference between
the peaks of each case is approximately equal to 40 K and
remains practically constant during the expansion stroke for
different engine speeds. As shown in Figure 5, the CFD
calculation accurately predicted the mean pressure during
the compression stroke. Therefore, it could be concluded
that the calculated mean gas temperature for different
engine speeds was predicted practically well.
Heat transfer over cylinder walls is an important para-
meter, which has a great effect on the performance of
internal combustion engines. Figure 7 presents the calcu-
lated wall heat transfer for different engine speeds. There is
a large difference between maximum values of heat transfer
at TDC. However, there is a similar trend for wall heat
losses at different engine speeds. The in-cylinder gas
Figure 5 Effect of rotor speed on the cylinder mean pressure. temperature and velocity increase when switching to higher
engine speeds which causes higher heat losses over the
combustion chamber walls during the compression stroke.
The numerical results for heat losses should be in accordance
with the physical process inside the combustion chamber. The
convection and conduction heat transfer from the hot core
inside the combustion chamber to periphery cells near the
cylinder wall was calculated from a 3-D numerical simulation.
According to Annand's law [26], the boundary cells transfer
heat to cylinder walls by convection. The physical process,
which was predicted correctly by considering both the
temperature and the flow field inside the combustion chamber,
affects heat losses over the cylinder walls, which confirms heat
and mass transfer inside the cylinder.
The temperature variation inside the cylinder correspond-
ing to 10° bTDC for 7000 rpm is presented in Figure 8(a).
The higher temperature was restricted in the center of the
cylinder and near the rotor bowl with a maximum value of
Figure 6 Effect of rotor speed on the cylinder mean temperature. approximately 1080 °C. The gas temperature decreased near
Three-dimensional simulation of a novel rotary-piston engine in the motoring mode 201

Figure 8 Variation of cylinder temperature (Kelvin) at (a) 10° bTDC

and (b) 20° aTDC.

Figure 9 Temporal distribution of (a) mass averaged swirl ratio and

the wall cylinder due to the heat transfer from the hot gas to (b) turbulent velocity.
cylinder cold walls. The temperature contour inside the
cylinder at 20° aTDC is presented in Figure 8(b). Hot core near the 90° bTDC when the cylinder has the maximum
was restricted inside the extended portion of the cylinder volume. During the end of the intake stroke, the swirl ratio
between the cam and bowl. Similar to Figure 8(a), the in- drops slowly due to the reduction of intake velocity into the
cylinder gas temperature decreased near to the cylinder cylinder. This reduction trend lasts near to 30° bTDC
walls. because of the cylinder wall friction. In the second phase
of the compression stroke, the swirl ratio increases when the
cylinder volume decreases. The in-cylinder flow is acceler-
3.1. Characteristics of intake stroke ated to maintain its angular momentum inside the small
rotor bowl size. After the swirl ratio reaches its maximum at
In this section, the flow characteristics during the intake TDC and during the expansion phase, the swirl ratio
stroke were investigated under the motoring condition decreases rapidly as a result of the reverse squish effect
independently of the engine speed. Therefore, the turbulent and the cylinder wall friction.
velocity field and the mean velocity are normalized by the Similar to the swirl ratio curves, the turbulent velocity
mean rotor speed. It is confirmed that the flow character- rose at the beginning of the intake stroke due to the shear
istics inside the cylinder are independent of the engine stresses related to the jet entering from the intake port.
operational condition [10,15,27]. This concept makes the Afterward, the cylinder turbulent velocity decreased up to
comparison of the flow fields in different engine speeds and near the end of the compression stroke, rose up to TDC, and
for different locations easy. decreased again in the expansion phase.
The mass averaged non-dimensional turbulent velocity Figure 10 presents the in-cylinder turbulent velocity and
field and the swirl ratio calculated by Eqs. (5) and (6) mass-averaged swirl ratio along the z-axis (distance
during the complete motoring condition are presented in between rotor surface and cam) during the intake stroke.
Figure 9. These parameters are the most proper variables to The small values of swirl ratio at the beginning of the intake
characterize the in-cylinder flow. stroke indicate that the small volume between the cam and
n the rotor is not adequate to progress a vortical flow inside
∑ virðiθÞ ∀ðθÞi ρðθÞi the cylinder. During the intake stroke, the strong interaction
i
SRðθÞavg ¼ n ð5Þ between intake jets impinged on the rotor lateral wall
60 ∑ ∀ðθ Þi ρðθ Þi
2πn
leading to the appearance of a twin vertical flow inside
i
the cylinder shown in Figure 11(a). Figure 11(b) presents
n 2 1=2 the turbulent velocity contour along the same surface. The
∑ 3 ki ∀ðθÞi ρðθÞi maximum values of turbulent velocity take place near the
u′ðθÞavg i
¼ n ð6Þ cam and the lateral wall where the impinging jets hit the
Vp
V p ∑ ∀ðθÞi ρðθÞi rotor. Streamlines along a radial plane are located at a
i
radius of 107 mm from the engine axis and z-plane surfaces
The results reveal that the swirl ratio is generated at the (Figure 12), and confirm that there are no vertical and
beginning of the intake stroke and reaches its maximum planner structures in these sections.
202 Mohammadreza Khani, Ghazaleh Esmaeelzade

Figure 11 (a) Streamlines and (b) turbulent velocity field in the axial
section at 165° bTDC.

Figure 12 Streamlines in (a) radial and (b) z-plane surfaces at 165°

bTDC.

shown in Figure 13(b) along the same surface. The

maximum values reached the upper region of the cylinder
near the cam and the intake port where the jet impinges.
However, the turbulence level declined at the bottom of the
cylinder and inside the bowl. Finally, at the end of the
intake stroke, the twin symmetry vortex perceived earlier
deformed due to a lower velocity of the flow entering the
combustion chamber.
Figure 10 Mass averaged axial distribution of (a) swirl ratio and
(b) turbulent velocity during intake stroke in the cylinder.
3.2. Compression stroke

Figure 13(a) presents the expanded twin vortex inside the The shear stresses and velocity fluctuation at the upper
combustion chamber at 125° bTDC, when the cylinder zone of the cylinder declined when the intake port closed at
height increased and the intake port was completely the end of the intake stroke. Figure 14 presents the turbulent
exposed. The jet impinging on the rotor surface bounced velocity fluctuation and swirl ratio along the z-axis during
back axially through the lateral wall and formed a symmetry the compression stroke. The twin vortex attained from the
twin clockwise and counterclockwise vortex inside the intake stroke remained inside the cylinder during the first
combustion chamber. The turbulent velocity contour is stage of the compression stroke. The downward sinusoidal
Three-dimensional simulation of a novel rotary-piston engine in the motoring mode 203

Figure 13 (a) Velocity streamlines and (b) turbulent velocity

fluctuation field at 125° bTDC.

motion of the cam during the compression stroke prompted

a small increase of the swirl ratio near the cam surface. The
vertical flow became approximately homogenous inside the
cylinder near the end of the compression stroke at 20°
bTDC.
The turbulent velocity fluctuation inside the cylinder
declined along the cylinder height with the same trend for
the rotor angle from 110° to 160°. At the end of the intake
stroke, the turbulent velocity fluctuation produced by
annular port jets during the intake stroke became uniform
inside the combustion chamber.

3.2.1. Tangential velocity field

The swirl motion inside the cylinder is studied in this
section by considering the solid body rotation assumption
near TDC when the squish effects are unimportant [28]. The
tangential velocity profiles calculated by interpolation along
seven radial lines at TDC are presented in Figure 15. The Figure 14 (a) Axial distribution of mass averaged swirl ratio and
radial lines are located at two different sections inside the (b) turbulent velocity fluctuation during the compression stroke inside
bowl. The first surface is located at 1.5 mm from the cam the cylinder.
surface and the other surface is located at half depth of the
bowl (see Figure 16). 3.2.2. Radial velocity field
The radial motion interaction with the swirl bends the The centrifugal flow convinces inside the cylinder during
velocity profile on the section near the cam. The tangential the compression stroke because of the cylinder rotation. The
velocity on the lower surface is linear because of the lower extent of the radial flow is dependent on the clearance space
squish effects. However, the tangential velocity drops at the between the rotor and cam near the TDC and the ratio of
periphery of the cylinder. The solid body rotation assump- cylinder width to the engine diameter. The temporal
tion may be valid at the lower part of the cylinder due to a variation of the radial velocity integrated on the rotor
weaker tangential flow. However, this is not acceptable for surface around TDC is presented in Figure 17. The
the periphery of the cylinder and at the top of the cylinder maximum value of the radial velocity is achieved near
where the squish effect is important. 10° before TDC. The values at the end of the compression
204 Mohammadreza Khani, Ghazaleh Esmaeelzade

Figure 18 Velocity streamlines at TDC.

Figure 19 Non-dimensional turbulent velocity field at TDC.

pattern from the intake stroke because of the lower squish

effect.

3.2.3. Turbulence field

Figure 19 presents the turbulent velocity fluctuation field
inside the cylinder at TDC with the maximum turbulence
level of approximately 0.062 near the periphery of the
cylinder. The turbulence velocity field is considered homo-
Figure 15 Tangential velocity profiles at TDC along the radial line. geneous inside the cylinder. The turbulent velocity level
during the intake stroke is two times the turbulence at TDC
during the compression stroke because of the intake port
impinging jet.

4. Conclusion
The novel rotary-piston engines were simulated in a
motoring condition for various speeds to properly achieve
an insight from the physical process inside the cylinder. The
experimental results that were measured for a conventional
reciprocating engine with the same compression ratio
validated the CFD calculation; however, it slightly over-
estimated the pressure trace near TDC. The spatial tem-
perature distribution, which was achieved inside the
cylinder, showed that maximum temperature was restricted
in the center of the cylinder and near the rotor bowl. In
addition, it was observed that the mean temperature and
Figure 16 Points located along seven radial lines in different pressure inside the cylinder increased at a higher engine
segments (1 mm and hb/2) of the cylinder. speed.
Also the heat losses from the cylinder walls were studied
in this paper and the maximum amount of heat loss was
predicted at TDC for each engine speed. The non-dimen-
sional turbulence velocity field and the swirl ratio were
investigated inside the cylinder during the intake and
compression strokes. The maximum amount of turbulent
velocity was achieved at the upper zone of the cylinder near
the cam at the beginning of the intake stroke. The swirl
level increased during the intake stroke and reached the
Figure 17 Radial velocity profile at the rotor surface. maximum near the end of the intake stroke. The twin
symmetry vortex detected at the beginning of the intake
stroke are greater than the first stage of the expansion stroke remained inside the cylinder during the compression
stroke. Figure 18 presents the velocity streamlines inside the stroke. However, the swirl ratio declined at the beginning of
cylinder and indicates that the flow preserved its symmetric the compression stroke when the intake port closed. The
Three-dimensional simulation of a novel rotary-piston engine in the motoring mode 205

turbulent velocity field became homogenous during the [11] B. Dillies, A. Ducamin, L. Lebrere, F. Neveu, Direct
compression stroke when the intake port closed. The injection diesel engine simulation: a combined numerical
compression stroke results were analyzed comprehensively and experimental approach from aerodynamics to combus-
near TDC. tion, SAE Technical Paper 970880, 1997.
Interested researchers can pursue further studies by [12] S.A. Orszag, V. Yakhot, W.S. Flannery, F. Boysan, D.
Choudhury, J. Maruzewski, B. Patel, Renormalization group
simulating the novel rotary-piston engine under the firing
modeling and turbulence simulations, in: R.M.C. So, C.G.
condition and analyzing the parameters affecting the engine Speziale, B.E. Launder (Eds.), Near-Wall Turbulent Flows,
efficiency. This will help to develop a new generation of 1993, pp. 1031–1046.
internal combustion engines. [13] V. Yakhot, S. Orszag, S. Thangam, T. Gatski, C. Speziale,
Development of turbulence models for shear flows by a
Acknowledgements double expansion technique, Phys. Fluids A 4 (7) (1992)
1989–1993.
The authors would like to thank IUST Energy, water and [14] V. Yakhot, S.A. Orszag, Renormalization group analysis of
turbulence, I: basic theory, J. Sci. Comput. 1 (1) (1986) 3–51.
environment research center for its support.
[15] S. Bopp, C. Vafidis, J.H. Whitelaw, The effect of engine
speed on the TDC flowfield in a motored reciprocating
References engine, SAE Technical Paper 860023, 1986.
[16] W. Brandstätter, R.J.R. Johns, G. Wigley, The effect of inlet
[1] Rad Max Engine, High Efficiency Diesel Engine, URL: port geometry on in-cylinder flow structure, SAE Technical
〈http://www.regtech.com〉. Paper 850499, 1985.
[2] J.B. Heywood, Fluid motion within the cylinder of internal [17] Ansys Fluent 12.0 Theory Guide, Ansys Inc., 2009.
combustion engines – The 1986 freeman scholar lecture, [18] R.I. Issa, Solution of the implicitly discretised fluid flow
ASME J. Fluids Eng. 109 (1) (1987) 3–35. equations by operator-splitting, J. Comput. Phys. 62 (1)
[3] P.S. Abraham, X. Yang, S. Gupta, T.W. Kuo, D.L. Reuss, V. (1986) 40–65.
Sick, Flow-pattern switching in a motored spark ignition [19] H.K. Versteeg, W. Malalasekera, An Introduction to Com-
engine, Int. J. Engine Res. 16 (3) (2015) 323–339. putational Fluid Dynamics: The Finite Volume Method,
[4] A. Agrira, D.R. Buttsworth, M.A. Said, Instantaneous heat Second ed, Pearson Education, 2007, pp. 40-90.
flux simulation of a motored reciprocating engine: unsteady [20] J.M. Salavert, Estudio teórico-experimental de la combustión
thermal boundary layer with variable turbulent thermal en motores diesel de inyección directa, de 2 litros de
conductivity, ASME J. Heat. Transf. 136 (3) (2014) cilindrada unitaria fuertemente sobrealimentados, 1st ed.,
031701–031709. Universidad Politécnica de Valencia, Spain, 1995.
[5] L. Fonseca, R. Coelho, R. Braga, B. Pena, R.M. Valle, [21] B.E. Launder, D.B. Spalding, Lectures in Mathematical
Numerical analysis of tumble, cross tumble and swirl ratios Models of Turbulence, Academic Press, London, 1972.
in a single cylinder research engine, Adv. Mater. Res. 1016 [22] S. Inoue, K. Kobayasi, F. Akatsuka, E. Fukumori, Calcula-
(2014) 617–623. tion of the in-cylinder flow and heat transfer in DI and IDI
[6] J.P. Keskinen, V. Vuorinen, O. Kaario, M. Larmi, Large diesel engines, SAE Technical Paper 890667, 1989.
eddy simulation of a piston-cylinder assembly: the sensitivity [23] S. Henriot, J.F. Le Coz, P. Pinchón, Three dimensional
of the in-cylinder flow field for residual intake and in- modelling of flow and turbulence in a four-valve spark
cylinder velocity structures, Comput. Fluids 122 (2015) ignition engine-comparison with LDV measurements, SAE
123–135. Technical Paper 890843, 1989.
[7] T. Wang, W. Li, M. Jia, D. Liu, W. Qin, X. Zhang, Large- [24] G. Woschni, A universally applicable equation for the
eddy simulation of in-cylinder flow in a DISI engine with instantaneous heat transfer coefficient in the internal combus-
charge motion control valve: proper orthogonal decomposi- tion engine, SAE Technical Paper 670931, 1967.
tion analysis and cyclic variation, Appl. Therm. Eng. 75 [25] E.G. Pariotis, G.M. Kosmadakis, C.D. Rakopoulos, Com-
(2015) 561–574. parative analysis of three simulation models applied on a
[8] Y. Mao, M. Buffat, D. Jeandel, Simulation of the turbulent motored internal combustion engine, Energy Convers.
flow inside the combustion chamber of a reciprocating engine Manag. 60 (2012) 45–55.
with a finite element method, ASME J. Fluids Eng. 116 (2) [26] W.J.D. Annand, Heat transfer in the cylinders of reciprocat-
(1994) 363–369. ing internal combustion engines, Proc. Inst. Mech. Eng. 177
[9] S. Aita, A. Tabbal, G. Munck, N. Montmayeur, Y. Takenaka, (1963) 973–996.
Y. Aoyagi, S. Obana, Numerical simulation of swirling port- [27] P.O. Witze, Measurements of the spatial distribution and
valve-cylinder flow in diesel engines, SAE Technical Paper engine speed dependence of turbulent air motion in an IC
910263, 1991. engine, SAE Technical Paper 770220, 1977.
[10] A. Chen, A. Veshagh, S. Wallace, Intake flow predictions of [28] F.R. Menter, Two-equation eddy-viscosity turbulence models
a transparent DI diesel engine, SAE Technical Paper 981020, for engineering applications, AIAA J. 32 (8) (1994) 1598–
1998. 1605.