Experimental study and numerical simulation on in-cylinder flow of small motorcycle engine

Applied Energy 255 (2019) 113863

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Experimental study and numerical simulation on in-cylinder flow of small motorcycle engine

T

Bambang Wahonoa,b, Ardhika Setiawana, Ocktaeck Limc,

⁎

a

Graduate School of Mechanical Engineering, University of Ulsan, San 29, Mugeo2-dong, Nam-gu, Ulsan 44610, South Korea Research Centre for Electrical Power and Mechatronics – Indonesian Institute of Sciences, Jl. Cisitu No 154/21D, Bandung 40135, Indonesia c School of Mechanical Engineering, University of Ulsan, San 29, Mugeo2-dong, Nam-gu, Ulsan 44610, South Korea b

HIGHLIGHTS

experimental study and numerical simulation on the in-cylinder flow of small motorcycle engine was investigated. • The evolution of the in-cylinder flow small motorcycle engine was illustrated well in this study. • The • Some flow characteristics in-cylinder small motorcycle engine was analyzed well and detail in this study. ARTICLE INFO

ABSTRACT

Keywords: Small engine In-cylinder flow Steady-state flow Pressure drop

This study aimed at experimental and numerically investigating in-cylinder flows of a small motorcycle engine under steady-state conditions. The experiment was conducted in the engine head for a variety of fixed valve lifts at two pressure drops (300 and 600 mmH2O). Besides, this study attempted to analyze the characteristics of incylinder flows in a small engine by applying CFD methods. Looking at the results, there was a good agreement between the results of experiment and simulation in terms of flow coefficient, air flow rate and discharge coefficient at two pressure drops. In both horizontal and vertical planes, both increased valve lift and pressure drop delivered increased velocity and vorticity magnitude. An increase in pressure drop at the beginning of valve lift opening appeared to have no effect on the swirl ratio until the valve lift reached 5 mm. However, the swirl ratio got a 25% reduction when the valve lift reached 6.25 mm. An increase in pressure drop at the intake stroke did not deliver a significant effect on the tumble ratio and accumulated air mass. However, the accumulated air mass increased 3.77% at compression stroke. An increase in pressure drop delivered a significant effect on turbulent kinetic energy (TKE), turbulent length scale and turbulent kinematic viscosity. TKE reached its peak (200%) at 470 °CA, while turbulent length scale and turbulent kinematic viscosity reached their peaks at 590 °CA where the intake valve was almost closed. The increase in turbulence in fact produced a more homogeneous incylinder air-fuel mixing. Besides, the increase in turbulence directly increased the rate of fire propagation. Further study would be expected to focus on modifying the design of intake port for improving the air flow characteristics of small engines. Then, this study was expected to reduce the number of experiments required for investigating optimized parameters in small engines.

1. Introduction Currently, the primary challenges for advances in the design of internal combustion engines are how to improve engine performance and meet stringent emission standards [1]. In the past few decades, engine efficiency has been vastly improved, while exhaust emissions have been greatly reduced by various methods, including exhaust gas recirculation (EGR) [2], higher compression ratio [3], fuel injection strategy [4] and

lean burn combustion [5]. However, demanded emission requirements continue to tighten. Among others, results of combustion processes are a critical contributor inside any combustion engines to meet these evertightened standards. Technically, combustion process is affected by the motion of air inside an engine cylinder [6]. The intake process controls various critical things of the in-cylinder airflow. In fact, structures of the flow are complex. Understanding in-cylinder airflow in a combustion engine is therefore required to explore ways for improving the

⁎ Corresponding author at: Department of Mechanical and Automotive Engineering, University of Ulsan, San 29, Mugeo2-dong, Nam-gu, Ulsan 680-749, South Korea. E-mail address: [email protected] (O. Lim).

https://doi.org/10.1016/j.apenergy.2019.113863 Received 9 July 2019; Received in revised form 25 August 2019; Accepted 3 September 2019 0306-2619/ © 2019 Elsevier Ltd. All rights reserved.

Applied Energy 255 (2019) 113863

B. Wahono, et al.

In the literature, studies on in-cylinder flows in the large engine using steady flow bench and CFD method have been widely conducted. El-Adawy et al. [23] conducted an experiment to evaluate in-cylinder flows of large engine by applying two methods, i.e., torque meter and paddle wheel steady flow bench. The study was performed in the fourvalve pent roof cylinder head of a diesel engine with a large bore (92.5 mm). They reported that a fairly good agreement level between the two methods for up to 6 mm valve lift. After the lift threshold, however, flows in the steady-state flow bench began to choke, delivering effects to the flows and the discharge coefficient value. Meanwhile, Fang and Singh [12] applied a commercial CFD “CONVERGE” to estimate the airflow for steady-state flow tests. They reported increases in mass flow rate as being proportional to increases in valve lift. In some cases, however, the mass flow rate is possible to get reduced when the valve lift increases due to segregated movements of the turbulent flow in the valve seat. In another work, Binjuwair and Ibrahim [25] evaluated the in-cylinder flow fields of a single-cylinder optical large engine with 88 mm bore by using realizable k-ε and Reynolds stress turbulence model (RSTM) against experimental PIV data. Their results showed the realizable k-ε turbulence model as being able to predict qualitative trends of kinetic energy profile of the turbulence but poorly able to predict the average velocity of RSTM fluctuations. Furthermore, Abdul Gafoor and Gupta [18] applied another commercial CFD “AVL FIRE” to investigate the effects of swirl ratio in conjunction to varied piston bowl geometries in diesel engines. The simulation was conducted in a single cylinder diesel engine with a large bore (85.73 mm). They concluded an increase in turbulent kinetic energy (TKE) along with an increase in the swirl ratio as a requirement for a better combustion. Moreover, combustion has been stated to greatly be influenced by geometry of the bowl with shallow depth and large diameter. Mohammadebrahim et al. [26] investigated flow characteristics in the intake port of a four-valve engine with 78.6 mm bore at a steady-state flow bench. Their study aimed at determining the effects of intake adaptor, paddlewheel diameter, adaptor length and diameter, test pressure, asymmetric valves and lifting adaptor roughness on the intensity of swirl and the measured flow coefficients. They reported the intensity of swirl to depend on the adaptor length and pressure. They also concluded intake adapter as one of the most effective parameters for testing flow coefficients. Yang et al. [27] have investigated the in-cylinder flow of a large-bore engine by using six models of modified flow box in steady-state port flow simulations. It reported estimated mass-flow rates to agree within 1% with measured data between intermediate and high valve lifts. At low valve lifts, slightly overpredicted mass-flow rate was observed. Perini et al. [14] applied KIVA to investigate fluid flows in the cylinder of a diesel engine with 82 mm bore. The results were helpful to investigate structures of the flow in an engine cylinder during ignition process in partial combustion mode. Meanwhile, Jemni et al. [16] conducted a numerical study and experiments study on the effects of a modified intake manifold on the in-cylinder flows. They used a heavy-duty engine with a large bore (137 mm). They found it to provide good knowledge on the structure of in-cylinder flow for testing the performance of a modified manifold. Experiment and simulation results confirmed the benefits of a modified manifold for the in-cylinder flows. Furthermore, Idris Saad and Saiful Bari [28] have used a guide vane swirl and tumble device (GVSTD) in compression ignition (CI) engines with large bore (104 mm) by applying vegetable oils to improve characteristics of the in-cylinder airflow. They reported parameters such as in-cylinder pressure, TKE and velocity to enhance the combustion by breaking-up more fuel molecules, resulting in a better mixing of vegetable oil fuel and air. Then, Yang et al. [29] investigated the intake port of an engine with 86 mm bore by comparatively using steady-state flow bench and numerical simulation. They reported the estimated discharge coefficients and swirl-tumble index of steady flow bench to have a good agreement with the results of numerical simulation. In above-mentioned works, researchers experimentally and numerically analyzed the in-cylinder flows for large engines only.

Fig. 1. Cylinder head of a small gasoline engine. Table 1 Specification of engine. Parameter

Value

Bore Stroke Displacement Diameter intake valve Diameter exhaust valve Maximum lift of intake valve Maximum lift of exhaust valve

57 mm 48.8 mm 125 cm3 22.03 mm 18.75 mm 6.46 mm 6.42 mm

performance of internal combustion engine. Two types of flow have been known to represent the in-cylinder air motion, i.e., swirl flow and tumble flow [7]. The swirl flow refers to a rotational movement of air related to the axis of an engine cylinder. A swirling flow is generated during an intake process by directing the airflow into an engine cylinder tangentially. On the other hand, the tumble flow refers to a perpendicular movement of air related to the axis of an engine cylinder [8]. In general, both flow types aim at increasing the mixing process inside an engine cylinder and building high turbulences at the top dead center in a compression stroke. Then, the ultimate goal is an improved management of combustion process, consequently enhancing combustion efficiency and reducing exhaust emissions of engine cylinder. Prior works have studied in-cylinder airflows by applying various methods, i.e., numerical simulations (e.g., AVL FIRE [9], STAR-CD [10,11] ANSYS-CFX [7], CONVERGE [12,13], KIVA [14], GT-Power [15], FloWorks [16], combined AVL FIRE-KIVA-3V [17], combined AVL FIRE-SIMPLE algorithm [18], combined KIVA- STAR-CD [19], etc.) and experimental methods (e.g., particle image velocimetry (PIV) [20,21], laser doppler anemometry (LDA) [22], steady-state flow bench [23], etc.). Numerical simulation offers an advantage over experiments in terms of producing more detailed flow information for all flow fields. To extend understandings on in-cylinder air motion of an engine, the flow is therefore simulated by applying computational fluid dynamics (CFD). Still, experiments are required as a preliminary verification for the steady flow problem. In practices, the preliminary verification is taken by using a steady flow bench [24], by which the measurement is a simplified for studying engine performance. Steady flow bench has become a standard measurement to characterize in-cylinder engine airflows in the automotive industry. Assumptions used in this method include a presumed close relation between steady flow measurements and the results achieved from a running engine cylinder. 2

Applied Energy 255 (2019) 113863

B. Wahono, et al.

Fig. 2. Schematic of the steady-state tests.

Technically, large engines contain enough spaces to more easily visualize and analyze in-cylinder flow motion than small engines. Besides, these works provide less comprehensive and less detailed explanations. There is also a lack of focuses on the characteristics of incylinder flow motion for small engines. A comprehensive investigation is therefore required to deliver a better understanding on the characteristics. This work is expected to act as a foundation to visualize incylinder flow field in small engines, especially motorcycles. This study was aimed at experimentally and numerically investigating in-cylinder flows of small motorcycle engines under steadystate conditions. This work was conducted on the engine head for several fixed valve lifts at two pressure drops (300 and 600 mmH2O). Besides, this study analyzed characteristics of the in-cylinder flow for small engines under motoring condition using CFD methods. Those characteristics included swirl ratio, TKE, turbulent length scale, tumble ratio, turbulent kinematic viscosity and in-cylinder air mass and flow patterns (velocity vector and vorticity magnitude) during both intake and compression strokes.

ports. In this study, pressure drops used to perform the tests was chosen based on the criteria of Reynolds number. Two pressure drops chosen to confirm the flow motion were 300 and 600 mmH2O. Then, a torque meter for steady-state flow bench was taken to measure the flow. 2.2. Steady-state flow bench test The steady-state flow bench is a tool to evaluate the resistance of a test piece, e.g., a cylinder head, against airflow. Besides, it is easy to apply and inexpensive to predict the ability of a cylinder head to convert the linear motion of intake flow to a rotational motion. Fig. 2 exhibits a schematic of the steady-state flow bench taken in this study. First, the air was driven by a blower fan to pass through surge tank, pressure-relief valve and laminar flow rate, and eventually entering the engine head. The intake valve was adjusted manually using a micrometer and positioned at a constant valve lift of 2–6.46 mm (maximum valve lift). The measurements were carried out at a constant pressure drop of 300 and 600 mmH2O across the intake valve. The pressure adjustment required for different valve lift settings was controlled through the feedback controller valve to pressure relief valve. Therefore, the pressure drop between the inlet of the intake port and the lower end of the cylinder can be kept constant so that the pressure drop can confirm enough turbulence. Pressure and temperature of the intake air were measured by pressure transducer and thermocouple, respectively. This tool was also accompanied by a honeycomb-type commutator to convert angular momentum of air into a rotational force. The honeycomb was located from the cylinder head at a depth of 1.75×

2. Setup of experiments 2.1. Engine Specification The cylinder head of four-stroke and four-valves single-cylinder small motorcycle engine 125 cm3 was examined (Fig. 1). Table 1 provides detailed specifications of the engine. Any calculation of port flow in the small engine was based on one dimensional characteristic for all 3

Applied Energy 255 (2019) 113863

B. Wahono, et al.

Fig. 3. Geometric model of the small engine.

0.04 0.035

flow rate (m^3/s)

0.03 0.025 0.02 0.015 experiment 300 mmH2O

0.01

Simulation 300 mmH2O Experiment 600 mmH2O

0.005

Simulation 600 mmH2O

0 2

3

4

5 valve lift (mm)

6

7

Fig. 4. Airflow rate in steady-state flow bench experiment and simulation.

inner diameter. A load cell connected to the commutator was used to transmit rotational force added by the air to the commutator. Finally, measurements were conducted on air flow rate, forces required to fix commutator to not get rotated, and rotational force with the load cell.

2.3. Governing equation To provide a good understanding on the suction capability of engine, this study analyzed the ability to breathe by observing both flow and discharge coefficients. In prior studies, different definitions have been taken for flow coefficient and discharge coefficient. This study particularly relied on a definition by Xu [30] due to its more frequent 4

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.6

flow coefficient

0.5 0.4 0.3 0.2

experiment 300 mmH2O Simulation 300 mmH2O

0.1

Experiment 600 mmH2O Simulation 600 mmH2O

0 2

3

4

5 valve lift (mm)

6

7

Fig. 5. Flow coefficient in steady-state flow bench experiment and simulation.

0.8 experiment 300 mmH2O Simulation 300 mmH2O

Coefficient of Discharge

0.7

Experiment 600 mmH2O Simulation 600 mmH2O

0.6 0.5 0.4 0.3 0.2 2

3

4

5 valve lift (mm)

6

7

Fig. 6. Discharge coefficients in steady-state flow bench experiment and simulation.

uses in the analysis of intake flow by vehicle companies. Flow coefficient, Cf, was expressed as follows.

Cf =

Q Aseat xVo

was stated as follows.

Aseat = (1)

2 Dseat

(3)

where D referred to the diameter of intake valve seat. Closely related to the flow coefficient, discharge coefficient, Cd, was hence expressed as follows.

where Q referred to measured volume flow rate, VO was the velocity head and Aseat was the inner seat area. Meanwhile, the velocity head, VO, was defined as follows.

Vo =

4

2x P

Cd =

(2)

Q Av xVo

where Av referred to the orifice area between seat and valve head.

where ρ referred to air density and ΔP was pressure drop. Besides, Aseat

5

(4)

Applied Energy 255 (2019) 113863

B. Wahono, et al.

a). 300 mmH2O

b). 600 mmH2O Fig. 7. Velocity vector for varied valve lifts and pressure drops at x-y plane (z = −4 mm).

a). 300 mmH2O

b). 600 mmH2O Fig. 8. Velocity vector for varied valve lifts and pressure drops at x-z plane (y = 11.47 mm).

6

Applied Energy 255 (2019) 113863

B. Wahono, et al.

a). 300 mmH2O

b). 600 mmH2O Fig. 9. Vorticity magnitude for varied valve lifts and pressure drops at x-y plane (z = −4 mm).

a). 300 mmH2O

b). 600 mmH2O Fig. 10. Vorticity magnitude for varied valve lifts and pressure drops at x-z plane (y = 11.47 mm).

7

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.02

0.01

0

Swirl ratio

300

350

400

450

500

550

600

650

700

750

-0.01

-0.02

-0.03

300 mmH2O 600 mmH2O -0.04

Crank angle (deg) Fig. 11. In-cylinder swirl ratio at different pressure drops.

3. Simulation model

3.2. Boundary condition

3.1. Geometric model

In this study, the boundary condition was set for a steady-state condition, which was equipped with an open-end cylinder. Besides, the position of initial intake valve and exhaust valve position were critical. In this work, the intake valve lift for the steady-state flow bench was set to a fixed position. The fluid for the steady-state model was air, which was a mixture of 23% oxygen and 77% nitrogen by mass. Air temperature was set at 300 K, while pressure was set at 1 bar. Then, the air density was calculated by the ideal gas law.

In this study, numerical analyses were executed by applying a commercial CFD package “CONVERGE”. The surface geometry and input files were first prepared before running a CONVERGE simulation. A stereolithography (STL) format file produced by most CAD software contained the surface geometry required by CONVERGE. Fig. 3 shows the geometric model of the observed small engine. The surface geometry was set up and exported by a graphical user interface pre-processor to the CONVERGE solver. Surface geometry with a unique boundary area was included in a surface definition file, while volume mesh was built automatically during the runtime of CONVERGE solver. CONVERGE package has an ability to eliminate the need for a computational grid by applying an innovative boundary approach. Producing the grid internally to the code at run time offered various advantages. First, it could manage moving surfaces by reproducing grids near the moving boundaries without specifying input files. Second, it permitted changing grids during a simulation. Then, it required less time than other CFD methods.

3.3. Computational method Conditions for simulations were set in the input files. The flow was simulated by applying the renormalized group (RNG) k-ε turbulence model and the law of the wall, which were defined in CONVERGE to identify all solid surfaces. The steady-state model was simulated under varied mass flow rates. The temperature of engine cylinder was initially set at 300 K for all rates. An implicit discretization procedure was applied to solve the problem of discrete Navier-Stokes equation on a cartesian grid. General equation used to construct the in-cylinder flow models included the mass, momentum and energy conversion equations. Then, implicit pressure for the splitting of operator (PISO) algorithm was used to solve the velocity-pressure coupling problem. This

8

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.01

0.005

0

Swirl ratio

3

3.5

4

4.5

5

5.5

6

6.5

7

-0.005

-0.01

-0.015

-0.02

300 mmH2O 600 mmh2O -0.025

valve lift (mm) Fig. 12. Variations of swirl ratio at different valve lifts.

algorithm was applied to the grid collocated by applying the Rhie-Chow scheme [31].

steady-state flow bench experiments and simulations at different valve lifts and pressure drops. The coefficients appeared to increase proportionally with valve lift as the effective flow region thru the valve also increased until 4.5 mm valve lift. Between 4.5 and 6.46 mm (maximum valve lift), the flow coefficient was almost constant. The flow coefficient increased with valve lift due to the effective flow area through the valve increased. The flow coefficient significantly affected engine’s breathing capacity due to a lot of air entering the cylinder on a higher valve lift. In lower valve lifts, however, the effects were quite small. Looking at Fig. 5, an increase in pressure drops did not significantly affect flow coefficients. There was a good agreement between experiment and simulation. The biggest deviation was only 15.87% at 2.176 mm valve lift and 300 mmH2O pressure drop.

4. Experimental and simulation results 4.1. Air flow rate Fig. 4 illustrates a comparison of air flow rate as a valve lift function between the results of experiments in steady-state flow benches and simulation results at 300 and 600 mmH2O pressure drops. It suggested air flow rate to increase when the valve lift increased for up to 4.5 mm. Between 4.5 and 6.46 mm (maximum valve lift), air flow rate was almost constant. In fact, a higher valve lift causes more air flow into the engine cylinder and hence higher engine breathing capacity. An increase in pressure drops delivered increases in air flow rate. The air flow rate was 42.86% higher for 600 mmH2O than that of 300 mmH2O. Experiments and simulation produce similar results for all pressure drops. The biggest deviation was 17.76% at 2.176 mm lift and 300 mmH2O drop.

4.3. Discharge coefficient Furthermore, Fig. 6 illustrates a comparison of discharge coefficients measured in the steady-state flow bench experiments and simulation for varied valve lifts at 300 and 600 mmH2O pressure drops. Looking at the figure, the discharge coefficient was not independent, indicating a flow limitation by the seat lips and valve on the valve lift. At a low valve lift, air added to the engine cylinder passed through both the seat lip and valve, resulting in a viscous shear effect. Since air

4.2. Flow coefficient Fig. 5 compares calculated flow coefficients based on the results of

9

Applied Energy 255 (2019) 113863

B. Wahono, et al.

1.8

300 mmH2O 1.6

600 mmH2O

1.4

Tumble Ratio

1.2 1 0.8 0.6 0.4 0.2 0

300

400

500

600

700

Crank angle (deg) Fig. 13. In-cylinder tumble at different pressure drops.

1.2

300 mmH2O 600 mmH2O

1

Tumble Ratio

0.8

0.6

0.4

0.2

0 3

3.5

4

4.5

5

5.5

6

Valve lift (mm) Fig. 14. Variations of tumble ratio at different valve lifts.

10

6.5

7

Applied Energy 255 (2019) 113863

B. Wahono, et al.

60

300 mmH2O 600 mmH2O 50

TKE (m^2/s^2

40

30

20

10

0

300

350

400

450

500

550

600

650

700

750

Crank angle (deg) Fig. 15. In-cylinder TKE at different pressure drops.

entering the cylinder increased due to increases in valve lift, the discharge coefficient slightly decreased with increases in the Reynolds number. It occurred due to decreases in the viscous shear effect. When the valve lifts higher, fluid inertia prevented the flow from rotating along with the valve seat, breaking the flow away and forming a free jet. On the other hand, an increase in pressure drop did not significantly decrease discharge coefficient. However, there was a slight drop of discharge coefficient at 3.179 mm valve lift and 600 mmH2O pressure drop, indicating a flow segregation from the valve. Experiment and simulation results appear similar. The biggest deviation was only 11.51% for 2.176 mm valve lift and 300 mmH2O pressure drop.

lifts. Symmetrical counter-rotating vortices also appeared in all images for high valve lifts, occurring for either 300 or 600 mmH2O pressure drop. At 6.1 mm and 6.46 mm valve lifts, flow velocities near cylinder edge were more obvious. The highest velocity, however, occurred at the center of the cylinder. In fact, higher valve lifts resulted in greater velocities at the center of the cylinder due to higher turbulences in the areas. Increased turbulences produced a more homogenous in-cylinder air-fuel mixing. Then, increases in turbulence directly increased the rate of fire propagation. 4.4.1.2. Effect of pressure drop. Fig. 7 also illustrates varying flow motions caused by an increased pressure drop from 300 to 600 mmH2O. Besides, increased pressure drops for the same valve lift appeared to establish vortices to maintain their shape in all images. The only difference was observed for a changing strength of the vortex, in which the greatest velocity occurred at the maximum valve lift (6.46) mm and 600 mmH2O pressure drop. It was about 67% larger than that of 300 mmH2O pressure drop. It appeared parallel with Bernoulli’s calculation where the maximum inlet speed for cases of pressure drops of 300 and 600 mmH2O were 60 m/s and 100 m/s, respectively.

4.4. Velocity vectors for in-cylinder flow field 4.4.1. Horizontal swirl plane 4.4.1.1. Effect of valve lift. Fig. 7 illustrates a comparison of in-cylinder flow field velocity vectors for different valve lifts, for two pressure drops (300 and 600 mmH2O) under motoring conditions, along a sectional x-y plane (z = −4 mm) passing through the intake valve and in-cylinder small engine. Swirl flow patterns were obviously varying, which was caused by adjusted valve lifts at 300 mmH2O (Fig. 7a) and 600 mmH2O pressure drops (Fig. 7b). Besides, flow velocity increased when the valve lift increased up to the maximum valve lift being set (6.46 mm). It appeared to differ between 6.1 mm and 5.7 mm valve lifts where the velocity of flow is smaller than that of 6.46 mm lifts. In general, the flow motion showed a lot of variation with increased valve

4.4.2. Vertical tumble plane 4.4.2.1. Effect of valve lift. Fig. 8 illustrates a comparison of in-cylinder flow field velocity vectors at varied valve lifts, for two pressure drops (300 and 600 mmH2O) under motoring conditions, along a sectional x-z plane (y = 11.47 mm) passing through the intake valve and in-cylinder

11

Applied Energy 255 (2019) 113863

B. Wahono, et al.

45

300 mmH2O 40

600 mmh2O

35

TKE (m^2/s^2)

30

25 20 15 10 5 0 3

3.5

4

4.5

5

5.5

6

6.5

7

Valve lift (mm) Fig. 16. Variations of TKE at different valve lifts.

small engine. The figure showed tumble flow patterns caused by varied valve lifts for both 300 and 600 mmH2O pressure drops. The flow pattern remained similar as expected, but flow velocity in the pattern increased along with increased valve lifts for both 300 mm and 600 mmH2O pressure drops. At 5.7 mm valve lift, a small vortex began to form in a clockwise direction close to the intake valve. It was clearer in 600 mmH2O pressure drop setting. In this step, air entered engine cylinder through the side of seat lips and the exhaust side of valve, which interacted with air entering near the side of cylinder wall. An interaction as such created a small vortex close to the intake valve. At 6.1 mm valve lift, more air entered the to the engine cylinder. In this stage, more vortices were formed, e.g., a vortex close to the intake valve and another vortex close to the exhaust valve. At 6.46 mm valve lift, one big vortex occurred in the center of the cylinder. Maximum flow velocity produced at the 6.46 mm valve lift was quite higher than those of 6.1 mm and 5.7 mm valve lifts. It was apparently in accordance to the calculated maximum flow velocities based on measured volumetric flow rate. The maximum flow velocities for 6.46 mm, 6.1 mm and 5.7 mm valve lifts were 100.56 m/s, 100.51 m/s and 100.48 m/s, respectively.

lift, i.e., reduces the area from which the vector field was generated. It was caused by a higher flow rate inside the cylinder, which caused an increase in the spreading of velocity across the cylinder. It occurred at all valve lifts, including 5.7 mm, 6.1 mm and 6.46 mm. 4.5. Vorticity magnitude for in-cylinder flow field 4.5.1. Horizontal swirl plane 4.5.1.1. Effect of valve lift. Vorticity has been stated as the velocity field curl and posited as a measure of local rotation of a fluid. Technically, it is taken to quantitatively express the initiation, transformation and extinction of vortices. Fig. 9 illustrates a comparison of the vorticity magnitudes of in-cylinder flow field at varied valve lifts, for two pressure drops (300 and 600 mmH2O) under motoring conditions, along a sectional x-y plane (z = −4 mm) passing through the intake valve and in-cylinder small engine. Vorticity at 5.7 mm valve lift and 300 mmH2O pressure drop appeared to dominate at the right side of the cylinder where the intake valve was located. Meanwhile, the vorticity magnitude was very small at the left side of the cylinder around the exhaust valve. At an increased valve lift to 6.1 mm, dominant vorticity was still on the right side of the cylinder and reached its peak at the center of the cylinder. At a 6.46 mm valve lift, the vorticity magnitude got bigger at the left side of the cylinder even though the biggest vorticity was still at the center of the cylinder. Vorticity changes at 300 mmH2O pressure drop were similar to those of 600 mmH2O pressure drop. In short, the strength of vorticity increased with increases in valve lift.

4.4.2.2. Effects of pressure drop. Fig. 8 also illustrates variations in tumble flow motion caused by an increasing pressure drop from 300 to 600 mmH2O. As expected, the flow pattern remained in the same orientation. However, the magnitude of velocity in this flow pattern increased. Related to the flow pattern, therefore, there was an obvious similarity between an increase in pressure drop and increases in valve

12

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.00018 0.00016

0.00014

Air mass (kg)

0.00012 0.0001 0.00008 0.00006 0.00004

300 mmH2O

0.00002

600 mmh2O 0 300

400

500

600

700

800

Valve lift (mm) Fig. 17. In-cylinder air mass at different pressure drops.

4.5.1.2. Effects of pressure drop. Fig. 9 also illustrates a variation of vorticity magnitudes for in-cylinder flow pattern generated by an increase in pressure drop across the cylinder head from 300 to 600 mmH2O for the same valve lift. Looking at the figure, the vorticity pattern remained in a similar orientation; however, the vorticity magnitude increased. In areas with have large vorticity, e.g., the center of and along the edge of the cylinder close to the intake valve, there was an obviously high local rotation, and occurred for all valve lifts (5.7 mm, 6.1 mm and 6.46 mm).

maximum level (6.46 mm), vorticity began to spread to several places although the maximum vorticity was still around the end of the intake valve. This process was similar to what happened at a 600 mmH2O pressure drop at 5.7 mm, 6.1 mm and 6.46 mm valve lifts. Increases in intake valve at both 300 and 600 mmH2O pressure drops resulted in an increased vorticity magnitude. The largest vorticity magnitude occurred around the end of both sides of the intake valve. 4.5.2.2. Effect of pressure drop. Fig. 10 illustrates variations of vorticity magnitude caused by an increased pressure drop from 300 to 600 mmH2O for the same valve lift. As expected, the vorticity pattern remained to take the same shape. However, vorticity magnitude in this flow pattern was 57% greater for the 6.46 mm valve lift at 600 mmH2O (91000 1/s) pressure drop than for 6.46 mm valve lift at 300 mmH2O pressure drop (58000 1/s).

4.5.2. Vertical tumble plane 4.5.2.1. Effect of valve lift. Fig. 10 illustrates a comparison of the vorticity magnitudes of in-cylinder flow field for different valve lifts, at two pressure drops (300 and 600 mmH2O) under motoring conditions, along a sectional x-z plane (y = 11.47 mm) passing through the intake valve and in-cylinder small engine. There were obvious variations of vorticity magnitude for in-cylinder flow structures produced by increases in valve lift at 300 mmH2O (Fig. 10a) and 600 mmH2O (Fig. 10a) pressure drops across the cylinder head. The vorticity at 5.7 mm valve lift and 300 mmH2O pressure drop began to form at the side of the intake valve and reached its peak at both intake valve valves, which was due to collisions of airflow at the end of the intake valve side with the cylinder wall when the intake valve began to open. Vorticity began to expand in the same position when the intake valve was opened larger to 6.1 mm. When the intake valve reached its

4.6. Swirl ratio The formation of swirl flow began with flows from the intake port into the engine cylinder, which produced an initial angular momentum, and interactions between the cylinder wall, intake port and piston surface. Fig. 11 illustrates a comparison of the swirl ratio for two different pressure drops (300 and 600 mmH2O) during the intake and compression strokes. Rotating flow was first formed in the initial phase of induction process. The flow was up and down until it achieved a peak

13

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.00012

300 mmH2O 600 mmh2O

0.0001

Air mass (kg)

0.00008

0.00006

0.00004

0.00002

0 3

4

5

6

7

Valve lift (mm) Fig. 18. Variation of air mass at different valve lifts.

in the negative axis around 500 °CA. After the point, it became almost stable until around 600 °CA. It possibly occurred due to a turbulent dissipation in the flow and the angular momentum disappeared into the wall. Later, the swirl ratio increased during the last stages of a compression stroke. It occurred at almost all pressure drops (except at 700 °CA crank angle and 300 mmH2O pressure drop) due to a tangential velocity of the swirl flow, which increased due to an interaction between the piston and the cylinder wall. Next, Fig. 12 illustrates effects of the pressure drop to the swirl flow for varied valve lifts. Increased pressure drops in a small engine at the beginning of an intake stroke, especially at the beginning of a valve lift opening, had no effect on the swirl ratio. It occurred until the valve lift reached 5 mm. After the threshold, however, it delivered a 25% reduction of the swirl ratio when the valve lift reached 6.25 mm. It was probably caused by a poor strength of the intake flow in the engine, even though the pressure drop was increased to the only point swirling vortices were generated. When the valve lift increased from 6.39 mm to 6.46 mm (maximum) at 300 mmH2O pressure drop, the swirl ratio decreased again, which was due to a disappearing angular momentum to the wall and a turbulent dissipation in the flow.

comparison of tumble ratio for two different pressure drops (300 and 600 mmH2O) under motoring conditions. Cross tumble (TRx) and normal tumble (TRy) were two important tumble vortices rotation in a flow motion, while Tumble Ratio (TR) was the total of these two tumbles. Three phases of tumble ratio were production, stabilization and destruction. In this study, production phase occurred in the early intake process up to 380 °CA and then decreased until zero at 400 °CA. Next, a stabilization began and reached its highest level at 490 °CA and dropped dramatically until 560 °CA. It was due to a decreased incoming momentum of flow to the engine cylinder through the valves and an increased cylinder volume to BDC. From this point, the tumble flowed upward again until 630 °CA due to a conservation of angular momentum. Then, the last phase (destruction) occurred at the end of compression stroke until 720 °CA. There was a dramatic decrease due to the dissipation effect. Besides, pressure drop from 300 to 600 mmH2O marginally affected increments of the tumble ratio in a small engine. The biggest increase of tumble ratio (2.99%) occurred at 490 °CA. It might be due to a poor strength of the intake flow in a small engine despite an increased pressure drop to the only point generating tumbling vortices, which was not sustainable during the stabilization phase. In short, increased pressure drop in a small engine did not deliver a significant effect on the increase of the tumble ratio. Fig. 14 illustrates the effect of the pressure drop to the tumble flow for varied valve lifts. Increased pressure drop in a small engine at the beginning of intake stroke until the end of compression stroke was observed to not affecting

4.7. Tumble ratio The tumble ratio should usually be approximated to offer a significant effect on the in-cylinder tumble flow. Fig. 13 illustrates a

14

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.0012

300 mmH2O 600 mmh2O 0.001

length scale (m)

0.0008

0.0006

0.0004

0.0002

0 300

400

500

600

700

800

Valve lift (mm) Fig. 19. In-cylinder length scale at different pressure drops.

the tumble ratio.

pressure drop from 300 to 600 mmH2O when the valve lift started to open until 5 mm. After the valve lift was open more than 5 mm, however, there was a significant increase in TKE along with an increase in pressure drop.

4.8. Turbulent kinetic energy (TKE) In a turbulent flow, TKE refers to the mean of kinetic energy per unit mass, which is indicated by the root-mean-square (RMS) measure of velocity fluctuation. A critical parameter to estimate the turbulent viscosity is TKE. Fig. 15 illustrates a comparison of TKE with the crank angle for two pressure drops (300 and 600 mmH2O) under motoring conditions during intake and compression strokes. Looking at the figure, there were two peaks in the TKE. The first peak (highest) occurred in the middle of the intake stroke around 460 °CA, which was due to a turbulence produced when air flowed through the intake valve curtain area. The second peak occurred after the first peak in the last stage of the compression stroke, which occurred when the piston moved further up in the cylinder and fell into a smaller space. Besides, increased pressure drops from 300 to 600 mmH2O were observed to affect increments of TKE in a small engine from the early intake stroke until the end of compression stroke. It reached its highest increment (200%) at 470 °CA, which might due to the increase in pressure drop, causing a significant increase in the inlet air velocity, especially when a transition dynamics of intake valve started closing from a full opening position. Increases in the incoming air velocity indicated a collapse of tumbling vortices. Then, the destruction of tumbling vortices increased turbulence, resulting in a higher TKE. In Fig. 16, TKE did not change significantly along with an increase in

4.9. Air mass In general, the speed of a gasoline engine is regulated by quality rather than quantity. Speed adjustments set by a throttle may get eliminated by quality settings. Fig. 17 illustrates a comparison of incylinder air mass versus CAD at two different pressure drops (300 and 600 mmH2O) under motoring conditions during intake and compression strokes. Accumulated air mass entered the engine cylinder from the early of intake stroke when the valve lift started to open, then continued to increase until the end of intake stroke at 540 °CA. Then, the air mass became stable until the end stages of compression stroke. Apparently, there was no significant effect of increased pressure drop from 300 to 600 mmH2O on the air mass accumulation in a small engine along the intake stroke. Fig. 18 provides a more detailed comparison of air mass for different valve lifts, which revealed no significant effect on air mass entering the engine cylinder despite an increased pressure drop. During compression stroke, however, there was a ± 3.77% increase of air mass due to increased pressure drop from 300 to 600 mmH2O. It might due to the higher inertia of incoming air than that of moving upward air mass near the piston.

15

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.0002

length scale (m)

0.00018

0.00016

0.00014

0.00012

300 mmH2O 600 mmh2O 0.0001 3

4

5

6

7

Valve lift (mm) Fig. 20. Variations of length scale at different valve lifts.

4.10. Turbulent length scale

rate tended to reduce due to a reduced in-flow work taken against the action of viscous stresses. When the dissipation rate decreased, the turbulent length scale would be expected to increase. Fig. 20 provides a more detailed illustration of a comparison of length scales at varied valve lifts and increased pressure drops. Looking at the figure, there was an increase in length scale when the valve lift was opened from 3.5 to 5 mm due to increased pressure drop from 300 to 600 mmH2O. When the valve lift was opened larger than 6.25 mm, however, the length scale decreased despite an increased pressure drop from 300 to 600 mmH2O. Then, it went up again after the valve lift reached its maximum (6.46 mm). These values indicated a fluctuated length scale at the beginning of stroke intake, which might due to the shape of intake port at air entrance near the valve lift. Despite the small fluctuation, it required a higher concern because the turbulent length scale was a critical parameter in determining turbulence. Modifying the shape of the intake port on the engine cylinder head might hence help improve the condition.

The turbulence length scale is a measure to approximate the characteristics of inlet turbulence during a CFD simulation. It measures the physical quantity of large eddies loaded with energy in a turbulent flow. Each eddy scale is appropriate to a specific turbulence length scale. Among others, a standardized eddy scale includes integral length scales. In this study, the integral length scale was applied to estimate the turbulent length scale. Fig. 19 illustrates a comparison of turbulent length scale versus CAD for two different pressure drops (300 and 600 mmH2O) under motoring conditions during intake and compression strokes. Looking at the figure, the turbulent length scale was almost constant at the beginning of intake stroke until 450 °CA. Then, it increased rapidly and achieved its peak when the intake valve was almost closed ( ± 590 °CA). Finally, it dropped dramatically until the last stage of compression stroke. In fact, it occurred for all pressure drops. Generally, the level of dissipation influenced the turbulent length scale. A small turbulent length scale represented a big dissipation rate, which occurred due to an increased in-flow work taken against the action of viscous stresses. Increased flow work performed on the action of viscous stresses could then increase the rate of dissipation. In fact, turbulent length scales in a small engine appeared insensitive to increased pressure drops, especially in the intake stroke up to 490 °CA. Above the point, turbulent length scale increased as pressure drop increased. The highest increase of length scale ( ± 44.5%) occurred at 595 °CA when the intake valve was almost closed. When the intake valve was closed and the pressure drop increased, the dissipation

4.11. Turbulent kinematic viscosity Turbulent kinematic viscosity (eddy viscosity) is a measure of the relative magnitudes of fluid viscosity and inertia. Turbulent kinematic viscosity is a parameterization for eddy momentum flux that works well when there is a small vortex in the flow. However, it works poorly when there is a large vortex. Fig. 21 illustrates a comparison of the turbulent kinematic viscosity with CAD for two pressure drops (300 and 600 mmH2O) under motoring conditions during intake and compression

16

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.0035

300 mmH2O 600 mmh2O

Turbulent Kinematic Viscosity (m^2/s)

0.003

0.0025

0.002

0.0015

0.001

0.0005

0 300

400

500

600

700

800

Valve lift (mm) Fig. 21. In-cylinder of turbulent kinematic viscosity at different pressure drops.

strokes. There was an apparent decrease in turbulent kinematic viscosity at the beginning of the intake stroke for up to 400 °CA. After the point, turbulent kinematic viscosity rose sharply until it reached its peak at around 590 °CA and dropped dramatically until the stroke compression ended. The trend was present for all pressure drops. In fact, turbulent kinematic viscosity in a small engine at the beginning of stroke intakes to 400 °CA decreased when the pressure drops increased from 300 to 600 mmH2O. The decrease might occur due to the shape of intake port at air entrance near the valve lift. It, however, did not occur for a long time because turbulent kinematic viscosity increased sharply after 400 °CA due to an increase in pressure drop. The opening of valve lift and the movement of piston produced more small eddies inside the cylinder, causing more increases in turbulent kinematic viscosity. Fig. 22 provides a more detailed comparison of turbulent kinematic viscosities for varied valve lifts and increased pressure drops. When the valve lift was opened between 3.5 and 5 mm, there was a decrease in turbulent kinematic viscosity due to an increased pressure drop from 300 to 600 mmH2O. When the valve lift was opened larger until maximum (6.46 mm), however, there was an increase in turbulent kinematic viscosity due to an increased pressure drop from 300 to 600 mmH2O. It implied turbulent kinematic viscosity fluctuations to occur at the beginning of intake stroke, which might due to the shape of intake port at air entrance near the valve lift. Despite the small fluctuation, it required more attention as turbulent kinematic viscosity was a critical parameter in determining turbulence. To improve this

condition, the shape of intake port on the engine cylinder head might need to be modified. 5. Conclusion This study aimed at comparing the results of experiments on incylinder flow of a small engine using steady-state flow bench and simulations based on adjusted pressure drops. Besides, this study illustrated a variation of in-cylinder flow motions of a small engine using CFD methods. Then, this study analyzed characteristics of in-cylinder airflow being performed in a four-stroke small engine under motoring condition at two different pressure drops (300 and 600 mmH2O). Looking at the results, this study suggested conclusions as follows.

• There was a good agreement between the results of steady-state flow • • •

17

bench experiments and simulations in terms of airflow rates, flow coefficients and discharge coefficients at 300 and 600 mmH2O pressure drops. In both horizontal and vertical planes, increased valve lift and pressure drop were observed to deliver an increased magnitude of velocity and vorticity. Increasing pressure drop at the beginning of valve lift opening appeared to have no effect on the swirl ratio until the valve lift reached 5 mm. After the threshold, it delivered a reduction effect on the swirl ratio for up to 75% at 6.25 valve lift. Increasing pressure drop at the beginning of intake stroke appeared

Applied Energy 255 (2019) 113863

B. Wahono, et al.

0.0008

300 mmH2O

Turbulent Kinematic Viscosity (m^2/s)

0.0007

600 mmh2O

0.0006

0.0005

0.0004

0.0003

0.0002

0.0001

0 3

4

5

6

7

Valve lift (mm) Fig. 22. Variations of turbulent kinematic viscosity at different valve lifts.

• •

to have no significant effect on the tumble ratio and accumulated air mass. However, there was a 3.77% increase in accumulated air mass during compression strokes. Increasing pressure drop delivered a significant effect on TKE, turbulent length scale and turbulent kinematic viscosity. TKE reached the highest value ( ± 200%) at 470 °CA. However, turbulent length scale and turbulent kinematic viscosity reached the highest value at around 590 °CA where the intake valve had almost closed. Increases in turbulence produced a more homogenous in-cylinder air-fuel mixing. Then, the increases in turbulence directly increased the rate of fire propagation.

[4] Mohan B, Yang W, kiang Chou S. Fuel injection strategies for performance improvement and emissions reduction in compression ignition engines—a review. Renew. Sustain. Energy Rev. 2013;28:664–76. [5] Ishiyama T, Kang J, Ozawa Y, Sako T. Improvement of performance and reduction of exhaust emissions by pilot-fuel-injection control in a lean-burning natural-gas dual-fuel engine. SAE Int J Fuels Lubr 2012;5(1):243–53. [6] Claudio F, Cristian C, Giulio C, Marco BG, Stefania F, Federico B, et al. Numerical evaluation of the applicability of steady test bench swirl ratios to diesel engine dynamic conditions. Energy Proc 2015;81:732–41. [7] Bari S, Saad I. CFD modelling of the effect of guide vane swirl and tumble device to generate better in-cylinder air flow in a CI engine fuelled by biodiesel. Comput Fluids 2013;84:262–9. [8] Wang T, Li W, Jia M, Liu D, Qin W, Zhang X. Large-eddy simulation of in-cylinder flow in a DISI engine with charge motion control valve: Proper orthogonal decomposition analysis and cyclic variation. Appl Therm Eng 2015;75:561–74. [9] Wu C, Deng K, Wang Z. The effect of combustion chamber shape on cylinder flow and lean combustion process in a large bore spark-ignition CNG engine. J Energy Inst 2016;89(2):240–7. [10] Krishna AS, Mallikarjuna JM, Davinder K, Babu YR. In-cylinder flow analysis in a two-stroke engine-a comparison of different turbulence models using CFD. SAE Tech Paper 2013. [11] Raj ARGS, Mallikarjuna JM, Ganesan V. Energy efficient piston configuration for effective air motion–a CFD study. Appl Energy 2013;102:347–54. [12] Fang T, Singh S. Predictions of flow separation at the valve seat for steady-state port-flow simulation. J Eng Gas Turbines Power 2015;137(11):111512. [13] Senecal PK, Pomraning E, Richards KJ, Som S. Grid-convergent spray models for internal combustion engine CFD simulations. In: ASME 2012 Internal combustion engine division fall technical conference; 2012. p. 697–710. [14] Perini F, Miles PC, Reitz RD. A comprehensive modeling study of in-cylinder fluid flows in a high-swirl, light-duty optical diesel engine. Comput Fluids 2014;105:113–24. [15] de Souza GR, de Castro Pellegrini C, Ferreira SL, Pau FS, Armas O. Study of intake manifolds of an internal combustion engine: a new geometry based on experimental results and numerical simulations. Therm Sci Eng Progr 2019;9:248–58. [16] Jemni MA, Kantchev G, Abid MS. Influence of intake manifold design on in-cylinder flow and engine performances in a bus diesel engine converted to LPG gas fuelled, using CFD analyses and experimental investigations. Energy 2011;36(5):2701–15. [17] Sharma CS, Anand TNC, Ravikrishna RV. A methodology for analysis of diesel engine in-cylinder flow and combustion. Progr Comput Fluid Dynam, Int J 2010;10(3):157–67. [18] Gafoor CPA, Gupta R. Numerical investigation of piston bowl geometry and swirl

Acknowledgements This research was financially supported by the Centre for Environmentally Friendly Vehicle (CEFV) as a Global Top Project of KMOE (2016002070009, Development of Engine System and Adapting Vehicle for Model 110 cc and 300 cc Correspond to EURO-5 Emission). Bambang Wahono acknowledges the support from the Program for Research and Innovation in Science and Technology (RISET-Pro) Ref. No: 272/RISET-Pro/FGS/VII/2016, Ministry of Research, Technology and Higher Education, Republic of Indonesia. References [1] Yilmaz N, Vigil FM, Benalil K, Davis SM, Calva A. Effect of biodiesel–butanol fuel blends on emissions and performance characteristics of a diesel engine. Fuel 2014;135:46–50. [2] Saravanan S. Effect of exhaust gas recirculation (EGR) on performance and emissions of a constant speed DI diesel engine fueled with pentanol/diesel blends. Fuel 2015;160:217–26. [3] Costa RC, Sodré JR. Compression ratio effects on an ethanol/gasoline fuelled engine performance. Appl Therm Eng 2011;31(2–3):278–83.

18

Applied Energy 255 (2019) 113863

B. Wahono, et al. ratio on emission from diesel engines. Energy Convers Manage 2015;101:541–51. [19] Qi YL, Dong LC, Liu H, Puzinauskas PV, Midkiff KC. Optimization of intake port design for SI engine. Int J Automot Technol 2012;13(6):861–72. [20] Jainski C, Lu L, Dreizler A, Sick V. High-speed micro particle image velocimetry studies of boundary-layer flows in a direct-injection engine. Int J Engine Res 2013;14(3):247–59. [21] Westerweel J, Elsinga GE, Adrian RJ. Particle image velocimetry for complex and turbulent flows. Annu Rev Fluid Mech 2013;45:409–36. [22] Bertsch M, Schreer K, Disch C, Beck KW, Spicher U. Investigation of the flow velocity in the spark plug gap of a two-stroke gasoline engine using laser-doppleranemometry. SAE Int J Engines 2012;5(1):34–41. [23] El-Adawy M, Heikal MR, Aziz ARA, Siddiqui MI, Wahhab HAA. Experimental study on an IC engine in-cylinder flow using different steady-state flow benches. Alexandr Eng J 2017;56(4):727–36. [24] Rabault J, Vernet JA, Lindgren B, Alfredsson PH. A study using PIV of the intake flow in a diesel engine cylinder. Int J Heat Fluid Flow 2016;62:56–67.

[25] Binjuwair S, Ibrahim S. Numerical analysis of flow structure inside a single cylinder of a 4 valve head under steady-state condition. SAE technical paper; 2013. [26] Mohammadebrahim A, Shafii MB, Hannani SK. The effect of various test parameters on the steady flow test results of a four-valve spark ignition engine: a tentative approach toward standardization. Adv Mech Eng 2013;5:482317. [27] Yang X, Chen Z, Kuo T-W. Pitfalls for accurate steady-state port flow simulations. J Eng Gas Turbines Power 2013;135(6):61601. [28] Saad I, Bari S. Guide vane swirl and tumble device to improve in-cylinder air flow of CI engine using vegetable oil. Proc Eng 2014;90:425–30. [29] Yang X, Kuo T-W, Guralp O, Grover RO, Najt P. In-cylinder flow correlations between steady flow bench and motored engine using computational fluid dynamics. J Eng Gas Turbines Power 2017;139(7):72802. [30] Xu H. Some critical technical issues on the steady flow testing of cylinder heads. SAE Tech Paper 2001. [31] Rhie CM, Chow WL. Numerical study of the turbulent flow past an airfoil with trailing edge separation. AIAA J 1983;21(11):1525–32.

19